U1 snRNP-ASF/SF2 interaction and 5′ splice site recognition: Characterization of required elements

Abstract

Members of the SR family of proteins, can collaborate with U1 snRNP In the recognition of 5′ splice sites in pre-messenger
RNAs. We have previously shown that purified U1 snRNP and ASF/SF2 form a ternary complex with pre-mRNA, which is dependent
on a functional 5′ splice site. In this manuscript we dissect the requirements for the formation of this complex. Sequences
in the pre-mRNA, domains in ASF/SF2 and components of the U1 snRNP particle are shown to be required for complex formation.
We had shown that sequences at the 5′ splice site of PIP7.A are necessary and now we show these are sufficient for complex
formation. Furthermore, we show that one functional RNA binding domain and the RS domain are both required for ASF/SF2 to
participate in complex formation. The RNA binding domains were redundant in this assay, suggesting that either domain can
Interact with the pre-messenger RNA. Finally, our experiments show no function for the U1-specific A protein in complex formation,
whereas a function for U1-specific C protein was strongly suggested. The study of the earliest interactions between pre-mRNA
and splicing factors suggests a model for 5' splice site recognition.

Full text

3260-3267
Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
Ul
snRNP-ASF/SF2
interaction
and
5'
splice
site
recognition:
characterization
of
required
elements
Sharon
F.
Jamison1,
Zvi
Pasmanl'2,
Jin
Wang4,
Cindy
Will5,
Reinhard
Luhrmann5,
James
L.
Manley4
and
Mariano
A.
Garcia-Blanco12,3,*
Departments
of
1Molecular
Cancer
Biology,
2Microbiology
and
3Medicine,
Levine
Science
Research
Center,
Duke
University
Medical
Center,
Durham,
NC
27710,
USA,
4Department
of
Biological
Sciences,
Columbia
University,
New
York,
NY
10027,
USA
and
51nstitut
fur
Molekularbiologie
und
Tumorforschung,
Emil-Mannkopff-Strasse
2,
D-35037
Marburg,
Germany
Received
April
21,
1995;
Revised
and
Accepted
July
4,
1995
ABSTRACT
Members
of
the
SR
family
of
proteins,
can
collaborate
with
Ul
snRNP
in
the
recognition
of
5'
splice
sites
in
pre-messenger
RNAs.
We
have
previously
shown
that
purified
Ul
snRNP
and
ASF/SF2
form
a
ternary
complex
with
pre-mRNA,
which
is
dependent
on
a
functional
5'
splice
site.
In
this
manuscript
we
dissect
the
requirements
for
the
formation
of
this
complex.
Sequences
in
the
pre-mRNA,
domains
in
ASF/SF2
and
components
of
the
Ul
snRNP
particle
are
shown
to
be
required
for
complex
formation.
We
had
shown
that
sequences
at
the
5'
splice
site
of
PIP7.A
are
necessary
and
now
we
show
these
are
sufficient
for
complex
formation.
Furthermore,
we
show
that
one
functional
RNA
binding
domain
and
the
RS
domain
are
both
required
for
ASF/SF2
to
participate
in
complex
forma-
tion.
The
RNA
binding
domains
were
redundant
in
this
assay,
suggesting
that
either
domain
can
interact
with
the
pre-messenger
RNA.
Finally,
our
experiments
show
no
function
for
the
Ul-specific
A
protein
in
complex
formation,
whereas
a
function
for
Ul
-specific
C
protein
was
strongly
suggested.
The
study
of
the
earliest
interactions
between
pre-mRNA
and
splicing
factors
suggests
a
model
for
5'
splice
site
recognition.
INTRODUCTION
Genes
in
vertebrates
are
complex
units,
usually
composed
of
multiple
and
sometimes
alternatively
utilized
exons
and
introns
(1).
Thus,
primary
transcripts
or
pre-messenger
RNAs
(pre-
mRNAs),
contain
many
5'
splice
site
and
3'
splice
site
pairs,
which
must
be
properly
recognized
and
processed.
Each
splice
site
pair
encompasses
an
intron
that
will
be
removed
by
spliceosomes,
macromolecular
enzymes
that
undergo
step-wise
assembly
with
pre-mRNAs
(2-6).
Early
recognition
of
splice
sites
is
conserved
from
yeast
to
vertebrates:
formation
of
the
commitment
complex,
the
first
stable
complex
formed
between
pre-mRNA
and
splicing
factors,
usually
requires
both
5'
splice
site
and
3'
splice
site
sequences
(6,7).
In
many
cases
pre-mRNAs
with
mutations
in
either
the
5'
or
the
3'
splice
site
show
defects
in
formation
of
the
commitment
complex
or
the
pre-spliceosome
(6,8,9).
Several
required
splicing
factors
are
known
to
interact
with
pre-mRNA
in
early
complexes:
members
of
the
SR
protein
family
(10),
U2AF
(11)
and
Ul
and
U2
snRNPs
(12,13).
These
factors
provide
a
platform
for
assembly
of
the
full
spliceosome.
In
cases
of
alternative
splicing,
regulatory
factors
are
found
interacting
with
the
constitutive
splicing
machinery
in
these
early
complexes
(14,15).
Alternative
splicing
decisions
are
established
early
in
the
case
of
the
Drosophila
developmental
regulators
Sxl
(16),
tra
and
tra-2
(17)
and
in
the
case
of
adenovirus
early
to
late
switch
(18).
Protein
factors
of
the
SR
family
have
been
shown
to
be
required
for
early
steps
of
spliceosome
assembly.
These
proteins
share
N-terminal
RNA
recognition
motifs
(RRM)
and
homologous
C-terminal
domains,
rich
in
serines
and
arginines,
RS
domains
(19).
SR
proteins
of
20,
30a/b,
40,55
and
75
kDa
can
be
purified
to
apparent
homogeneity
by
two
sequential
salt
precipitations.
SRp30a,
ASF/SF2,
was
shown
to
be
required
for
constitutive
splicing
and
to
be
involved
in
alternative
selection
of
proximal
5'
splice
sites
(20-22).
A
second
member
of
the
SR
family,
SRp30b
or
SC35,
was
identified
first
as
a
spliceosome
component
and
later
was
shown
to
be
a
required
splicing
factor
(23).
S100
extracts
are
deficient
in
SR
proteins
and
because
of
this
cannot
splice
pre-mRNAs.
Addition
of
a
mixture
of
purified
SR
proteins
or
addition
of
any
one
of
six
SR
proteins
SRp2O,
SRp3Oa,
SRp3Ob,
SRp4O,
SRpS5
or
SRp75
can
reconstitute
pre-mRNA
splicing
in
S100
extracts,
suggesting
that
SR
proteins
have
overlapping
activities
(24,25).
There
is,
however,
evidence
that
supports
the
idea
of
preferred
functions
for
each
SR
protein
(25-27).
Several
lines
of
evidence argue
that
the
SR
proteins
may
be
the
first
components
to
bind
the
pre-mRNA.
First,
pre-mRNAs
that
are
pre-incubated
with
SC35,
ASF/SF2
or
SRpS5
can
be
spliced
in
the
presence
of
competitor
RNA
(26).
Secondly,
ASF/SF2,
Ul
snRNP
and
pre-mRNA
form
stable
and
specific
complexes,
but
only
when
ASF/SF2
is
added
first
to
the
pre-mRNA
(28).
Thirdly,
SR
proteins
potentiate
the
formation
of
E
complex,
which
may
be
*To
whom
correspondence
should
be
addressed
at:
Department
of
Molecular
Cancer
Biology,
Box
3686,
DUMC,
Durham,
NC
27710,
USA
Q--D/
1995
Oxford
University
Press

Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
3261
equivalent
to
the
commitment
complex
(15).
Moreover,
SR
proteins
collaborate
with
Ul
snRNP
in
5'
splice
site
recognition:
purified
ASF/SF2
and
Ul
snRNP
form
complexes
on
PIP7.A
pre-mRNA,
but
not
on
a
5'
splice
site
mutant
pre-mRNA
(28).
This
protein
has
been
postulated
to
enhance
the
affinity
of
U
1
snRNP
for
5'
splice
sites
(29).
This
effect
is
likely
to
be
mediated
directly
by
an
interaction
between
ASF/SF2
and
70
kDa
protein
of
U
l
snRNP,
which
requires
the
arginine-serine
rich
domains
in
both
ASF/SF2
and
U1-70
kDa
protein
(28,30).
In
keeping
with
its
very
early
role
in
5'
splice
site
recognition,
ASF/SF2
has
altemative
splicing
activity
both
in
vitro
(20,22)
and
in
vivo
(31).
Ul
snRNP
is
the
earliest
snRNP
to
assemble
on
pre-mRNAs
(5,32)
defining
the
5'
splice
site
via
RNA-RNA
and
RNA-pro-
tein
interactions
(33,34).
The
Ul
snRNP
is
also
the
only
spliceosomal
snRNP
required
for
commitment
complex
forma-
tion
(6,7).
The
subsequent
arrival
of
U2
snRNP
and
the
interactions
between
U2
and
Ul
snRNPs
may
be
mediated
by
U2AF
and
SC35
(11,35).
Direct
and
indirect
interactions
among
these
factors
has
been
suggested
by
two-hybrid
experiments
(30).
For
several
pre-mRNAs,
the
requirement
for
U
1
snRNP
can
be
relieved
by
very
high
levels
of
SR
proteins
(36,37).
This
implies
that
the
interactions
of
Ul
snRNP
with
U2
snRNP
and
other
factors
arriving
later
may
be
mediated
by
SR
proteins.
Cleavage
at
a
5'
splice
site
during
the
first
transesterification
reaction
can
be
determined
by
subsequent
interactions
between
this
site
and
US
or
U6
snRNPs
(38-42).
The
Ul
independent
reactions
have
lost
fidelity
for
the
authentic
5'
splice
site,
however,
and
show
a
preponderance
of
cryptic
site
utilization
(36,37).
Therefore,
it
is
possible
to
dissect
the
early
roles
of
the
SR
proteins
and
U
1
snRNP,
the
former
nucleate
the
assembly
of
the
splicing
machinery
and
the
latter
provides
specific
selection
of
authentic
5'
splice
site
sequences.
Previously
we
have
shown
that
highly
purified
Ul
snRNP
and
ASF/SF2
cooperated
to
form
complexes
with
pre-mRNAs
(28).
These
complexes
could
distinguish
between
a
functional
5'
splice
site
and
an
inactive
site
where
the
GU
dinucleotide
was
mutated
to
AU.
Moreover,
we
showed
that
this
interaction
required
the
arginine-serine
domains
of
ASF/SF2
and
probably
also
an
arginine-serine
rich
domain
of
the
70
kDa
protein
of
Ul
snRNP.
Here
we
characterize
these
complexes
further:
we
show
that
5'
splice
site
sequences
are
necessary
and
sufficient
for
complex
formation.
We
also
show
that
deletions
in
the
RNA
binding
domains
of
ASF/SF2
abolish
complex
formation.
In
addition
our
data
demonstrate
that
the
C
protein
of
U
I
snRNP
is
required
for
this
interaction.
The
data
presented
here
and
previous
findings
suggest
a
model
for
5'
splice
site
recognition
and
spliceosome
assembly
mediated
by
cooperation
between
SR
proteins
and
U1
snRNP.
the
sequence
AGGACAGAGC
or
the
sequence
AGUACUAU-
CU
respectively
replacing
the
pPIP7.A
sequence:
AGGA-
CAAACU,
which
is
located
in
the
3'
exon
bases
197-207
in
the
RNA.
These
changes
were
introduced
by
replacing the
XhoI-PstI
fragment
with
mutated
oligodeoxyribonucleotides.
The
sequences
of
all
plasmid
inserts
and
deletions
were
verified
by
dideoxy-
sequencing
using
Sequenase
2.0
(US
Biochemical).
All
RNAs
were
synthesized
using
T7
RNA
polymerase
(Stratagene)
as
described
previously
(44,45)
except
that
RNAs
were
labeled
with
[a-32P]UTP
at
a
final
specific
activity
of
300
Ci/mmol
in
the
labeling
reaction.
PIP7.A
BamHI,
PIP7.A
Sall,
PIP7.A
PstI
and
PIP7.A
A5'E-BamHI
RNAs
were
synthesized
by
linearizing
pPIP7.A
or
pPIP7.A
A5'E
with
the
indicated
restriction
endonucleases.
Otherwise
all
RNAs
were
synthesized
off
HindIll
linearized
templates.
Competitor
RNAs
were
labeled
in
a
reaction
mixture
containing
-3.75
mCi
of
[a-32P]UTP
per
mmol.
ASF/SF2 and
Ul
snRNP
Recombinant
ASF/SF2
and
mutant
derivatives
were
described
previously
(28,46).
To
construct
the
double
mutant
with
deletions
in
both
RBDs,
pDS-ASF-NR-g
=
35-63,
pDS-ASF-NR-d
=
172-198
(46)
were
digested
with
BstBI
and
ApaI
and
religated
following
treatment
with
Klenow.
Proteins
were
expressed
and
purified
as
described
(46).
Purification
of
U
1
snRNPs
was
accomplished
by
using
anti-m3G
affinity
chromatography
as
described
previously
(47).
In
order
to
produce
Ul
AC,
Ul
AA
or
core
Ul
snRNP
particles
partially
purified
U
1
snRNPs
were
fractionated
on
Mono-Q
columns
at
various
temperatures
as
described
by
Bach
et
al.
(47).
Complex
reactions
and
native
gel
electrophoresis
ASF/SF2
and
Ul
snRNP
complex
formation
assays
were
described
previously
(28).
Briefly,
ASF/SF2
protein
(240
ng;
7.3
pmol),
Ul
snRNP
(800
ng;
2.7
pmol),
pre-mRNA
(16
pg;
2.4
fmol),
62
mM
KCl
and
1.6
mM
MgCl2
were
incubated
for
5
min
at
30°C.
Heparin
was
added
to
a
final
concentration
of
0.5
mgml-l
before
loading
onto
native
gels.
Samples
were
loaded
on
non-denaturing
4%
acrylamide:bisacrylamide
(80:1)
gels
con-
taining
50
mM
Tris-50
mM
glycine
(3).
Electrophoresis
was
at
14
Vcm71
for
2.5
h.
Complexes
were
visualized
by
autoradio-
graphy
with
hyperfilm
(Amersham).
MATERIALS
AND
METHODS
Antisera
Plasmids
and
RNAs
pPIP7.A
was
described
previously
(43).
pPIP7.A
A5E
was
constructed
by
removing
the
EcoRl-KpnI
fragment
of
pPIP7.A,
making
ends
blunt
with
T4
DNA
polymerase
in
the
presence
of
deoxynucleotide
triphosphates
and
ligating,
pPIP7.A
Del
I
was
constructed
by
removing
the
SacI-XhoI
fragment
of
pPIP7.A.
In
pPIP7.A
APy,
a
mutant
XhoI-PstI
fragment
derived
from
oligonucleotides,
replaced
the
PIP7.A
XhoI-PstI
fragment
and
changed
the
polypyrimidine
tract
from
TTTCCC1T-llTITIT
to
ATACACATATATAT.
pPIP7.A(AG)n
and
pPIP7.A(AAG)n
have
The
normal
human
serum
(Jm)
and
the
systemic
lupus
erythema-
tosus
patient
serum
(AW),
a
kind
gift
of
Dr
J.
Keene,
were
both
described
in
Jamison
et
al.
(6).
Reactions
containing
ASF/SF2,
Ul
snRNP
and
radiolabeled
PIP7.A
and
supplemented
with
Heparin,
were
incubated
with
1:160
(v/v)
dilution
of
antisera
in
Buffer
D
(48)
in
the
presence
of
600-fold
molar
excess
of
competitor
PIP7.A
RNA,
5
mg/ml
BSA
and
1%
(w/v)
NP40
for
5
min
at
30°C.
These
were
then
loaded
onto
native
gels
as
described
above.

3262
Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
A
PIP7.A
PIP7.Apmst
m
PIP7.Asal
I
PIP7.ABanil
IC}
ks-EZnZZ
+
AaEZ
+
B
tilsnRNI
-+
-+1-
+-+
ASP
I
-
H-t+-H-I+I-
ITN
A
i
Bamli
Sal
P
st
I
P1
P7
Del
I
-
sri
_
PIP7.Aneii
P1
P7.A
A5'E
PIP7.AA5'E-
Bandil/
P
P7I.A(AG)n
G;
___
A
-Z
Z
+
_N..
_ _
+
MW~~~~~~~~~~~~~~3.E
2
4
8
9
1
PIP7.A(AAG)n
I.
A_
MAmEZ-E
+
PIP7.AAPpy
Ir-
A
Mr-EZZ
+
C
JI
F
InRNP
I
IT+
-
I
ASF-I
+tLi
+TLIt
IIRiNA
PIP
7
5JA
E
D
L
++
0-
0-
04
CL
C-
Cl
Z7Z7
I'
=
=
{
r
:
C-
010c
~-X
Lt
-
-
orl
E
lsnRNP
-
+
ASP-i
-
+
RNA
AK-~~~~~~~~,P)
xv
2K
10
Q
I
PY,J~
-
1'TI
-
ori
so
4
If.\
:'-;~~~~~~~~~~
~
~~~~~~~~~~~~~~~~~~~~~~~~~~~
....\.\
RN;A-~
fUR
I23
-
6
1
2
3
4
5.
6
-
RNA
23
1
8
1
2
3
4
5
Figure
1.
Sequences
at
the
5'
splice
site
are
necessary
and
sufficient
for
U
l
snRNP
and
ASF/SF2
binding.
(A)
A
schematic
showing
the
pre-mRNAs
used
to
analyze
the
cis-element
requirement
for
the
ASF/SF2
and
U
l
snRNP
dependent
gel
shift.
PIP7.A
RNA
has
been
described
previously
by
Gil
et
al.
(43).
The
deletion
mutants
are
indicated
clearly
in
the
diagram.
PIP7.A(AG),,
substituted
the
sequence
5'-AGGACAAACU-3'
in
the
second
exon
of
PIP7.A
for
5'-AGGACAGAGC-3'.
This
sequence
was
among
those
selected
by
ASF/SF2
in
in
vitro
selection
experiments.
In
PIP7.A(AA4G),
this
sequence
was
mutated
in
order
to
make
it
pyrimidine-rich:
5'-AGUACUAUCU-3'.
The
polypyrimidine
tract
in
the
intron
of
PIP7.A
was
substituted
by
a
tract
containing
an
adenine
at
every
other
position
in
PIP7APy.
(B)
Deletion
mutants
of
PIP7.A
RNA,
which
are
described
in
(A),
were
assayed
for
complex
formation
in
the
presence
of
ASF/SF2
and
purified
Ul
snRNP
(even
number
lanes).
The
RNAs
were
loaded
in
odd
number
lanes
in
this
non-denaturing
gel.
ASF-
l
is
wild
type
recombinant
ASF/SF2.
(C)
A
deletion
in
the
5'
exon
(PIP7.A
AS'E)
of
PIP7.A
was
tested
for
complex
formation
in
the
presence
of
ASF/SF2
and/or
U1
snRNP
(lanes
4-6).
PIP7.A
is
shown
as
a
control
(lanes
1-3).
(D)
A
short
RNA
spanning
the
5'
splice
site
consensus
sequence
(PIP7.A
AS'E-BamHI)
of
PIP7.A
was
tested
for
complex
formation
in
the
presence
of
ASF/SF2
(ASF-
1)
(lane
1
)
or
RS,
an
ASF
deleted
of
the
RS
domain
(lane
2),
U1
snRNP
(lane
3).
U1
snRNP
and
ASF-l
(lane
4)
or
U
I
snRNP
and
RS
(lane
5).
(E)
Mutant
PIP7.A
RNAs
with
changes
in
either
the
second
exon,
PIP7(AG),,
and
PIP7(AAG%,
or
in
the
intronic
polypyrimidine
tract,
PIP7APy,
were
assayed
as
in
B
and
C
(lanes
6-8);
RNAs
are
shown
in
lanes
1-4.
Ori
denotes
the
origin
of
electrophoresis
and
complexes
are
indicated
by
arrowheads.
--
II
.6-
Aga
-I

Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
3263
RESULTS
A
The
5'
splice
site
is
necessary
and
sufficient
for
complex
formation
We
first
tested
the
effect
of
deletions
in
PIP7.A
pre-mRNA
in
order
to
identify
the
minimum
sequence
required
for
formation
of
the
ASF/SF2
and
Ul
dependent
complexes
(28)
(Fig.
IA).
Efficient
formation
of
the
complexes
was
seen
using
the
full
length
PIP7.A
pre-mRNA
(Fig.
IB,
lane
8)
and
with
all
deletions
that
retained
the
5'
splice
site,
PIP7.ABamHI,
PIP7.ASalI,
PIP7.APstI
(Fig.
lB,
lanes
2,
4
and
6).
A
hemi-intron
RNA
missing
the
5'
splice
site,
PIP7.ADelI,
however,
could
no
longer
form
complexes
(Fig.
IB,
lane
10).
A
deletion
that
removed
most
of
the
5'
exon
sequence
but
left
the
5'
splice
site
consensus
sequence
intact,
PIP7.AAS'E,
also
formed
complexes
efficiently
(Fig.
IC,
lane
6).
Furthermore,
a
26
nucleotide-long
RNA
spanning
the
5'
splice
site
consensus
sequence,
PIP7.AA5'E-BamHI,
was
competent
to
form
complexes
(Fig.
lD).
Our
previously
published
report
indicates
that
the
5'
splice
site
is
necessary
for
ternary
complex
formation.
The
data
above
showed
that
sequences
at
or
around
the
5'
splice
site
are
necessary
and
sufficient
for
complex
formation.
We
also
tested
the
possibility
that
inserting
a
purine-rich
sequence
in
the
downstream
exon
might
improve
complex
formation.
Purine-rich
elements
have been
shown
to
bind
SR
proteins
and
the
sequence
5'-AGGACAGAGC-3'
which
we
introduced
in
the
second
exon
of
PIP7.A(AG)n,
corresponds
to
one
of
two
high
affinity
consensus
sequences
derived
from
a
combina-
torial
selection
experiment
with
ASF/SF2
(R.
Tacke
and
J.
Manley,
submitted).
As
a
control
for
the
purine-rich
sequence,
we
constructed
a
mutant
sequence:
5'-AGUACUAUCU-3'
in
the
second
exon
of
RNA
PIP7.A(AAG)n.
When
tested
in
our
assay,
both
PIP7.A(AG)n
and
PIP7.A(AAG)n
formed
complexes
with
the
same
efficiency
as
PIP7.A
(Fig.
IE,
lanes
5-7).
It
is
also
noteworthy
that
these
two
RNAs
were
spliced
equally
well
(data
not
shown),
which
is
consistent
with
the
idea
that
purine-rich
elements
display
their
function
primarily
on
weak
introns.
A
mutation
in
PIP7.A
that
changed
the
polypyrimidine
tract
by
inserting
adenines
at
every
other
position,
PIP7.AAPy,
was
also
able
to
form
complexes
(Fig.
lE,
lane
8).
The
very
slight
advantage
seen
for
PIP7.AAPy
over
PIP7.A
seen
in
this
experiment
was
not
reproducible.
These
data
indicated
that
in
the
presence
of
a
strong
5'
splice
site,
other
elements
are
not
necessary
for
recognition
by
ASF/SF2
and
Ul
snRNP.
H
ASF-1
16
RBD1
98
Gly106
RBD2
RS
1
/
\/\
/
\
/
\
198
248
RNP-2
RNP-1
RNP-2
RNP-1
18-23
54-61
123-128
158-172
0
S--
**
I_:~
\/
/ /
M
EE
-
-------
_
..
RS
1-35
35-63
108-1
37
148-172
35-63,
172-198
B
1'1--'--'
en
|
|
+
-
on
A
functional
RNA
binding
domain
and
an
arginine-serine
domain
are
both
required
for
recognition
of
the
5'
splice
site
ASF/SF2
has
three
well
characterized
primary
sequence
motifs.
Two
RNA
recognition
motifs
are
found
in
the
N-terminal
200
amino
acids
(Fig.
2A).
These
motifs
have
been
empirically
shown
to
bind
RNA
in
isolation
and
have
been
thus
named
RNA
binding
domains
1
and
2
(RBD1
and
RBD2)
(46,49).
The
C-terminal
50
amino
acids
constitute
the
arginine-serine
(RS)
domain,
which
we
had
previously
shown
is
required
for
5'
splice
site
recognition
(28).
We
decided
to
test
mutants
in
the
RNA
binding
domains
of
ASF/SF2
to
determine
their
requirements
in
5'
splice
site
recognition
in
collaboration
with
Ul
snRNP.
Two
mutants
A35-63
and
A1-35,
which
disrupt
the
RNP-l
and
RNP-2
of
RBDl
respectively,
did
not
diminish
the
capacity
of
the
ASF/SF2
pre-mRNA_
u:W
1
2
3
4
5 6
7
8
Figure
2.
One
functional
RBD
in
ASF/SF2
is
sufficient
for
5'
splice
site
recognition.
(A)
Schematic
showing
ASF/SF2
deletibn
mutants
that
were
used,
similar
levels
of
protein
were
confirmed
by
limiting
dilution
series
and
silver
staining
of
purified
recombinant
proteins.
(B)
Full
length
recombinant
and
His
tagged
ASF/SF2,
ASF-1,
(1-248)
or
mutants
of
ASF/SF2
were
assayed
for
complex
formation
with
Ul
snRNP
and
PIP7.A
pre-mRNA.
Deletions
in
RBD1
were
35-63
and
1-35,
deletions
in
RBD2
were
148-172
and
108-137
and
a
deletion
in
both
RBDs
was
35-63,
172-198.
Complexes
were
detected
by
native
gel
electrophoresis
as
described
in
the
Materials
and
Methods.
Ori
denotes
the
origin
of
electrophoresis
and
complexes
are
indicated
by
arrowheads.
EE::..
\/

3264
Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
to
interact
with
Ul
snRNP
and
the
5'
splice
site
(Fig.
2B,
lanes
4
and
5).
The
same
was
observed
when
either
the
RNP-
1
or
RNP-2
sequences
of
RBD-2
were
disrupted
in
mutants
A
148-172
and
A108-137
respectively
(Fig.
2B,
lanes
6
and
7).
When
both
RBD-
1
and
RBD-2
were
disrupted,
however,
no
complex
formation
was
observed
(Fig.
2B,
lane
8).
Thus,
we
conclude
that
only
one
of
the
two
RBDs
in
ASF/SF2
is
required
for
5'
splice
site
recognition.
A
PRF
f
-
+
-
-
POST]
Ul
_
+,
-
5'
io'
20'
30'
45'
60'
Iactors
sr
ASF-
I
U1l
snRNP
+
ASF-I
o
SnRNP
The
Ul
snRNP
stabilizes
the
interaction
of
ASF/SF2
with
pre-mRNA
We
have
shown
previously
that
5'
splice
site
recognition
by
purified
ASF/SF2
and
Ul
snRNP
was
absolutely
dependent
on
the
order
of
addition
(28).
These
experiments
indicated
that
ASF/SF2
must
interact
with
the
pre-mRNA
first,
although
this
interaction
could
not
be
detected
by
gel
shift
assays.
Presumably
ASF/SF2
promotes
subsequent
Ul
snRNP
binding.
We
show
here
that
the
complexes
seen
on
native
gels
were
remarkably
stable
(Fig.
3A).
Incubations
with
either
Ul
snRNP
or
ASF/SF2
alone
did
not
result
in
efficient
complex
formation,
which
is
consistent
with
our
previous
data
(Fig.
3A,
lanes
1-4).
Preincubation
with
1500-fold
molar
excess
of
unlabeled
competitor
PIP7.A
inhibited
formation
of
the
ASF/SF2
and
Ul
snRNP
dependent
complexes
(lane
5).
If
these
complexes
were
allowed
to
form,
however,
they
were
recalcitrant
to
challenge
with
the
same
amount
of
competitor
for
up
to
1
h
of
incubation
(lanes
6-11).
Therefore
the
temary
complexes
are
remarkably
stable.
It
was
possible
that
stable
complexes
formed
between
ASF/SF2
and
pre-mRNA
that
were
not
resistant
to
electrophoresis
and
thus
invisible
in
our
gel
assay.
To
test
this
possibility
we
performed
the
experiment
described
in
Figure
3B.
Addition
of
competitor
before
incubation
of
PIP7.A
with
ASF/SF2
resulted
in
no
complexes,
as
expected
(Fig.
3C,
lane
6).
Incubation
with
ASF/SF2
followed
sequentially
by
mock-addition
of
competitor
and
incubation
with
Ul
snRNP
resulted
in
formation
of
complexes
(lane
7).
Addition
of
competitor,
however,
for
5
or
15
min,
after
incubation
with
PIP7.A
and
ASF/SF2
prevented
complex
formation
upon
subse-
quent
incubation
with
Ul
snRNP
(Fig.
3C,
lanes
8
and
9).
It
is
unlikely,
therefore,
that
the
interaction
between
ASF/SF2
and
PIP7.A
was
stable
prior
to
the
addition
of
U
I
snRNP.
Complexes
did
not
form
in
the
absence
of
ASF/SF2
when
PIP7.A
and
Ul
snRNP
were
incubated
together
whether
or
not
competitor
was
added
(lanes
2-5).
As
a
positive
control,
PIP7.A
was
incubated
with
both
ASF/SF2
and
U1
snRNP
prior
to
addition
of
competitor,
demonstrating
the
formation
of
stable
complexes
(lanes
11-13).
Together,
these
results
suggest
an
early
but
unstable
interaction
between
ASF/SF2
and
PIP7.A
pre-mRNA
that
is
essential
for
subsequent
complex
formation.
We
characterized
the
binding
of
Ul
snRNP
and
ASF/SF2
further
by
nuclease
footprinting.
RNase
TI
and
RNase
A
gave
reproducible
cleavage
patterns
on
5'
end
labeled
PIP7.A
RNA
(data
not shown).
Addition
of
increasing
concentrations
of
ASF/SF2
resulted
in
equivalent
protection
from
RNase
TI
cleavage
at
all
guanines
(G).
No
preference
for
protection
of
G55
and
G56
at
the
5'
splice
site
was
observed,
whereas
U1
snRNP
specifically
protected
the
RNase
TI
cleavage
sites
at
the
5'
splice
site,
G55
and
G56
(data
not
shown).
preniRNA-
I
2
4
5
6
7
0
9
1
101
B
i
F'
I9:
*
..'
.m
3...C
.,
"
m
electrophoresis
5mr.
30:C
t.
30
C
5m.1
30
C
5
ni.
30'C
;
o_o'];s
~~~~~~~~~electrophoresis
5ml
30'C
t
30
C
5r3i.
30CC
5
nm.
30
C
-
.
_______
:
__
electrophoresis
2n.R;
m4'C
Sm
30
C
t
30CC
m5
r1,
30
C
C
Precompeftion
-
-
+
+
-
Factos)
before
comp.
-
,,,
U
I
snRNP
f7
ASF-I
.
Ul
snRNP+ASF-I
Timeof
Postcompefition
_
-
5I
-
515'
I-_
-515'
-1
5'
151
Factos)
after
comp.
-
AS
F-I
Z
U
1
snRNP
-
on-
AD
pre-mRNA-'-
A
I
2
3
4
5
6
7
8
9 10
11
12
13
Figure
3.
The
ASF/SF2
and
Ul
snRNP
dependent
complexes
are
stable.
(A)
Complex
formation
was
assayed
by
native
gel
electrophoresis
as
described
above.
In
this
experiment,
a
1500-fold
molar
excess
of
unlabeled
PIP7.A
pre-mRNA
was
added
either
before
or
after
the
complexes
were
formed
'PRE'
and
'POST'
respectively.
PIP7.A
RNA
was
incubated
with
either
ASF-
I
or
U
I
snRNP
(lanes
1-4)
or
both
together
(lanes
5-11).
Unlabeled
PIP7.A
was
added
before
complex
formation
(lane
5)
or
after
complex
formation
(lanes
6-1
1).
(B)
A
schematic
that
explains
the
experiment
shown
in
panel
C.
The
time
of
addition
of
excess
unlabeled
PIP7.A
RNA
(competitor
PIP7.A
relative
to
the
addition
of
ASF/SF2
and
Ul
snRNP
is
indicated.
't'
indicates
where
times
of
incubation
were
varied
in
the
experiment
(see
C).
(C)
Complex
formation
was
assayed
as
described
above,
competitor
RNA
was
added
before
ASF/SF2
or
U
I
snRNP
in
lanes
2,
6
and
10.
ASF/SF2
or
Ul
snRNP
or
both
were
added
to
labeled
PIP7.A
RNA
before
the
competitor
in
lanes
3-5,
7-9
or
11-13
respectively.
A
minus
sign
for
time
of
post-competition
indicates
that
no
competitor
RNA
was
added
to
those
reactions.
Ori
denotes
the
origin
of
electrophoresis
and
complexes
are
indicated
by
arrowheads.
-4
-4

Serum
-Ism
Jim
|Sm|Jm
ASFI+Ul
I
+
RNrA
_
At$
-
ori
z
_
z
-
0
D
D
c)
+
+ +
S)s
"
's
..
..
pre-mRNA-_
_
1
2
3
4
5
6
Figure
4.
Ul
snRNP
is
found
in
the
complexes
after
electrophoresis.
ASF/SF2
and
UI
snRNP
were
incubated
with
labeled
PIP7.A
RNA
as
described
above.
After
complex
formation,
the
reactions
were
incubated
with
Buffer
D
(lane
4)
or
Sm
(lane
5)
or
a
Jm
(lane
6)
(see
Materials
and
Methods).
The
free
RNAs
were
also
incubated
with
the
antisera
as
a
negative
control
(lanes
1-3).
These
reactions
were
then
subjected
to
non-denaturing
gel
electrophoresis
as
described
above.
Ori denotes
the
origin
of
electrophoresis
and
complexes
are
indicated
by
arrowheads.
Role
of
Ut
snRNP
proteins
in
recognition
of
the
5'
splice
site
We
had
previously
shown
that
in
order
to
form
complexes
both
ASF/SF2
and
Ul
snRNP
were
required
(28).
It
was
not
clear,
however,
if
U1
snRNP
was
in
the
complexes
in
the
native
gels.
To
date
it
has
been
difficult
to
identify
a
Ul
snRNP
pre-mRNA
complex
in
native
gels
when
using
mammalian
nuclear
extracts.
In
contrast
yeast
CCl
and
CC2
complexes
are
Ul
snRNP-con-
taining
complexes
observed
in
native
gels
(50).
To
determine
whether
or
not
Ul
snRNP
was
in
the
U1-ASF/SF2
dependent
complexes,
we
tested
the
effect
of
a
highly
specific
anti-U
snRNP
antiserum
on
their
mobility.
A
mobility
shift
of
the
complexes
was
observed
with
this
human
antiserum
(Sm)
directed
at
U
snRNP
proteins
(Fig.
4,
lane
5),
whereas
a
normal
human
serum
(Jm)
did
not
shift
these
complexes
(Fig.
4,
lane
6).
Further
evidence
that
Ul
snRNP
was
in
the
complex
was
obtained
by
observing
complex
mobility
changes
with
variant
Ul
snRNP
particles
(see
below
and
Fig.
5).
A
Ul
snRNP
particle
missing
the
Ul
specific
A
protein
(Ul
AA)
was
capable
of
collaborating
with
ASF/SF2
to
bind
PIP7.A
(Fig.
5,
lane
7).
The
resulting
complex
had
a
faster
mobility
than
the
one
formed
with
Ul
snRNP
in
the
native
gels.
The
complex
formed
by
Ul
AA
required
a
functional
5'
splice
site
because
it
was
not
observed
with
PIP7.A
5'
AU
where
the
first
base of
the
intron
has
been
changed
to
an
A
(data
not
shown).
Moreover,
this
complex
did
not
form
with
ASF
ARS,
a
truncated
ASF/SF2
missing
the
RS
domain
(data
not
shown).
A
Ul
snRNP
particle
missing
the
Ul
specific
C
protein
U1
AC
did
not
form
complexes
with
mobilities
similar
to
the
Ul
AA
and
Ul
snRNP
complexes
(Fig.
5,
lane
8).
A
small
shift,
however,
was
reproducibly
observed
when
PIP7.A
was
incubated
with
ASF/
_
* *
Free
RNA
1
2
3
4
5
67
8
9
Figure
5.
Ul
snRNP
specific
C
protein
is
necessary
for
complex
formation.
Complex
formation
was
assayed
as
above
except
that
incubations
contained
incomplete
Ul
snRNP
particles:
Ul
snRNP
missing
the
UI-specific
A
protein
(Ul
AA),
U1-specific
C
protein
(U/AC)
or
all
Ul-specific
proteins
(core).
Incubations
with
either
ASF/SF2
(lane
1)
or
U1
snRNP
particles
(lanes
2-5)
as
well
as
incubations
with
both
(lanes
6-9)
are
shown.
Ori
denotes
the
origin
of
electrophoresis
and
complexes
are
indicated
by
arrowheads.
SF2
and
U1
AC.
This
band
was
not
observed
with
a
pre-mRNA
mutated
in
the
5'
splice
site
or
when
the
ASF
ARS
was
used,
suggesting
that
it
required
an
interaction
between
ASF/SF2
and
Ul
snRNP
(data
not
shown).
The
shifted
band,
however,
has
a
mobility
that
is
not
consistent
with
the
presence
of
Ul
snRNP
(data
not
shown).
A
Ul
snRNP
preparation
known
as
core
Ul,
which
is
missing
the
A,
C
and
70
kDa
proteins,
resulted
in
low
levels
of
complexes,
mostly
with
the
same
mobility
as
those
seen
with
Ul
AC
(Fig.
5,
lane
9).
Results
with
the
core
Ul
preparation
were
more
variable
leading
us
to
believe
it
was
a
mixed
population
of
Ul
snRNP
particles
(51).
These
results
and
our
previously
published
data
are
consistent
with
the
idea
that
the
U1-C
and
70
kDa
proteins,
but
not
the
U1-A
protein,
are
required
for
complex
formation.
DISCUSSION
We
have
shown
previously
that
purified
SR
protein
ASF/SF2
collaborates
with
purified
Ul
snRNP
to
recognize
a
functional
5'
splice
site
(28).
Here
we
show
that
a
complex
formed
by
ASF/SF2,
Ul
snRNP
and
pre-mRNA
requires
only
sequences
at
and
around
the
5'
splice
site.
Only
one
functional
RNA
binding
domain
and
the
RS
domain
in
ASF/SF2
are
required
for
5'
splice
site
recognition
in
cooperation
with
Ul
snRNP.
Furthermore,
we
show
that
Ul
snRNP
C
protein
is
required
for
proper
5'
splice
site
recognition,
whereas
Ul
A
protein
is
dispensable.
Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
3265
0
--No

3266
Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
Recognition
and
selection
of
5'
splice
sites
Recognition
and
selection
of
5'
splice
site
sequences
and
eventual
selection
of
authentic
sites
is
a
two
phase
process.
The
early
phase
is
determined
by
SR
proteins
and
Ul
snRNP
and
the
late
phase
is
probably
determined
by
the
U5
*
U4/U6
tri-snRNP.
Watson-
Crick
base
pairing
between
the
5'
end
of
Ul
snRNA
and
the
sequences
at
the
5'
splice
site
is
required
for
recognition
of
the
splice
site
(7,33).
The
binding
of
Ul
snRNP
to
the
5'
splice
sites
is
well
documented
(32,51)
and
a
requirement
for
U
1
snRNP
at
the
level
of
splicing
commitment
has
been
shown
for
both
Saccharomyces
cerevisiae
and
human
cells
(6,50).
Therefore
there
is
little
doubt
that
the
interaction
of
Ul
snRNP
with
sequences
at
the
5'
splice
site
is
normally
involved
in
very
early
recognition
of
these
sites.
There
is,
however,
a
growing
body
of
evidence
that
U
1
snRNP
is
neither
absolutely
required
(36,37)
nor
is
it
sufficient
(52,53)
for
5'
splice
site
recognition.
The
former
conclusion
comes
from
studies
in
which
excess
SR
proteins
could
relieve
blocks
in
extracts
where
Ul
snRNP
was
biochemically
depleted
(36)
or
debilitated
with
an
antisense
2'-O-methyl
oligoribonucleotide
directed
to
the
5'
end
of
the
Ul
snRNA
(37).
The
SR
proteins
had
already
been
shown
to
be
required
splicing
factors
that
acted
at
the
earliest
measurable
steps
in
the
splicing
pathway
(26)
and
could
collaborate
with
Ul
snRNP
to
specifically
recognize
a
functional
5'
splice
site
(28).
None
of
these
reports,
however,
predicted
that
SR
proteins
could
replace
the
splicing
requirement
for
Ul
snRNP
and
therefore
all
previous
models
of
SR
function
must
be
modified
to
explain
the
data
of
Crispino
et
al.
(36)
and
Tarn
and
Steitz
(37).
Our
data
both
in
this
report
and
in
Kohtz
et
al.
(28)
help
establish
such
a
model
of
SR
function.
Tarn
and
Steitz
have
offered
an
alternative
(37)
but
similar
explanation
based
on
the
ability
of
SR
proteins
to
directly
recognize
5'
splice
sites
(54).
Our
model
of
SR
protein
action
does
not
deny
a
possible
preference
for
5'
splice
site
sequences
but
it
does
not
require
it.
What
is
the
mechanism
of
action
of
SR
proteins?
We
envision
that
SR
proteins
bind
the
pre-mRNA
first
and
promote
a
stabilization
of
U
I
snRNP
binding
at
5'
splice
sites.
It
is
clear
that
purified
Ul
snRNP
can
bind
pre-mRNA
in
the
absence
of
SR
proteins
as
documented
by
RNase
TI
footprinting
(data
not
shown).
This
U1-5'
splice
site
interaction
does
not
withstand
native
gel
electrophoresis,
however,
whereas
a
U1-5'
splice
site
interaction
in
the
presence
of
ASF/SF2
results
in
the
formation
of
stable
complexes.
In
agreement
with
this,
addition
of
ASF/SF2
to
nuclear
extracts
increases
the
apparent
affinity
of
U1
snRNP
to
all
5'
splice
site
sequences
in
a
pre-mRNA
(29).
Our
data
also
suggest
that
Ul
snRNP
changes
the
interaction
between
ASF/SF2
and
pre-mRNA.
Our
data
argue
that
Ul
snRNP
and
SR
proteins
mutually
enhance
the
affinity
of
the
other
for
5'
splice
sites.
We
can
define
then
a
role
for
U1
snRNP
as
a
5'
splice
site
dependent
stabilizer
of
SR
protein-pre-mRNA
interactions.
This
is
a
role
that
could
be
replaced
by
addition
of
a
large
excess
of
SR
proteins.
This
role
must
be
one
acted
out
early
by
Ul
snRNP
given
that
an
excess of
SR
proteins
can
replace
the
requirement
for
U
1
snRNP
in
splicing
commitment
(26).
Following
initial
pre-mRNA
binding
and
5'
splice
site
defini-
tion
in
collaboration
with
Ul
snRNP,
SR
proteins
promote
an
interaction
across
the
intron
(14)
with
the
splicing
factor
U2AF
(11).
This
heterodimeric
factor
has
two
subunits,
one
of
which,
a
65
kDa
polypyrimidine
tract
binding
subunit
with
an
N-terminal
arginie-serine
domain,
is
absolutely
required
for
splicing
(55).
U2AF,
probably
in
combination
with
SR
proteins
(23),
promotes
the
binding
of
U2
snRNP
to
branch
point
sequences
(1
1).
The
U6
and
U5
snRNAs
can
recognize
the
5'
splice
site
a
second
time
and
situate
the
spliceosome
so
as
to
cleave
the
phosphodiester
bond
at
the
exon-intron
junction
(38-42,56-59).
This
would
represent
a
second
and
possible
independent
recognition
of
5'
splice
site
sequences.
One
function
of
U1
snRNP,
which
apparently
cannot
be
duplicated
by
SR
proteins
is
the
selection
of
the
authentic
5'
splice
site
among
several
good
candidate
sites
(37).
RNA
binding
domains
and
RNA
binding
by
SR
proteins
The
SR
proteins
have
partially
redundant
activity
in
the
complementation
of
splicing
deficient
S100
extracts
(24,25).
In
regulation
of
alternative
splicing
or
preference
for
proximal
5'
splice
sites
at
least
two
SR
proteins,
ASF/SF2
and
SC
35,
can
also
have
similar
activity
(35).
This
apparent
conservation
of
function
is
paralleled
by
conservation
of
structure.
Zahler
etal.
(25)
divide
the
SR
family
into
two
subgroups,
those
containing
one
RBD,
SRp2O
and
SRp3Ob
(SC35)
and
those
with
one
classic
RBD
and
a
distant
homologue
of an
RBD:
SR30a(ASF/SF2),
SRp4O,
SRpS5
and
SRp75
(24).
Both
RBD1
and
RBD2
in
isolation
can
bind
a
95
nt
SV40
pre-mRNA
fragment
(46)
and
can
be
cross-linked
to
a
human
f3-globin
pre-mRNA
(49),
albeit
with
low
affinity
relative
to
proteins
containing
both
domains.
Mutations
in
either
RBD
can
diminish
the
constitutive
splicing
activity
of
ASF/SF2,
however,
the
mutational
analysis
has
not
clearly
established
separable
functions
for
the
domains.
Our
data
demonstrate
that
disruption
of
one
or
another
of
the
RBDs
does
not
diminish
the
ability
of
ASF/SF2
to
recruit
Ul
snRNP
to
the
5'
splice
site.
On
the
other
hand
disruption
of
both
completely
prevents
this
activity.
The
RBD
redundancy
may
be
consistent
with
evolution
of
the
SR
family,
in
which
some
members
only
have
one
RBD
while
others
have
two
(24).
These
results,
coupled
with
previous
experiments
(43;
Tacke
and
Manley,
submitted),
suggest
that
low
affinity,
non-specific
interactions
with
the
RNA
are
sufficient
for
complex
formation
with
Ul
snRNP.
It
is
important
to
note,
however,
that
all
of
the
RBD
mutants
tested
here
are
completely
inactive
as
essential
splicing
factors
(46),
indicating
that
the
ability
to
form
ternary
complexes
is
not
sufficient
for
splicing
activity.
Given
that
SR
proteins
interact
with
several
different
components
of
the
spliceosome
it
is
not
surprising
that
the
splicing
assay
is
more
sensitive
than
our
assay
to
single
RBD
mutations.
Our
previous
work
(28)
and
this
report
establish
that
the
RS
domain
is
absolutely
required
for
Ul
snRNP
recruitment
to
5'
splice
sites.
This
central
activity
of
RS
domain
is
conserved
throughout
the
family
and
may
reflect
an
ability
to
interact
with
the
U
1
70
kDa
protein
by
most
SR
members
(28,30).
ASF/SF2
has
been
shown
to
bind
SV40
and
adenovirus
pre-mRNAs
in
a
5'
splice
site
dependent
fashion
(54),
suggesting
an
ability
for
the
protein
to
recognize
the
site.
Recently,
in
vitro
selection
experiments
have
revealed
binding
preferences
for
ASF/SF2
and
SC35
to
distinct
purine-rich
sequences
(Tacke
and
Manley,
submitted).
The
sequences
recognized
by
ASF/SF2
are
very
similar
to
those
described
in
certain
exon
splicing
enhancer
elements
(14,60,61).
In
footprinting
experiments,
we
found
no
evidence
of
preferential
affinity
for
sequences
at
the
5'
splice
site
of
PIP7.A
RNA.
It
is
possible
that
purine-rich
sites
or
5'
splice
sites
nucleate
the
initial
binding,
which
can
then
be
observed
over
most
of
the
pre-mRNA.
The
5'
splice
site
consensus
in
mammals
AG:GURAGU
is
purine-rich
and
thus
may
favor
the
binding
of

Nucleic
Acids
Research,
1995,
Vol.
23,
No.
16
3267
SR
proteins,
as
originally
suggested
by
Zuo
and
Manley
(54).
Our
view
of
SR
protein
function,
however,
does
not
require
indepen-
dent
binding
to
5'
splice
sites
or
to
purine-rich
sequences.
As
long
as
these
proteins
bind
the
pre-mRNA
and
recruit
U6/4
and
U5
snRNPs
to
the
vicinity
of
an
intron,
these
latter
factors
could
'search'
the
pre-mRNA
to
select
an
active
5'
splice
site.
In
the
presence
of
U1
snRNP,
this
process
is
more
selective;
the
functional
concentration
of
SR
proteins
at
the
authentic
5'
splice
site
is
high
and
subsequent
reactions
are
thus
driven
to
the
correct
site.
An
attractive
model,
consistent
with
all
the
data,
is
that
for
introns
with
weak
splicing
signals,
high
affinity
SR-protein
binding
sites,
and
hence
a
high
concentration
of
SR
proteins,
are
required
for
recruitment
of
snRNPs
and
activation
of
splicing.
In
contrast,
when
splicing
signals
are
strong,
U1
may
require
less
'assistance'
from
SR
proteins
and
the
non-specific
binding,
such
as
we
observed
with
the
PIP7.A
pre-mRNA,
may
be
sufficient
to
allow
stable
complex
formation.
Indeed,
in
this
case
the
snRNP,
in
addition
to
helping
define
the
5'
splice
site,
may
function
to
stabilize
binding
of
SR
proteins,
which
are
then
required
for
subsequent
steps
in
spliceosome
assembly.
ACKNOWLEDGEMENTS
We
thank
J.
Keene
(Duke
University)
for
much
encouragement
and
for
providing
us
with
Sm
antisera.
We
are
grateful
to
Sabina
Sager
for
assistance
in
the
preparation
of
this
manuscript.
This
work
was
funded
by
research
grants
from
NIH
(M.A.G.-B.
and
J.L.M).
M.A.G.-B.
is
funded
by
a
JFRA
from
the
ACS
and
would
also
like
to
acknowledge
the
support
of
the
Keck
Foundation
to
the
Levine
Science
Research
Center.
REFERENCES
I
Sharp,
P.A.
(1994)
Cell,
77,
805-815.
2
Moore,
J.M.,
Query,
C.C.
and
Sharp,
P.A.
(1993)
In
R.F.
Gesteland
and
J.F.
Atkins
(eds),
The
RNA
World.
Cold
Spring
Harbor
Laboratory
Press,
Cold
Spring
Harbor,
NY,
pp.
303-357.
3
Konarska,
M.M.
and
Sharp,
P.A.
(1986)
Cell,
46,
845-855.
4
Pikielny,
C.W.,
Rymond,
B.C.
and
Rosbash,
M.
(1986)
Nature,
324,
341-345.
5
Ruby,
S.W. and
Abelson,
J.
(1988)
Science,
242,
1028-1035.
6
Jamison,
S.F.,
Crow,
A.
and
Garcia-Blanco,
M.A.
(1992)
Mol.
Cell.
Biol.,
12,4279-4287.
7
Seraphin,
B.
and
Rosbash,
M.
(1989)
Gene,
82,
145-151.
8
Lamond,
A.I.,
Konarska,
M.M.
and
Sharp,
P.A.
(1987)
Genes
Dev.,
1,
532-543.
9
Roscigno,
R.F.,
Weiner,
M.
and
Garcia-Blanco,
M.A.
(1993)
J.
Biol.
Chem.,
268,
11
222-11
229.
10
Zahler,
A.M.,
Lane,
W.S.,
Stolk, J.A.
and
Roth,
M.B.
(1992)
Genes
Dev.,
6,
837-847.
11
Zamore,
P.D.
and
Green,
M.R.
(1989)
Proc
Natl
Acad
Sci
USA,
86,
9243-9247.
12
Black,
D.L.,
Chabot,
B.
and
Steitz,
J.A.
(1985)
Cell,
42,
737-750.
13
Krainer,
A.R.
and
Maniatis,
T.
(1985)
Cell,
42,
725-736.
14
Tian,
M.
and
Maniatis,
T.
(1993)
Cell,
74,
105-114.
15
Staknis,
D.
and
Reed,
R.
(1994)
MoL
Cell.
Biol.,
14,
2994-3005.
16
Valcarcel,
J.,
Singh,
R.,
Zamore,
P.D.
and
Green,
M.R.
(1993)
Nature,
362,
171-175.
17
Tlan,
M.
and
Maniatis,
T.
(1992)
Science,
256,
237-240.
18
Kreivi,
J.P.
and
Akusjarvi,
G.
(1994)
Nucleic
Acids
Res.,
22,
332-337.
19
Kenan,
D.J.,
Query,
C.C.
and
Keene,
J.D.
(1991) Trends
Biochem
Sci.,
16,
214-220.
20
Ge,
H.
and
Manley,
J.L.
(1990)
Cell,
62,
25-34.
21
Krainer,
A.R.,
Conway,
G.C.
and
Kozak,
D.
(1990)
Genes
Dev.,
4,
1158-1171.
22
Krainer,
A.R.,
Conway,
G.C.
and
Kozak,
D.
(1990I
Cell,
62,
35-42.
23
Fu,
X.D.
and
Maniatis,
T.
(1992)
Science,
256,
535-538.
24
Zahler,
A.M.,
Neugebauer,
K.M.,
Stolk,
J.A.
and
Roth,
M.B.
(1993)
Mol.
Cell.
Biol.,
13,4023-4028.
25
Zahler,
A.M.,
Neugebauer,
K.M.,
Lane,
W.S.
and
Roth,
M.B.
(1993b)
Science,
260,
219-222.
26
Fu,
X.D.
(1993)
Nature,
365,
82-85.
27
Kim,
Y.J.,
Zuo,
P.,
Manley,
J.L.
and
Baker,
B.S.
(1992)
Genes
Dev.,
6,
2569-2579.
28
Kohtz,
J.D.,
Jamison,
S.F.,
Will,
C.L.,
Zuo,
P.,
Luhrmann,
R.,
Garcia-Blanco,
M.A.
and
Manley,
J.L.
(1994)
Nature,
368,
119-124.
29
Eperon,
I.C.,
Ireland,
D.C.,
Smith,
R.A.,
Mayeda,
A.
and
Krainer,
A.R.
(1993)
EMBO
J.,
12,
3607-3617.
30
Wu,
J.Y.
and
Maniatis,
T.
(1993)
Cell,
75,
1061-1070.
31
Caceres,
J.F.,
Stamm,
S.,
Helfman,
D.M.
and
Krainer,
A.R.
(1994)
Science,
265,
1706-1709.
32
Mount,
S.M.,
Pettersson,
I.,
Hinterberger,
M.,
Karmas,
A.
and
Steitz,
J.A.
(1983)
Cell,
33,
509-518.
33
Zhuang,
Y.
and
Weiner,
A.M.
(1986)
Cell,
46,
827-835.
34
Zhuang,
Y.,
Leung,
H.
and
Weiner,
A.M.
(1987)
Mo.
Cell.
Biol.,
7,
3018-3020.
35
Fu,
X.D.,
Mayeda,
A.,
Maniatis,
T.
and
Krainer,
A.R.
(1992)
Proc.
Natl
Acad.
Sci.
USA,
89,
11
224-11
228.
36
Crispino,
J.D.,
Blencowe,
B.J.
and
Sharp,
P.A.
(1994)
Science,
265,
1866-1869.
37
Tarn,
W.Y.
and
Steitz,
J.A.
(1994)
Genes
De.v,
8,
2704-2717.
38
Newman,
A.J.
and
Norman,
C.
(1991)
Cell,
65,
115-124.
39
Newman,
A.J.
and
Norman,
C.
(1992)
Cell,
68,
743-754.
40
Lesser,
C.F.
and
Guthrie,
C.
(1993)
Science,
262,
1982-1988.
41
Kandels-Lewis,
S.
and
Seraphin,
B.
(1993)
Science,
262,
2035-2039.
42
Konforti,
B.B.
and
Konarska,
M.M.
(1994)
Genes
Dev.,
8,
1962-1973.
43
Gil,
A.,
Sharp,
P.A.,
Jamison,
S.F.
and
Garcia-Blanco,
M.A.
(1991)
Genes
Dev.,
5,
1224-1236.
44
Garcia-Blanco,
M.A.,
Jamison,
S.F.
and
Sharp,
P.A.
(1989)
Genes
Dev.,
3,
1874-1886.
45
Melton,
D.A.,
Krieg,
PA.,
Rebagliati,
M.R.,
Maniatis,
T.,
Zinn,
K.
and
Green,
M.R.
(1984)
Nucleic
Acids
Res.,
12,
7035-7056.
46
Zuo,
P.
and
Manley,
J.L.
(1993)
EMBO
J,
12,
4727-4737.
47
Bach,
M.,
Bringmann,
P.
and
Luhrmann,
R.
(1990)
Methods
Enzymol.,
181,
232-257.
48
Dignam,
J.D.,
Lebovitz,
R.M.
and
Roeder,
R.G.
(1983)
Nucleic
Acids
Res.,
11,
1475-1489.
49
Caceres,
J.F.
and
Krainer,
A.R.
(1993)
EMBO
J.,
12,
4715-4726.
50
Seraphin,
B.
and
Rosbash,
M.
(1989b)
Cell,
59,
349-358.
51
Heinrichs,
V.,
Bach,
M.,
Winkelmann,
G.
and
Luhrmann,
R.
(1990)
Science,
247,
69-72.
52
Siliciano,
PG.
and
Guthrie,
C.
(1988)
Genes
Dev.,
2,
1258-1267.
53
Seraphin,
B.
and
Rosbash,
M.
(1990)
Cell,
63,
619-629.
54
Zuo,
P.
and
Manley,
J.L.
(1994)
Proc.
Natl
Acad.
Sci.
USA,
91,
3363-3367.
55
Zamore,
P.D.,
Patton,
J.G.
and
Green,
M.R.
(1992)
Nature,
355,
609-614.
56
Cortes,
J.J.,
Sontheimer,
E.J.,
Seiwert,
S.D.
and
Steitz,
J.A.
(1993)
EMBO
J.,
12,
5181-5189.
57
Wyatt,
J.R.,
Sontheimer,
E.J.
and
Steitz,
J.A.
(1992)
Genes
Dev.,
6,
2542-2553.
58
Sawa,
H.
and
Shimura,
Y.
(1992)
Genes
Dev.,
6,
244-254.
59
Sontheimer,
E.J.
and
Steitz,
J.A.
(1993)
Science,
262,
1989-1996.
60
Watakabe,
A.,
Tanaka,
K.
and
Shimura,
Y.
(1993)
Genes
Dev.,
7,
407-418.
61
Inoue,
K.,
Hoshijima,
K.,
Higuchi,
I.,
Sakamoto,
H.
and
Shimura,
Y.
(1992)
Proc.
Natl
Acad.
Sci.
USA,
89,
8092-8096.