The EMBO Journal vol.10 no.9 pp.2645-2651, 1991
An intact Box C sequence in the U3 snRNA is required
for binding of fibrillarin, the protein common to the
major family of nucleolar snRNPs
Susan J.Baserga, Xiangdong W.Yang and
Department of Molecular Biophysics and Biochemistry. Howard
Hughes Medical Institute, Yale School of Medicine, 333 Cedar St,
New Haven, CT 06511, USA
Communicated by J.A.Steitz
The mammalian U3 snRNP is one member of a recently
described family ofnucleolar snRNPs which also includes
U8, U13, U14, X and Y. All of these snRNPs are
immunoprecipitable by anti-fibrillarin autoantibodies,
suggesting the existence ofa common binding site for the
34 kDa fibrillarin (Fb) protein. Two short nucleotide
sequences, called Boxes C and D, present in each ofthese
RNAs are the most likely sites for fibrillarin binding. We
have developed a HeLa in vitro assembly system for
binding offibrillarin to human U3 snRNA. Reconstitution
of the input RNA is specific in our assay since four of
the other nucleolar small RNAs (U8, U13, X and Y)
which have Boxes C and D become immunoprecipitable
by anti-fibrillarin whereas two RNAs which lack these
sequences (5S and 5.8S) do not. Deletion analyses of the
U3 snRNA demonstrate that the presence of Box C but
not Box D is required for fibrillarin binding. Moreover,
seven single or double site-specific mutations in the U3
Box C abolish binding. The role ofthe Box C-fibriliarin
interaction in the biogenesis of the Fb snRNPs
Key words: fibrillarin/nucleolus/snRNPs/U3 snRNA
U1, U2, U3, U4, U5 and U6 constitute the major small
nuclear ribonucleoprotein particles (snRNPs) in mammalian
cells because of their high abundance (0.2-1 x 106
copies/cell). Each snRNP consists of one highly structured
RNA and as many as 10 different proteins. With the
exception of U3, all of these snRNPs participate in pre-
mRNA splicing and can be found in spliceosomes in the
nucleoplasm. The spliceosomal snRNPs, except for U6, are
recognize a common epitope on several of the particle
proteins (reviewed in Luhrmann, 1988; Zieve and Sauterer,
1990). In contrast, the U3 snRNP participates in rRNA
processing (Kass et al., 1990; Savino and Gerbi, 1990) and
is located in the nucleolus (Nakamura et al., 1968; Weinberg
and Penman, 1968; Prestayko et al., 1970; Tyc and Steitz,
1989). It is immunoprecipitable byantibodies directed against
an abundant 34 kDa nucleolar protein (Lischweetal., 1985)
called fibrillarin, and not by anti-Sm sera. The U3 snRNP
therefore belongs to a distinctly different class than the
spliceosomal snRNPs by virtue of its function, subcellular
location, and protein composition.
The biogenesis of the snRNPs of the Sm class has been
well-studied. The snRNAs, except for U6, possess a short
conserved sequence [PuA(U)nGPu] which was originally
shown to be the binding site for the Sm proteins (B', B, D,
E, F, G) by RNase digestion of native snRNP particles
(Liautard et al., 1982). Specific mutations studied both in
vivo and in vitro confirmed the importance of this sequence
for binding of the Sm proteins (Hernandez and Weiner,
1986; Mattaj, 1986; Hamm et al., 1987; Surowy et al.,
1989). In addition to the common Sm proteins, each snRNP
binds specific proteins. For example, the Ul snRNP contains
three specific proteins called the 70K, A and C. The
arrangement of these proteins in the U1 RNP has been
determined using: (i) in vitro assembly of in vitro synthesized
Ul RNA with HeLa cell extracts (Patton et al., 1987; Patton
and Pederson, 1988; Lutz-Reyermuth and Keene, 1989;
Scherly et al., 1989; Surowy et al., 1989); (ii) in vitro
assembly of in vitro synthesized Ul RNA with Xenopus
oocyte extracts (Hamm et al., 1987, 1988; Scherly et al.,
1989); (iii) direct binding of Ul RNA to U1-specific proteins
synthesized in vitro (Spritz et al., 1987; Query et al., 1989;
Scherly et al., 1989; and Surowy et al., 1989); and (iv)
expression of mutated U1 genes injected into Xenopus
oocytes (Hamm et al., 1990b). Further studies of the
assembly of the Ul snRNP in Xenopus oocytes and in
mammalian cells have localized the steps in its biogenesis
to various cell compartments (Feeney et al, 1989; Fischer
and Liihrmann, 1990; Hamm et al., 1990a; Neuman de
Vegvar and Dahlberg, 1990). After transcription by RNA
polymerase II in the nucleus, Ul RNA quickly exits to the
cytoplasm where it binds pre-assembled Sm proteins and its
7-methyl guanosine cap undergoes modification to become
trimethylated (2,2,7-trimethyl guanosine). Only then is the
nascent Ul particle imported into the nucleus and its
assembly completed by binding ofthe U1-specific proteins.
Import and trimethylation are not affected by deletion of the
binding sites for the specific proteins from U 1 RNA, whereas
the Sm binding site is crucial for these processes.
In contrast to the U1 snRNP, very little is known about
the assembly of the U3 snRNP. The U3 snRNP consists of
one 217 nucleotide-long trimethyl guanosine-capped RNA
and six proteins (74, 59, 34, 30, 13 and 12.5 kDa; Parker
and Steitz, 1987). The 34 kDa protein, fibrillarin, is common
to at least five other nucleolar snRNPs (U8, U13, U14, X
and Y; Tyc and Steitz, 1989; Liu and Maxwell, 1990).
Except for fibrillarin, the protein componentsof these other
nucleolar snRNPs are not yet known. The U3, U8, U13 and
U14 RNAs possess two short nucleotide sequences called
Boxes C and D (see Figure la) which are also conserved
throughout U3 evolution (Wise and Weiner, 1980; Hughes
et al., 1987; Jeppesen et al., 1988; Tyc and Steitz, 1989).
RNAs X and Y also contain Box D and the most highly
conserved first six nucleotides of the Box C sequence
(UGAUGA and UGAAGA, respectively; K.Tyc, personal
communication). Immunoprecipitationwith anti-fibrillarin
S.J.Baserga, X.W.Yang and J.A.Steitz
Y vCUU CUCUCCOU
/C-GGAAAGUGAGGG U U
U - A
m3GpppAAGAC - GAGCACCGAAAA-U-H
3 CTCOTGTCCTTMGGTG 5 oligonucleotide
to residues 64-79
5' PCR oigo
T7prorn. IGGG 31 basesU3
3' PCR oligo
Fig. 1. a. Primary and secondary structure of human U3 RNA. The U3 structure is based on analyses using chemical and enzymatic probes (Parker
and Steitz, 1987). The conserved Box C and Box D regions (Jeppesen et al., 1988) are indicated. A U3 fragment containing nucleotides 80-217
was generated by RNase-H directed digestion of native U3 using the oligonucleotide shown (see text). b. Cloning of the human U3 snRNA for in
vitro synthesis. A human U3 RNA was cloned under the control of a T7 RNA polymerase promoter by PCR. A schematic of the 5' and 3'
oligonucleotides is shown. The 5' oligonucleotide has at its 3' end 31 nucleotides of U3 RNA sequence, preceded by three guanosines to facilitate
the start of T7 transcription, the T7 RNA polymerase promoter, an EcoRI site for cloning, and 6 guanosines to serve as a 'clamp' for EcoRI
cleavage. The 3' oligonucleotide has at its 3' end 44 bases complementary to the U3 RNA preceded by an RsaI site (to yield the exact 3' end of the
U3 RNA), a Sacl site for cloning and six cytidines to serve as a clamp for cleavage by Sacl.
antibodies following RNase digestion of the U3 snRNP has
suggested that Boxes C and D interact (directly or indirectly)
with the fibrillarin protein (Parker and Steitz, 1987).
As the first step toward a better understanding of the
assembly of the U3 snRNP, we have undertaken a mutational
analysis of conserved Boxes C and D to establish their role
in fibrillarin binding. Because U3 and five other nucleolar
snRNAs share sequence similarity, specific protein binding
properties and subcellular location, our results are relevant
to this entire class of snRNPs, which we propose be called
the Fb snRNPs.
Binding of U3 RNA to fibrillarin in HeLa cell
We first tested the ability of fibrillarin to assemble with U3
RNA in vitro using native U3 isolated from HeLa cells and
labeled with 32pCp. After incubation for 45 min at 37°C in
a HeLa whole cell extract (WCE), fibrillarin binding was
determined by immunoprecipitation with anti-fibrillarin
antibodies; since the antibody does not bind the RNA alone
(data not shown; Parker and Steitz, 1987), the amount of
U3 RNA immunoprecipitated is an indication of assembly.
Box C is required for fibrillarin binding to RNA
Fig. 2. Specific binding of fibrillarin to U3 RNA in HeLa whole cell
extracts (WCE). 32pCp-labeled RNAs from HeLa cells were incubated
in HeLa WCE for 45 min at 37°C and then analyzed for fibrillarin
binding by immunoprecipitation with either an anti-fibrillarin serum
(cFb) or normal human serum (NHS) as a control. The pellets (lanes
1-5) and corresponding supernatants (lanes 6-10) were analyzed on
an 8% denaturing polyacrylamide gel. Lanes
precipitated with anti-fibrillarin antibodies. Lanes 2 and 7: U3 RNA
precipitated with NHS. Lanes 3 and 8: 5S RNA precipitated with anti-
fibrillarin antibodies. Lanes 4 and 9: U3 RNA 3' fragment nt 80-217
(see text) precipitated with anti-fibrillarin antibodies. Lanes 5 and 10:
U3 RNA fragment nt 80-217 (see text) precipitated with NHS.
1 and 6: U3 RNA
Fig. 3. Five Fb snRNAs bind fibrillarin in vitro. 32pCp RNAs from
HeLa cells were incubated in HeLa WCE for 45 min at37°C and
then analyzed for fibrillarin binding by immunoprecipitation with anti-
fibrillarin serum. The pellets (lanes 1-7) andcorresponding
supernatants (lanes 8-14) were analyzedon an 8% denaturing
polyacrylamide gel. Lanes 1 and 8: U3 RNA. Lanes 2 and 9: U8
RNA. Lanes 3 and 10: U13 RNA. Lanes 4 and 11: 5.8S rRNA.
Lanes 5 and 12: 5S rRNA. Lanes 6 and 13: RNA X. Lanes 7 and 13:
As shown in Figure 2, lane 1, fibrillarin binds to U3 RNA
added to the HeLa WCE. The interaction is specific since
normal human serum (NHS) does not immunoprecipitate U3
(Figure 2, lane 2) and 32pCp-labeled 5S RNA isolated from
antibodies (Figure 2, lane 3). The percentage reconstitution
(pellet/pellet + supernatant) of U3 RNA varied from 10 to
50% in different experiments.
To determine which part of U3 participates in fibrillarin
binding,ashortened U3 RNA
oligonucleotide-directed RNase H digestion of 32pCp-
labeled full-length U3 isolated from HeLa cells. The
oligonucleotide is complementary to bases 64-79 ofthe U3
RNA (see Figure la); upon digestion with RNase H an RNA
containing the 3' 138 nucleotides ofU3 is created. This U3
fragment, containing conserved Boxes C and D (see
Figure la), binds fibrillarin to approximately the same extent
as the intact full-length U3 (Figure 2, lane 4). This suggests
that the sequences required for fibrillarin binding are
contained in the 3' 138 nucleotides of U3.
Several other types of cell extracts were tested for
fibrillarin binding to U3 RNA. The HeLa WCE was found
to be several-fold more active than HeLa S100 or HeLa
Dignam extract (Dignam et al., 1983; Heintz and Roeder,
1984). Fibrillarin binding in HeLa WCE was comparable
to that observed in mouse ascites S100 (data not shown),
an extract known to be active in the first cleavage step of
rRNA processing (Kass et al., 1990). Reconstitution with
fibrillarin does not require exogenously added ATP or
creatine phosphate (data not shown). The WCE was most
active when the protein concentration exceeded 30 mg/ml
(data not shown).
is not immunoprecipited by anti-fibrillarin
Other Fb snRNAs also bind fibrillarin in vitro
The results shown in Figure 2 demonstrate that U3 binds
fibrillarin in WCE, that this interaction is specific, and that
the 3' 138 residues of U3, which contain Boxes C and D,
are sufficient. Since there are at least five other snRNAs
in mammalian cells which are both immunoprecipitable by
anti-fibrillarin sera and contain Box C and D homologs (Tyc
and Steitz, 1990; K.Tyc, personal communication), we tested
their ability to bind fibrillarin in vitro. Four of these (U8,
U13, X, and Y) and two RNAs which do not bind fibrillarin
(5S, 5.8S) were isolated from HeLa cells, labeled with
32pCp and incubated
precipitation results are shown in Figure 3. All ofthe RNAs
which belong to the Fb snRNP class (U3, U8, U13, X, Y)
become bound by fibrillarin in WCE (lanes 1-3, 6-7) while
neither 5.8S rRNA nor 5S rRNA does (lanes 4 and 5).
in HeLa WCE. The immuno-
Are conserved Boxes C and D in U3 RNA necessary
for fibrillarin binding?
In order to transcribe and assay mutant U3 RNA molecules
for their interaction with the fibrillarin protein, we cloned
a human U3 cDNA under the control of a T7 RNA
polymerase promoter. This was accomplished starting with
purified HeLa U3 RNA using a procedure involving PCR
which has general cloning applications (see Figure lb). The
U3 cDNA clone was designed so that run-offtranscription
of RsaI digested DNA would yield an RNA with a 3' end
identical to that of the wild-type U3 and a 5' end withonly
three additional guanosine residues. Furthermore, after
subcloning into Ml3mpl8 and subsequent mutagenesis, the
S.J.Baserga, X.W.Yang and J.A.Steitz
replicative form M13 DNA can be used directly as a template
for T7 transcription without subcloning back into the parent
To assess the roles ofBoxes C and D in fibrillarin binding,
we first tested deletions of the 3' end of the U3 RNA, as
shown in Figure 4a. We compared RNAs transcribed from
the U3 cDNA clone cut with RsaI (full-length), BstUI (AD),
Fnu4HI (ACD1) or DdeI (ACD2) (see Figure la). Upon
Fig. 4. a. Conserved region Box C is necessary for fibrillarin binding
in vitro. Truncations of the U3 RNA were generated by digestion of
the template DNA with the indicated restriction enzymes. Digestion by
RsaI yields a U3 RNA with the correct 3' end after T7 transcription.
Digestion with BstUI yields an RNA missing Box D (AD) which is
204 nucleotides in length. Digestion with Fnu4HI yields an RNA 138
nucleotides long that is missing Boxes C and D (ACD1). Digestion
with DdeI yields an RNA 104 nucleotides long missing Boxes C and
D (ACD2). There are three extra guanosines at the 5' end of each of
these transcripts. b. The 32pCp RNAs represented in (a) were
incubated in HeLa WCE for 45 min at 37°C and analyzed for binding
to fibrillarin by immunoprecipitation with anti-fibrillarin antibodies
(csFb or NHS). The pellets (lanes 1-5) and corresponding
supernatants (6-10) from the immunoprecipitations were run on an
8% denaturing polyacrylamide gel. The U3 RNA (lane 1) and the U3
RNA lacking Box D (AD; lane 2) bind to fibrillarin whereas two
RNAs which lack both Boxes C and D do not (ACD1, lane 3, and
ACD2, lane 4).
immunoprecipitation after incubation in WCE (Figure 4b),
only the full-length U3 and the U3 AD were observed to
bind fibrillarin (lanes 1-2). The two RNAs with Box C and
Box D deleted (ACD1 and ACD2) did not (lanes 3-4).
These results suggest that Box D is not involved in fibrillarin
binding, whereas Box C remains a candidate.
The possibility that fibrillarin binding is directed by Box
C was further tested by analyzing the effect of specific base
mutations in Box C. Seven U3 RNAs with either single or
double base substitutions (outlined in Table I) were incubated
in WCE and immunoprecipitated
antibodies. Figure 5 shows that while wild-type U3 RNA
binds fibrillarin (lane 1), all seven RNAs with Box C
mutations bind fibrillarin to a much lesser extent (lanes 2-8)
(ranging from 5 to 10% of wild-type binding). This level
is about the same as seen for the background binding of
fibrillarin to U3 RNA upon immunoprecipitation with NHS
(lane 9). We conclude that an intact Box C is required for
fibrillarin binding. In contrast, four single base substitutions
or a single base deletion in Box D and a substitution of the
loop at U3 residues 134-138 (see Figure
and methods) have no effect on fibrillarin binding (data not
1 and Materials
We have developed an in vitro assembly system for binding
of human U3 RNA to fibrillarin using HeLa WCE.
Reconstitution of the input RNA with fibrillarin is specific
in our assay since four other Fb snRNAs (U8, U13, X, Y)
which have Boxes C and D become immunoprecipitable by
anti-fibrillarin serum whereas two RNAs which lack these
sequences (5S and 5.8S) do not. Deletion analyses using
fragments of in vivo or in vitro transcribed U3 RNA
demonstrate that the presence of Box C but not Box D is
required for fibrillarin binding. Moreover, seven single or
double site-specific mutations in Box C abolish binding. We
conclude that an intact Box C is necessary for fibrillarin
binding to U3 snRNA.
Since the Box C sequence is conserved among all known
mammalian Fb snRNPs (Tyc and Steitz, 1989; Liu and
Maxwell, 1990), it seems likely that it plays a similar role
in directing the assembly of the fibrillarin protein into each
of these nucleolar snRNPs. For U3, U8 and U13, computer
modelling suggests that the Box C sequence occurs in a
stem-loop. Thus, these Fb snRNPs share both sequence
identity for Box C and surrounding structural similarity. The
precise length of the Box C sequence in RNAs X and Y
remains to be defined; these RNAs do contain six nucleotides
Table I. Site-directed mutagenesis of Box C in U3 snRNA
Oligonucleotide used for mutagenesis
Box C sequence
Box C is required for fibrillarin binding to RNA
of the Box C
respectively; K.Tyc, personal communication).
evidence that Box D is not necessary for fibrillarin binding
comes from studies of a trypanosome RNA, RNAB, of
unknown function (Hartshorne and Agabian, 1991). This
snRNA bears a putative six nucleotide Box C sequence
immunoprecipitable by anti-fibrillarin sera.
In accord with the conservation of Box C among Fb
snRNAs, the fibrillarin protein is highly conserved among
species. The yeast (Schimmang et al., 1989; Henriquez et
al., 1990), Xenopus (Lapeyre et al., 1990) and human
(Jansen et al., 1991) fibrillarin genes have been cloned.
Human and yeast fibrillarin are 70% identical and 80%
similar while human and Xenopus fibrillarin
identical. Furthermore, human fibrillarin can functionally
complement a yeast fibrillarin mutant (Jansen et al., 1991).
This conservation argues for an important function for
We have shown that Box C is necessary for the binding
of fibrillarin to U3 snRNA. Is the presence of Box C alone
sufficient? We suspect that it may not be and that some aspect
of the 3' stem -loop structure of U3 snRNA may also be
critical. In Figure 2 we observed that shortened U3 RNAs
containing only the 3' 138 residues can bind to fibrillarin
in our in vitro assay. However, a 3' end fragment created
by RNase H digestion of native U3 snRNA that is only 97
residues long (U3 nt 121-217) does not assemble (data not
shown), even though it contains the Box C sequence. This
suggests that the region between nucleotides 80 and 121 of
U3 (see Figure la)
conformation on the Box C sequence which may be
necessary for fibrillarin binding. An alternative possibility
is that fibrillarin interacts directly with other parts of the U3
snRNA and that these sequences stabilize
However, such sequences might be expected to also appear
in the other Fb snRNAs, whose conservation is limited to
Boxes C and D.
Does fibrillarin bind directly to Box C? In our assay we
cannot distinguish between direct binding of fibrillarin to
Box C and indirect binding through another Fb snRNP
protein. There are at least five other proteins in the U3
snRNP, some of which may be common to the other Fb
snRNPs. The observation that U3 immunoprecipitability by
anti-fibrillarin antibodies is reduced in 0.5 M NaCl (Parker
and Steitz, 1987; Tyc and Steitz, 1989) suggests that the
sequence (UGAUGA and UGAAGA,
are 81 %
is required to confer a particular
interaction may indeed be indirect, but U8 and U13 remain
immunoprecipitable at the same salt concentration, arguing
against this explanation. Experiments testing association of
the U3 RNA with fibrillarin in the absence of other proteins
will be necessary to resolve this question.
interactions have tested large deletions or substitutions in the
RNA instead of point mutations (for reviews, see Luhrmann,
1988; Zieve and Sauterer, 1990). In the studies where point
mutations in the RNA have been analyzed, different proteins
have exhibited quite different sensitivities to mutation in their
binding sites. For example, Scherly et al. (1990) studied the
binding of the U lA and U2B" proteins to their respective
RNAs, Ul and U2, and found that exchange of only two
nucleotides between the two RNAs reverses the binding
specificity of these proteins. Yuo and Weiner (1989) found
that two single nucleotide substitutions in stem-loop I of
human U1 RNA decreased binding of the U1-specific
proteins. In contrast, Jones and Guthrie (1990) demonstrated
by saturation mutagenesis of the Sm binding site in yeast
U5 that function of the Sm site is relatively insensitive to
mutation. Our studies on the binding of fibrillarin to the U3
snRNA suggest that Box C is highly sensitive to mutation
since a single nucleotide change can diminish binding.
Although we have defined the minimum length of the Box
C sequence necessary for fibrillarin binding, we have not
defined the precise boundaries of the required nucleotides
nor have we ruled out the possibility that in other Fb snRNPs
a Box C sequence smaller than the nine nucleotide U3 Box
C can suffice. Our finding that Box C is sensitive to mutation
is supported by elegant studies on the essential elements for
U14 snRNA function in yeast (Jarmolowski et al., 1990).
Nitrous acid mutations
distributed over three domains including Boxes C and D and
a third region complementary to 18S rRNA.
The Sm proteins, which bind to
[PuA(U)nGPu] in spliceosomal snRNAs, are said to be the
'core proteins' of the Sm snRNPs. 'Core' in this case refers
to: (i) proteins which form a precursor particle in the absence
of an snRNA (Fisher et al., 1985); (ii) proteins which
assemble onto the RNA in the cytoplasm (reviewed in
Mattaj, 1988); and (iii) proteins which are common to a
family of snRNPs. In addition, the Sm proteins play an active
role in directing the cap trimethylation and nuclear targeting
of the Sm snRNPs. Whether fibrillarin conforms to the
definition of a 'core' protein for the Fb snRNPs remains to
in defective U14 snRNAs
a short sequence
Fig. 5. An intact box C is required for fibrillarin binding to U3. 32pCp-labeled wild-type U3 and seven U3 RNAs with single or double base
mutations in Box C (see Table I)were incubated in HeLa WCE for 45 min at37°Cand analyzed for binding to fibrillarin by immunoprecipitation
with anti-fibrillarin antibodies (aFb). The pellets (lanes 1-9) and corresponding supernatants (10-18) were analyzed on an 8% denaturing
polyacrylamide gel. The bindingof the mutated U3 RNAs to fibrillarin (lanes 2-8) was 5-10% of the binding of the U3 RNA (lane 1).
Background binding (lane 9) with NHS was -5% of the binding of U3 RNA with anti-fibrillarin.
S.J.Baserga, X.W.Yang and J.A.Steitz
be seen.Althoughit is common to six nucleolar snRNPs,
it is notyetknown whether it pre-assembles with other
proteinsto form aprecursor particlethat then associates with
these snRNAs in thecytoplasm.The results reported here
are aprerequisiteto studies that address the necessity of
fibrillarinbindingforcap trimethylationof U3, U8 and U 13
snRNAs and for the nucleolar localization of all the Fb
Materials and methods
Boehringer-Mannheim, New England
polymerasefrom Perkin-Elmer Cetus.Oligonuclotidesweresynthesized
on anApplied Biosystems oligonucleotide synthesizer by Dr John Flory,
YaleUniversity School of Medicine.
Biolabs and Pharmacia;
Sera frompatientswith scleroderma containinganti-fibrillarin antibodies
were obtained from DrJoseph Craft, YaleUniversitySchool of Medicine.
# 1875 and #1746, which arespecificfor fibrillarin by RNA and protein
immunoprecipitation,were used in thisstudy. Normal human serum was
generouslydonatedbyScott Seiwert in ourlaboratory.
Purification of RNAs from HeLa cells
The U3, U8, U13, X and Y RNAs,
immunoprecipitable byanti-fibrillarin sera, werepurified from HeLa cells
and32pCp-labeled accordingto standardprocedures (Tyc and Steitz, 1989).
A U3 RNAcontaining onlynucleotides 80-217(the3' end) was created
in vitrobyRNase H-directeddigestionof32pCp-labeled U3 RNA using
anoligonucleotide complementaryto nucleotides 64-79 (see Figure la;
Kass et al., 1990). RNAs and U3 RNA fragments were purified on
denaturing polyacrylamide gelsfor use in the assembly assay.
all of which
Extractpreparationand fibrillarin binding assay
Whole cell extracts (WCE) were prepared from HeLa cells using the
procedureofManleyet al.(1980).At the finalstep, the extract was dialyzed
into 20 mM HEPESpH 7.9,50 mMKCI, 0.1 mM EDTA, 2 mM MgCl2,
17%glyceroland 2 mM DTT.
RNA was assembled with fibrillarin in HeLa WCE. Generally, 5000
c.p.m.ofpurifiedU3 RNApreparedas outlined above was added to 15 al
water, givingaproportionof 60% extract in the reconstitution mixture.
The RNA dissolved in water was heated to 90°C and cooled slowly over
15 min before addition. The reaction was incubated at 30°C for 45 min
and then theinputU3 RNA was assayed for binding to fibrillarin by
immunoprecipitationwith anti-fibrillarin antibodies.
Il RNasin. The volume was brought to 25/lwith
Either sera(5 1d)from sclerodermapatients containing antibodies to fibrillarin
ornormal human serum was added to 2.5 mg protein A-Sepharose (PAS)
in NET-2 (150 mM NaCl, 50 mM Tris pH 7.5, 0.05% NP-40) and
washed fourtimes, and then the reconstitution reaction was added to the
antibody-PAS pelletin 0.5 ml NET-2 and nutated for 2 h at 4°C. The
immunoprecipitateswereprocessed according to standard procedures (Steitz,
1989)and the RNA recovered from bothpellets and supernatants analyzed
on 8% denaturing polyacrylamide gels. Results were quantified on a
1 h at roomtemperature. The antibody-PAS pellets were
Cloningof the human U3 cDNA for in vitro RNA synthesis
The human U3 RNA was cloned behind a T7 polymerase promoter using
thepolymerasechain reaction(PCR)as follows. Two deoxyoligonucleotides
correspondingto the 5' and 3' ends of the U3 RNA were synthesized (see
Figure lb). The 3'oligonucleotide has 44 nucleotides complementary to
the U3 RNA 3' endpreceded byan RsaI site (four nucleotides) and a SacI
site(six nucleotides). The 5'oligonucleotide contains 31 nucleotides of U3
sequence preceded bya T7promoter (17 nucleotides), and an EcoRI site
(six nucleotides).At their extreme 5' termini the oligonucleotides end with
either sixcytidinesor sixguanosines, respectively, as a 'clamp' to facilitate
cleavage bythe restriction endonucleases. The EcoRI and Sacl sites are
forcloningthe PCRfragment; digestionat theRsaIsitegivesa U3 RNA
with a correct 3' endupon transcription.Threeguanosineswere included
between the T7promoterand the in vivo U3 RNAsequencetoprovide
astrong trancriptionalstart site for the T7polymerase.
U3 RNA was extracted from
precipitation (Tycand Steitz, 1989), gel purified,and used for first strand
synthesis. The U3 RNApelletand 100pmol3' endprimerwere annealed
in a total volume of5yl. TheRNA-oligonucleotidemixture was heated
to 90°C for 2 min and then cooledslowlyfor 45 min. Reversetranscription
wasperformedin 50 mM TrispH 8.3,6 mMMgCl2,40 mMKCI, 1 mM
DTT, 1 mMdeoxynucleotidesand 13 units reversetranscriptase (Pharmacia)
for 1 h at 37°C. One-fourth of this reaction was then used as thetemplate
forPCR, performed accordingtoToczyskiand Steitz(1991). Thirty cycles
of PCR were carried out on a Perkin-Elmer Cetusthermocycler. The
annealing was at 60°C, extension at 72°C and denaturation at 94°C.
The PCR reaction was ethanol precipitated and fractionated on an 8%
denaturing polyacrylamide gel.Thesingle correctly-sizedband was excised,
eluted, and then cut with EcoRI and Sacl. The cut PCRproductwas cloned
into the EcoRI-SacI site ofpSP64 (Promega Biotech). The entire insert
was sequenced and found to be correct. Run-offtrancriptionafterdigestion
with RsaIgives a 3' end identical to the in vivo U3 RNA but with 3 additional
guanosines at the 5' end.
gof HeLa cells by anti-fibrillarin
The EcoRI-HindIII fragmentof the U3pSP64 constuct was subcloned into
for site-directedmutagenesis. Mutagenesis wasperformed according to the
method of Zoller and Smith (1983) as modified by Kunkel (1987). Seven
single or double base substitutions in the U3 Box C region and the
oligonucleotides that were used in the synthesis are depicted in Table I.
Using comparable methods, four Box D mutations (U208C, C209A, U2OC,
G211U) as well as a deletion/substitution mutation of the U3 loop at
nucleotides 134-138 (UUGGC to ACACA) were also made (see Figure la).
Potential mutations were screened by direct DNA sequencing; the mutation
rate varied from 10-90%. All clones were sequenced in their entirety.
Replicative form DNA was prepared according to standard procedures
(Maniatis et al., 1982).
sites of M13 mpl8 togenerate asingle-stranded template
Togenerate full-length U3 RNAs, 5 yg of either plasmid or RF DNA were
digestedwith RsaI. Forgeneration of 3' end deletions, DNA was cut with
BstUI, Fnu4HI or DdeI (see Figures
performed according to Melton et al.(l984) using 1 ,ug of DNA template.
No radioactive nucleotides were used. After digestion with DNase the RNA
transcript was purified over a G-50 spin column. The RNA was labeled
with 32pCp in 50 mM HEPES pH 8.3, 20 mM MgCl2, 3 mM DTT,
400tiM ATP, 40/ACi 32pCp (3000 Ci/mmol) and 70 units of RNA ligase
for 1 h at 370C. The RNA was ethanol precipitated and run on an 8%
denaturing polyacrymide gel. RNA was extracted from the gel by standard
procedures (Maniatis et al., 1982) and then used in assembly reactions.
1 and 4a). Transcriptions were
We thank Michelle Caizergues-Ferrer, Peter Glazer, Gregg Morin, David
Toczyski, KazioTyc and David Wassarman for helpful discussions, Karen
Montzka for the initial HeLa whole cell extracts and Joe Craft for anti-
fibrillarin antibodies. We thank David Tollervey for generously sharing
results before publication. This research was supported by grant GM26154
from the National Institutes of Health. S.J.B. is a fellow of the Leukemia
Society of America.
Dignam,J.D., Lebovitz,R.M. and Roeder,R.G. (1983) Nucleic Acids Res.,
Feeney,R.J., Sauterer,R.A., Feeney,J.L. and Zieve,G.W. (1989) J. Biol.
Chem., 264, 5776-5783.
Fischer,U. and Luhrmann,R. (1990) Science, 249, 786-790.
Fisher,D.E., Conner,G.E., Reeves,W.H., Wisniewolski,R. and Blobel,G.
(1985) Cell, 42, 751-758.
Hamm,J., Darzynkiewicz,E., Tahara,S.M. and Mattaj,I.W. (1990a) Cell,
Hamm,J., Dathan,N.A., Scherly,D. and Mattaj,I.W. (1990b) EMBO J.,
Hamm,J., Kazmaier,M. and Mattaj,I.W. (1987) EMBO J., 6, 3479-3485.
Hamm,J., van Santen,V.L., Spritz,R.A. and Mattaj,I.W. (1988) Mo. Cell.
Biol., 8, 4787-4791.
Hartshorne,T. and Agabian,N. (1991) J. Cell. Biochem., suppl. 15D, 67.
Heintz,N. and Roeder,R.G. (1984) Proc. NatI. Acad. Sci. USA, 81,
Henriquez,R., Blobel,G. and Aris,J.P. (1990)
J. Biol. Chem., 265,
Box C is required for fibrillarin binding to RNA
Hernandez,N. and Weiner,A. (1986) Cell, 47, 249-258.
Hughes,J.M.X., Konings,D.A.M. and Cesareni,G. (1987) EMBO J., 6,
Jansen,R.P., Hurt,E.C., Kern,H., Lehtonen,H., Carmo-Fonseca,M.,
Lapeyre,B. and Tollervey,D. (1991) J. Cell Biol., 113, 715-729.
Jarmolowski,A., Zagorski,J., Li,H. and Fournier,M.J. (1990) EMBO J.,
Jeppesen,C., Stebbins-Boaz,B. and Gerbi,S.A. (1988) Nucleic Acids Res.,
Jones,M.H. and Guthrie,C. (1990) EMBO J., 9, 2555-2561.
Steitz,J.A. and Sollner-Webb,B.
Kunkel,T.A., Roberts,J.D. and Zakour,R.A. (1987) Methods Enzymol.,
Lapeyre,B., Mariottini,P., Mathieu,C., Ferrer,P., Amaldi,F., Amalric,F.
and Caizergues-Ferrer,M. (1990) Mol. Cell. Biol., 10, 430-434.
Liautard,J.P., Sri-Widada,J., Brunel,C. and Jeanteur,P. (1982) . Mol. Biol.,
Lischwe,M.A., Ochs,R.L., Reddy,R., Cook,R.G., Yeoman,L.C. Tan,
Liu,J. and Maxwell,E.S. (1990) Nucleic Acids Res., 18, 6565-6571.
Liimnn,R. (1988) In Bimstiel,M.L. (ed.), Structure and Function ofMajor
and Minor Small Nuclear Ribonucleoprotein Particles. Springer Verlag,
Heidelberg, pp. 71 -99.
Lutz-Reyermuth,C. and Keene,J.D. (1989) Mol. Cell. Biol., 9, 2975 -2982.
Maniatis,T., Fritsch,E.F. and Sambrook,J. (1982) Molecular Cloning: A
Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring
Manley,J.L., Fire,A., Cano,A., Sharp,P.A. and Gefter,M.L. (1980) Proc.
Natl. Acad. Sci. USA, 77, 3855-3859.
Mattaj,I.W. (1988) In Birnstiel,M.L. (ed.), Structure and Function ofMajor
and Minor Small Nuclear Ribonucleoprotein Particles. Springer Verlag,
Heidelberg, pp. 100-114.
Melton,D.A., Krieg,P.A., Rebagliati,M.R., Maniatis,T., Zinn,K. and
Green,M.R. (1984) Nucleic Acids Res., 12, 7035-7056.
Nakamura,T., Prestayko,A.N. and Busch,H. (1968) J. Biol. Chem., 243,
Neuman de Vegvar,H.E. and Dahlberg,J.E. (1990) Mol. Cell. Biol., 10,
Parker,K.A. and Steitz,J.A. (1987) Mol. Cell. Biol., 7, 2899-2913.
Patton,J.R. and Pederson,T. (1988) Proc. Natl. Acad. Sci. USA, 85,
Patton,J.R., Patterson,R.J. and Pederson,T. (1987) Mol. Cell. Biol., 7,
Prestayko,A.W., Tonato,M. and Busch,H. (1970) J. Mol. Biol., 47,
Query,C.C., Bentley,R.C. and Keene,J.D. (1989) Mol. Cell. Biol., 9,
Savino,R. and Gerbi,S.A. (1990) EMBO J., 9, 2299-2308.
Scherly,D., Boelens,W., van Venrooij,W.J., Dathan,N.A., Hamm,J. and
Mattaj,I.W. (1989) EMBO J., 8, 4163-4170.
Scherly,D., Boelens,W., Dathan,N.A., van Venrooij,W.J. and Mattaj,I.W.
(1990) Nature, 345, 502-506.
Schimmang,T., Tollervey,D., Kern,H., Frank,R. and Hurt,E.C. (1989)
EMBO J., 8, 4015-4024.
Strunk,K., Surowy,C.S., Hoch,S.O., Barton,D.E. and
Francke,U. (1987) Nucleic Acids Res., 15, 10373- 10391.
Steitz,J.A. (1989) Methods Enzymol., 180, 468-481.
Surowy,C.S., van Santen,V.L., Scheib-Wixted,S.M. and Spritz,R.A. (1989)
Mol. Cell. Biol., 9, 4179-4186.
Toczyski,D.P.W. and Steitz,J.A. (1991) EMBO J., 10, 459-466.
Tyc,K. and Steitz,J.A. (1989) EMBO J., 8, 3113-3119.
Weinberg,R.A. and Penman,S. (1968) J. Mol. Biol., 38, 289-304.
Wise,J.A. and Weiner,A.M. (1980) Cell, 22, 109-118.
Yuo,C.Y. and Weiner,A.M. (1989) Genes Dev., 3, 697-707.
Zieve,G.W. and Sauterer,R.A. (1990) Crit. Rev. Biochem. Mol. Biol., 25,
Zoller,M.J. and Smith,M. (1983) Methods Enzvmol., 100, 468-500.
Received on April 19, 1991