Introns Increase Transcriptional Efficiency in Transgenic Mice

Abstract

Experiments were designed to test the effect of introns on gene expression in transgenic mice. Four different pairs of gene constructs, which were identical except that one member of each pair lacked all introns, were compared for expression of mRNA after introduction into the murine germ line by microinjection of fertilized eggs. The expression of two chimeric genes, made by fusing either the mouse metallothionein I or the rat elastase 1 promoter/enhancer to the rat growth hormone gene, was assayed in fetal liver or pancreas, respectively, while two natural genes, an oligonucleotide-marked mouse metallothionein I gene and the human beta-globin gene, were assayed in fetal liver. In each case there was, on average, 10- to 100-fold more mRNA produced from the intron-containing construct. Moreover, mRNA levels were proportional to the relative rates of transcription that were measured in isolated nuclei. However, when the expression of the two mouse metallothionein I gene-based constructs was tested after transfection into cultured cells, little difference was observed. These observations suggest that introns play a role in facilitating transcription of microinjected genes and that this effect may be manifest only on genes exposed to developmental influences.

Full text

Proc.
Nati.
Acad.
Sci.
USA
Vol.
85,
pp.
836-840,
February
1988
Developmental
Biology
Introns
increase
transcriptional
efficiency
in
transgenic
mice
(intervening
sequences/gene
expression/gene
transfer)
RALPH
L.
BRINSTER*,
JAMES
M.
ALLENt,
RICHARD
R.
BEHRINGER*,
RICHARD
E.
GELINAS*,
AND
RICHARD
D.
PALMITERt
*Laboratory
of
Reproductive
Physiology,
School
of
Veterinary
Medicine,
University
of
Pennsylvania,
Philadelphia,
PA
19104;
tHoward
Hughes
Medical
Institute
and
Department
of
Biochemistry
SL-15,
University
of
Washington,
Seattle,
WA
98195;
and
1IDepartment
of
Molecular
Medicine,
Fred
Hutchinson
Cancer
Research
Center,
1124
Columbia
Street,
Seattle,
WA
98104
Contributed
by
Ralph
L.
Brinster,
October
13,
1987
ABSTRACT
Experiments
were
designed
to
test
the
effect
of
introns
on
gene
expression
in
transgenic
mice.
Four
differ-
ent
pairs
of
gene
constructs,
which
were
identical
except
that
one
member
of
each
pair
lacked
all
introns,
were
compared
for
expression
of
mRNA
after
introduction
into
the
murine
germ
line
by
microiqjection
of
fertilized
eggs.
The
expression
of
two
chimeric
genes,
made
by
fusing
either
the
mouse
metallothio-
nein
I
or
the
rat
elastase
1
promoter/enhancer
to
the
rat
growth
hormone
gene,
was
assayed
in
fetal
liver
or
pancreas,
respectively,
while
two
natural
genes,
an
oligonucleotide-
marked
mouse
metallothionein
I
gene
and
the
human
.8-globin
gene,
were
assayed
in
fetal
liver.
In
each
case
there
was,
on
average,
10-
to
100-fold
more
mRNA
produced
from
the
intron-containing
construct.
Moreover,
mRNA
levels
were
proportional
to
the
relative
rates
of
transcription
that
were
measured
in
isolated
nuclei.
However,
when
the
expression
of
the
two
mouse
metallothionein
I
gene-based
constructs
was
tested
after
transfection into
cultured
cells,
little
difference
was
observed.
These
observations
suggest
that
introns
play
a
role
in
facilitating
transcription
of
microinjected
genes
and
that
this
effect
may
be
manifest
only
on
genes
exposed
to
developmental
influences.
Most
mammalian
genes
coding
for
mRNA
are
interrupted
by
noncoding
sequences
known
as
introns,
many
of
which
are
larger
than
the
exons;
thus,
the
entire
gene
may
span
tens
or
even
hundreds
of
kilobases
(1).
For
some
genes,
introns
clearly
separate
functional
or
structural
domains
of
the
proteins
encoded
by
the
exons.
This
observation,
and
the
fact
that
similar
domains
can
be
found
within
different
proteins,
has
led
to
the
suggestion
that
one
function
of
introns
may
be
to
accelerate
the
evolution
of
proteins
with
different
properties
(2,
3).
Introns
also
allow
differential
joining
of
exons
during
splicing,
which
can
-result
in
the
synthesis
of
variant
proteins
with
new
properties
(4,
5).
The
possibility
that
introns
may
be
necessary
for
efficient
processing
and
transport
of
mRNA
to
the
cytoplasm
has
been
examined.
Hamer
and
Leder
(6)
showed
that
a
series
of
simian
virus
40
viruses
that
contained
various
combinations
of
simian
virus
40
and
mouse
j3-globin
splice
sites
produced
stable
mRNAs
only
if
at
least
one
splice
site,
derived
from
either
the
virus
or
P-globin,
was
retained.
In
another
exper-
iment,
a
simian
virus
40
mutant
missing
a
late
gene
intron
failed
to
produce
stable
transcripts
but
could
be
rescued
by
addition
of
an
heterologous
intron
(7,
8).
These
early
obser-
vations
suggested
that
splicing
was
obligatory
for
mRNA
accumulation
in
the
cytoplasm.
As
a
consequence,
the
first-generation
cDNA
expression
vectors
usually
included
a
heterologous
intron
in
addition
to
promoter
and
polyaden-
ylylation
sequences
(9).
Subsequently,
it
was
discovered
that
deletion
of
introns
does
not
always
result
in
loss
of
mRNA
production.
For
instance,
several
viral
genes,
includ-
ing
those
for
ElA
protein
(10)
and
polyoma
large
tumor
(T)
antigen
(11)
function
without
introns.
Likewise,
the
genes
for
bean
phaseolin
(12),
yeast
actin
(13),
and
chicken
thymi-
dine
kinase
(14)
produce
approximately
equal
amounts
of
mRNA
regardless
of
whether
introns
are
present.
These
experiments,
and
the
fact
that
many
second-generation
cDNA
expression
vectors
lack
introns
yet
allow
good
expression,
suggest
that
only
in
a
few
special
cases
are
introns
necessary
for
mRNA
accumulation,
at
least
when
tested
in
culture
cells.
The
ability
to
redirect
expression
of
genes
to
various
cell
types
in
transgenic
animals
by
combining
the
regulatory
elements
of
one
gene
with
the
coding
region
of
another
has
facilitated
a
number
of
unique
experimental
approaches
to
developmental,
physiological,
and
pathological
processes
(15,
16).
Because
of
the
greater
availability
of
cDNAs
and
because
some
natural
genes
are
too
large
to
manipulate
conveniently,
many
chimeric
genes
based
on
cDNA
expres-
sion
vectors
have
been
constructed.
Although
many
worked
well
when
tested
in
cell
culture
and
some
have
expressed
in
transgenic
mice,
our
experience
is
that
there
are
a
large
number
of
cDNA-based
constructs
that
are
not
expressed
or
are
expressed
poorly
in
transgenic
mice
(ref.
16;
and
our
unpublished
observations).
Furthermore,
addition
of
hetero-
logous
introns
to
cDNA
constructs
usually
did
not
rescue
their
expression.
Because
many
of
these
experiments
were
not
well
controlled,
we
tested
several
different
genes
with
and
without
their
introns,
but
with
identical
5'
and
3'
flanking
regions,
to
ascertain
whether
introns
have
a
con-
sistent
effect
on
expression
of
genes
in
transgenic
mice.
MATERIALS
AND
METHODS
Gene
Constructs.
The
MTrGH
gene
construct
(Fig.
1)
has
1.8
kilobases
(kb)
of
the
5'
flanking
sequence
of
the
gene
for
mouse
metallothionein
I
(mMT-I)
(dashed
line)
fused
at
the
Bgl
II
site
(position
+
64)
to
the
Xho
I
site
of
the
gene
for
rat
growth
hormone
(rGH)
(17,
18).
The
6.8-kb
EcoRI-BamHI
fragment
was
isolated
for
microinjection.
The
ErGH
gene
construct
has
4.5
kb
of
the
5'
flanking
region
of
the
gene
for
rat
elastase
I
(dashed
line)
fused
at
an
artificial
Xho
I
site
(position
+
10)
to
the
Xho
I
site
of
the
gene
for
rGH
(19).
The
8.5-kb
HindIII
fragment
was
isolated
for
microinjection.
For
transcription-rate
measurements,
a
plasmid
containing
the
rGH
sequences
between
Xho
I-Xba
I
and
Ava
I-Nsi
I
was
constructed
(solid
bars
beneath
the
ErGH
construct),
thereby
eliminating
the
repeat
sequences
that
lie
in
intron
B.
The
mMT-I
gene
(17,
20)
was
marked
by
insertion
of
a
pair
of
oligonucleotides
(the
mRNA
strand
sequence
is
5'
Abbreviations:
rGH,
rat
growth
hormone;
mMT-I,
mouse
metallo-
thionein
I;
BHK
cells,
baby
hamster
kidney
cells.
836
The
publication
costs
of
this
article
were
defrayed
in
part
by
page
charge
payment.
This
article
must
therefore
be
hereby
marked
"advertisement"
in
accordance
with
18
U.S.C.
§1734
solely
to
indicate
this
fact.

Proc.
Natl.
Acad.
Sci.
USA
85
(1988)
837
rGH
I
a
*..
rE-I
'
rGH
z
LL.
L.
A
0
E
CS
MT1
w
m
=
c
P-globin
I
FIG.
1.
Diagram
of
gene
constructs
injected
into
fertilized
mouse
eggs.
The
intronic
version
of
each
of
the
genes
is
shown,
drawn
approximately
to
scale;
the
intronless
versions
are
the
same
except
that
cDNA
sequences
were
substituted
between
the
first
and
last
exons.
CACGGTTGACTAAGCTA)
into
the
Bgl
II
site
(position
+64).
The
EcoRI-BamHI
fragment
that
included
350
base
pairs
(bp)
of
pBR322
(dashed
line)
was
microinjected.
The
human
8-globin
gene
is
included
on
a
Hpa
I
(-
815)-Xba
I
(+
3300)
fragment
that
was
microinjected
(21).
Transgenic
mice
were
identified
by
dot
hybridization
using
nick-
translated
probes:
for
rGH,
a
385-bp
Xho
I-Pvu
II
fragment
that
includes
most
of
exons
1
and
2
and
intron
A
was
used;
for
mMT-I,
pBR322
was
used;
and
for
f-globin,
the
Hpa
I-Xba
I
fragment
was
used.
Transgenic
Mice.
Each
of
the
genes
shown
in
Fig.
1
(with
or
without
introns)
was
microinjected
into
fertilized
mouse
eggs,
and
transgenic
fetuses
or
mice
were
identified
by
dot
hybridization
(22).
Total
nucleic
acids
were
isolated
by
the
NaDodSO4/proteinase
K
method
(23).
The
amount
of
mRNA
in
liver
or
pancreas
was
measured
by
solution
hybridization
with
oligonucleotides
complementary
to
rGH
mRNA
(5'
GCATTGGCAAACAGACTGGACAAGGG-
CATG),
the
marked
mMT-I
mRNA
(5'
GATCTAGCT-
TAGTCAACCGTG),
or
human
,B-globin
mRNA
(5'
CCA-
CAGGGCAGTAACGGCAGA)
as
described
(24).
Single-
stranded
phage
M13
clones
carrying
the
complementary
sequence
were
used
as
hybridization
standards.
mRNA
molecules
per
cell
were
calculated
after
determining
the
fraction
of
total
nucleic
acid
that
was
DNA:
day
16
fetal
liver,
50%;
day
15
fetal
pancreas,
33%;
and
week
6
adult
pancreas,
11%
(25).
Transcription
"Run
On"
Experiments.
Nuclei
were
iso-
lated
(26)
and
used
to
measure
the
relative
rate
of
transcrip-
tion
essentially
as
described
(27).
Nuclei
(100
j.g
of
DNA)
were
incubated
in
100
,ul
in
the
presence
of
transcription
buffer
containing
100
,uCi
(1
Ci
=
37
GBq)
of
[32P]UTP
for
45
min
at
26°C.
RNA
was
isolated
as
described
(27),
and
three
different
quantities
(7.5,
15,
and
30
,ul
containing
7-13
x
10W
cpm/,ul)
were
hybridized
to
immobilized
filter
disc
contain-
ing
1.2
,ug
of
rGH
in
pUC
(see
Fig.
1),
mouse
albumin
cDNA
in
pBR322,
mMT-I,
and
pUC19
for
48
hr
at
45°C,
washed
extensively,
treated
with
RNases
A
and
T,
and
then
assayed
(27).
The
relative
rate
of
transcription
was
calculated
as
cpm
specifically
bound
per
106
cpm
of
input
RNA
and
is
ex-
pressed
as
ppm.
Because
the
rGH
plasmid
has
1.17
kb
of
DNA
that
could
hybridize
to
the
MTrGH
transcripts
but
only
0.63
kb
of
that
would
be
present
in
intronless
tran-
scripts,
the
latter
values
were
multiplied
by
1.85
for
compar-
ison.
The
albumin
cDNA
has
a
1.3-kb
insert.
The
average
rate
of
mMT-I
transcription
(1.1
kb
of
hybridizable
se-
quence)
was
69
ppm
for
mice
1-3
and
33
ppm
for
mice
4-6
(not
shown).
Tissue
Culture
Transfection.
Plasmids
containing
the
mMT-I
or
MTrGH
genes
with
or
without
introns
(10
,g)
were
mixed
with
8
jug
of
carrier
herring
sperm
DNA
and
2
jzg
of
thymidine
kinase-neomycin-resistance
gene
as
a
select-
able
gene,
precipitated
with
calcium
phosphate,
and
added
to
semiconfluent
cultures
of
baby
hamster
kidney
(BHK)
cells
on
100-mm
Petri
dishes
in
10
ml
of
Dulbecco's
modified
Eagle's
medium
containing
10%
fetal
calf
serum
(28).
After
6
hr,
the
medium
was
changed,
and
the
cells
were
allowed
to
recover
for
24
hr
before
G418
was
added
(750
,ug/ml)
to
select
for
cells
that
had
stably
incorporated
the
foreign
genes.
After
10
days,
when
all
nontransfected
cells
had
died
and
the
plates
were
nearly
confluent
with
stably
transformed
cells,
the
cells
were
split
onto
two
plates.
One
plate
served
as
a
control,
whereas
100
,uM
ZnSO4
was
added
to
the
other
to
induce
the
mMT-I
gene
promoter.
The
cells
were
har-
vested
8
hr
later,
and
total
nucleic
acid
was
isolated
for
solution
hybridization
analysis
of
mMT-I
and
rGH
mRNA
levels
as
described
above.
Each
transfection
was
performed
in
triplicate,
and
the
results
represent
the
means
±
SEM.
The
amount
of
foreign
DNA
taken
up
by
the
cell
populations
was
estimated
by
quantitative
dot
hybridization
using
nick-
translated
pUC
sequences
as
a
probe.
These
DNA
values
varied
by
<2-fold
compared
with
the
cells
transfected
with
genes
with
or
without
introns.
RESULTS
Effect
of
Introns
on
MTrGH
Expression
in
Fetal
Liver.
The
introns
were
removed
from
a
genomic
clone
of
the
rGH
gene
by
substituting
cDNA
sequences
between
the
first
and
last
exons.
Then
the
entire
gene
and
3'
flanking
sequences
were
fused
to
the
mMT-I
gene
promoter
and
its
5'
flanking
sequences
(Fig.
1).
Fig.
2A
shows
that
9
of
11
transgenic
samples
from
mice
with
the
normal
gene
had
significant
rGH
mRNA
levels,
whereas
only
7
of
15
of
the
intronless
samples
had
detectable
mRNA.
Furthermore,
comparison
of
the
amount
of
rGH
mRNA
in
the
two
sets
of
data
revealed
6-fold
more
mRNA
in
the
intron-containing
samples
than
in
the
intronless
ones.
Thus,
both
the
frequency
of
expression
and
the
average
level
of
mRNA
accumulation
were
depressed
with
the
intronless
construct;
the
product
of
these
two
values
represents
the
overall
efficiency
of
expression,
which
for
this
pair
of
constructs
differed
by
about
10-fold
as
summarized
in
Table
1.
As
a
control
for
possible
differences
in
mRNA
degradation
during
sample
preparation,
we
quan-
titated
the
amount
of
endogenous
mMT-I
mRNA
and
found
no
significant
difference
in
samples
from
the
two
constructs
(data
not
shown).
Furthermore,
there
was
no
obvious
differ-
ence
in
gene
copy
number
in
the
samples
from
these
two
constructs
or
any
of
the
other
pairs
described
below
(data
not
shown).
The
endogenous
mMT-I
gene
is
inducible
by
heavy
metals
(29),
and
various
mMT-I
fusion
genes
in
transgenic
mice
are
inducible
by
zinc
or
cadmium
(23,
30).
Thus,
in
repeating
the
experiment,
we
placed
the
foster
mothers
on
a
zinc
diet
to
induce
both
maternal
and
fetal
mMT-I
gene
expression.
The
data
from
the
zinc-treated
fetal
liver
samples
(Fig.
2B)
are
similar
to
untreated
samples
except
that
the
average
amount
of
rGH
mRNA
is
2-fold
higher.
In
this
experiment,
7
of
9
of
the
transgenic
samples
containing
the
normal
gene
had
high
levels
of
rGH
mRNA,
whereas
only
4
of
20
samples
from
the
intronless
construct
had
rGH
mRNA,
and
the
average
level
of
mRNA
in
samples
where
it
was
detected
was
much
lower.
The
overall
efficiency
of
expression
of
the
intronless
con-
struct
was
only
1%
of
the
normal
gene
(Table
1).
To
determine
whether
the
difference
in
accumulation
of
rGH
mRNA
from
the
two
constructs
was
due
to
a
difference
in
the
rate
of
transcription
or
the
rate
of
mRNA
degradation,
transcription
'"run-on"
experiments
were
performed
with
nuclei
isolated
from
three
fetal
liver
samples
(from
the
experiment
shown
in
Fig.
2B)
of
mice
with
either
the
intron
?E
0
MTrGH
M
mMTI
)x
c
I
ErGH
Developmental
Biology:
Brinster
et
al.

838
Developmental
Biology:
Brinster
et
al.
_,-
50
0
I
0
v
725
E
0
SI
3
0
0
0
O
-0-0
-
2
8
INTRONS
+
-
GENE
MTrGH
100
0
0
C
D
800o
0
400
.0
I1
4.--
o
MtrH
Er-
o
rt_
2
16
l
3
7
l
MTrGHt
ErGH
ErGH
ORGAN
Liver-Zn
Liver
+Zn
AGE
E
dl6
,
E
dl6
Pancreas
Pancreas
E
dl5
6wk
|
F
s15OF
1io4
50_
c
+
MT-I
Liver
E
d16
FIG.
2.
Effect
of
introns
on
transgene
expression.
The
num-
*
ber
of
samples
that
gave
no
de-
tectable
hybridization
signal
(<1
to
10
mRNA
molecules
per
cell,
depending
on
specific
activity
of
0
the
probe
and
the
amount
of
nu-
16
19
cleic
acid
used)
are
indicated
by
+
-
the
numbers
below
each
scatter
h(-globir
diagram.
In
B,
the
foster
mothers
were
reared
on
a
diet
that
in-
Uver
cluded
25
mM
ZnSO4
in
the
E
d16
water.
or
intronless
constructs.
The
relative
rate
of
transcription
paralleled
the
relative
mRNA
abundance
in
all
samples
(Fig.
3).
Of
particular
significance,
the
relative
rates
of
transcrip-
tion
were
low
for
the
three
intronless
samples;
thus,
the
presence
of
introns
has
a
primary
influence
on
transcription,
not
on
mRNA
stability.
Because
the
plasmid
used
to
mea-
sure
rGH
transcripts
was
derived
primarily
from
the
last
three
exons
of
the
gene,
we
cannot
eliminate
the
possibility
that
transcriptional
attenuation
occurs
upstream
of
exon
3.
As
an
internal
control
in
this
experiment,
we
measured
the
relative
rate
of
albumin
transcription;
it
was
marginally
higher
(1.3-fold)
in
the
intron-containing
samples
(Fig.
3)
as
were
the
rates
of
endogenous
mMT-I
gene
transcription
(2-fold)
and
mMT-I
mRNA
accumulation
(1.3-fold).
We
conclude
that
the
relative
rate
of
transcription
(either
initia-
tion
or
elongation
beyond
exon
3)
is
increased
when
trans-
genes
containing
introns
are
introduced
into
mice.
Effect
of
Introns
on
ErGH
Gene
Expression
in
the
Pancreas.
In
the
previous
experiments,
a
promoter
was
used
that
is
expressed
in
many
different
cell
types.
We
also
tested
the
cell-specific
elastase
I
gene
promoter/enhancer
(19)
fused
to
the
rGH
gene
constructs
with
and
without
introns.
The
accumulation
of
rGH
mRNA
in
fetal
(day
14
or
15)
and
adult
(week
6)
pancreases
of
transgenic
mice
made
with
the
two
constructs
are
shown
in
Fig.
2
C
and
D.
The
patterns
of
expression
are
similar
to
the
fetal
liver
data.
When
the
two
sets
of
data
(Fig.
2
C
and
D)
were
combined,
rGH
mRNA
was
detected
in
12
of
20
samples
from
the
intron-containing
construct
but
in
only
3
of
17
samples
from
the
intronless
construct.
Moreover,
the
average
level
of
rGH
mRNA
was
Table
1.
Effect
of
introns
on
gene
expression
in
transgenic
mice
Average
mRNA
Fig.
2
Introns
Frequency
of
concentrationt
Overall
Construct
panel
present
Tissue
(age)*
expression
molecules
per
cell
efficiencyt
MTrGH
A
+
Liver
(E16)
-Zn
9
/
11
365
300
-
Liver
(E16)
-
Zn
7
/
15
63
29
B
+
Liver(E16)
+Zn
7
/9
770
599
-
Liver(E16)
+Zn
4
/20
30
6
ElrGH
C
+
Pancreas
(E15)
5
/
6
715
595
-
Pancreas
(E15)
2
/
5
110
44
D
+
Pancreas
(6
wk)
7
/
14
1365
680
-
Pancreas
(6
wk)
1
/
12
720
60
MT-I
E
+
Liver
(E16)
38
/
44
19.3
16.7
-
Liver
(E16)
10
/
28
2.6
0.94
hPG
F
+
Liver
(E16)
7
/
23
442
134
-
Liver
(E16)
0
/
19
0
0.0
Summary
of
data
from
Fig.
2.
*E16
and
E15,
embryonic
day
16
and
15.
tThe
average
mRNA
concentration
is
based
on
those
mice
that
had
detectable
levels
of
mRNA
(solid
symbols
in
Fig.
2),
whereas
overall
efficiency
is
the
product
of
the
average
mRNA
concentration
and
the
frequency
of
expression.
All
paired
groups
showed
significant
differences
(P
<
0.05)
by
the
Mann-Whitney
U
test
except
for
group
C,
which
was
P
<
0.1.
Proc.
Natl.
Acad
Sci.
USA
85
(1988)
75
re
A
Tof
B
-
400k
200-J
20(0-

Proc.
Natl.
Acad.
Sci.
USA
85
(1988)
839
40
I-
E
va
C
0
0.
U)
C_
p
0
.O-
'4
0-
0
mouse
2
3
gene
MTrGH
2400
looo0
1-
lat
0
1200
Eo
I,
E
Er
C:
4
5
6
MTr
GH
aintrons
FIG.
3.
Effect
of
introns
on
the
rate
of
transcription
of
the
MTrGH
gene
in
transgenic
mice.
Fetal
livers
from
three
transgenic
mice
bearing
the
MTrGH
gene
with
introns
(mice
1-3)
and
three
without
introns
(mice
4-6)
were
used
for
the
determination
of
rGH
mRNA
(open
histograms).
(A)
Relative
rate
of
rGH
gene
transcrip-
tion
and
rGH
mRNA
abundance.
(B)
Relative
rate
of
albumin
gene
transcription.
lower
in
samples
containing
the
intronless
construct.
Thus,
expression
in
two
different
tissues
with
two
different
pro-
moters
reveals
the
same
phenomenon.
No
Effect
of
Introns
on
Expression
of
Constructs
Transfect-
ed
into
Cultured
Cells.
The
MTrGH
constructs
with
and
without
introns
were
also
transfected
into
BHK
cells.
Fig.
4B
shows
that
there
was
no
significant
difference
in
either
the
uninduced
or
induced
levels
of
rGH
mRNA
measured
after
transfection
with
either
construct.
Quantitative
dot
hybridizations
of
DNA
from
the
transfected
cells
indicated
no
significant
difference
in
the
average
number
of
integrated
gene
copies
with
the
two
constructs
(data
not
shown).
Thus,
in
contrast
to
the
results
obtained
in
transgenic
mice,
MTrGH
gene
expression
in
cultured
cells
seems
to
be
insensitive
to
the
presence
of
introns.
Although
it
may
be
significant
that
the
cells
were
selected
for
expression
whereas
the
mice
were
not,
there
is
always
a
significant
effect
of
introns
on
transgene
expression
even
when
one
compares
only
those
mice
that
express
the
gene
(Table
1).
Effect
of
Introns
on
Expression
of
the
mMT-I
Gene
in
Fetal
Liver.
In
the
previous
experiments,
we
tested
two
different
chimeric
genes.
In
these
experiments
we
tested
a
marked
natural
gene.
The
data
(Fig.
2E)
closely
resemble
those
obtained
with
the
chimeric
constructs;
both
the
frequency
of
expression
and
the
accumulation
of
mMT-I
mRNA
were
much
lower
with
the
intronless
construct
compared
with
results
with
the
intact
gene.
Only
10
of
28
mice
with
the
intronless
construct
gave
detectable
mMT-I
mRNA
com-
pared
to
38
of
44
with
the
intron-containing
construct,
and
there
was
less
mRNA
in
the
samples
containing
the
intron-
less
construct
by
a
factor,
on
average,
of
7.5,
as
summarized
in
Table
1.
I,,
500
0
CD
Co
200
E
+
introns
-
introns
FIG.
4.
Effect
of
introns
on
expression
of
mMT-I
and
MTrGH
genes
transfected
into
tissue
culture
cells.
m,
Control;
a,
induced
with
ZnSO4.
These
two
constructs
were
also
tested
by
transfection
into
BHK
cells
(Fig.
4A).
In
the
absence
of
induction,
there
was
less
mMT-I
mRNA
from
the
intronless
construct;
however,
after
induction
there
was
only
a
1.4-fold
difference
in
the
amount
of
mMT-I
mRNA
accumulation.
These
differences
are
small
compared
to
the
differences
observed
in
transgenic
mice.
Thus,
the
results
support
the
previous
conclusion
obtained
with
MTrGH
that
introns
have
little
influence
on
expression
of
a
gene
when
stably
integrated
into
chromo-
somes
of
cultured
cells.
Effect
of
Introns
on
Globin
Gene
Expression.
We
also
tested
the
effect
of
introns
on
the
expression
of
the
human
3-globin
gene
in
transgenic
mice.
Normal
and
intronless
human
,-globin
gene
constructs
(Fig.
1)
were
microinjected
into
fertilized
eggs,
and
day
16
fetal
livers
were
isolated
for
quantitation
of
human
13-globin
mRNA.
Fig.
2F
shows
that
in
the
absence
of
introns
there
was
no
expression
in
any
of
19
transgenic
mice,
whereas
7
of
23
of
the
samples
containing
the
intact
gene
had
measurable
amounts
of
human
P-globin
mRNA.
DISCUSSION
The
general
conclusion
from
these
studies
with
matched
pairs
of
genes
with
or
without
introns
is
that
introns
improve
transcriptional
efficiency
10-
to
100-fold
in
transgenic
mice,
but
they
have
little
effect
on
expression
when
transfected
into
cultured
cells.
Both
the
number
of
transgenic
mice
that
have
detectable
levels
of
expression
and
the
average
level
of
mRNA
expression
are
consistently
depressed
in
the
absence
of
introns
(Table
1).
At
the
outset,
it
seemed
reasonable
to
expect
that
introns
might
have
an
important
effect
on
mRNA
processing.
This
notion
predicts
that
nuclear
mRNA
precursors
are
relatively
unstable
and
that
association
of
the
precursors
with
"splic-
osomes"
would
lead
to
nuclear
stabilization
and/or
more
efficient
transport
into
the
cytoplasm.
This
model
suggests
that
the
average
rate
of
transcription
would
be
comparable
regardless
of
the
presence
of
introns,
but
the
accumulation
of
mRNA
would
differ.
However,
the
nuclear
"run-on"
experiments
presented
in
Fig.
3
indicate
that
mRNA
abun-
dance
is
proportional
to
the
relative
rate
of
transcription.
Furthermore,
one
might
expect
an
effect
of
introns
on
some
basic
aspect
of
RNA
processing
or
transport
to
be
revealed
after
transfection
into
tissue
culture
cells,
contrary
to
what
was
observed.
Although
tissue-specific
factors
may
affect
A
rGH
NM
I-F-
L
I
L.
-
B
albumin~J
0
I
Developmental
Biology:
Brinster
et
al.
r-
A
MT-
I
r1mm"
.l
no
rown=

840
Developmental
Biology:
Brinster
et
al.
splicing,
this
is
not
a
likely
explanation
for
these
results
because
the
rGH
gene
was
not
assayed
in
the
cell
type
in
which
this
gene
is
normally
expressed
either
in
vivo
or
in
culture.
These
observations
lead
us
to
suspect
that
a
major
effect
of
introns,
when
assayed
in
transgenic
mice,
is
on
some
aspect
of
transcription
rather
than
RNA
processing.
Transcriptional
enhancers
can
function
in
a
position-
and
orientation-independent
manner
(31);
they
function
when
placed
within
introns
(32),
and
they
naturally
reside
within
the
introns
of
some
genes
(33, 34).
Therefore,
one
might
imagine
that
each
of
the
genes
we
chose
to
study
has
one
or
more
previously
unidentified
enhancers
in
its
introns.
In-
deed,
our
previous
analysis
of
the
human
P-globin
gene
implicated
the
presence
of
important
enhancer-like
elements
near
the
boundary
of
intron
B/exon
3
as
well
as
3'
of
the
structural
gene
(35).
Similar,
all-or-none
expression
results
were
obtained
by
infection
of
murine
erythroleukemia
cells
with
retroviruses
containing
the
human
p-globin
gene
with
or
without
introns
(M.
A.
Bender,
A.
D.
Miller,
and
R.E.G.,
unpublished
data).
Thus,
the
observation
presented
here,
that
the
intronless
human
f-globin
gene
is
not
expressed
at
all,
is
most
easily
interpreted
as
the
loss
of
an
essential
enhancer
element.
By
extension
of
the
enhancer
argument,
one
might
postu-
late
that
the
rGH
gene
and
the
mMT-I
gene
also
have
important
control
elements
within
their
introns.
Several
observations
suggest
that
this
is
unlikely.
First,
the
5'
flanking
regions
of
both
mMT-I
and
rGH
genes
have
been
shown
to
confer
proper
tissue-specific
expression
and/or
regulation
to
heterologous
reporter
genes
(29,
30,
36-38).
Second,
because
the
rGH
gene
was
assayed
in
two
cell
types
that
do
not
normally
express
it
(hepatocytes
and
pancreatic
acinar
cells)
and
introns
had
a
positive
effect
in
each,
it
seems
unlikely
that
a
tissue-specific
enhancer
can
be
in-
volved.
Third,
if
a
general
enhancer-like
element
were
present
in
the
introns,
then
the
introns
might
be
expected
to
affect
expression
after
transfection
in
cultured
cells.
The
experiments
in
BHK
cells
do
not
support
this
latter
view,
although
this
cell
line,
and
perhaps
all
transformed
cell
lines,
may
have
alternative
mechanisms
of
stimulating
transcrip-
tion
that
are
independent
of
general
enhancers.
Thus,
it
might
be
informative
to
test
the
effect
of
introns
after
introduction
of
genes
into
primary
cell
lines.
Another
more
provocative
possibility
is
that
introns
con-
tain
DNA
sequences
that
are
recognized
at
some
stage
during
development
but
are
not
required
after
transfection
into
established
cell
lines.
If
this
hypothesis
is
correct,
then
one
might
expect
to
find
conserved
recognition
sequences
within
introns.
A
mechanistically
different
possibility
is
that
introns
(or
exons)
contain
sequences
that
are
important
for
phasing
nucleosomes
relative
to
important
promoter
elements
(39,
40).
According
to
this
hypothesis,
the
position
and
orienta-
tion
of
introns
would
be
critical,
a
property
that
distinguishes
this
idea
from
those
invoking
enhancer-like
properties
to
introns.
One
might
explain
why
introns
have
little
effect
after
transfection
into
cultured
cells
by
suggesting
that
critical
phasing
of
nucleosomes
is
established
during
development
or
that
nucleosomes
in
cultured
cells
are
not
phased
as
rigidly
as
they
are
in
vivo
(41).
Clearly,
introns
have
an
important
influence
on
the
func-
tion
of
genes
introduced
into
the
germ
line
of
mice
by
microinjection.
The
results
suggest
that
they
may
help
maintain
transcriptional
activity
during
development.
Aside
from
the
practical
aspect
of
considering
their
presence
during
the
preparation
of
gene
constructs
destined
for
expression
in
transgenic animals,
these
results
provide
new
insight
into
the
role
of
intron
sequences
during
development
and
evolution.
We
thank
Mary
Avarbock,
David
Lo,
and
Eric
Sandgren
for
microinjection
of
some
of
the
gene
constructs;
Felicity
Oram
for
help
in
identifying
the
transgenic
mice;
Gary
Stuart
for
constructing
the
intronless
mMT-I
gene;
Dusty
Miller
for
providing
the
rGH
cDNA
construct;
and
Gary
Merrill,
Roger
Perlmutter,
and
our
colleagues
for
valuable
discussions
during
the
conduct
of
these
experiments
and
preparation
of
the
manuscript.
This
work
was
supported
by
Grants
HD-09172
and
HD-19018
from
the
National
Institutes
of
Health.
1.
Naora,
H.
&
Deacon,
N.
J.
(1982)
Proc.
Natl.
Acad.
Sci.
USA
79,
61%-6200.
2.
Gilbert,
W.
(1978)
Nature
(London)
271,
501.
3.
Sudhof,
T.
C.,
Russell,
D.
W.,
Goldstein,
J.
L.,
Brown,
M.
S.,
Sanchez-
Pescador,
R.
&
Bell,
G.
T.
(1985)
Science
228,
893-895.
4.
Breitbart,
R.
E.,
Nguyen,
H.
T.,
Medford,
R.
M.,
Destree,
A.
T.,
Madhavi,
V.
&
Nadal-Ginard,
B.
(1985)
Cell
41,
67-82.
5.
Leff,
S.
E.,
Rosenfeld,
M.
G.
&
Evans,
R.
M.
(1986)
Annu.
Rev.
Biochem.
55,
1091-1118.
6.
Hamer,
D.
&
Leder,
P.
(1979)
Cell
18,
1299-1302.
7.
Gruss,
P.,
Lai,
C.-J.,
Dhar,
R.
&
Khoury,
G.
(1979)
Proc.
Natl.
Acad.
Sci.
USA
76,
4317-4321.
8.
Gruss,
P.
&
Khoury,
G.
(1980)
Nature
(London)
286,
634-637.
9.
Mulligan,
R.
C.
&
Berg,
P.
(1980)
Science
209,
1422-1427.
10.
Carlock,
L.
&
Jones,
N.
C.
(1981)
Nature
(London)
294,
572-574.
11.
Treisman,
R.,
Novak,
J.,
Favaloro,
J.
&
Kamen,
R.
(1981)
Nature
(London)
292,
595-600.
12.
Chee,
P.
P.,
Klassy,
R.
C.
&
Slightom,
J.
L.
(1986)
Gene
41,
47-57.
13.
Ng,
R.,
Domdey,
H.,
Larson,
G.,
Rossi,
J.
J.
&
Abelson,
J.
(1985)
Nature
(London)
314,
183-184.
14.
Gross,
M.
K.,
Kainz,
M.
S.
&
Merrill,
G.
F.
(1987)
Mol.
Cell
Biol.
7,
4576-4581.
15.
Brinster,
R.
L.
&
Palmiter,
R.
D.
(1986)
Harvey
Lect.
80,
1-38.
16.
Palmiter,
R.
D.
&
Brinster,
R.
L.
(1986)
Annu.
Rev.
Gen.
20,
465-499.
17.
Durnam,
D.
M.,
Perrin,
F.,
Gannon,
F.
&
Palmiter,
R.
D.
(1980)
Proc.
Natl.
Acad.
Sci.
USA
77,
6511-6515.
18.
Barta,
A.,
Richards,
R.
I.,
Baxter,
J.
D.
&
Shine,
J.
(1981)
Proc.
Nati.
Acad.
Sci.
USA
78,
4867-4871.
19.
Ornitz,
D.
M.,
Palmiter,
R.
D.,
Hammer,
R.
E.,
Brinster,
R.
L.,
Swift,
G.
H.
&
MacDonald,
R.
J.
(1985)
Nature
(London)
313,
600-603.
20.
Glanville,
N.,
Durnam,
D.
M.
&
Palmiter,
R.
D.
(1981)
Nature
(London)
292,
267-2%.
21.
Collins,
F.
S.
&
Weissman,
S.
M.
(1984)
Prog.
Nucleic
Acid
Res.
Mol.
Biol.
31,
315-437.
22.
Brinster,
R.
L.,
Chen,
H.
Y.,
Trumbauer,
M.
E.,
Yagle,
M.
K.
&
Pal-
miter,
R.
D.
(1985)
Proc.
Natl.
Acad.
Sci.
USA
82,
4438-4442.
23.
Palmiter,
R.
D.,
Chen,
H.
Y.
&
Brinster,
R.
L.
(1982)
Cell
29,
701-710.
24.
Townes,
T.
M.,
Lingrel,
J.
B.,
Chen,
H.
Y.,
Brinster,
R.
L.
&
Palmiter,
R.
D.
(1985)
EMBO
J.
4,
1715-1723.
25.
Durnam,
D.
M.
&
Palmiter,
R.
D.
(1983)
Anal.
Biochem.
131,
385-393.
26.
Mulvihill,
E.
R.
&
Palmiter,
R.
D.
(1977)
J.
Biol.
Chem.
252,
2060-2068.
27.
McKnight,
G.
S.
&
Palmiter,
R.
D.
(1979)
J.
Biol.
Chem.
254,
9050-9058.
28.
Searle,
P.
F.,
Stuart,
G.
W.
&
Palmiter,
R.
D.
(1985)
Mol.
Cell.
Biol.
5,
1480-1489.
29.
Hamer,
D.
H.
(1986)
Annu.
Rev.
Biochem.
55,
913-952.
30.
Palmiter,
R.
D.,
Norstedt,
G.,
Gelinas,
R.
E.,
Hammer,
R.
E.
&
Brin-
ster,
R.
L.
(1983)
Science
222,
809-814.
31.
Serfling,
E.,
Jasin,
M.
&
Schaffner,
W.
(1985)
Trends
Genet.
1,
224-230.
32.
Hammer,
R.
E.,
Swift,
G.
H.,
Ornitz,
D.
M.,
Quaife,
C.
J.,
Palmiter,
R.
D.,
Brinster,
R.
L.
&
MacDonald,
R.
J.
(1987)
Mol.
Cell.
Biol.
7,
2956-2%7.
33.
Banerji,
J.,
Olson,
L.
&
Schaffner,
W.
(1983)
Cell
33,
729-740.
34.
Gillies,
S.
D.,
Morrison,
S.
L.,
Oi,
V.
T.
&
Tonegawa,
S.
(1983)
Cell
33,
717-728.
35.
Behringer,
R.
R.,
Hammer,
R.
E.,
Brinster,
R.
L.,
Palmiter,
R.
D.
&
Townes,
T.
M.
(1987)
Proc.
Natl.
Acad.
Sci.
USA
84,
7056-7060.
36.
Stuart,
G.
W.,
Searle,
P.
F.,
Chen,
H.
Y.,
Brinster,
R.
L.
&
Palmiter,
R.
D.
(1984)
Proc.
Natl.
Acad.
Sci.
USA
81,
7318-7322.
37.
Nelson,
C.,
Crenshaw,
E.
B.,
III,
Franco,
R.,
Lira,
S.
A.,
Albert,
V.
R.,
Evans,
R.
M.
&
Rosenfeld,
M.
G.
(1986)
Nature
(London)
322,
557-562.
38.
Lefevre,
C.,
Imagawa,
M.,
Dana,
S.,
Grindlay,
J.,
Bodner,
M.
&
Karin,
M.
(1987)
EMBO
J.
6,
971-981.
39.
Eissenberg,
J.
C.,
Cartwright,
I.
L.,
Thomas,
G.
H.
&
Elgin,
S.
C.
(1985)
Annu.
Rev.
Genet.
19,
485-536.
40.
Richard-Foy,
H.
&
Hager,
G.
L.
(1987)
EMBO
J.
6,
2321-2328.
41.
Chambon,
P.
(1977)
Cold
Spring
Harbor
Symp.
Quant.
Biol.
42,
1209-1234.
Proc.
Natl.
Acad.
Sci.
USA
85
(1988)