ABSTRACT
The function in splicing of a heterodimeric nuclear cap binding complex (yCBC)
from the yeast
Saccharomyces cerevisiae
has been examined. Immunodepletion of splicing extracts with antibodies directed against one component of the
complex, yCBP80, results in the efficient co-depletion of the second component, yCBP20, producing CBC-deficient splicing extract. This extract exhibits strongly reduced
splicing efficiency and similar reductions in the assembly of both spliceosomes
and of the earliest defined precursors to spliceosomes, commitment complexes.
The addition of highly purified yCBC substantially restores these defects.
These results, together with other data, suggest that CBCs play a highly
conserved role in the recognition of pre-mRNA substrates at an early step in the splicing process.
The removal of intervening sequences (introns) from pre-mRNAs is an essential step in the expression pathway of many genes.
Introns are recognized by a subset of the splicing factors and assembled into a
large multi-component ribonucleoprotein complex called the spliceosome, where they are
accurately removed from the pre-mRNA. The spliceosome consists of U1, U2, U4/U6 and U5 small nuclear
ribonucleoproteins (snRNPs), together with numerous non-snRNP splicing factors. Within the spliceosome two
trans
-esterification reactions occur which result in production of the mature
mRNA and intron lariat. Efficient splicing of an intron depends on recognition
of the
cis
-acting nucleotide sequences present in the pre-mRNA at the 5'-splice site, branch point and 3'-splice site (
1
-
3
).
Both
in vivo
studies in yeast and the development of extracts capable of
in vitro
splicing from both yeast and higher eukaryotes has allowed several steps of
spliceosome assembly to be characterized and some of the components required
for these steps to be defined. In yeast the first complex to be detected is the
`commitment complex', which requires the 5'-splice site and branch point sequences and forms in the absence of
ATP (
4
-
6
). This complex is so named as its formation commits the pre-mRNA to the subsequent spliceosome assembly and splicing pathway. Two
components present in this complex and required for its formation are the U1
snRNP, which base pairs with the 5'-splice site (
6
-
8
), and MUD2, which is thought to be the yeast homologue of the mammalian
splicing factor U2AF (U2 snRNP auxiliary factor) (
9
). A complex with similar characteristics and components to the yeast commitment
complex has also been detected and characterized in mammalian extracts, the
early or E complex (
10
,
11
). The commitment or E complexes are the substrate for the addition of the U2
snRNP, a step which requires the hydrolysis of ATP. The U4/U6.U5 tri-snRNP then joins this complex to form the mature spliceosome in which the
chemical reactions of splicing occur (
1
-
3
,
12
).
Much has been learned of the contributions of the spliceosomal U snRNAs to
spliceosome assembly, RNA recognition and cleavage of the splice sites,
however, the role of non-snRNP splicing factors in recognition of the pre-mRNA is less well characterized. The cap structure, which is
characteristic of RNA polymerase II transcripts, has been shown to play an
important role in pre-mRNA splicing in higher eukaryotes (
13
-
17
). This activity of the cap is mediated by a nuclear cap binding protein complex
(CBC). CBC has been previously characterized and cloned. It consists of two
subunits, CBP80 and CBP20, both of which are required for cap binding (
18
-
20
). CBC is required for nuclear export of U snRNAs from the nucleus of
Xenopus
oocytes (
19
) and efficient pre-mRNA splicing in HeLa cell nuclear extracts, where it functions early in
spliceosome assembly (
18
). A more detailed analysis of the splicing defect has shown that CBC is
required for efficient association of U1 snRNP with the 5'-splice site of the pre-mRNA in what may be one of the earliest steps in pre-mRNA recognition (
21
).
The homologues of CBP80 and CBP20 have been identified in the yeast
Saccharomyces cerevisiae
. For reasons of clarity the yeast homologues of CBP80 and CBP20 or yeast CBC
will be prefixed by the letter y. yCBP80 was found as part of the GenEMBL
database, where it was present as multiple submissions of a known gene sequence
(18,22, and references therein). The peptide sequence of yCBP80 shows ~33% similarity to human CBP80.
yCBP20 was characterized independently in two studies. Görlich
et al
. (
22
) found a significant amount of yCBC in association with SRP1p (
23
,
24
), a subunit of the yeast nuclear protein import receptor (
25
). This observation may have implications for the role of CBC in RNA export and
there is evidence that the vertebrate homologue of the yCBC-Srp1p complex may be a functional intermediate in the cycling of CBC
between the nucleus and cytoplasm (
22
). Purified yCBC shows a similar, but not identical, cap binding specificity to
the human complex. In parallel, the yeast
MUD13
gene was found in a genetic screen which was designed to identify components of
the commitment complex (
26
) and shown to encode the yeast homologue of CBP20 (
27
). It was shown that
mud13
mutant strains were deficient for pre-mRNA splicing both
in vivo
and
in vitro
(
27
).
In this report we demonstrate that immunodepletion of yCBC from yeast
in vitro
splicing extracts results in inefficient assembly of both commitment complexes
and spliceosomes, as well as to a reduction in pre-mRNA splicing. These defects can be partially restored by the re-addition of yCBC. Our data are in good agreement with, and
complementary to, data obtained by Colot
et al
. (
27
), who show that extracts prepared from a
mud13
mutant strain, and thus lacking yCBP20, are defective for both commitment
complex assembly and splicing. Taken together with the data from human cell
extracts of Lewis
et al
. (
21
), these data show that the function of CBC in promoting early steps in
recognition of pre-mRNA is conserved between yeast and mammals.
Plasmids pBS195 (
28
) and SP6-actin (
29
) were a kind gift from Dr Bertrand Séraphin. pBS195 was linearized with
Dde
I and transcribed by T7 RNA polymerase and SP6-actin was linearized with
Msp
I and transcribed by SP6 RNA polymerase. Transcriptions used 0.5 [mu]g cleaved DNA as template with nucleotides at 500 [mu]M, except for UTP, which was at 50 [mu]M, and 1 mM m
7
GpppG with 40 [mu]Ci [[alpha]-
32
P]UTP (Amersham) in the reaction using Promega enzymes and buffers.
Splicing extracts were made as described by Lin
et al
. (
30
), but were not dialysed prior to immunodepletion. After immunodepletion they
were dialysed into buffer D50 (50 mM KCl, 20 mM HEPES, pH 7.5, 20% glycerol, 1
mM DTT) and flash frozen in liquid nitrogen.
Native gel analysis was as described by Séraphin and Rosbash (
5
). All native gels were prepared with glycerol at a final concentration of 5%.
Splicing reactions (
30
) were incubated for 40 min at 25oC and used m
7
GpppG capped substrates (
29
). For splicing add-back reactions, equivalent volumes of yCBC dialysis buffer (20 mM Tris, pH
7.5, 250 mM sucrose, 50 mM NaCl, 5% glycerol) were added to each reaction to
control for non-specific buffer effects. Spliced products were recovered and resolved by
8% denaturing PAGE.
A 1 ml column of affinity-purified immobilized rabbit antibodies raised against a peptide consisting
of the N-terminal 10 amino acids of yCBP80 (
22
) was loaded with yeast high speed supernatant, prepared as described (
22
) and adjusted to 50 mM HEPES-KOH, pH 7.5, 50 mM Tris-HCl, pH 7.5, 50 mM potassium acetate, 200 mM NaCl, 5 mM [beta]-mercaptoethanol, 10 mg/ml leupeptin, 5 mg/ml each of
chymostatin and elastitinal, 10% glycerol. Bound material was eluted with 1
mg/ml antigenic peptide in 50 mM Tris, pH 7.5, 1 M NaCl at room temperature.
Peak fractions were pooled and further purified on a Superdex 200 column
equilibrated in 50 mM Tris, pH 7.5, 100 mM NaCl, 3 mM mercaptoethanol. Peak
fractions, identified first by protein content and then by SDS-PAGE analysis, were dialysed into 20 mM Tris, pH 7.5, 250 mM sucrose, 50
mM NaCl, 5% glycerol for adding back to depleted extracts.
Immobilized immunopurified anti-yCBP80 antibodies raised against recombinant yCBP80 (
22
) or, as a control, a 1:1 mix of protein A and protein G was pre-washed with an equal volume of yeast splicing extract. For depletion, the
ratio of extract to beads was 3:1. Extracts were depleted in batch with gentle
rotation for 90 min. The beads were then pelleted at 2000
g
for 1 min and the supernatant recovered and stored on ice. The immobilized
antibodies were regenerated by washing twice with 10 vol 100 mM glycine, pH 3,
and the beads were then equilibrated in buffer D250 (250 mM KCl, 20 mM Tris, pH
7.5, 20% glycerol, 1 mM DTT), pelleted and the supernatant aspirated off. The
beads were then washed once with 0.5 vol either depleted or mock depleted
extract and a second round of depletion done for 90 min. The supernatants were
then dialysed against buffer D50. Depletion was checked by Western blotting.
Extracts were resolved by 12.5% SDS-PAGE, blotted onto nitrocellulose and the filter blocked in 1* PBS, 5% dried milk, 1% Triton X-100. The blot was then probed using either anti-yCBP80 or anti-yCBP20 antibodies at 1:5000 dilution in 1* PBS, 1% dried milk, 1% Triton-X100 for 45 min then washed twice
with a large volume of buffer. The secondary antibody was used at a dilution of
1:1000 in the same buffer as above. The blot was developed using ECL (Amersham
International).
Functional yCBC was purified from yeast extracts using an antibody raised
against an N-terminal peptide of yCBP80 (
22
). The peptide-eluted material (Fig.
1
, left panel) shows a very high enrichment in the subunits of yCBC (yCBP80 and
yCBP20) as determined by sequence analysis of the two proteins and by cap-specific binding in a gel mobility shift assay (
22
). Further purification by size exclusion chromatography, in addition to
demonstrating that yCBP80 and yCBP20 co-fractionate, results in fractions containing an almost homogenous
preparation of yCBC (Fig.
1
, right panel).
In mammalian splicing extracts, CBC is required to promote efficient association
of U1 snRNP with the 5'-splice site (
21
). In order to determine whether the function of yCBC in pre-mRNA recognition was conserved, yeast splicing extracts were
immunodepleted using immobilized antiserum raised against recombinant yCBP80
(see Materials and Methods). The levels of depletion were assayed by Western
blot using antiserum raised against either yCBP80 or yCBP20. As can be seen in
Figure
2
, yCBP80 is efficiently depleted from the extract [compare mock depleted (M)
lane 1 with depleted (D) lane 2]. yCBP20 is efficiently co-depleted, as would be expected from its co-fractionation with yCBP80 (compare lanes 1 and 2), reinforcing the
conclusion that, as is the case in human cell extracts (
18
), the greater parts of yCBP80 and yCBP20 are present as a complex.
In mammalian nuclear extracts, depletion of CBC results in inhibition of the
first step of pre-mRNA splicing by ~90%. We were interested to determine whether this was also paralleled
in the yeast splicing extract depleted of yCBC. Uniformly labelled m
7
GpppG-capped actin pre-mRNA was therefore added to either mock depleted or yCBC depleted
extracts. Depletion of yCBC resulted in a strong inhibition of pre-mRNA splicing (Fig.
5
, lanes 1 and 4). Quantitation of this and other experiments has shown that the
inhibition observed was routinely between 70 and 80%. Splicing could be
restored to the depleted extracts by addition of increasing amounts of purified
yCBC (lanes 5-8) to a level similar to the restoration of spliceosome formation (see
above). No general stimulation of splicing was observed upon addition of yCBC
to the mock depleted control reactions.
The experiments presented here show that yCBC is required for efficient
commitment complex assembly, spliceosome assembly and pre-mRNA splicing
in vitro
. The inhibition of commitment complex formation observed in depleted extracts
leads to a strong reduction in spliceosome formation and a reduction in the
levels of splicing observed. These data complement the work by Colot
et al
. (
27
), who showed not only that mutation or deletion of the
MUD13
gene, which encodes Mud13p, the yeast homologue of CBP20, affected splicing
efficiency
in vivo
, but also that extracts prepared from
mud13
mutant strains show impaired ability to form commitment complex (both CC1 and
CC2) and to carry out pre-mRNA splicing. Interestingly, if yCBC is removed either by immunodepletion
(this paper) or genetically by disruption of the
MUD13
gene (
27
), the small amount of commitment complex which is formed has a lower mobility
than the wild-type commitment complex. Since CBC appears to be a component of CC1 and
CC2 (
27
), one might have expected the opposite result, i.e. that complexes formed in
extracts lacking CBC would exhibit a higher mobility. This suggests that the
presence of CBC in the commitment complex may, either directly or indirectly,
induce a conformational change that results in a more compact RNP. Addition of
yCBC to depleted extract both increases the quantity of commitment complexes
formed to ~50% of that seen in mock depleted extract and restores the mobility of
these complexes to that observed in wild-type extracts, demonstrating that both changes are specific results of the
lack of CBC.
As would be expected, the defect observed in commitment complex formation in the
absence of yCBC is reflected in deficiencies in both the formation of
spliceosomes and in pre-mRNA splicing. The effects of yCBC depletion in each case could be
partially (to a level of ~50%) restored by the re-addition of highly purified yCBC. There are two possible reasons for
the incompleteness of the recovery of activity. The first is that we may be
unable to add sufficient yCBC to completely regain activity. This is suggested
by the facts that yCBC is probably rate limiting for CC1 and CC2 formation in
our extracts (Fig.
3
) and that progressive increases in restoration of activity were seen in all
three assays up to the maximal levels of yCBC added back. Unfortunately, for
technical reasons, we were unable to obtain active preparations of yCBC at
higher concentrations. The second possible reason is that we may partially co-deplete a second factor involved in commitment complex formation, such
that this factor becomes limiting when enough yCBC is added to the depleted
extract. As previously mentioned, a fraction of up to ~30% of yCBC is associated with Srp1p, the yeast nuclear protein import
receptor (
22
). However, immunodepletion of Srp1p from splicing extract had no detectable
effect on splicing (data not shown), indicating that this trimeric complex does
not have a specific function in splicing.
The overall conservation of the amino acid sequence of CBP80 and CBP20 between
humans and
S.cerevisiae
(
18
,
22
,
27
) and other higher eukaryotes (our unpublished data) would suggest that the
function of CBC should be conserved. Taken together, the evidence from both
yeast (this paper and
27
) and mammalian systems (
21
) strongly supports the model that CBC binding to the cap structure of a pre-mRNA facilitates an early step in splicing complex assembly, most likely
U1 snRNP association with the 5'-splice site. Although this does not explain how CBC facilitates U1
snRNP association with the 5'-splice site, it does suggest that the mechanism will be conserved.
Further work will be required to determine whether yCBC functions by directly
interacting with a known component of the commitment complex, like the U1
snRNP, or whether additional factors are required to mediate the effect of yCBC
on complex formation.
We thank Puri Fortes, Elisa Izaurralde, Hildur Colot and Michael Rosbash for
discussions and exchange of unpublished information, Chiara Gamberi and
Bertrand Sraphin for criticism of the manuscript, and Zoi Lygerou for help in
preparing splicing extract. J.D.L. was the recipient of an EU HCM fellowship.
REFERENCES
Return