ABSTRACT
Binding of the signal recognition particle (SRP) to signal sequences during
translation leads to an inhibition of polypeptide elongation known as
translation arrest. The arrest activity is mediated by a discrete domain
comprised of the Alu portion of SRP RNA and a 9 and 14 kDa polypeptide
heterodimer (SRP9/14). Although very few nucleotides in SRP RNA are conserved throughout evolution, the remarkable conservation of G24, which resides in the region of SRP9/14 interaction,
suggests that it is essential for translation arrest. To understand the
functional significance of the G24 residue, we made single base substitutions
in SRP RNA at this position and analyzed the ability of the mutants to bind
SRP9/14 and to reconstitute functional SRPs. Mutation of G24 to C reduced
binding to SRP9/14 by at least 50-fold, whereas mutation to A and U reduced binding
~
2- and 5-fold respectively. The mutant RNAs could nevertheless assemble into SRPs at high subunit concentrations. SRPs reconstituted with mutant RNAs were not significantly
defective in translation arrest assays, indicating that the conserved guanosine does not interact directly with the translational machinery. Taken together, these results demonstrate
that G24 plays an important role in the translation arrest function of SRP by
mediating high affinity binding of SRP9/14.
The mammalian signal recognition particle (SRP) is a ribonucleoprotein composed
of six polypeptides and a 300 nt RNA (7SL RNA) that recognizes signal sequences
of nascent secreted and membrane proteins and targets them as ribosome-associated intermediates to the endoplasmic reticulum (ER) (reviewed in
1
). The binding of SRP to signal sequences leads to a transient inhibition of
further polypeptide elongation known as `translation arrest'. This
translational block is relieved after interaction between SRP and the SRP
receptor anchored in the ER membrane. Interaction between SRP and its receptor
also catalyzes release of the nascent chain from SRP and its insertion into a
protein translocation complex or `translocon' composed of the Sec61 complex and
an additional protein called TRAM (
2
,
3
). Although some organisms contain a much smaller SRP (
4
,
5
), the basic elements of the SRP pathway as well as the translocon are highly
conserved throughout evolution (
1
,
6
,
7
).
Extensive dissection of SRP has provided an integrated model of its structure
and function. Signal sequence recognition, translation arrest and nascent
polypeptide translocation activities reside in distinct domains of SRP (
8
). Signal sequences are recognized by a 54 kDa polypeptide (SRP54) as they
emerge from translating ribosomes (
9
,
10
). Translation arrest is mediated by a domain located at the opposite end of the
rod-like SRP. Release of translation arrest and translocation of the nascent
chain across the ER membrane requires activity of the S domain of the particle,
which consists of the SRP 19, 54, 68 and 72 kDa subunits plus the ~150 nt of 7SL RNA to which they are bound (
11
,
12
). The translation arrest domain consists of the Alu-homologous region of 7SL RNA, which folds into a tRNA-like cruciform structure (Fig.
1
), and the associated 9 and 14 kDa protein heterodimer SRP9/14 (
13
). It has been proposed that the Alu domain tRNA-like structure of 7SL RNA might effect translation arrest by blocking the
access of incoming aminoacylated tRNA molecules (
14
,
15
). This suggests that although SRP9/14 is required for translation arrest, the
Alu portion of SRP RNA may play a direct role in this activity.
Oligonucleotide primers containing the promoter for T7 RNA polymerase precisely
juxtaposed to the first 31 nt of the human 7SL RNA gene
7L30.1
(
30
) or scB1 were used in standard PCR amplification reactions to produce templates
for transcription of full-length 7SL RNA and scB1 RNA as described (
25
). Mutagenic primers contained an A, T or C instead of a G at position 24 of the
7SL RNA gene and created or destroyed a diagnostic restriction site. After purification by phenol/chloroform extraction and ethanol precipitation, the PCR products containing G24 or point mutations were used directly for T7 polymerase- directed RNA synthesis as described (
25
). Free NTPs were removed by gel filtration using High Capacity Quick Spin RNA Columns (Boehringer Mannheim) according to the manufacturer's instructions. RNA was further purified by phenol/chloroform extraction followed
by ethanol precipitation and was stored at -20oC as a precipitate until just before use.
RNA electrophoretic mobility shift assays (EMSA) using the 145 nt Alu domain of
7SL RNA and highly purified human SRP9/14 have been described previously (
28
). Templates for the 145 nt Alu domain of 7SL RNA were constructed by PCR
amplification of an S domain deletion mutant of the 7SL RNA gene previously
used for SRP9/14 binding studies (
28
,
30
). All templates were transcribed by T7 RNA polymerase (Promega) in parallel reactions that contained a pre-mixed solution of NTPs including [[alpha]-
32
P]GTP to ensure that each RNA was labeled to the same specific activity (
25
). All EMSA reactions contained 10 ng poly(G), which was used as a non-specific competitor (
28
). Quantitation was performed using a PhosphorImager and accompanying ImageQuant software (Molecular Dynamics).
SRP protein subunits were purified from a high salt extract of canine pancreatic
rough microsomes as described (
31
). To reconstitute SRPs, the proteins were added at a final concentration of 2 [mu]M together with 6 [mu]M synthetic human 7SL RNA into a buffer containing 50 mM HEPES, pH 7.5,
500 mM potassium acetate, 5.5 mM magnesium acetate, 0.5 mM EDTA, 1 mM DTT and 0.01% (w/v) Nikkol (octaethyleneglycol mono-
N
-dodecylether; Nikko Chemical Corp., Tokyo, Japan). Whereas all components are typically
added at equimolar concentrations (
32
,
33
), we used an excess of synthetic RNA because it assembles into particles less
efficiently than native RNA. Reconstitution reactions were incubated on ice for
10 min and then at 37oC for an additional 10 min. Subsequent purification steps were performed at
4oC. Intact SRP particles were purified away from free subunits and incomplete particles by spin filter chromatography using Ultrafree-MC DEAE anion exchange membrane units (Millipore). Typical 50 [mu]l reconstitution reactions were first diluted with 150 [mu]l 50 mM HEPES, pH 7.5, 10% (v/v) glycerol, 0.5 mM EDTA, 1 mM DTT,
0.01% Nikkol (UDEAE buffer) containing 500 mM potassium acetate and 5.5 mM magnesium acetate and then with 200 [mu]l 0.5 mM magnesium acetate in UDEAE buffer to yield final potassium acetate and
magnesium acetate concentrations of 250 and 3 mM respectively. The samples were then centrifuged for 2 min at 16 000
g
max
to remove any insoluble material. Supernatants were loaded onto DEAE filter units pre-equilibrated with 250 mM potassium acetate and 3 mM magnesium acetate in UDEAE buffer and
centrifuged at 4500
g
max
for ~5 min until all the solution had passed through the membranes. The
membranes were then washed by repeating the centrifugation with 400 [mu]l UDEAE buffer containing 350 mM potassium acetate and 4 mM magnesium acetate. Intact SRPs were eluted by centrifuging 30 [mu]l UDEAE buffer containing 600 mM potassium acetate and 6.5 mM magnesium acetate through the membranes twice. Nearly all the SRP
eluted in the first step. Incomplete SRPs remained bound to the membranes under
these elution conditions. The concentration of SRPs was determined by
comparison of SRP68 staining intensity with that of a bovine serum albumin
standard curve on a Coomassie Brilliant Blue stained SDS-PAGE gel. The stoichiometry of SRP9 and SRP14 in reconstituted SRPs was verified by examination of
silver stained SDS-PAGE gels (
34
).
Translation arrest and translocation assays were performed as described (
31
,
35
) with the following minor modifications. Cyclin[Delta]90 and preprolactin capped mRNAs were used at final concentrations of 3 and 2 [mu]g/ml respectively. Wheat germ translation reactions were precipitated
with 10% (w/v) trichloroacetic acid, re-dissolved with SDS sample buffer (5% w/v SDS, 105 mM Tris-HCl, pH 8.2, 12.5% v/v glycerol, 1 mM EDTA, 100 mM DTT), heated to 65oC for 30 min and analyzed by electrophoresis on 14% acrylamide
Tris-glycine SDS-PAGE gels. After the gels were fixed and dried, translation
products were quantitated using a Fuji BAS2000 PhosphorImager. Translation
arrest activity was calculated as described (
36
) except that cyclin[Delta]90 (
37
) was used instead of globin as the non-secretory control. Percent processing was calculated by a method similar to that
described previously (
36
) except that any variations in sample loading were corrected by normalizing the
radioactivity in the preprolactin and prolactin bands to that of the cyclin[Delta]90 band in the same lane.
The EMSA was previously shown to be useful for monitoring relative affinities of
SRP9/14 for the Alu domain of 7SL RNA and Alu-related RNAs such as small cytoplasmic (sc) B1 RNA (
28
). The equilibrium
K
d
value of 2 * 10
-10
M for the interaction between SRP9/14 and 7SL-Alu RNA determined using this assay agreed with the value determined in
solution by others (
38
). The affinity of SRP9/14 for scB1 RNA was previously found to be between 5- and 10-fold lower than for 7SL-Alu RNA (
28
). To determine the relative affinities of SRP9/14 for the Alu domain of 7SL
RNAs containing different nucleotides at position 24, we synthesized
32
P-labeled transcripts of the Alu region of 7SL (hereafter referred to as 7SL-Alu; see Materials and Methods) which varied only at this position.
In the experiment shown in Figure
2
A, purified scB1 [
32
P]RNA and 7SL-Alu [
32
P]RNAs were incubated with SRP9/14 alone (odd numbered lanes) or with SRP9/14
plus non-radioactive scB1 RNA competitor (even numbered lanes) and binding was analyzed by EMSA.
PhosphorImager quantitation of the radioactivity in the RNP bands is shown below the lanes of Figure
2
A and is expressed as a percentage of the radioactivity contained in the wild-type 7SL G24 RNP band in lane 3. SRP9/14 exhibited the highest affinity
for wild-type (G24) 7SL-Alu RNA (lane 3), followed by A24 (lane 7), U24 (lane 9) and C24
(lane 5). SRP9/14 exhibited 2- and 5-fold lower affinities for 7SL-Alu A24 and 7SL-Alu U24 RNAs respectively. 7SL-Alu U24 RNA and scB1 RNA exhibited similar binding
affinity (compare lanes 1 and 9), probably because scB1 RNA contains a U at the
analogous position (
25
,
28
). The order of relative affinities of SRP9/14 for 7SL RNAs containing
substitutions at position 24 determined here correlate with the frequency with
which the individual bases are found in the SRP RNA database (13; Zwieb at http://pegasus. uthct.edu/SRPDB/SRPDB.html).
The relative affinities reported above were supported by the results obtained
with scB1 RNA, which was used as a reference to calibrate binding affinity.
Comparison of lanes 1 and 3 revealed that ~5-fold fewer SRP9/14-containing RNP complexes formed with scB1 [
32
P]RNA than with 7SL G24 [
32
P]RNA. This result demonstrates that this assay accurately reflects differences
in binding affinity, since it was previously determined that SRP9/14 exhibits a
4- to 9-fold lower affinity for scB1 than for 7SL-Alu RNA (
28
). We also used unlabeled scB1 RNA as a competitor (even numbered lanes).
Unlabeled scB1 RNA did not compete as well with 7SL G24 [
32
P]RNA (lane 4) or 7SL A24 [
32
P]RNA (lane 8) as it did with scB1 [
32
P]RNA (lane 2), while it competed well with 7SL U24 [
32
P]RNA (lane 10). These results agree with the relative affinities of wild-type 7SL-Alu RNA and scB1 RNA for SRP9/14 determined previously (
28
).
The observation that introduction of mutations at position 24 of 7SL RNA
reduced, but did not abolish, binding of SRP9/14 implies that reconstitution of
SRPs with mutant RNAs should be possible provided that subunit concentrations
are sufficiently high. SRP reconstitutions are typically performed with subunit
concentrations in the 1 [mu]M range, which allows addition of reconstituted particles to biochemical
reactions at physiological concentrations (~10 nM) (
4
1
). Because the
K
d
for the interaction between SRP9/14 and 7SL RNA is 5000-fold lower than the concentration of subunits in the reconstitution
reactions, we reasoned that even the most severe mutation at position 24 (G -> C), which raises the
K
d
to ~10
-8
M, would still be compatible with SRP assembly.
To determine whether SRP RNAs that contain substitutions at position 24 could be
used to assemble complete SRPs, we synthesized full-length 7SL RNAs containing G24, A24 and C24 and used them in
reconstitution reactions (see Materials and Methods). Following incubation of
SRP proteins with the 7SL RNAs, assembled SRPs were isolated by DEAE chromatography. Equal portions of the samples were analyzed by SDS-PAGE followed by silver staining (Fig.
3
). The observation that similar amounts of each SRP polypeptide were present in
each sample indicates that wild-type and mutant RNAs assembled equally efficiently into complete
particles. Hence, the mutations in SRP RNA at position 24 did not prevent
binding of SRP9/14 (or any other subunit) to the RNA under standard reconstitution conditions. It is particularly noteworthy that SRP reconstituted with 7SL C24 RNA contains
nearly as much SRP9 and SRP14 as SRP reconstituted with wild-type RNA. Thus once the mutant SRPs were assembled, they were sufficiently
stable to be recovered after purification with minimal loss of SRP9/14.
To test for activity, reconstituted SRPs were added to wheat germ
in vitro
translation reactions and translation arrest activity was measured by specific
inhibition of preprolactin synthesis (
36
). The concentration of SRP required to cause 50% translation arrest can be used
to define relative activity (
31
). By this criterion, purified SRP reconstituted with wild-type synthetic 7SL RNA supports the same level of translation arrest
activity (Fig.
4
A) as SRP reconstituted with native RNA (
31
). Furthermore, an SRP containing 7SL A24 RNA exhibited translation arrest
activity that was nearly indistinguishable from wild-type (Fig.
4
A). SRP containing 7SL C24 RNA, however, was ~2-fold less active than the other two SRPs. This reduction in
translation arrest activity may be explained by the fact that binding of
SRP9/14 to the 7SL C24 RNA was slightly less efficient (compare Fig.
3
, lanes G24 and C24) and that a small amount of SRP9/14 might have dissociated
from the mutant RNA during the assay. In any case, the effect of the G24 -> C mutation on translation arrest activity was extremely small compared
with the effect this mutation had on binding of SRP9/14.
In this study we have explained the functional significance of one of the few
nucleotides in SRP RNA that is highly conserved throughout evolution. Given
that the Alu domain is known to interact with the translation machinery to
produce an inhibition of polypeptide elongation, a reasonable hypothesis is
that the conserved G24 residue participates in a key interaction between the
ribosome and SRP. Our results demonstrate clearly, however, that whereas
substitution of A or C in this position significantly reduces binding of
SRP9/14, the mutations have little or no effect on the ability of SRP to
mediate translation arrest. Introduction of a C at position 24 of 7SL-Alu RNA reduced the equilibrium binding affinity for SRP9/14 by at least
50-fold, as determined by our assay, but 7SL RNA with this mutation
nonetheless assembled into a functional SRP provided that subunits were present
at high concentrations. The relatively small (<2-fold) decrease in translation arrest observed in experiments with the C24
mutant can probably be attributed to a slight reduction in the amount of
SRP9/14 successfully bound to the particle or to a slight loss of SRP9/14
during the experiment. Thus, the data are most consistent with the
interpretation that the role of G24 in 7SL RNA is to promote efficient binding
to SRP9/14.
One interesting question that arises from this work is whether G24 facilitates
binding of SRP9/14 by directly interacting with the proteins or by promoting a
folded structure of the RNA that the proteins recognize. The report that G24
was not observed to be a point of contact in chemical protection studies
supports the latter hypothesis (
20
). The observation of a fast-migrating form of the 7SL RNA-SRP9/14 RNP complex when G was present at position 24 (Fig.
2
A) suggests that this residue may play a role in the overall structure of the
Alu domain. Although the G4 residue in
S.pombe
SRP RNA that is equivalent to G24 resides in a loop that is part of a single
hairpin instead of an internal loop between two hairpins, genetic analysis is
consistent with the notion that it contributes significantly to the structure
of the Alu domain (
18
). In any case, it is noteworthy that this single nucleotide appears to exert a
very significant influence over the binding of SRP9/14, despite evidence that a
large number of contacts are made between the RNA and protein heterodimer (
17
).
The data presented here, together with results from studies on domain IV of SRP
RNA (
16
), indicate that the highly conserved residues in SRP RNA that have been studied
thus far are required for protein binding and not for interaction with
ribosomes. Indeed, previous work suggests that the interaction between SRP and
the ribosome is complex and may require the participation of all the subunits (
4
2
). Consistent with this notion, a 7SL-Alu-SRP9/14 RNP fragment of SRP is unable to compete with SRP in
translation arrest assays (
11
). Given that 7SL RNA appears to undergo conformational changes in different
phases of the SRP cycle (
20
), one of its functions may be to allow the protein subunits to change their
relative position with respect to one another. The proteins may have evolved as
part of an RNP, so that in bacteria which have lost a significant portion of
the primordial SRP, the remaining protein(s) must still bind to RNA to adopt a
functional conformation.
Given that SRP9/14 binds to SRP RNA with a
K
d
of ~10
-10
M and that the concentration of SRP subunits in the cell has been measured to
be ~10
-8
M (
4
1
), the mutation G24 -> C, which reduces equilibrium binding to <= 2 * 10
-8
M, is likely to impair function
in vivo
. However, an interesting puzzle is why G24 is so highly conserved if mutation
to A or U reduces the affinity of the RNA for SRP9/14 by only a few-fold. The notion that U at position 24 is compatible with SRP9/14 binding
is supported by the observation that scB1 and scAlu RNAs maintain stable
association with SRP9/14 despite the fact that scB1 RNA harbors a U at the G24-homologous position (
21
,
23
,
25
,
28
). One explanation for the conservation of G24 is that 7SL RNA or SRP has an additional function
in vivo
that has not been measured in
in vitro
assays.
The observation that 7SL RNAs that contain substitutions at position 24 can
assemble into functional particles despite a decreased affinity for SRP9/14
provides an explanation for previous findings that scB1 and scAlu RNAs increase
in response to elevated intracellular SRP9/14 levels (
23
). Although these RNAs are present at only ~0.1% the level of 7SL RNA, the high concentration of SRP9/14 should drive
the assembly of RNP complexes. Moreover, the observation that 7SL RNA
containing U at position 24 is active in translation arrest is consistent with
the possibility that scB1 and scAlu RNAs play a regulatory role in translation.
We are grateful to M.L.Lanigan for assistance with preparation of the manuscript
and to B.Peculis and E.Englander for critical reading and helpful comments.
K.H. was supported by the NIH Research Scholars Program of the Howard Hughes
Medical Institute. D.-Y.C. was supported by an Interpersonnel Act between NICHD and the
Department of Biochemistry, University of Maryland Medical School.
*To whom correspondence should be addressed at: Building 6, Room 416, 9000
Rockville Pike, Bethesda, MD 20892-2753, USA. Tel: +1 301 402 3567; Fax: +1 301 480 9354; Email:
maraia@ncbi.nlm.nih.gov
REFERENCES
Return