Metabolism of pre-messenger RNA splicing cofactors: modification of U6 RNA is dependent on
its interaction with U4 RNA
Metabolism of pre-messenger RNA splicing cofactors: modification of U6 RNA is dependent on its interaction with U4 RNA
Dennis B.
Zerby
and
Jeffrey R.
Patton*
Department of Pathology, School of Medicine, University of South Carolina,
Columbia
, SC 29208,
USA
Received May 13, 1996;
Revised and Accepted August 5, 1996
ABSTRACT
The requirements for the formation of pseudouridine (
[Psi]
) in U4 and U6 RNAs, cofactors in the splicing of pre-messenger RNA, were investigated
in vitro
using HeLa nuclear (NE) and cytoplasmic (S100) extracts. Maximal
[Psi]
formation for both RNAs was extract order-dependent. Maximal
[Psi]
formation in U4 RNA required incubation in S100 followed by the addition of NE,
paralleling the
in vivo
maturation pathway of U4 RNA. In contrast, maximal formation of
[Psi]
in U6 RNA required incubation in NE followed by the addition of S100 extract.
Since U6 RNA does not exit the nucleus
in vivo
the contribution of S100 was investigated. In experiments where the extracts
were treated with micrococcal nuclease to digest endogenous snRNAs, the
efficient formation of
[Psi]
in U6 RNA was dependent on the presence of U4 RNA, but not in U5 RNA or tRNA.
When mutant U4 RNAs that inhibit or strengthen the interaction between U4 RNA
and U6 RNA were substituted for wild-type U4 RNA, the results confirmed the need for the interaction between
these two RNAs for
[Psi]
formation in U6 RNA. U6 RNA isolated from glycerol gradients after incubation
in extracts had four times as much
[Psi]
when associated with U4 RNA.
INTRODUCTION
Small nuclear ribonucleoprotein particles (snRNPs) are essential cofactors in
the splicing of premessenger RNA (pre-mRNA;
1
). The snRNPs are composed of highly conserved small nuclear RNAs (snRNAs), U1,
U2, U4, U5 and U6, and associated proteins. A common core of proteins, the Sm
proteins, are found in all the snRNPs, while additional proteins are specific
to particular snRNPs (
2
,
3
). The snRNPs involved in splicing are highly modified, and these modifications
include base methylations, 3'-end processing, 5'-end capping and pseudouridine ([Psi]) formation (
4
,
5
). U4 and U6 RNAs are extensively base paired and form one snRNP which enters
the spliceosome complexed with U5 snRNP as part of a tri-snRNP. This interaction between U4 and U6 RNA is essential for
spliceosomal assembly and subsequent spliceosomal function (
6
-
8
). Several lines of evidence indicate U6 RNA is a catalytic component of the
spliceosome, while U4 RNA appears to function in ferrying U6 into the
spliceosome, and perhaps in keeping U6 RNA inactive until needed in splicing (
9
).
The formation of [Psi] in these snRNAs has been the subject of several reports. Using
in vitro
transcribed snRNAs and extracts from HeLa cells, there is evidence for multiple
[Psi] synthase activities that specifically recognize U1, U2 and U5 snRNAs (
10
,
11
). In addition, for U2 snRNA which contains 13 [Psi] residues, the formation of [Psi] is not dependent upon earlier [Psi] formation at another site (
11
). U5 snRNA contains three [Psi] residues but only two sites were modified in HeLa S100 (essentially
cytoplasmic) extracts. Modification at the third site required HeLa nuclear
extract (NE), again suggesting multiple [Psi] synthase activities for a single snRNA (
12
). Efficient [Psi] modification at all sites in U5 snRNA requires Sm protein binding, while [Psi] modification at some sites in U2 RNA does not require particle
assembly (
11
,
13
).
The function of [Psi] in snRNAs is unknown, but it is important to note that [Psi] is found in regions of snRNAs that are necessary for snRNP function
in the splicing of pre-mRNA (
14
,
15
). When U2 snRNA that contains no modifications was added to U2-depleted extracts, pre-mRNA splicing was not restored, however, U2 RNA isolated from HeLa
cells was able to restore splicing. When either unmodified or fully modified U5
snRNA was added to U5-depleted extracts, splicing was restored (
16
). This difference may be due to the fact that U2 snRNP contains a large
percentage of [Psi] in contrast to U5 snRNP, and its absence may have a greater impact on the
function of U2 snRNP than on the function of U5 snRNP (
16
).
In tRNA, [Psi] appears to be required for the efficient reading of codons during the
translation process (
17
).
HisT
in
Escherichia coli
codes for a [Psi] synthase that modifies certain positions in the anticodon of tRNAs. The
hisT
gene product is necessary for normal growth of
E.coli
on minimal media since the
hisT
mutant has excessive need for uracil that interferes with cell division (
18
,
19
). Most [Psi] residues in rRNAs are found near the functional centers of the ribosome (
20
,
21
). Recently, it has been shown that acetylated [Psi], but not acetylated uridine, can transfer an acetyl group to the N-terminus of a peptide, suggesting that [Psi] participates in the acyl transfer reaction in the ribosome (
22
).
In this report, the formation of [Psi] in U4 and U6 RNA was studied and was found to be HeLa extract order-dependent, with at least two [Psi] synthase activities required for [Psi] formation in U4 RNA. In addition, the efficient formation
of [Psi] in U6 RNA is dependent on its interaction with U4 RNA.
MATERIALS AND METHODS
SP6 transcriptions of the
Dra
I-cut human U4 RNA and U4 RNA mutants pSP6-U4, pSP6-U4 [Delta]StemI (deleting nucleotides 56-63), pSP6-U4 [Delta]StemII (deleting nucleotides 1-16), pSP6-U4 [Delta]5'Stem-loop (deleting nucleotides 19-55) and of the
Xba
I-cut human U4 mutant pSP6-U4 [Delta]Sm (deleting nucleotides 91-145) were performed as described (
10
-
12
). The U4 clones were a generous gift from Albrecht Bindereif, Humbolt
University, Germany (
6
). Human U6 RNA was transcribed
in vitro
using T7 polymerase and
Dra
I-cut pHU6-1 as described (
23
) and human U5 RNA was transcribed using SP6 RNA polymerase and
Bfa
I-cut pHU5a2 (
12
). Pre-tRNA
Ser
was transcribed with T7 RNA polymerase and
Ava
I-cut pUC19pSer and was a gift from C. Guerrier-Takada and S. Altman, Yale University (unpublished). The
in vitro
transcription reactions contained, depending upon the application, [[alpha]-
32
P]UTP (50 [mu]Ci, 800 Ci/mmol) or [5-
3
H]UTP (1-50 [mu]Ci, 17 Ci/mmol), 50 [mu]M GTP, 250 [mu]M ATP and CTP. When U4 RNA was synthesized, m7GpppG was
included in the reaction mixture at 1 mM. Low specific activity [5-
3
H]UTP labeled RNAs were made with 50 [mu]M UTP and 1 [mu]Ci [5-
3
H]UTP to facilitate the determination of the amount of RNA synthesized. RNAs to
be used as substrates in the
3
H release/Norit A charcoal binding assay (
19
) were synthesized in the presence of 50 [mu]Ci [5-
3
H]UTP with no additional UTP added.
The
in vitro
modification reactions were carried out as previously described in 300 [mu]l total volume using HeLa S100 and NE extracts (
10
-
12
,
24
). Briefly, the reaction mixture contained 60% HeLa S100 and/or NE extract (
25
) by volume, 0.5 mM ATP, 20 mM creatine phosphate, 3.2 mM MgCl
2
and 2 mM dithiothreitol. RNAs were incubated 30 min at 37oC in either HeLa S100 or NE or a combination of the two extracts. That
incubation was followed by the addition of another aliquot of reaction/extract
mix (see particular experiments for details), and incubated for an additional
2.5 h at 37oC. Extracts to be micrococcal nuclease (MN) treated also contained 1 mM
CaCl
2
and were treated with 1 U/[mu]l MN for 30 min at 37oC. The MN was subsequently inhibited by adding EGTA to 10 mM and poly
A/poly C to a final concentration of 1 mg/ml, prior to the addition of
32
P-labeled RNA.
32
P-labeled RNA was purified on a 10% polyacrylamide/8.3 M urea gel after incubation in the reactions. To determine site specific [Psi] formation, the gel-purified,
32
P-labeled RNA was RNase T1 digested, electrophoresed, and the fragments
eluted from a 20% polyacrylamide/8.3 M urea gel. The fragments were nuclease P1
digested and analyzed by thin layer chromatography (TLC) on cellulose plates in
2-propanol:concentrated HCl:water (70:15:15 v/v/v) (
10
-
12
,
24
). For total [Psi], gel purified RNA was subjected directly to nuclease P1 digestion and TLC.
The formation of [Psi] on [5-
3
H]UTP labeled RNA was assayed in a
3
H release/Norit A charcoal binding assay, where the release of
3
H to solvent water is an indication of [Psi] formation (
12
,
19
).
Glycerol gradients (10-30%) were prepared and centrifuged at 4oC in an SW41 rotor for 18 h at 40 000 r.p.m. (
12
). The buffer for the gradients contained 150 mM KCl, 20 mM Tris-HCl (pH 7.6), 1.5 mM MgCl
2
, 0.5 mM phenylmethylsulfonyl fluoride and 0.5 mM dithiothreitol. After
centrifugation the gradients were fractionated from the bottom.
RESULTS
HeLa extract order-dependence for
[Psi]
formation
U4 and U6 RNAs contain three [Psi] residues each, located at positions 4, 72 and 79 in human U4 RNA, and at
positions 31, 40 and 86 in U6 RNA (Fig.
1
) (
3
). In order to determine the conditions necessary for maximum
in vitro
[Psi] formation, the extract-order dependence for [Psi] formation in U4 RNA was investigated. [5-
3
H]UTP labeled U4 RNA was incubated in HeLa NE or S100 extract or a combination
of the extracts, followed by another aliquot of NE or S100 or a combination of
the two extracts, and further incubation. The amount of
3
H that does not bind to Norit A charcoal,
3
H released from uridine to the bulk solvent when the C[middot]C glycosidic bond is formed, is a function of [Psi] formation (
10
,
12
,
19
). The highest levels of [Psi] formation in U4 RNA were observed for incubation in S100 followed by
incubation in NE, and S100 incubation followed by the addition of another
aliquot of S100, at 19.0 and 15.3%, respectively (Table
1
). The fact that the highest amount of [Psi] was observed with HeLa cytoplasmic extract followed by nuclear extract was
expected given that after U4 RNA is transcribed it exits into the cytoplasm, is
assembled into an RNP and modified, and then enters the nucleus, where it may
undergo further modification before being incorporated into the spliceosome.
Requirement of U4 RNA for
[Psi]
formation on U6 RNA
In order to determine if U6 RNA requires the interaction of U4 RNA for [Psi] formation, it was necessary to eliminate endogenous snRNAs from the
extracts using micrococcal nuclease (MN). MN is a non-specific nuclease that digests all RNAs in the extracts. By eliminating
the endogenous snRNAs from the extracts we can determine if U4 RNA that we add
back to the extract contributes to [Psi] formation on U6 RNA. MN at a final concentration of 1 U/[mu]l was added to the extracts, incubated for 30 min at 37oC, and then the MN was inhibited with EGTA. Poly A and poly C RNA
were also added to counter the effects of substrate masking (
27
). Then [
32
P]U6 RNA was added with or without increasing amounts of [5-
3
H]U4 RNA to the MN treated NE, incubated for 30 min, followed by the addition of
MN treated S100, and incubated for an additional 2.5 h. The isolated [
32
P]U6 RNA was gel purified, subjected to a nuclease P1 digestion, and
chromatographed on TLC plates to determine total [Psi] formation (Fig.
2
). With no U4 present in the extracts the amount of [Psi] found in the U6 RNA is quite low (0.3% of theoretical; Table
3
), but as the amount of
3
H-labeled U4 RNA added to the reaction was increased, the formation of [Psi] in U6 RNA also increased to a high of 8.2% when a 25* molar excess of U4 RNA was added to the treated extracts
(Table
3
).
DISCUSSION
The requirements and conditions for [Psi] formation in the snRNAs involved in splicing are beginning to be
elucidated. For U4 RNA, optimal [Psi] formation is observed when U4 RNA is first incubated in S100 followed by
the addition of NE. This is reminiscent of the metabolism of U4
in vivo
, since after transcription U4 RNA exits the nucleus, is assembled into a
ribonucleoprotein particle, and the 5' cap hypermethylated, before returning to the nucleus. The fact that
optimal [Psi] formation for U4 RNA is not observed with the opposite extract incubation
condition (NE followed by S100), suggests that U4 RNA needs to form an snRNP
before [Psi] formation occurs, that the Sm proteins may be necessary for efficient [Psi] formation. The formation of [Psi] in U5 RNA
in vitro
was shown to be dependent on RNP formation, since U5 mutant RNA that does not
contain an Sm binding site does not form [Psi] (
12
,
28
). Interestingly, the formation of [Psi] in a U4 [Delta]Sm mutant was 25% of wild-type U4 RNA so it would appear some [Psi] formation can occur in U4 RNA in the absence of Sm
protein binding (Zerby and Patton, unpublished data).
The fact that a combination of NE and S100 extracts is needed for optimal [Psi] formation in U4 RNA, together with the fact that [Psi] formation at position 4 of U4 RNA is increased 4-fold by the addition of NE (after incubating in S100 first),
suggests there are at least two [Psi] synthase activities required for [Psi] formation in U4 snRNA, with one of the activities probably located
within the nucleus. The procedure used to isolate HeLa S100 and NE (
25
) makes it more likely to have leakage of nuclear components into the cytoplasm
than to have leakage of cytoplasmic components into the nucleus. Another
possibility is that a cofactor necessary for [Psi] synthase activity is enriched in the nuclear extract. Alternatively, it is
possible that an inhibitor of [Psi] formation in U4 RNA is preferentially found in the S100 and incubation
with NE releases that inhibition. The determination of the actual number of [Psi] synthase activities required for [Psi] formation in both U4 and U6 RNAs, as well as the compartmentalization
of the activities, awaits future experiments that will employ both point
mutants of the two RNAs and microinjection techniques.
[Psi] formation in U6 RNA is optimal with incubation in nuclear extract followed
by the addition of S100. This was surprising since U6 RNA does not exit the
nucleus
in vivo
. Since the [Psi] formation in U6 RNA is low, we will need to consider additional
manipulations of the
in vitro
system components to boost the level of [Psi] formation in this essential splicing cofactor.
Recently it was shown that certain small nucleolar RNAs (snoRNAs) function as
guide RNAs in the ribose methylation of pre-rRNA (
29
). However, even though several tRNA [Psi] synthases have been cloned (
30
-
32
), the requirement for the interaction of one RNA with another in order for [Psi] formation to occur has not been suggested. The need for U6 RNA to interact
with U4 RNA would imply the substrate for [Psi] formation in U6 RNA is the U4-U6 snRNP. Which is intriguing in light of the fact that U6 snRNA
appears to be a catalytic component of the spliceosome, that U4 snRNP may
function as an inhibitor of U6 snRNA catalytic function (
9
), and the possible involvement of [Psi] in chemical reactions (
20
,
21
). The interaction between these two RNAs might be a possible way of regulating
both [Psi] formation and U6 RNA catalytic function by not allowing [Psi] formation in U6 RNA until associated with U4 RNA.
ACKNOWLEDGEMENTS
The authors would like to thank Drs Marty Jacobson and Thoru Pederson (Worcester
Foundation, Shrewsbury, MA) for the pHU6-1 plasmid, Drs C. Guerrier-Takada and Sidney Altman (Yale University) for the pUC19pSer
plasmid, and Dr Albrecht Bindereif (Humbolt University, Germany) for the U4
wild-type and mutant plasmids. This work was supported by the Stefan Mironescu
Grant for Research from the University of South Carolina School of Medicine.
REFERENCES
1 Baserga,S.J. and Steitz,J.A. (1993) In Gesteland,R.F. and Atkins,J.F. (eds), The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, Plainview, NY, pp. 359-381.
3 Reddy,R. and Busch,H. (1988) In Birnstiel,M.L.(ed.), Structure and Function of Major and Minor Small Nuclear Ribonucleoprotein Particles. Springer-Verlag, Heidleberg, pp.1-37.
9 Moore,M.J., Query,C.C. and Sharp,P.A. (1993) In Gesteland,R.F. and Atkins,J.F .(eds), The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, Plainview, NY, pp. 303-357.
14 Steitz,J.A., Black,D.L., Gerke,V., Parker,K.A., Kramer,A., Frendewey,D. and Keller,W. (1988) In Birnstiel,M.L. (ed.), Structure and Function of Major and Minor Small Nuclear Ribonucleoprotein Particles. Springer-Verlag, Heidleberg, pp. 115-154.
15 Ares,M., Jr and Weiser,B. (1995) Prog. Nucl. Acid. Res. Mol. Biol., 50, 131-159.
