Content uploaded by Béatrice Clouet-d'Orval
Author content
All content in this area was uploaded by Béatrice Clouet-d'Orval on Jul 20, 2018
Content may be subject to copyright.
JMB—MS 718 Cust. Ref. No. YAN 013/95
[SGML]
J. Mol. Biol.
(1995) 251, 591–602
b-Tropomyosin Pre-mRNA Folding Around a
Muscle-specific Exon Interferes with Several Steps of
Spliceosome Assembly
Pascal Sirand-Pugnet, Patrice Durosay, Be´atrice Clouet d’Orval
Edward Brody and Joe¨lle Marie*
Centre de Ge´ne´tique The chicken b-tropomyosin pre-mRNA is spliced in a tissue-specific
manner. Internal exons 6B and 6A are specifically used in skeletal muscleMole´culaire, Centre National and non-skeletal muscle cells, respectively. Pre-mRNA secondary structurede la Recherche Scientifique around exon 6B has been shown to be part of the mechanism that inhibitsassociated with the Universite´ exon 6B to 7 splicing in HeLa nuclear extract. We analyse the influence ofde Paris VI, Gif-sur-Yvette pre-mRNA folding on the different steps of spliceosome assembly under91190 France different conditions. At 3 mM MgCl2, conditions that favour RNA structure
formation, the interactions of U1, U2, U4, U5 and U6 small nuclear
ribonucleoprotein particles (snRNPs) with the pre-mRNA are all affected.
The study of several mutants destabilising some proposed stem-loop
structures shows that the in vitro splicing activation is correlated with an
increased binding of snRNPs on pre-mRNA molecules. At 1 mM MgCl2,
conditions that allow a partial relaxation of the inhibitory structure, U1
snRNP binding on exon 6B 5' splice site occurs very efficiently. Nonetheless,
if this first step of spliceosome assembly is derepressed, U2, U4, U5 and U6
snRNP interaction processes remain inhibited. Altogether, these results
suggest that the choice between exon 6A and 6B donor sites is a complex
process not simply directed by a difference in the efficiency of interaction
between U1 snRNP and alternative 5' splice sites.
71995 Academic Press Limited
Keywords: spliceosome assembly; snRNP; pre-mRNA structure;
*Corresponding author b-tropomyosin; alternative splicing
Introduction
Pre-mRNA splicing is a general post-transcrip-
tional process that requires both cis and trans-acting
elementsthatinteractina complexdynamicpathway
(reviewed by Green, 1991; Moore et al., 1993). The
spliceosome (Brody & Abelson, 1985; Grabowski
et al., 1985; Moore et al., 1993) contains five small
nuclear ribonucleoprotein particles (snRNP: U1, U2,
U4,U5 andU6) anda numberof non-snRNP factors,
of which only a few have been characterised. This
complex assembles on pre-mRNA, at least partly in
response to the presence of conserved sequences
located at the exon/intron junctions and the branch
site. Spliceosome assembly occurs in an ordered
pathway, involving first 5' splice site recognition by
U1, then U2 stable binding at the branch point to
forma complexknownas the pre-complexA. U4,U5
and U6 snRNPs interact with this pre-complex as a
tri-particle to form the spliceosome B (for a review,
see Moore et al., 1993). Biochemical and genetic data
indicate that the interactions of trans-acting factors
with pre-mRNA or with each other evolve during
assembly and transesterification reactions. In par-
ticular, several interactions between snRNPs are
remodelled during the splicing process: before 5'
splice site cleavage, 5' splice site U1 binding is
weakened and U4/U6 snRNA base-pairing is
unwound, presumably to allow an interaction
between the 5' region of U6 snRNA and the 3' end
of U2 snRNA (Weiner, 1993; Wise, 1993). Exper-
iments performed in both mammalian and yeast
systems during the last ten years have revealed
spliceosome formation as a way to bring 5' and 3'
splice sites close to each other. This model had been
hypothesised as early as 1980 (Lerner et al., 1980;
Roger & Wall, 1980) and is now supported by a
growingnumber ofresults:first, U1 snRNPhasbeen
Present address: E. Brody, SUNY BUFFALO, Dept. of
Biological Sciences, Cooke Hall, Buffalo, NY 14260,
USA.
Abbreviations used: snRNP, small nuclear
ribonucleoprotein; hnRNP, heterogeneous nuclear
ribonucleoprotein; Ig, immunoglobulin.
0022–2836/95/350591–12 $12.00/0 71995 Academic Press Limited
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
592
shown to interact with the 3' splice site. This
interaction does not require its base-pairing with the
5' consensus sequence (Barabino et al., 1990; Tatei
et al., 1987; Zillmann et al., 1987). More recently, in
yeast, the 3' splice site selection has been shown to
be influenced by deletions in a particular domain of
the U1 snRNA (Goguel et al., 1991). Finally, a
base-pairing has been found between 5' U1 snRNA
and the 3' intron dinucleotide AG, in Saccharomyces
pombe (Reich et al., 1992). Second, other results
indicatethatU5andU6snRNPinteractwithboththe
3' and the 5' splice sites (Kandels-Lewis & Seraphin,
1993; Lesser & Guthrie, 1993; Sontheimer & Steitz,
1993). In addition to these results, U1 and U2 may
interact in a direct and/or indirect way during
complex formation (Daugeron et al., 1992). In
mammalian systems, SC35 and ASF/SF2 have been
proposed to mediate this interaction (Fu et al., 1992;
Wu & Maniatis, 1993). Finally, communication
between the 5' and the 3' splice site has been
documented as a very early event occurring in the
mammalian commitment complex (E complex) that
formsintheabsenceof ATP(Michaud&Reed, 1993).
These studies tend to elucidate the way a splicing
reaction occurs; they also bring into light the
potential complexity of its regulation in alternative
splicing. The use of alternative splice sites results in
the production of different mature mRNAs from a
single type of precursor. The different splicing
patterns are usually a function of cell type or
developmental stage (Maniatis, 1991; Smith et al.,
1989). Only in a few cases have specific trans-acting
factors involved in alternative splicing regulation
been characterised (reviewed by Moore et al., 1993).
Splice site choice may not always require the
presence or absence of specific factors but can be
mediated by differences in the activities or amounts
of general splicing factors: in vitro and in vivo
experiments have shown that the ratio between
SF2/ASF and hnRNP A1 directly influences the
choicebetweendifferent5'splice sites (Caceresetal.,
1994;Ge &Manley, 1990;Krainer et al., 1990; Mayeda
& Krainer, 1992).
If our knowledge about trans-acting factors is still
rather poor, we know more concerning cis elements
involved in the regulation of alternative splicing.
Several studies mention the existence of such
elements, located in introns or in exons, outside the
consensus sequences. Their interaction with specific
or constitutive factors may modulate the relative
strength of alternative splice sites. In several cases,
pre-mRNA secondary structure has been shown to
affect splice site selection. In the adenovirus E3
transcription unit, the selection of a 5' splice site and
the non-utilisation of a cryptic one is dependent on
an exonic secondary structure in vitro (Kister et al.,
1993). In the E1A transcription unit, the formation of
an intronic hairpin structure allows the in vitro
splicing reaction to occur, presumably by decreasing
the physical distance between the 3' splice site and
an unusually far upstream branch point (Chebli
et al., 1989). Such a mechanism has been reported in
the Kluyveromyces lactis actin gene (Deshler & Rossi,
Figure 1. A diagram of the chicken b-tropomyosin gene
region showing the two mutually exclusive exons 6A and
6B. Open boxes represent exons. Exon 6B is specific for
skeletal muscle and exon 6A is used in other cell types.
Horizontallinesrepresenttheintronsequence.Thelengths
of the different elements are indicated. The main
pre-mRNAs used in this work are presented below. Both
the pSma and pPmac pre-mRNAs contain exon 6B, IVS B7
andexon7sequences.ThepPmac5' regioncontainsthelast
167 nucleotides of the IVS AB (see Clouet d’Orval et al.
(1991a,b) for the RNA structure in this region). The pSma
5' end is composed of the last nine nucleotides of IVS AB.
1991),where anRNA secondarystructurebrings the
3'splicesite close tothe branch point,andsequesters
a cryptic 3' AG acceptor site in vivo. Moreover, the
productionofthe secreted ormembraneformofIgM
may be associated with the formation of an RNA
secondary structure that traps an alternative 3' splice
site (Watakabe et al., 1989). Recent work has con-
firmed the importance of pre-mRNA structure on
invivoandinvitrosplicinginSaccharomycescerevisiae.
Hairpins that sequester the 5' splice site interfere
with early steps of spliceosome assembly, including
U1 snRNP binding (Goguel et al., 1993).
An internal region of the chicken b-tropomyosin
gene contains two exons that are spliced in a
mutually exclusive way. Mature RNA formed in
non-muscle and smooth muscle cells contains exon
6A, while exon 6B is chosen in skeletal muscle cells
(Libri et al., 1989). The mechanism by which these
exons are differentially used in a specific cellular
context has been under investigation in our
laboratory for several years. Recent studies demon-
strate that both exons are specifically regulated. In
non-muscle cells, the exon 6A donor site is activated
by a pyrimidine-rich sequence located in the 5'
region of the intron between exons 6A and 6B (IVS
AB:Balvayet al.,1992; Gallegoet al., 1992) when two
different cis-acting elements repress exon 6B
recognition: (1) sequences in the 3' end of IVS AB as
well as the first nucleotides of the exon 6B are
involvedin the repressionof the exon6B 3'splicesite
(Gallego et al., 1992; Goux-Pelletan et al., 1990; Libri
et al., 1992). (2) An RNA secondary structure
involving IVS AB, exon 6B and the intron between
exons 6B and 7 (IVS B7) has been associated with
exon 6B to 7 splicing inhibition in vitro (Clouet-
d’Orval et al., 1991a,b). Moreover, part of this
secondarystructure has been shownto influencethe
use of exon 6B in vivo (Libri et al., 1991). Native gel
assaysindicate that IVSB7 splicing inhibition invitro
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
593
is associated with a decrease of spliceosome
assembly efficiency (Clouet-d’Orval et al., 1991b).
Here, we have focused on the mechanism by which
the IVS AB region involved in the pre-mRNA
structurearoundexon6Binterfereswith spliceosome
assembly. Using a spliceosome-affinity selection
approach, we show that the pre-mRNA structure
interferes with several steps of complex formation:
U2 snRNP binding at the branch site and U4, U5, U6
tri-snRNP interaction with pre-complex A. More-
over, U1 snRNP binding on the exon 6B 5' splice site
isdependentonthe MgCl2concentration,suggesting
that structural constraints can efficiently modulate
the5'splicesiteaccessibility. The correlationbetween
pre-mRNA structure occurrence, snRNP binding
level and in vitro splicing efficiency is confirmed by
the study of several activating mutations.
Results
IVS AB 3' region inhibits the splicing of IVS B7
in highly active nuclear extracts
It has been shown that the pyrimidine-rich 3'
regionof IVSABis directly involvedin theformation
of a secondary structure including exon 6B and
intron IVS B7 (Clouet d’Orval et al., 1991a,b). We
found, using a mutagenesis approach, a correlation
between the stability of this structure and the
inhibition of IVS B7 splicing in HeLa cell nuclear
extracts (Clouet-d’Orval et al., 1991b). Modifications
proposed by Abmayr et al. (1988) in the protocol of
preparation of the extracts have led to production of
HeLa cell nuclear extracts exhibiting a strongly
enhanced splicing activity. This observation has led
us to test the splicing capacity of these extracts on a
pre-mRNA where the inhibitory secondary struc-
ture can form. Two transcripts were used for this
study (Figure 1). These two precursors contain
sequencesspanningfromexon6Btoexon7anddiffer
in their 5' extremity: in pPmac pre-mRNA, the exon
6B sequence is preceded by the last 167 nucleotides
ofIVS AB.Thissequence allowssecondary structure
formation and causes the inhibition of in vitro
splicing of IVS B7 in classical Dignam nuclear
extracts (Clouet-d’Orval et al., 1991b). The 5' end of
pSma pre-mRNA contains only the last nine
nucleotides of the IVS AB, which do not allow the
formation of the secondary structure. This precursor
is spliced very efficiently in the presence of Dignam
nuclear extracts. Splicing reactions were performed
with these transcripts using highly active nuclear
extracts. Knowing that Mg2+ stabilises RNA struc-
tures, two MgCl2concentrations were tested
(Figure 2(a)). The splicing efficiency was calculated
as the molar fraction of messenger RNA produced at
each time during the reaction (Figure 2(b)).
At 1 mM MgCl2, pSma pre-mRNA is 58% spliced
after 60 minutes of incubation, whereas pPmac
pre-mRNA is only 24% spliced. This inhibition of
pPmac splicing at 1mM MgCl2had been observed
with Dignam nuclear extracts (Clouet-d’Orval et al.,
Figure2.Invitro splicingof pSma andpPmacprecursors
in highly active nuclear extracts. (a) Splicing product
analysis. Labelled transcripts were incubated in the
presence of HeLa cell nuclear extracts under splicing
conditions as described in Materials and Methods, with a
finalMgCl2concentrationof 1and 3 mM.Splicingproducts
were analysed on a denaturing 6% polyacrylamide gel.
Pre-mRNA, mRNA and IVS B7 positions are indicated.
The 1and 3 mM MgCl2sampleswere run on separate gels,
which explains the slightly different migration of the
species. (b) Splicingefficiencyof pSma and pPmac at 1 and
3 mM MgCl2. Pre-mRNAs and mRNAs were quantified
and the percentage of spliced product calculated at
different times during the reaction.
1991b). Nonetheless, it is noticeable that the
activation resulting from the use of nuclear extract
prepared according to the Abmayr et al. (1988)
protocol is much greater for pPmac than for pSma
(compare with Figure 5 of Clouet-d’Orval et al.,
1991b). This result suggests that a major difference
between the two types of nuclear extracts may be a
difference in the concentration of factors that
interfere with pre-mRNA structure, perhaps factors
exhibiting helicase activities. Consistent with this
hypothesis, the kinetic lag observed for pPmac (but
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
594
not for pSma) until 30 minutes of incubation may
reflect the time required for these activities to
partially unwind the inhibitory structure. However,
these activities are not sufficient to produce a
complete derepression of pPmac pre-mRNA.
At 3 mM MgCl2, both pSma and pPmac splicing
efficiencies are decreased in comparison with the
results obtained at 1mM MgCl2(34% and 5%,
respectively,at60 minutes). However, theincrease of
MgCl2concentration has a much more deleterious
effect on pPmac splicing. This observation is con-
sistent with the idea of an inhibitory pre-mRNA
structure that is stabilised when the salt concen-
trationisincreased. Altogether,these results indicate
that, even in highly active nuclear extracts, the
pre-mRNA structure around IVS AB can inhibit the
splicing of IVS B7. These nuclear extracts were used
in all the experiments presented below.
In vitro
splicing inhibition of IVS B7 is
correlated with a decreased interaction
of all snRNPs at 3 mM MgCl2
In order to study the mechanism by which the
pre-mRNA structure inhibits the in vitro splicing of
IVS B7, we have compared the ability of the
repressed (pPmac) and the derepressed (pSma)
transcript to support spliceosome assembly. To
begin, this comparison was carried out under
conditions where the structural inhibition is
particularly important (3 mM MgCl2). First, we
investigatedspliceosomeformationby usinganative
gel assay. As previously reported (Clouet-d’Orval
et al., 1991b), both A and B complex formation are
reduced for the pPmac transcript (data not shown).
This suggests an inhibition at an early stage of the
spliceosome assembly.
U1interactionwiththe 5'splicesite isknowntobe
one of the earliest events in spliceosome assembly
(Guthrie, 1991). In several cases, a defect in this step
has been shown to be responsible for the lack of 5'
splice site utilisation (Goguel et al., 1993; Grabowski
et al., 1991). In order to test whether U1 fixation on
the 5' splice site was affected by the formation of the
inhibitory structure, spliceosome assembly was
investigated at 3 mM MgCl2using affinity selection
experiments (Figure 3) as described by Ryder et al.
(1990). As has been described for other pre-mRNAs,
snRNP binding on transcripts occurs following an
ordered pathway: U1 snRNP interaction with
precursoris a nearlyinstantaneousprocess and does
not require incubation at 30°C (in our experiments,
thezero time-pointsare performedonice). After five
minutesincubationat30°C,complexesassembled on
pre-mRNA contain U1 and U2, U4, U5 and U6
snRNPs. The maximal level of these last snRNPs is
reached only after 15 minutes of incubation. This
experiment, performed with both pSma and pPmac
transcripts, reveals different patterns in snRNP
binding efficiency. Although highly efficient binding
of all the snRNPs is observed on pSma pre-mRNA,
the repressed transcript pPmac exhibits a decreased
levelofU2, U4, U5andU6aswell as U1snRNPs.The
low level of U2, U4, U5 and U6 is consistent with the
inefficient formation of splicing complexes observed
in the native gel assay. U1 interaction appears to
be inhibited also. This result suggests that under
conditions favouring pre-mRNA folding the splicing
inhibition is correlated with an alteration of the 5'
splice site recognition. The binding of U1 at the 5'
splice site is a determinant step that conditions the
following steps of spliceosome assembly. Conse-
quently, the level of U2, U4, U5 and U6 is expected
to be decreased as a result of the poorly efficient
interaction of U1.
Destabilisation of pre-mRNA structure
increases both the
in vitro
splicing and
snRNP binding at 3 mM MgCl2
In pSma pre-mRNA, the nearly complete deletion
of the IVS AB sequences leads to a strong activation
oftheinvitrosplicingreaction.In ordertoconfirmthe
correlation between pre-mRNA structure, splicing
efficiency and snRNP interaction level, we have
tested different mutant transcripts that lead to a
partial derepression of IVS B7 splicing.
Inthe pBR-Sma pre-mRNA,the IVS ABsequences
of pPmac (167 nt) are replaced by a fragment of
pBR322 (123 nt). Pre-mRNA folding into the
inhibitory structure is then abolished but the global
length of the transcript is nearly conserved. This
mutationstronglyactivatesin vitrosplicingefficiency
(Figure 4(b) and (c): 19% of spliced RNA at 60
minutescomparedwith 6% forpPmac) as wellas the
binding level of all the snRNPs (Figure 3(a) and (b)).
InpPmac16, a hairpinstructure at the 5'endof the
exon 6B (stem I) is disrupted by replacement of 16
nucleotides (Figure 4(a)). Mutation 16 has only a
smallactivatingeffectonsplicing(8% of splicedRNA
at 60 minutes; Figure 4(b) and (c)). This stimulation
iscorrelatedwith a smallactivationofU1,U2,U4, U5
and U6 binding levels (Figure 3(a) and (b)).
Mutation21 destabilisesa stem-looplocated inthe
region of the 5' splice site downstream of exon 6B
(stemII;Figure4(a). It isnoticeablethatthismutation
modifies nucleotides +8 to +18 after the 5' splice site
and therefore might not alter the positions involved
in the Watson-Crick base-pairing with the 5'
extremity of U1 snRNA. The results presented in
Figures4and3 showthatthis mutationactivatesboth
splicing (15% of spliced RNA at 60 minutes) and
snRNP binding. Mutation 21 exhibits a stronger
effect than mutation 16 both on splicing and on U2,
U4, U5 and U6 interaction. Concerning U1 snRNP,
wehavenoticedbutdonotunderstandwhymutation
21 presents a relatively low signal (except at 45 min-
utes, in this particular experiment) in comparison
with U2, U4, U5 and U6 snRNP binding. This may
imply a weak interaction of U1 snRNP with this pre-
mRNA leading to U1 snRNP being washed away
during the affinity selection process.
Altogether, these results indicate that the pPmac
pre-mRNA structure inhibits the binding of all the
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
595
snRNPs at 3 mM MgCl2. Mutations that destabilise
this structure partially derepress both the in vitro
splicing and the snRNP binding process. Nonethe-
less, it is noticeable that a strict correlation cannot be
made between splicing efficiency and snRNP
bindingcapacity, especially for thepPmac21 andthe
pBR-Sma mutants, which exhibit large stimulation
of all snRNPs binding (except for U1 snRNP on
pPmac 21), but only a partial restoration of the
in vitro splicing capacity. This suggests that some
steps, posterior to the interaction of the tri-particle
U4, U5, U6 may remain at least partly inhibited in
these mutants.
U1 snRNP interacts efficiently with the pPmac
5' splice site at 1 mM MgCl2
At 1 mM MgCl2, the splicing efficiency of pPmac
is significantly enhanced in comparison with
reactions carried out at 3 mM MgCl2(Figure 2). In
ordertotest if thisstimulationwascorrelatedwithan
activation of the binding of the snRNPs, we have
Figure 3. Analysis of snRNA in affinity selected spliceosomes assembled at 3mM MgCl2. (a) In vitro transcripts were
prehybridised with biotinylated 2'-O-allyl RNA oligonucleotide F7 B and incubated for 0 to 45 minutes with HeLa cell
nuclear extracts under splicing conditions as described in Materials and Methods. Affinity selected snRNAs were
separated on a 10% polyacrylamide denaturing gel and analysed by Northern blot. The variations of level among the
different pre-mRNAs are not important because this parameter is not used in the normalisation of the samples (see
MaterialsandMethods). C, Control,the same as pSma45 minutes, exceptthat oligonucleotideF7 Bwasomitted.Labelled
pre-mRNA and probed snRNA positions are indicated. (b) Comparison of the snRNA composition of splicing complexes
formed on pSma, pPmac, pPmac 16, pPmac 21 and pBR-Sma pre-mRNAs at 3 mM MgCl2. snRNA levels were quantified
and pre-mRNA X/pSma ratios established for the different snRNAs.
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
596
Figure 4. (a) Secondary structure
in the exon 6B 5' splice site region.
Thestructure presentedis part ofthe
previously published pPmac pre-
mRNA structure (Clouet d’Orval
et al., 1991a). Exon 6B and intron
sequences are written in bold upper-
case and plain lowercase characters,
respectively. The position of the exon
6B donor site is indicated by an
arrow. Sequences that have been
replaced in mutants 16 and 21 are
indicated.(b)Invitrosplicingassayof
pSma, pPmac and the mutant pre-
cursors pBR-Sma, pPmac 16 and
pPmac21. The reactionswere carried
out as described in Materials and
Methods, with a final concentration
of 3 mM MgCl2. (c) Quantification of
the in vitro splicing kinetics of the
activating mutants. Precursor and
messenger RNAs were quantified
and the percentage of spliced pro-
duct calculated at different times.
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
597
Figure 5. (a) Affinity selection of spliceosomes
assembled on pSma and pPmac pre-mRNAs at 1 mM
MgCl2. Pre-mRNAs were prehybridised with oligonucle-
otide F7 B (except pSma in control C) and incubated for
various times under splicing conditions. snRNAs present
in assembled complexes were analysed by Northern blot
(see Materials and Methods for experimental and
normalisation procedures). (b) Comparison of the snRNA
composition of splicing complexes formed on pSma and
pPmac pre-mRNAs at 1mM MgCl2. snRNA levels were
quantified and pPmac/pSma ratios established for the
different snRNAs.
Figure 6. Affinity selection of complexes assembled on
a 5' splice site mutant. Pre-mRNAs were prehybridised
with oligonucleotide F7 B (except pPmac in control C) and
incubated for various times under splicing conditions.
snRNAs present in assembled complexes were analysed
by Northern blot (see Materials and Methods for
experimental and normalisation procedures). The pPmac
ponctranscript is equivalent to pPmac, except thatthe first
guanosine base of IVS B7 has been mutated to cytosine.
5' splice site located 47 nucleotides downstream
of the exon 6B donor site. This cryptic site
(GAG/GGGAGC) matches six bases out of nine
of the mammalian 5' splice site consensus se-
quence ((C/A)AG/GU(A/G)AGU). The presence of
guanosine instead of the highly conserved uracil at
position+2 ofthis crypticsite probably explainswhy
it is not used for the in vitro splicing assay. However,
this does not preclude binding to U1 snRNP. The
Drosophila P-element gene has provided a clear
exampleinwhichthe bindingofU1snRNPto cryptic
5'splicesitesplaysan importantroleinthe inhibition
of the splicing of the pre-mRNA third intron by
preventing U1 binding on the accurate 5' splice site
insomaticcell extracts (Siebelet al.,1992). Sincesuch
a process is involved in the regulation of some
pre-mRNA alternative splicing, we have tested
whether U1 binding to the repressed transcript
pPmac occurs only at the accurate 5' splice site
(AAG/GUAUGA). Thus, a mutation was introduced
to replace the consensus guanosine at position +1 of
the IVS B7 by cytosine (pPmac ponc). Because this
consensus guanosine is involved in Watson-Crick
base-pairing with the 5' end of U1 snRNA, a
mutation at this position is expected to produce a
dramatic decrease of U1 binding if it actually occurs
at this site. Mutating the 5' splice site in pPmac leads
to a complete inhibition of the in vitro splicing
reaction (data not shown) and abolishes all snRNP
binding on the pre-mRNA (Figure 6). This result
strongly supports the idea that U1 binds pPmac
pre-mRNA only at the correct 5' splice site. In
addition, we have performed spliceosome selection
experiments in the absence of ATP. Under these
conditions,where spliceosomeassemblyisrestricted
to the early steps occurring before the stable binding
of U2 snRNP at the branch point, U1 interacts
efficiently with both pSma and pPmac, in a stable
performed affinity selection experiments at 1 mM
MgCl2. The result is presented in Figure 5(a). Under
theseconditions,U1 snRNP interactsefficiently with
both the repressed transcript pPmac and the
derepressedpSma. Onthecontrary,theinteractionof
U2, U4, U5 and U6 on pPmac pre-mRNA is only
poorly stimulated. The fact that the quantitative
binding of U1 on pPmac does not lead to an efficient
binding of the other snRNPs may be the result of:
(1) a specific inhibition of the binding process of U2,
and U4, U5 and U6; (2) an efficient but non-func-
tional interaction of U1 snRNP with a cryptic
site. Concerning the first hypothesis, an inhibition
process occurring after the interaction of U1 snRNP
with the 5' splice site has been described in the
Saccharomyces cerevisiae RPL32 pre-mRNA (Vilardell
& Warner, 1994) and in the HIV system (Kjems &
Sharp, 1993). The other hypothesis that U1 snRNP
mayinteractwith a crypticsiteiscompatiblewith the
sequence analysis of IVS B7 that reveals a cryptic
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
598
complexthatdoesnotevolveduringincubation(data
not shown). Therefore, even when no further step of
spliceosome assembly can modulate levels of U1
bound to pre-mRNA, its interaction is quantitatively
equivalent in both the repressed and derepressed
transcripts. Altogether, these experiments indicate
that, when spliceosome assembly is carried out at
1 mM MgCl2, the 5' splice site of pPmac pre-mRNA
is fully accessible to U1 snRNP. However, this first
stepdoesnoteffectivelystimulatethe followingsteps
of spliceosome formation.
The measurement of the pPmac/pSma ratio
reveals that U2 and U4/U5/U6 snRNP
interaction processes with the pPmac
transcript are both inhibited at 1 mM MgCl2
As the interaction of U1 snRNP with pPmac
pre-mRNA iseffective at 1 mMMgCl2, wehavetried
to characterise the steps that remain inhibited in
these conditions. The previous result indicates that
the interaction of U2 with the branch point is less
efficient on pPmac than on pSma transcript. In order
to test if, in these conditions, U2 snRNP binding
was the only limiting step, we have measured and
compared the pPmac/pSma ratio for the different
snRNPs. If U2 snRNP binding is the only limiting
step in the spliceosome assembly on pPmac
precursor, U4, U5 and U6 binding levels should not
be affected more than the binding level of U2. If this
were the case, the pPmac/pSma ratio should be
identical for U2 and for U4, U5, U6. The
pPmac/pSma ratio calculated for each snRNA is
presented in Figure 5(b). The U1 binding process is
effective and even slightly more efficient on pPmac
than it is on pSma (especially at the initial time).
Relative to U1, measurement of the pPmac/pSma
ratio for U2 reveals that U2 interaction with pPmac
pre-mRNA is reduced by 20% after five minutes of
incubation and 40% after 15 minutes. The U4, U5
and U6 binding levels are even more dramatically
decreased in the pPmac construct (about 75%
decreased compared with pSma). The signal of U2
snRNAweobservewithpSma and pPmac is thought
to be the result of an interaction with the IVS B7
branch point. Nonetheless, the IVS AB sequence in
the pPmac transcript contains the branch point that
is used in skeletal muscle cells to splice exon 5 to 6B.
In order to eliminate the possibility that this branch
point contributes to the U2 binding level, we have
mutated this branch site. This mutation does not
reduce the binding level of U2 snRNP (data not
shown). This indicates that the signal of U2 bound
on pPmac pre-mRNA most certainly reflects an
interaction with the IVS B7 branch point. Therefore,
the noticeable difference between the binding
efficiency of U2 and U4, U5, U6 snRNPs on pPmac
indicates that both the U2 and the tri-particle U4,
U5, U6 interactions with pPmac pre-mRNA are
inhibited.
In conclusion, this work indicates that the IVS AB
region involved in the pre-mRNA folding around
exon 6B inhibits the in vitro splicing of IVS B7 by
interfering with several steps in spliceosome
assembly. First, under conditions that stabilise the
pre-mRNA structure, the recognition of the 5' splice
site is significantly reduced. Second, in conditions
that allow an efficient binding of U1 snRNP with the
5' splice site, the interaction of U2 snRNP with the
branch point and the association of the tri-particle
U4, U5, U6 with the pre-spliceosome both remain
inhibited.
Discussion
Spliceosome affinity selection experiments have
been used toinvestigatethemechanismbywhichthe
pre-mRNA structure sequestering b-tropomyosin
exon 6B interferes with IVS B7 intron splicing in the
presence of HeLa cell nuclear extract. These
experiments have led to the identification of several
steps of spliceosome assembly that are inhibited by
the IVS AB region involved in the pre-mRNA
structure.
Influence of pre-mRNA structure on the
recognition of the IVS B7 5' splice site
by U1 snRNP
When splicing reactions are performed under
conditions that stabilise the RNA structures (3mM
MgCl2), the binding of U1 snRNP to the 5' splice site
downstream of exon 6B is notably reduced in pPmac
precursor. This result is consistent with previous
data showing that both the secondary and tertiary
structure around this 5' splice site are different in
pSma and pPmac pre-mRNAs at 3 mM MgCl2
(Clouet d’Orval et al., 1991a). This observation had
led us to speculate that the in vitro splicing
derepression of pSma pre-mRNA would be
parallelled by an increased accessibility of the 5'
splice site domain. Furthermore, we have observed
that diminishing the MgCl2concentration induces a
complete restoration of U1 snRNP interaction with
the 5' splice site of pPmac pre-mRNA. This result
suggests that modifications of the pre-mRNA
environment have important effects on spliceosome
assembly around IVS B7. Under biological con-
ditions, such modifications may be induced by
specific variations of Mg2+ concentration or by some
tissue-specific factors. They may also come from
variationsinthe relativeconcentrationof constitutive
splicing factors exhibiting antagonist effects on
pre-mRNA structures stability and/or on the
recruitmentofothersplicingfactors.Agood example
to illustrate the importance of the relative concen-
trationof somesplicingfactorshasbeen providedby
recent experiments describing, both in vitro and
invivo, theinfluenceofthe ratiobetween thesplicing
factors ASF/SF2 and hnRNP A1 on the choice
between two alternative 5' splice sites (Caceres et al.,
1994; Mayeda et al., 1993). In our in vitro system, the
influence of the structural environment of the 5'
splice site downstream of exon 6B has also been
investigated by introducing a mutation that desta-
bilisesstemII (pPmac21). Surprisingly, this mutation
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
599
does not restore a binding of U1 snRNP similar to
that of pSma. This result suggests that the ability of
this5' splicesite tobe recognisedby U1snRNP is not
simply controlled by the stability of stem II; other
elementsinfluence the structuralenvironmentof this
site. However, our results indicate that this mutation
as well as the pPmac 16 stimulates both the in vitro
splicing and the binding of U2, U4, U5 and U6
snRNPs. The partial activation induced by these
mutations is consistent with the idea of a structural
inhibition in which the global rigidity of the
pre-mRNA is determined by several elements
involving secondary and presumably tertiary
interactions (Clouet d’Orval et al., 1991a).
Influence of the IVS AB 3' region on the
interaction of U2 and U4, U5, U6 snRNPs
Experimentsperformedat 1 mMMgCl2revealthat
although the binding of U1 to pPmac pre-mRNA is
totally restored, the splicing efficiency as well as the
interaction level of the other snRNPs remains
significantly lower than those of pSma. Therefore,
our results suggest that the pre-mRNA structure
interferes with steps posterior to the recognition of
the 5' splice site by U1 snRNP, diminishing the
bindingefficiency ofU2 andU4,U5 andU6 snRNPs.
Theprecisemechanismby which the inhibitionofU2
and U4, U5, U6 binding occurs remains unclear. In
contrast to U1, which can bind to the 5' splice site in
the absence of other factors or ATP, both the
interaction of U2 at the branch site and the
association of the tri-particle U4, U5, U6 with the
pre-complex A require ATP hydrolysis and several
trans-acting factors. Whether the pre-mRNA struc-
turedirectly inhibits theinteractionof thesnRNPsor
whether this effect is transmitted through other
factors is not known. Stable binding of U2 snRNP on
the branch point sequence is dependent on the
preliminary interaction of the splicing factor U2AF
with the 3' splice site (Zamore et al., 1992). The
mammalian splicing factors SF3a 60kd and its yeast
equivalent PRP9 are also involved in the U2 snRNP
binding process and may be intrinsic components of
the particle (Brosi et al., 1993; Staknis & Reed, 1994).
Aninhibitionof theinteractionof these factorswould
lead to a decreased level of U2 and U4, U5 and U6
snRNPs. Other factors, especially ASF/SF2 and
SC35 have been shown to promote the formation of
the U2-containing pre-complex A. These factors may
have a determinant role, in association with others
andespeciallyU1andU2snRNPs,intheprocessthat
bridges5'and 3' splicesites (Fu &Maniatis,1992;Wu
& Maniatis, 1993). Not all the participants of this
process have been characterised so far, but U1
snRNP seems to play an important role, since its
presenceisrequiredforastablefixationof U2 snRNP
on the pre-mRNA (Barabino et al., 1990). It is
noteworthy that this promoting activity is not
dependent on U1 snRNA base-pairing with the 5'
splice site (Seiwert & Steitz, 1993). Such communi-
cation between splice sites seems to appear very
earlyin the spliceosomeformation,presumably even
before the ATP-dependent interaction of U2 with the
branch point (Michaud & Reed, 1993). Then, this
communication is maintained during the further
steps of spliceosome assembly. This is suggested by
recentdata describingboththeinteractionof U1 with
the 3' splice site and the involvement of U5 and U6
in the 5' splice site recognition. In our case, one can
hypothesise that in the repressed precursor pPmac,
pre-mRNA structure may interfere with this
communication between 5' and 3' splice sites,
resulting in a decreased level of U2 and U4, U5 and
U6 snRNPs. Another possibility could be a direct
effect of the structure on 3' splice site recognition.
However, this seems unlikely, for the following
reason:the regulationof splicingoftheexon5 toexon
7 region can be expressed in terms of the choice
between the 6A and 6B donor sites for a splicing
reaction involving the exon 7 acceptor site. This
implies that, in all cell types, the exon 7 acceptor site
isefficiently used eitherwith theexon 6A orwiththe
exon6B donorsite. This appearsdifficult toreconcile
with an intrinsic inhibition at this site by the
pre-mRNA structure. On the contrary, the other
hypothesis proposes that in a non-muscular context,
the pre-mRNA structure would interfere with the
communication between the exon 6B donor site and
the exon 7 acceptor site. In this hypothesis, the 3'
splice site by itself would remain fully accessible to
the splicing factors, allowing a constructive com-
munication with the exon 6A donor site. Further
analysis of the factors known to be involved in the
communication between U1 snRNP bound at the 5'
splice site and U2 snRNP at the 3' splice site will be
undertaken to test this hypothesis.
Materials and Methods
Nuclear extracts
HeLa cell nuclear extracts were provided by A. Miller
(Computer Cell Center, Parc de la Sablonnie`re, 681, B-7000
Mons, Belgium) and prepared according to the method of
Dignam et al. (1983) with the modifications made by
Abmayr et al. (1988).
Plasmid constructions
All constructions usedare derivedfrom a 1.7 kb chicken
b-tropomyosin genomic clone spanning exon 4 to 7 (Libri
et al., 1989). pSma, pPmac and pPmac 16 plasmids have
been described (Clouet-d’Orval et al., 1991b). A pBR322
HpaII fragment of 123 nucleotides (nucleotides 411 to 533)
was filled-in with Klenow fragment and inserted into the
SmaI site of the pSma vector to generate the pBR-Sma
construct (J. Marie, unpublished results). Site-directed
mutagenesis was performed according to the Amersham
handbook procedure to obtain the pPmac 5' ponc mutant.
Mutagenesis was achieved on the pSP65 700 plasmid
(Goux-Pelletanet al.,1990) usinga33 meroligodeoxynucle-
otide 5' GCTCCACCAGGTCATAGCTTGTCAGCCTG-
GGCC 3' containing the desired mutation (underlined).
Subcloning was performed by replacing a BstEII-HindIII
fragment of pSma and pPmac plasmids by the
BstEII-HindIII fragment containing the mutation. The
same type of procedure was used to substitute the
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
600
sequence from nucleotides 857 to 867 by the complemen-
tary sequence to generate the mutant pPmac 21 (B. Clouet
d’Orval & J. Marie, unpublished results; Clouet d’Orval
et al., 1991a,b).
In vitro
transcription
Uniformly labelled SP6 transcripts were synthesised in
a 100 ml reaction containing 40mM Tris-HCl (pH 7.5),
6 mM MgCl2, 2 mM spermidine, 10 mM NaCl, 10 mM
DTT, 0.3 mM NTP, 1 mM cap analogue (m7 G (5')ppp(5')
G), RNasin (50 units), SP6 RNA polymerase (150 units),
1mCi of [a-32P]UTP 800 Ci/mmol (Amersham), 3 mg of
linearised plasmid. Reactions were incubated for two
hours at 37°C. RNA transcripts were then precipitated
and gel-purified. These low-labelled transcripts (1000 to
2000 cts/min per pmol corresponding to the number of
uracil residues in the precursor) were used in the affinity
selection experiments. For analytic in vitro splicing
experiments, high specific radioactivity transcripts (7 to
15 ×105cts/min per pmol) were synthesised in a 20 ml
reaction, with 20mCi of [a-32P]UTP.
In vitro
splicing and affinity selection experiments
The streptavidin-agarose affinity selection procedure is
adapted from that described by Ryder et al. (1990): 40 ml
of streptavidin-agarose beads suspension (GIBCO-BRL)
were preblocked by addition of 10 mg of Escherichia coli
tRNA and 40 mg of acetylated bovine serum albumin
(BSA), and by rotating for 15 minutes at 4°C with 200 ml of
WB 50 (WB 50 in 20 mM Hepes/NaOH (pH 7.9), 0.01%
Nonidet P-40, 50mM KCl). Beads were then washed
successivelywith 200 ml ofWB 50, 200 mlof WB 350(as WB
50 butwith 350 mMKCl) andfinally resuspended in100 ml
of WB350. Before initiating thesplicing reaction, 3 pmol of
32P-labelled transcripts was hybridised to 10 pmol of
2'-O-allyl oligonucleotide F7 B in a 5 mM MgCl2solution
by heating to 56°C for five minutes and then cooling to
room temperature. The F7 B oligonucleotide is a 26-mer 5'
TCCUGAUUUCCUCCUCGUACUUGCCT3' (T, dTTP;C,
biotinylated dCTP; N, 2'-O-allyl NTP) complementary to
the exon 7 3' end. Large-scale splicing reactions (100ml)
were then performed in standard conditions: 1 or 3mM
MgCl2, 20 mM creatine phosphate, 0.5 mM ATP, 3% (w/v)
polyvinyl alcohol, 0.25 units/ml RNasin, 50% (v/v) HeLa
cell nuclear extracts. Spliceosome assembly was initiated
by incubation at 30°C for 0, 5, 15 or 45 minutes. The 100 ml
splicing reaction was mixed on ice with the pre-blocked
streptavidin-agarose. After three hours of rotation at 4°C,
beads were washed five times with 200 ml of WB 350,
digested for 30 minutes with proteinase K (0.75mg/ml
final concentration), heated for five minutes at 85°C,
vortexedforfivesecondsand pelletedbycentrifugationfor
four minutes at 6000revs/min. Supernatant containing
RNA was precipitated with 2.5M ammonium acetate,
0.1 mg/ml Escherichia coli tRNA and 2.5 volumes of
ethanol. Pellets were resuspended in 10 ml of formamide,
0.05% xylene cyanol and the appropriate volume was
loaded onto a denaturing gel (see below). Small-scale
splicingreactions(25 ml) wereperformedwith 0.02 pmolof
highly labelled transcript under the conditions used for
splicing product analysis.
Calculation of the
in vitro
splicing efficiency
Foreach time of thein vitrosplicingreaction,pre-mRNA
andmRNA signalwasmeasuredona MolecularDynamics
PhosphoImager. The intensity of the background was
subtractedfrom thisvalue and the result wasthen divided
by the number of uracil residues of each species to obtain
a normalised value. The efficiency of splicing was
calculated as the ratio between the mRNA value and the
sum of the mRNA and pre-mRNA value.
Normalisation procedure for affinity
selection experiments
During the snRNA elution process, a proportion of
the pre-mRNA is eluted. Since this proportion varies
according to the pre-mRNA, the normalisation of the
samplestobe loadedontoa denaturingpolyacrylamidegel
for Northern blot analysis is calculated as a function of the
number of labelled pre-mRNA molecules bound to
streptavidin-agarose after washes and before elution.
Moreover, the specific radioactivity is variable from one
precursor to another, depending on the number of uracil
residues, which explains the noticeable differences be-
tween pre-mRNA signals in Figures 3(a) and 5(a).
Gel electrophoresis and Northern blot analysis
RNAs were separated by electrophoresis in a 10%
acrylamide (2:38 (w/w), bis-acrylamide to acrylamide),
8 M urea denaturing gel. RNAs were electroblotted onto
Genescreen membranes for three hours at 350 mA in a
25 mM phosphate (pH 6.5) buffer. After transfer, RNAs
were UV cross-linked (Stratalinker) and hybridised with
32P-labelledantisenseRNAprobes, complementarytoeach
of the snRNAs (Konarska & Sharp, 1987), following a
standard procedure (Konarska, 1989). Quantification of
the signals was carried out on a Molecular Dynamics
PhosphoImager.
Acknowledgements
We thank A. I. Lamond for help with the spliceosome
purification procedure and M. M. Konarska for sending us
theplasmidsrequired for the transcriptionof RNAprobes
complementary to snRNAs. We thank J. Banroques, A.
Expert-Bezancon,M.E.Gallego,K.Tannerand C. Thermes
for carefully reading the manuscript and for constant
helpful suggestions. This work was supported by the
CNRS, INSERM (contrat externe no. 881002), Association
FrancaisecontrelesMyopathies(A.F.M.),Associationpour
la Recherche sur le Cancer (A.R.C.) and Ligue Nationale
Contre le Cancer (L.N.C.C.).
References
Abmayr,S.M.,Reed,R.& Maniatis,T.(1988).Identification
of a functional mammalian spliceosome containing
unspliced pre-mRNA. Proc. Natl Acad. Sci. USA,85,
7216–7220.
Balvay, L., Libri, D., Gallego, M. & Fiszman, M. (1992).
Intronic sequence with both negative and positive
effects on the regulation of the b-tropomyosin
transcript. Nucl. Acids Res. 20, 3987–3992.
Barabino, S. M., Blencowe, B. J., Ryder, U., Sproat, B. S. &
Lamond, A. I. (1990). Targeted snRNP depletion
reveals an additional role for mammalian U1 snRNP
in spliceosome assembly. Cell,63, 293–302.
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
601
Brody, E. & Abelson, J. (1985). The ‘‘spliceosome’’: yeast
pre-messengerRNA associates witha 40 S complex in
a splicing-dependent reaction. Science,228, 963–967.
Brosi, R., Gro¨ning, K., Behrnens, S. E., Lu¨hrmann, R. &
Kra¨mer, A. (1993). Interaction of mammalian splicing
factor SF3a with U2 snRNP and relation of its 60-kd
subunit to yeast PRP9. Science,262, 102–105.
Caceres, J. F., Stamm, S., Helfman, D. M. & Krainer, A. R.
(1994). Regulation of alternative splicing in vivo by
overexpression of antagonistic splicing factors.
Science,265, 1706–1709.
Chebli,K., Gattoni,R., Schmitt, P.,Hildwein,G.&Stevenin,
J. (1989). The 216-nucleotide intron of the ElA
pre-mRNA contains a hairpin structure that permits
utilizationof unusually distantbranch acceptors. Mol.
Cell. Biol. 9, 4852–4861.
Clouet d’Orval, B. C., d’Aubenton-Carafa, Y., Marie, J. &
Brody, E. (1991a). Determination of an RNA structure
involved in splicing inhibition of a muscle-specific
exon. J. Mol. Biol. 221, 837–856.
Clouet d’Orval, B., d’Aubenton-Carafa, Y., Sirand-Pugnet,
P., Brody, E. & Marie, J. (1991b). RNA structure
represses utilization of a muscle specific exon in Hela
cell nuclear extracts. Science,252, 1823–1828.
Daugeron,M.C., Tazi,J., Jeanteur, P.,Brunel, C. & Cathala,
G (1992). U1-U2 interaction induced by an RNA
complementary to the 5' end sequence of U1 snRNA.
Nucl. Acids Res. 20, 3625–3630.
Deshler, J. O. & Rossi, J. J. (1991). Unexpected point
mutations activatecryptic 3' splicesites byperturbing
a natural secondary structure within a yeast intron.
Genes Dev. 5, 1252–1263.
Dignam, J. D., Lebovitz, R. R. & Roeder, R. G. (1983).
Accurate transcription initiation by RNA polymerase
IIinasolubleextractfrom isolatedmammaliannuclei.
Nucl. Acids Res. 11, 1475–1489.
Fu, X. D. & Maniatis, T. (1992). The 35-kDa mammalian
splicing factor SC35 mediates specific interactions
between U1 and U2 small nuclear ribonucleoprotein
particles at the 3' splice site. Proc. Natl Acad. Sci. USA,
89, 1725–1729.
Fu, X. D., Mayeda, A., Maniatis, T. & Krainer, A. R. (1992).
GeneralsplicingfactorsSF2and SC35 haveequivalent
activities in vitro, and both affect alternative 5' and 3'
splice site selection. Proc. Natl Acad. Sci. USA,89,
11224–11228.
Gallego, M. E., Balvay, L. & Brody, E. (1992). Cis-acting
sequences involved in exon selection in the chicken
b-tropomyosin gene. Mol. Cell. Biol. 12, 5415–5425.
Ge,H.&Manley, J. L. (1990).A protein factor,ASF,controls
cell-specific alternative splicing of SV40 early
pre-mRNA in vitro. Cell,62, 25–34.
Goguel, V., Liao, X., Rymond, B. C. & Rosbash, M. (1991).
U1 snRNP can influence 3' splice site selection as well
as 5' splice site selection. Genes Dev. 5, 1430–1438.
Goguel, V., Wang, Y. & Rosbash, M. (1993). Short artificial
hairpins sequester splicing signals and inhibit yeast
pre-mRNA splicing. Mol. Cell. Biol. 13, 6841–6848.
Goux-Pelletan, M., Libri, D., D’Aubenton-Carafa, Y.,
Fiszman, M., Brody, E. & Marie, J. (1990). In vitro
splicing of mutually exclusive exons from the chicken
beta-tropomyosin gene: role of the branch point
location and very long pyrimidine stretch. EMBO J. 9,
241–249.
Grabowski, P. J., Seiler, S. R. & Sharp, P. A. (1985). A
multicomponentcomplexisinvolvedin the splicingof
messenger RNA precursors. Cell,42, 345–353.
Grabowski, P. J., Nasim, F. U. H., Kuo, H. C. & Burch, R.
(1991). Combinatorial splicing of exon pairs by
two-site binding of U1 small nuclear ribonucle-
oprotein particle. Mol. Cell Biol. 11, 5919–5928.
Green, M. R. (1991). Biochemical mechanisms of
constitutive and regulated pre-messenger-RNA splic-
ing. Annu. Rev. Cell Biol. 7, 559–599.
Guthrie, C. (1991). Messenger RNA splicing in yeast—
clues to why the spliceosome is a ribonucleoprotein.
Science,253, 157–163.
Kandels-Lewis,S. & Seraphin,B. (1993).Role of U6 snRNA
in 5' splice site selecion. Science,262, 2035–2039.
Kister,L., Domenjoud,L., Gallinaro, H. &Jacob,M. (1993).
Acisactingselector of a 5'splice site. J. Biol.Chem. 268,
21955–21961.
Kjems, J. & Sharp, P. A. (1993). The basic domain of Rev
from Human Immunodeficiency Virus Type I
specifically blocks the entry of U4/U6. U5 small
nuclear ribonucleoprotein in spliceosome assembly.
J. Virol. 67, 4759–4776.
Konarska,M.M. (1989). Analysis ofsplicingcomplexesand
smallnuclear ribonucleoproteinparticlesbynativegel
electrophoresis. In Methods in Enzymology (Dahlberg,
J. E. & Abelson, J. N., eds), vol. 180, pp. 442–453,
Academic Press Inc., San Diego, CA.
Konarska,M. M. & Sharp,P.A. (1987). Interactionbetween
smallnuclearribonucleoproteinparticlesinformation
of spliceosomes. Cell,49, 763–774.
Krainer, A. R., Conway, G. C. & Kazak, D. (1990). The
essential pre-mRNA splicing factor SF2 influences 5'
splice site selection by activating proximal sites. Cell,
62, 35–42.
Lerner, M. R., Boyle, J. A., Mount, S. M., Wolin, S. L. &
Steitz, J. A. (1980). Are snRNPs involved in splicing?
Nature,283, 220–224.
Lesser, C. F. & Guthrie, C. (1993). Mutations in U6 snRNA
that alter splice site specificity—implications for the
active site. Science,262, 1982–1988.
Libri, D., Lemonnier, M., Meinnel, T. & Fiszman, M. Y.
(1989).Asinglegene codesforthe bsubunitof smooth
andskeletalmuscletropomyosininthechicken. J.Biol.
Chem. 264, 2935–2944.
Libri, D., Piseri, A. & Fiszman, M. (1991). Tissue specific
splicing in vivo of the btropomyosin gene:
dependance on an RNA secondary structure. Science,
252, 1842–1845.
Libri,D., Balvay,L. & Fiszman, M.Y. (1992).In vivosplicing
of the b-tropomyosin pre-mRNA: a role for branch
point and donor site competition. Mol. Cell. Biol. 12,
3204–3215.
Maniatis, T. (1991). Mechanisms of alternative pre m-RNA
splicing. Science,251, 33–34.
Mayeda, A. & Krainer, A. R. (1992). Regulation of
alternative pre-messenger-RNA splicing by hnRNP-
A1 and splicing factor-SF2. Cell,68, 365–375.
Mayeda, A., Helfman, D. & Krainer, A. R. (1993).
Modulation of exon skipping and inclusion by
heterogenous nuclear ribonucleoprotein A1 and
pre-mRNAsplicingfactorSF2/ASF.Mol.Cell.Biol.13,
2993–3001.
Michaud, S. & Reed, R. (1993). A functional association
between the 5' and 3' splice sites is established in the
earliest prespliceosome complex (E) in mammals.
Genes Dev. 7, 1008–1020.
Moore, M. J., Query, C. C. & Sharp, P. A. (1993). Splicing
of precursors to mRNA by the spliceosome. In The
RNA World (Gesteland, R. F. & Atkins, J. F., eds),
vol. 13, pp. 303–357, Cold Spring Harbor Laboratory
Press, Plainview, NY.
Reich, C. I., Hoy, R. W. V., Porter, G. L. & Wise, J. A. (1992).
Mutations at the 3' splice site can be suppressed by
JMB—MS 718
Spliceosome Assembly Inhibition by RNA Structure
602
compensatory base changes in U1 snRNA in fission
yeast. Cell,69, 1159–1169.
Roger,J.& Wall,R. (1980).A mechanism for RNA splicing.
Proc. Natl Acad. Sci. USA,77, 1877–1879.
Ryder, U., Sproat, B. S. & Lamond, A. I. (1990).
Sequence-specific affinity selection of mammalian
splicing complexes. Nucl. Acids Res. 18, 7373–7379.
Seiwert, S. D. & Steitz, J. A. (1993). Uncoupling two
functions of the U1 Small Nuclear ribonucleoprotein
particle during in vitro splicing. Mol. Cell Biol. 13,
3135–3145.
Siebel, C. W., Fresco, L. D. & Rio, D. C. (1992). The
mechanism of somatic inhibition of Drosophila
P-element pre-mRNA splicing: multiprotein com-
plexes at an exon pseudo-5' splice site control U1
snRNP binding. Genes Dev. 6, 1386–1401.
Smith, C. W. J., Patton, J. G. & Nadal-Ginard, B. (1989).
Alternative splicing in the control of gene expression.
Annu. Rev. Genet. 23, 527–577.
Sontheimer,E. J. & Steitz,J.A. (1993). The U5and U6 small
nuclear RNAs as active site components of the
spliceosome. Science,262, 1989–1996.
Staknis, D. & Reed, R. (1994). Direct interactions between
pre-mRNA and six U2 small nuclear ribonucle-
oproteinsduringspliceosomeassembly. Mol. Cell. Biol.
14, 2994–3005.
Tatei, K., Takemura, K., Tanaka, H., Masaki, T. & Ohshima,
Y. (1987). Recognition of 5' and 3' splice sites in
pre-mRNA studied with a filter binding technique.
J. Biol. Chem. 262, 11667–11674.
Vilardell, J. & Warner, J. R. (1994). Regulation of splicing
at an intermediate step of the formation of the
spliceosome. Genes Dev. 8, 211–220.
Watakabe,A.,Inoue,K.,Sakamoto,H. &Shimura,Y.(1989).
A secondary structure at the 3' splice site affects the
in vitro splicing reaction of mouse immunoglobulin m
chain pre-mRNA. Nucl. Acids Res. 17, 8159–8169.
Weiner, A. M. (1993). Messenger RNA Splicing and
autocatalytic introns—distant cousins or the products
of chemical determinism. Cell,72, 161–164.
Wise, J. A. (1993). Guides to the heart of the spliceosome.
Science,262, 1978–1979.
Wu, J. Y. & Maniatis, T. (1993). Specific interactions
between proteins implicated in splice site selection
andregulatedalternativesplicing.Cell,75, 1061–1070.
Zamore, P. D., Patton, J. G. & Green, M. R. (1992). Cloning
and domain structure of the mammalian splicing
factor U2AF. Nature,355, 609–614.
Zillmann, M., Rose, S. D. & Berget, S. M. (1987). U1
small nuclear ribonucleoproteins are required early
during spliceosome assembly. Mol. Cell. Biol. 7,
2877–2883.
Edited by N. Yaniv
(Received 3 April 1995; accepted 16 June 1995)
