ABSTRACT
In metazoans, the E complex is operationally defined as an ATP-independent spliceosomal complex that elutes as a single peak on a gel
filtration column and can be chased into spliced products in the presence of an
excess of competitor pre-mRNA. The A complex is the first ATP-dependent functional spliceosomal complex. U1 snRNP first binds tightly to the 5
'
splice site in the E complex and U2 snRNP first binds tightly to the branch
site in the A complex. In this study, we have generated and characterized a
monoclonal antibody (mAb 4G8) directed against SAP 62, a component of U2 snRNP
and a subunit of the essential mammalian splicing factor SF3a. We show that
this antibody is highly specific for SAP 62, detecting only SAP 62 on Western
blots and immunoprecipitating only SAP 62 from nuclear extracts. The anti-SAP 62 antibody also immunoprecipitates U2 snRNP and the A complex.
Significantly, however, we find that the E complex is also efficiently
immunoprecipitated by the anti-SAP 62 antibody. This antibody does not cross-react with any E complex-specific components, indicating that SAP 62 itself is
associated with the E complex. To determine whether other U2 snRNP components
are associated with the E complex, we used antibodies to the U2 snRNP proteins
B
''
and SAP 155. These antibodies also specifically immunoprecipitate the E
complex. These observations indicate that U2 snRNP is associated with the E complex. However, we find that U2 snRNP is not as tightly bound in
the E complex as it is in the A complex. The possible significance of the weak
association of U2 snRNP with the E complex is discussed.
The pre-mRNA splicing reaction requires the formation of a series of highly dynamic spliceosomal complexes. These complexes are composed of U1, U2, U4, U5 and U6 small nuclear RNAs (snRNAs) and >50
proteins (reviewed in
1
). The E, A and B complexes contain unspliced pre-mRNA and the C complex contains the products of catalytic step I of the
splicing reaction (exon 1 and lariat exon 2). Two complexes, one containing the
excised lariat intron (i complex) and the other containing the spliced exons (D
complex), are the products of catalytic step II of the splicing reaction.
The ATP-independent E complex was originally identified as a spliceosomal complex
that elutes in a single peak on a gel filtration column and can be efficiently chased into spliced products in the presence of excess competitor pre-mRNA (
2
-
4
). The E complex is thought to be the mammalian counterpart of the yeast commitment complex, which was identified prior to the E complex and shown to be a functional intermediate in the yeast spliceosome assembly pathway (
5
). The E complex is also likely the same as a mammalian complex identified on
density gradients and shown to commit pre-mRNAs to the splicing pathway (
6
).
Highly purified mammalian E complex contains U1 snRNP, U2AF
35
, U2AF
65
and several spliceosome-associated proteins (SAPs) (
4
; reviewed in
7
). U2AF
65
binds to the pyrimidine tract at the 3' splice site in the E complex and U1 snRNA base pairs to the 5' splice site. Both U1 snRNP and U2AF are destabilized from the pre-mRNA during the E -> B complex transition, possibly as early as A complex
assembly (
8
). The ATP-dependent A complex contains U2 snRNP and an essential duplex is formed
between U2 snRNA and the branch point sequence (BPS). This duplex is thought to
specify the branch site adenosine as the nucleophile for the first
transesterification reaction (
9
). U2 snRNP can also bind stably to the BPS in the A3' complex, which assembles on a pre-mRNA lacking the 5' splice site (
10
).
Two multimeric splicing factors, SF3a and SF3b, are components of U2 snRNP and
are required for A complex assembly. SF3a consists of the spliceosome-associated proteins SAPs 61, 62 and 114 and SF3b is thought to consist of
SAPs 49, 130, 145 and 155 (
11
,
12
; reviewed in
7
). All of these proteins, except SAP 130, UV crosslink to pre-mRNA in the A complex to a 25 nt region upstream of the branch site (
13
,
14
). The interactions of SF3a/b with this region, designated the anchoring site,
are thought to play a central role in the tight binding of U2 snRNP to pre-mRNA (
14
).
U2 snRNP remains stably bound to pre-mRNA in the B complex and the U4/U5/U6 tri-snRNP first binds to pre-mRNA at this time (
10
; reviewed in
1
). U6 snRNA interacts with intron sequences at the 5' splice site, replacing U1 snRNA (
15
-
20
). Base pairing interactions between U2 and U6 snRNAs, in conjunction with the
U6 snRNA-5' splice site duplex, are thought to position the BPS near the 5' splice site for catalytic step I of the splicing reaction
(
21
; for a review see
22
). U5 snRNP also interacts with exon sequences at both the 5' and 3' splice sites and this interaction may hold the exons together for ligation (for reviews see
22
,
23
).
In this study, we have generated a monoclonal antibody (mAb 4G8) that is highly
specific for the SF3a subunit SAP 62. Unexpectedly, we found that this
antibody, as well as antibodies to other U2 snRNP components, efficiently
immunoprecipitates the E complex. These data, together with other observations,
indicate that U2 snRNP is weakly bound to pre-mRNA in the E complex and tightly bound in the subsequent spliceosomal complexes. The possible functional significance of the loose association of U2 snRNP with the E complex is discussed.
Rat liver nuclei were isolated as described (
24
). The resulting nuclear pellet was extracted with 10 mM EGTA in 10 mM HEPES, pH
7.4, and the EGTA-extracted proteins were used as the antigen source for the preparation of
monoclonal antibodies (
25
). The resulting hybridoma supernatants were screened by immunofluorescence for nuclear labeling and selected hybridomas were cloned. mAb 4G8 is of the IgG1 subtype (ELISA class test kit, Gibco BRL).
Ascitic fluid of mice was collected as described (
25
). IgG was purified using an antibody purification kit (Pierce) and the eluted
antibody was dialyzed into phosphate-buffered saline (PBS).
Cells grown to 50-80% confluency on coverslips were washed twice with PBSCM (PBS containing
1 mM CaCl
2
and 1 mM MgCl
2
) and fixed in 3% paraformaldehyde, PBSCM for 30 min at room temperature. After
washing with 0.3% NH
4
Cl, PBSCM and permeabilization in PBSCM containing 0.3% Triton X-100, cells were incubated with antibody as indicated, washed with 0.2% Triton X-100, PBSCM and incubated with sheep anti-mouse IgG conjugated to FITC (or rhodamine) (
26
). After washing with 0.2% Triton X-100, PBSCM, cell were mounted with 0.1% phenylenediamine and 80% glycerol
in PBS or fixed by other methods as indicated.
HeLa cells grown on poly(L-lysine)-coated coverslips were permeabilized for 3 min at 4oC by incubation with hypotonic buffer (10 mM HEPES-KOH, pH 7.9, 10 mM KCl, 1.5 mM MgCl
2
, 0.5 mM DTT, 0.1% Triton X-100). After washing with the same buffer, KCl was added to a final
concentration of 0.15, 0.3 or 0.4 M by stepwise application from a stock
solution of 2 M KCl in hypotonic buffer. Cells were fixed in methanol for 5 min
on ice and processed for immunofluorescence. For Western blot analysis, cells
grown in suspension were extracted as described above, centrifuged at 12 000
g
for 5 min and the supernatant and pellets were subjected to SDS-PAGE and Western blot analysis.
pAdML, pAd5' and pAd3' were as described previously (
3
,
4
,
27
). Assembly of the H complex and the spliceosomal complexes E, E3', E5', A3' and B was carried out as described (
2
-
4
). Reaction mixtures were incubated for 5 min for E or A complex assembly and
for 15 min for B complex assembly. Reaction mixtures for assembly of the E
complex lacked ATP and MgCl
2
. Complexes were isolated by gel filtration and, where indicated, also by affinity purification (
2
,
27
). For gel shift assay, reaction mixtures were adjusted to 0.5 mg/ml heparin and separated on a 4% native gel (
10
).
Western analysis of proteins from cellular extracts was performed as described (
26
). Filters containing immobilized proteins were incubated in mAb 4G8 tissue
culture supernatant or a 1:500 dilution of ascitic fluid. Rabbit anti-mouse IgG and [
125
I]protein were used for detection (Fig.
1
B). For Western analysis of spliceo- somal proteins, total protein from affinity purified spliceosomes
assembled on AdML pre-mRNA was separated by SDS or two- dimensional gel electrophoresis and transferred to nitrocellulose. Blots were probed with mAb 4G8 ascitic fluid at 1:1000 dilution,
with anti-B'' antibody (mAb 4G3) tissue culture supernatant (undiluted) or with anti-SAP 155 polyclonal antibodies at 1:1000 dilution. Detection was with horseradish
peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (Amersham). For immunoprecipitation of metabolically labeled SAP 62, subconfluent
monolayers of cells were labeled with [
35
S]methionine and/or [
35
S]cysteine (
28
). Cell extracts were then used for immunoprecipitations with mAb 4G8.
To map the mAb 4G8 epitope, derivatives of SAP 62 (
29
) were produced by
in vitro
translation. pSAP 62(1-257) was constructed by deleting the sequence between the
Stu
I and
Bam
HI sites in pSAP 62. pSAP 62(212-464) was constructed by deleting the sequence in pSAP 62 between the
Nco
I and
Rsr
II sites. For immunoprecipitations, mAb 4G8 or mAb 2G2 was coupled to protein A-Trisacryl and rotated with 1 [mu]l of
in vitro
translation reaction mixture in 100 [mu]l PBS for 1 h at 4oC and washed with cold buffer (10 mM Tris-HCl, pH 7.4, 500 mM NaCl, 1% NP40).
Aliquots (200 [mu]l) of gel filtration isolated complexes were rotated at 4oC with 3 [mu]l purified mAb 4G8 coupled to protein A-Trisacryl. For immunoprecipitation of the E complex in high
salt, the buffer was adjusted to 300 mM NaCl. After washing, total RNA in the
bound and unbound fractions was extracted and fractionated on 8% denaturing
gels. Immunoprecipitation of complexes using the anti-SAP 155 antibody or the anti-B'' antibody (mAb 4G3) was performed as above except that
25 [mu]l [alpha]SAP 155 rabbit serum or 500 [mu]l mAb 4G3 from tissue culture supernatant respectively were used.
Preimmune serum or rabbit secondary antibody were used as negative controls. To
immunoprecipitate U2 snRNP, mAb 4G8 was coupled to beads as above and rotated for 3 h at 4oC with either 100 [mu]l nuclear extract or 100 [mu]l nuclear extract that had been pre-incubated under splicing conditions (+ATP, +MgCl
2
) for 5 min at 30oC. The secondary antibody alone was used as a control. After washing in PBS, 300 mM NaCl, 1% NP40, total RNA was extracted, fractionated on an 8% denaturing gel and visualized by ethidium bromide staining.
During a screen of a panel of monoclonal antibodies generated against nuclear
components (see Materials and Methods), we identified an antibody, mAb 4G8,
that specifically labels subnuclear regions in HeLa cells (Fig.
1
A, panel A). Significantly, the 4G8 antigen co-localizes with the splicing factor SC35 (
30
), as well as with snRNP core proteins which bear the Sm epitope (
31
) (Fig.
1
A, panels A-D). The structures labeled by these splicing factors are referred to as
nuclear speckles and are thought to be sites of storage and/or assembly of spliceosomal components (
30
,
32
-
34
).
In order to identify the 4G8 antigen and determine whether a single antigen is
responsible for the speckled immunoflouresence staining pattern, mAb 4G8 was
used to stain cells extracted in increasing concentrations of salt after
permeabilization in Triton X-100 (Fig.
1
B, top panel). Total protein was prepared from the cell pellets and supernatants after the salt extractions and then used for Western analysis with the 4G8 antibody (Fig.
1
B, bottom panel). At salt concentrations up to 0.15 M KCl, a speckled immunostaining pattern similar to that obtained after conventional permeabilization in isotonic buffer was observed (Fig.
1
A, panels A-C). Much less 4G8 antigen was detected in the speckled pattern after incubation with 0.3 or 0.4 M KCl (Fig.
1
B, panels d and e). As shown in Figure
1
B, mAb 4G8 detects a single 60 kDa protein on a Western blot and the majority of this protein is associated with the cell pellet at salt concentrations up to 0.15 M (lanes 1-4); in contrast, the 60 kDa protein is largely extracted at 0.3 M KCl and above (lanes 5-10). These data indicate that the mAb 4G8 antigen is a 60 kDa protein. In
addition, the correlation between the immunofluorescence patterns and the Western analysis at the different salt concentrations indicates that the 60 kDa protein and not a cross-reacting protein(s) is the antigen responsible for the speckled staining pattern.
To determine whether the 60 kDa mAb 4G8 antigen corresponds to a splicing factor, a Western blot of spliceosomal proteins fractionated on a
two-dimensional gel was carried out. As shown in Figure
2
A, 4G8 detects a single protein and an ink stain of this two-dimensional blot (data not shown) showed that this protein is the U2 snRNP
component SAP 62.
In vitro
-translated SAP 62 (Fig.
2
B, lane 1) is also specifically immunoprecipitated by mAb 4G8 (Fig.
2
B, lane 4).
In vitro
-translated SAP 62 is not immunoprecipitated by a control monoclonal
antibody (2G2) or by the secondary antibody alone (R) (Fig.
2
B, lanes 5 and 6 respectively). Thus, mAb 4G8 specifically recognizes both the
native as well as the denatured form of SAP 62.
To characterize the association of SAP 62 with spliceosomal complexes, we used
the 4G8 antibody for immunoprecipitation studies (
3
) (see Materials and Methods). We found that mAb 4G8 efficiently
immunoprecipitates the B complex, but not the H complex [Fig.
4
A, compare the relative levels of pre-mRNA in the pellet (P) and supernatant (S); lanes 1 and 2 versus lanes 5
and 6]. The secondary antibody alone does not immunoprecipitate either complex
(Fig.
4
A, lanes 3, 4, 7 and 8). Unexpectedly, we found that mAb 4G8 also efficiently
immunoprecipitates the E complex (Fig.
4
A, lanes 9 and 10; see below for additional controls with non-specific antibodies). The E complex is immunoprecipitated somewhat less
efficiently than the B complex (Fig.
4
A, compare the ratio of pre-mRNA in lanes 9 and 10 with that in lanes 1 and 2), indicating that SAP 62
is bound less tightly to the E complex or is present in only a sub-population of these complexes (see below for further data indicating that
the antigen detected by 4G8 in the E complex is SAP 62).
In this study, we have characterized a monoclonal antibody (mAb 4G8) that
recognizes SAP 62, a U2 snRNP component and subunit of the essential splicing
factor SF3a (
11
,
12
,
29
). SAP 62 is the only protein detected by mAb 4G8 on immunoblots and is the only
protein immunoprecipitated from total cell lysates, indicating that the
antibody is highly specific for SAP 62. Our data indicate that the epitope
recognized by mAb 4G8 is located in the C-terminal portion of SAP 62, which consists of 23 repeats of a proline-rich sequence (GVHPPAP) (
29
). As was observed with other U2 snRNP components (
31
,
32
,
36
) and splicing factors (for a review see
37
), SAP 62 is localized in nuclear speckle domains
in vivo
.
U2 snRNP (and SAP 62) first binds stably to pre-mRNA in spliceosomal complex A. Not unexpectedly, we found that mAb 4G8
efficiently immunoprecipitates this complex. Surprisingly, however, we found
that this antibody also efficiently immunoprecipitates the E complex. In
previous work, analysis of the total protein composition of affinity purified
spliceosomal complexes by silver staining revealed that U2 snRNP-specific proteins (A', B'' and SF3a/b subunits) are present in the E complex,
but at very low levels (
4
,
27
). Consistent with this result, B'', SAP 155 and SAP 62 are all detected at low levels on Western
blots of affinity purified E complex. However, on the basis of these data
alone, it was not possible to distinguish between the possibility that a low
level of U2 snRNP is tightly associated with a very small population of the E
complex or that U2 snRNP is present in most if not all of the E complexes in the population, but dissociates during affinity purification. Our new analysis, in which we were able to use a gentle
immunoprecipitation assay, supports the latter conclusion.
Further evidence that U2 snRNP is associated with the E complex comes from the
observation that low levels of U2 snRNA are detected in affinity purified E
complex (
3
,
4
,
27
). However, U2 snRNA appears to be less tightly bound in the E complex than in
the A complex, as this snRNA is detected at higher levels in E complex affinity
purified in low (100 mM) than in high (300 mM) salt, whereas U2 snRNA is
present in the A complex at both salt concentrations (
3
). This salt labile association of U2 snRNP with the E complex was also seen in our study; we found that the
anti-SAP 62 antibody immunoprecipitates the E complex in low but not in high
salt.
In previous work, five U2 snRNP components (SF3a/b subunits) were shown to crosslink to pre-mRNA upstream of the branch site to a 25 nt region called the anchoring
site, and functional studies indicated that the interactions of these proteins
with the anchoring site are required to anchor U2 snRNP to the pre-mRNA in the A complex (
14
). Significantly, no crosslinking of any of the SF3a/b subunits to pre-mRNA can be detected in the E complex (
10
). If SF3a/b-pre-mRNA interactions are not involved in the association of U2 snRNP with the E complex, how is this snRNP associated? One
possibility is through interactions with U1 snRNP. Mattaj and co-workers (
35
) showed that antibodies to U2 snRNP A' could co-immunoprecipitate low levels of U1 snRNP along with U2 snRNP from
oocyte extracts. Consistent with this observation, we found that low levels of U1 snRNP are co-immunoprecipitated with U2 snRNP from HeLa extracts using the 4G8 antibody. The factors that
mediate the interactions between U1 and U2 snRNPs are not known, but changes to
U2 snRNA or U2 snRNP structure disrupt the interaction between the two snRNPs (
35
).
Another possible interaction that could be responsible for the association of U2
snRNP with the E complex is a protein-protein interaction between U2AF and SAP 155. This interaction was
detected in a Far Western assay and by yeast two-hybrid assays (O.Gozani and R.Reed, unpublished results). As U2AF is
present in the E complex, U2 snRNP may associate with E complex via the SAP 155-U2AF interaction. Our data also show that U2 snRNP is associated with E5' and E3' complexes, which assemble on RNAs lacking a 3' or a 5' splice site respectively. It is possible
that U2 snRNP binds via distinct interactions in these two complexes. U1 snRNP
is also detected in both the E5' and E3' complexes (
38
) and the interactions in each complex involve distinct domains on U1 snRNP (
39
).
The E complex has been operationally defined as a complex that elutes in a
single peak on a gel filtration column and can be quantitatively chased into
the A complex (
2
-
4
). Whether the E complex is actually a mixture of complexes that contain
different combinations of U1 and U2 snRNPs cannot be strictly ruled out on the
basis of present information. However, considering the operational definition
of the E complex, our data indicate that, at a minimum, 30% of the E complex
population is associated with U2 snRNP (based on the efficiency of
immunoprecipitation of the E complex using anti-U2 snRNP antibodies). Thus, at least a portion, if not all, of the E
complex that can be chased into the A complex contains U2 snRNP. SR proteins
are another example of a splicing factor that is detected in gel filtration
isolated E complex but not in complexes purified under more stringent
conditions (
38
). The SR proteins, though loosely associated with the E complex, nevertheless
play a role in its assembly (
38
).
The question of most obvious importance is whether U2 snRNP is bound in a
functional manner to the E complex. We have used several approaches to address
this question but, so far, have not been able to obtain a definitive answer. A
straightforward approach to the problem is to inactivate U2 snRNP or deplete it
from nuclear extracts and then determine whether pre-assembled E complex can be chased into the A complex. However, we are
unable to obtain an extract specifically lacking U2 snRNP that is sufficiently
active for these studies.
In yeast, U2 snRNP was not found to associate with the commitment complex (
5
,
40
), even under conditions that should detect a weak association of this snRNP (
41
). Thus, it is possible that there are differences in the early steps of
spliceosome assembly between yeast and mammals (
2
-
4
,
39
-
41
). However, these differences may be subtle and may only reflect quantitative differences rather than fundamental differences in the mechanism. Consistent with this view, U2 snRNP binding in yeast requires the 5' splice site and the U1 snRNA-5' splice site duplex (
5
,
41
). In mammals, the efficiency of U2 snRNP binding is increased by the presence
of an upstream 5' splice site (
42
) and by the presence of U1 snRNP (though the U1 snRNA-5' splice site duplex is not required for this stimulatory effect) (
39
). In addition, there is one example in yeast in which ATP-independent binding of U2 snRNP to pre-mRNA occurs (
41
). In this case, a mutation in U1 snRNA that weakens the U1 snRNA-5' splice site duplex results in a low level of ATP-independent U2 snRNP binding (
41
). Liao
et al
. (
41
) propose that this may occur because an ATP-dependent disruption of the U1 snRNA-5' splice site duplex may be required for U2 snRNP binding. It
is possible that the weak U2 snRNP binding that we detect in the E complex is
in some way related to this observation in yeast.
If U2 snRNP is not bound to the E complex in a functional manner, our detection
of this association may nevertheless reflect an important interaction that
occurs at some other point during the normal spliceosome assembly pathway. For example, an interaction between U1 and U2 snRNPs may be involved in forming the association between the
5' and 3' splice sites in the A complex. If, on the other hand, U2 snRNP is
functionally bound in the E complex, the early steps of spliceosome assembly
may involve U2 snRNP binding weakly in the E complex and then undergoing an ATP-dependent conformational change that results in stable U2 snRNP binding.
We are grateful to W. J. van Venrooij (University of Nijmegen, The Netherlands),
Tom Maniatis (Harvard University) and Ann Beyer for providing anti-Sm, anti-SC35 and anti-B'' (4G3) monoclonal antibodies respectively. We also
thank members of the Hong and Reed laboratories for critical reading of the
manuscript.
*To whom correspondence should be addressed. Tel: +1 617 432 2844; Fax: +1 617
432 1144; Email: rreed@warren.med.harvard.edu
REFERENCES
Return