ABSTRACT
The interaction of protein SRP54M from the human signal recognition particle
with SRP RNA was studied by systematic site-directed mutagenesis of the RNA molecule. Protein binding sites were
identified by the analysis of mutations that removed individual SRP RNA helices
or disrupted helical sections in the large SRP domain. The strongest effects on
the binding activity of a purified polypeptide that corresponds to the
methionine-rich domain of SRP54 (SRP54M) were caused by changes in helix 8 of the SRP
RNA. Binding of protein SRP19 was diminished significantly by mutations in
helix 6 and was stringently required for SRP54M to associate. Unexpectedly,
mutant RNA molecules that resembled bacterial SRP RNAs were incapable of
interaction with SRP54M, showing that protein SRP19 has an essential and direct
role in the formation of the ternary complex with SRP54 and SRP RNA. Our
findings provide an example for how, in eukaryotes, an RNA function has become
protein dependent.
Signal recognition particle (SRP) is a cytosolic ribonucleoprotein complex that
assists in the co-translational translocation of proteins across lipid bilayers [reviewed by
(
1
)]. Simple bacterial SRPs consist of a 54 kDa polypeptide [SRP54, also named P48
or
ffh
, (
2
-
5
)], that is bound to SRP RNA. Mammalian SRPs are composed of a 300 nucleotide
RNA (MW 97 142), and the six polypeptides SRP9, SRP14, SRP19, SRP54, SRP68, and
SRP72 (
6
). The prominent role of SRP54 in protein secretion is underscored by its
appearance in all phylogenetic groups, its high degree of conservation (
7
), and its close proximity to the signal peptide (
8
,
9
). The M-domain of SRP54 (SRP54M) contains many methionine residues (
4
,
5
) believed to bind not only with the SRP RNA (
10
,
11
), but also with the signal peptide (
12
). SRP54M is sufficient for the interaction with SRP RNA (
11
).
In mammals, protein SRP54 associates with the larger of the two SRP domains (
13
,
14
), but only after SRP19 has bound to the SRP RNA (
6
). Since there are no perceivable interactions between the free SRP19 and SRP54 polypeptides, their assembly appears to be directed by
an SRP19-induced conformational change in the RNA (
11
,
15
-
17
). Protein SRP19 may displace SRP RNA helix 6 and uncover the SRP54 binding
sites. Alternatively, SRP19 may play a more active role in shaping the SRP54-binding site.
To determine details of the RNP assembly in the human SRP, we measured the
binding activities of mutated SRP RNAs toward purified SRP19 and SRP54M. We
established that SRP54M binds predominantly to helix 8 of the human SRP RNA,
and that the association is strictly dependent on protein SRP19. Furthermore, SRP RNA helix 6 does not interfere sterically with the binding of SRP54M,
suggesting that SRP19 plays an intimate role in the formation of the SRP54
binding site.
Oligonucleotides used in PCR reactions, site-directed mutagenesis, or sequencing were synthesized on an Applied Biosystems (PCR-mate) DNA synthesizer (trityl-on) using [beta]-cyanoethyl-phosphoramidite chemistry. Products were
purified and detritylated on an oligonucleotide purification cartridge as supplied by the manufacturer.
The DNA region corresponding to the M-domain of human SRP54 (amino acid residue position 297-504; GenBank Accession number U51920) was amplified via PCR with
oligonucleotides 5'-ACTTCTTG
For large scale purification of the SRP54M protein, competent
E.coli
cells (BL21-DE3-pLysS, Novagen) were prepared, freshly transformed with phSRP54M-DNA (
20
), and incubated overnight at 37oC on 10 cm diameter agar plates containing LB and 100 [mu]g/ml Ampicillin (Sigma) and 37 [mu]g/ml chloramphenicol (Fisher Scientific). Two 2 l Erlenmeyer flasks, each with 400 ml of LB, 100 [mu]g/ml Ampicillin and 37 [mu]g/ml chloramphenicol were inoculated with all colonies from
two plates each, and the cultures
were grown at 37oC with vigorous shaking for ~45 min until the A
600
was between 0.5 and 0.8. Both cultures were used to seed a 20 l fermenter (Bioflo IV, New Brunswick Scientific) containing 11.2 l of LB medium (with enough nutrients for 12 l), 240 mg of Ampicillin, 960
mg of Methicillin (USB), 410 mg of chloramphenicol and 1.2 ml of Antifoam 289
(Sigma) kept at 37oC with aeration. The vessel pressure was 20 p.s.i., and the speed of the
stirrer was set to 600 r.p.m. Expression of SRP54M was induced when the A
600
of the culture reached 0.6-0.8 (after ~2.5 h) by adding IPTG (Gold Biotechnologies) to a final concentration of 1
mM, after which growth was continued for 2 h. Cells were subjected to centrifugation for 15 min at 5824
g
at 4oC (4000 r.p.m. in a Sorvall, H6000 rotor). Approximately 30 g of cells were resuspended in 210 ml of lysis buffer (50 mM NaPO
4
,
pH 8.0, 300 mM NaCl, 5 mM DTT) and frozen at -70oC.
All subsequent manipulations for protein purification were carried out at 4oC. The frozen cells were thawed on ice and sonicated (Fisher, Model 300) at
60% of maximum output using an `intermediate' tip, for 10 pulses, 15 s each.
The lysate was submitted to centrifugation for 10 min at 14 500
g
(Sorvall SS34, 10 000 r.p.m.). The resulting supernatant was submitted to centrifugation for 4 h at 80 000
g
(Beckman, VTi 50, 30 000 r.p.m.). The supernatant (~116 ml) was diluted by adding 5 vol of a buffer that contained 50 mM NaPO
4
, pH 8.0, and 5 mM DTT to reduce the NaCl concentration to 50 mM, and was loaded
onto a Biorex 70 (BioRad) cation exchange column (2.5 cm diameter, 28 cm long,
total bed volume of 138 ml) equilibrated in 50 mM NaPO
4
, pH 8.0, 50 mM NaCl, 5 mM DTT (equilibration buffer) and connected to an FPLC system (Pharmacia) at a flow rate of 1 ml/min. The column was washed with 300 ml of equilibration buffer, after which a linear
gradient from 50 mM to 1 M NaCl (total gradient volume of 150 ml) was applied
in 50 mM NaPO
4
, pH 8, 5 mM DTT at a flow rate of 1 ml/min. Analysis of aliquots from the
collected fractions by SDS PAGE and Coomassie blue G250 staining showed the
elution of SRP54M at ~250 mM NaCl and, to a minor degree, at ~305 mM. Fractions from each peak were pooled separately (~9 ml for the weakly-bound material and 21 ml for the strongly-bound material), concentrated to a volume of 2.5 ml by centrifugation at 3500
g
(Sorvall SS34, 5000 r.p.m.) using Amicon Centricon 10. The protein
concentration was determined by a modification of the Bradford (Bio-Rad) protein assay (
21
) as described (
22
). The protein was stored at a concentration of 7 mg/ml in 50 mM NaPO
4
, pH 8.0, 250 mM NaCl, 5 mM DTT, and 50% glycerol at -20oC until further use. Purity of the preparation was determined by
densitometric scanning of the Coomassie blue stained gels after SDS PAGE using
an Abaton 300/GS scanner and NIH Image software (
23
).
Plasmids coding for authentic human SRP RNA or for mutant RNAs that lacked individual RNA helices was described earlier (
17
). Mutant [Delta]H67 was generated by PCR site-directed mutagenesis (
24
) using oligonucleotide 5'-CGGTTCACCCCTTGCCGAACTTAGTG-3' as a mutagenic primer. Details of the construction of
other mutations in the large domain of the SRP covering residues 100-252 were
communicated previously (
25
).
T7-polymerase was prepared as described (
26
) with modifications kindly provided by Arthur Zaug, University of Colorado, Boulder. Plasmids were
restricted with
Dra
I, or
Bam
HI (for mutant DNAs of [Delta]35 and H6), concentrated by phenol-chloroform extraction and ethanol precipitation, and the DNAs were dissolved in 10 mM Tris-HCl, pH 7.5, 1 mM EDTA at a concentration of 1 [mu]g/[mu]l. Run-off transcriptions were carried out for 2 h
at 37oC in 200 mM HEPES-KOH, pH 7.5, 30 mM MgCl
2
, 2 mM spermidine, 40 mM DTT, 6 mM of each rNTP, 0.3-1 [mu]g/[mu]l of restricted DNA, and an amount of T7-polymerase that was optimized to maximize RNA yields. After transcription, 1/10 vol of 0.5 M EDTA, pH 8 and 0.4 vol of 9 M LiCl were added, the
samples were incubated on ice for 30 min, and the RNAs were concentrated by
centrifugation in a microfuge for 10 min. The pellets were washed with 1 ml of
ice-cold 2.5 M LiCl, followed by centrifugation and a wash with 80% ethanol.
The pellets were dried and the RNA was dissolved in a small volume of water and
stored at -20oC. Transcribed H6-RNAs were precipitated by adding 0.05 vol of 5 M NaCl, 1/10
vol of 0.5 M EDTA, pH 8, and 3 vol of ethanol. Incubation was at -70oC for 30 min, after which the RNA was collected, washed and
dissolved in water as described above for the larger RNAs.
The RNA concentrations of the samples were determined by electrophoresis of
appropriately diluted sample aliquots on 2% agarose gels followed by staining
with ethidium bromide and calculated from a standard curve obtained with known
amounts of
E.coli
5S rRNA (Boehringer) separated on the same gel.
RNAs were incubated in 100 [mu]l binding buffer (50 mM Tris-HCl, pH 7.9, 300 mM KOAc, 5 mM MgCl
2
, 1 mM DTT and 10% glycerol) at a concentration of 0.4 [mu]g/[mu]l at 60oC for 10 min followed by a gradual cooling to room temperature for ~30 min. Two [mu]l aliquots of purified human protein SRP19 (
22
) and/or purified human protein SRP54M at appropriate concentrations (see
Results) were added by gently mixing with a pipette tip followed by an
incubation at 37oC for 10 min. Sequential binding of proteins was carried out by adding 1 [mu]l of SRP54M to the RNA SRP19-complex followed by another 10 min incubation at 37oC.
RNA protein complexes were loaded onto a small (20 [mu]l bed volume) DEAE-Sepharose column prepared in a 200 [mu]l pipette tip (T-200, 0.4 mm, Phenix) with the insert cut from a barrier tip insert (National Scientific), equilibrated in 300 mM KOAc, 50 mM Tris-HCl, pH 7.9, 5 mM MgCl
2
and 1 mM DTT. Each column was washed with 200 [mu]l of the equilibration buffer. The flowthrough and the wash were collected
in the same tube (F). Bound material (E) was eluted with 200 [mu]l of a buffer containing 1 M KOAc, 50 mM Tris-HCl, pH 7.9, 5 mM MgCl
2
and 1 mM DTT. A 1/10th vol of 100% TCA was added to all samples followed by an
incubation on ice for 30 min to precipitate the proteins and the RNA protein
complexes. The precipitate was collected by centrifugation at room temperature
for 10 min, resuspended in 10 [mu]l SDS loading buffer [62.5 mM Tris-HCl, pH 6.8, 2% SDS, 0.5% (v/v) [beta]-mercaptoethanol, 0.00125% (w/v) Bromophenol blue, 10%
(v/v) glycerol and 100 mM Tris-base], mixed, and heated for 3 min at 95oC followed by electrophoretic separation of the polypeptides on 15% polyacrylamide SDS- containing gels. Proteins were stained with Coomassie blue G250, the gel
was destained, and scanned electronically without image enhancements. The areas
of the peaks were measured using NIH Image software (
23
).
The three-dimensional model of the human SRP RNA structure (
27
) was obtained from the SRP database (
7
) in the PDB file-format (
28
) at the Internet address http://pegasus.uthct.edu/SRPDB/SRPDB.html, and was
displayed on an Silicon Graphics Indigo 2 Extreme workstation using the program
DRAWNA (
29
). Experimental data were incorporated into the model by color-coding those mutated sections of the SRP RNA that had significant effects
on the binding of proteins SRP19 and SRP54M as described in Results. The model
was accepted as provided with no additional structural constraints.
The M-domain of human SRP54 (SRP54M) corresponding to amino acid residue positions 297-504 of the full-length polypeptide was cloned under control of the bacteriophage T7 promoter by
PCR amplification of the corresponding DNA fragment (see Materials and
Methods). A soluble polypeptide of the expected size was synthesized in
E.coli
by induction with IPTG. The protein was purified by differential centrifugation and chromatography on Biorex 70
(see Fig.
1
, and Materials and Methods). SDS PAGE showed that the majority of SRP54M eluted
at 250 mM salt, but ~10% of the protein bound under slightly more stringent conditions (305 mM).
Material from both the low- and high-salt fractions actively bound to the [Delta]35 RNA-SRP19 complex (not shown). Scanning of the Coomassie-stained SDS polyacrylamide gels showed that the
protein was ~77% pure.
Nucleotide changes were introduced into the gene for the human SRP RNA in a two-step polymerase chain reaction (PCR) protocol (
24
,
25
). This approach allowed us to generate any desired alteration by using only one mutagenic oligonucleotide for each mutant in parallel PCRs. Among the mutations (see Table
1
) were a deletion of the small domain (mutant [Delta]35), deletions of individual RNA helices (mutants [Delta]H6, [Delta]H67 and [Delta]H8), and a mutation that retained only helix 6
(mutant H6). Mutant [Delta]H67 lacked those nucleotides (positions 128-175) that are absent in the bacterial sequences of a global
alignment of SRP RNAs (
7
,
30
). The other 22 mutations in the large domain disrupted sections of helices 5-8
(mutants 5A-5F, 6A-6F, 7A and 8A-8F), or compensated the sectional disruptions in helix 6
(mutants C1-C3) by altering both basepaired RNA strands. Two mutants, 6T and 8T,
affected the RNA tetranucleotide loops (tetraloops) in helix 6 and helix 8,
respectively. (See Table
1
and Figure
2
for the location of the mutations in the secondary structure of the human SRP
RNA.) Appropriately restricted template DNAs were used as templates for
in vitro
run-off transcriptions with T7 polymerase to produce RNA molecules of the
expected quantity and length (see Materials and Methods).
Binding to SRP54M was completely abolished when helix 8 was deleted, or when the
central and distal portions of helix 8 were altered (mutants 8B, 8C, 8D and
8E). Therefore, we considered most of helix 8, namely positions 176-197 and 202-214, as the predominant binding site of SRP54M. The slightly
reduced binding to mutant 8T RNA (72+-14%) is explained by its somewhat lower capacity to associate with
protein SRP19 (88+-3.5%); therefore, the apical GAAA-tetraloop of helix 8 was unlikely to be in contact with SRP54M.
To visualize the binding sites of the two proteins in the context of the
assembled SRP, we incorporated the results from the site-directed mutagenesis experiments into a model of the three- dimensional structure of the SRP RNA, proposed previously (
27
). Our model is characterized by a `straight' helix 6, and a `bent' helix 8 that
is caused by a hypothetical tertiary interaction between the helix 8 tetraloop
and helix 5 (198-GA-199 with 232-GU-233 in the human SRP RNA). Figure
5
shows a color coded view of the large domain with strong and mild mutational
effects on the binding of SRP19 and SRP54M. Colors appear only in those regions
of the RNA where mutations had a conspicuous impact on protein binding. (The
criteria for the classification of the protein binding activities are described in the legend to Figure
5
.) Separate binding sites were supported by the complete separation of the strong
mutational effects for SRP19 and SRP54M.
RNA site-directed mutagenesis was used to identify binding sites for proteins SRP19
and SRP54M on the SRP RNA. Since an RNA that corresponded to the large domain
of the SRP (mutant [Delta]35) was fully active in binding both polypeptides, the detailed
mutational analysis was confined to a portion of helix 5, and to helices 6, 7
and 8 (Fig.
2
). Compensatory mutations that preserved base pairing were introduced only in
helix 6, because it is a more-or-less continuous RNA stack that lacks pronounced internal loops.
Most of the SRP RNA mutations (with the exception of mutant [Delta]H67), had been used earlier with radioactively labeled protein SRP19
translated
in vitro
in a wheat germ cell-free system (
17
,
25
). In those experiments, binding of an unknown amount of SRP19 was measured with
the respective mutant RNAs in excess. Furthermore, the samples contained an
undefined mixture of components including constituents from the wheat germ
SRPs. Despite these impediments, the earlier findings agree well with results
obtained in a defined system. In both cases, helix 6 was found to be the major
binding site for protein SRP19 with a preference for the tetraloop and the 5'-portion of the helix. The influence of the distal portion of helix
8 on SRP19 binding (e.g. mutant 8C) was confirmed. In addition, we now found
mild effects caused by mutations throughout helix 5; of those, only mutant 5C
was shown previously to have an effect.
There was a small but very reproducible (
n
= 11) difference in the affinities of SRP54M for SRP RNA (77+-5.6%) and the [Delta]35 RNA (92+-4.4%). This result may be explained by interactions
between the small and the large SRP domains. Although we do not know which site
of the small domain may cause this effect, in the large domain, nucleotides
located at the 3'-side of helix 6 appear involved, as evidenced by the undiminished
binding capacity of mutants 6E and C1 (Fig.
4
B).
The lower affinity of the unmutated SRP RNA and most mutant RNAs may result from incomplete localized renaturation. However, this interpretation does not explain the high affinities of mutants 6E and C1.
Furthermore, when we analyzed conformational differences between the SRP RNA
and the various mutant RNA by electrophoresis on non-denaturing polyacrylamide gels, we found no correlation of the binding
affinity with the appearance of unfolded molecules (not shown).
SRP RNA helix 8, which is conserved in all SRPs, was identified as the major
binding site for SRP54. This result confirmed experimentally what was expected from the comparative sequence analysis of the bacterial SRP components (
7
,
30
). Enzymatic and chemical probing showed that P48, the bacterial homologue of
SRP54, protects a region in 4.5S RNA that is equivalent to helix 8 of the
eukaryotic SRP RNA (
31
). In the
Mycoplasma mycoides
SRP, P48 protected helix 8 from digestion by RNase A (
32
). In addition, translocation of nascent polypeptides was promoted by a heterologous complex that consisted of mammalian SRP54 and
bacterial 4.5S RNA (
33
).
Our results show that the human SRP RNA has lost the ability to bind to SRP54M
when SRP19 is absent. Surprisingly, none of the mutant RNA molecules, including
mutants [Delta]H6 and [Delta]H67 which lack helix 6 and thereby closely resemble the bacterial
4.5S RNA, bound even trace amounts of SRP54M. This would have been expected if
protein SRP19 simply displaces SRP RNA helix 6 and uncovers a preexisting SRP54
binding sites. A full-length SRP54 polypeptide that we expressed in
E.coli
and insect cells (unpublished results) was incapable of binding to [Delta]H67 RNA, demonstrating that the G-domain of SRP54 cannot restore the RNA binding activity (not
shown).
We conclude that SRP19 is involved fundamentally in the assembly of the SRP by
actively shaping the binding site of SRP54. A process is favored whereby SRP19
configures helix 8 as is evidenced by the ability of the protein to influence
the distal portion (mutant 8C). However, upon SRP19's interaction with RNA, the
SRP54M binding site may be composed of nucleotide and amino acid residues.
Thus, the sequential assembly of SRP19 and SRP54 in the large SRP domain
provides a compelling example for the increased complexity of the eukaryotic
SRP and for an RNA function being `taken over' by proteins.
We thank Shaun D. Black for critical reading of the script and Kerfoot P. Walker
III for exceptional technical help. This work was supported by NIH grant GM-49034 to C.Z.
*To whom correspondence should be addressed. Tel: +1 903 877 7689/7676; Fax: +1
903 877 5876; Email: zwieb@jason.uthct.edu
REFERENCES
Return