Research Article

Crystal Structure of the Ribonucleoprotein Core of the Signal Recognition Particle

See allHide authors and affiliations

Science  18 Feb 2000:
Vol. 287, Issue 5456, pp. 1232-1239
DOI: 10.1126/science.287.5456.1232


The signal recognition particle (SRP), a protein-RNA complex conserved in all three kingdoms of life, recognizes and transports specific proteins to cellular membranes for insertion or secretion. We describe here the 1.8 angstrom crystal structure of the universal core of the SRP, revealing protein recognition of a distorted RNA minor groove. Nucleotide analog interference mapping demonstrates the biological importance of observed interactions, and genetic results show that this core is functional in vivo. The structure explains why the conserved residues in the protein and RNA are required for SRP assembly and defines a signal sequence recognition surface composed of both protein and RNA.

Cells communicate with their surroundings by means of proteins that either reside within cellular membranes or are secreted to the outside. One of the principal pathways for targeting these proteins uses the SRP to catalyze cotranslational transport of nascent secretory and membrane proteins to the endoplasmic reticulum (ER) in eukaryotes and to the plasma membrane in prokaryotes. The well-characterized eukaryotic SRP, an evolutionarily conserved ribonucleoprotein complex, recognizes the NH2-terminal signal sequence of targeted proteins as they emerge from the ribosome. Binding to the SRP arrests polypeptide elongation and mediates docking of the translating ribosome with receptors on the ER in a guanosine 5′-triphosphate (GTP)–dependent process (Fig. 1A) (1, 2).

Figure 1

(A) The protein translocation cycle catalyzed by the eukaryotic SRP. A nascent polypeptide chain attached to the ribosome (red) contains a signal sequence (SP) that is recognized by the SRP (yellow), targeting the complex to the translocon (TC) embedded in the ER membrane in a GTP-dependent process [adapted from (40)]. (B) Schematic of the secondary structure of the SRP RNA from humans and E. coli. The universally conserved domain IV of the RNA is highlighted in green. (C) Nucleotide sequences of SRP RNA domain IV from representative organisms of the three kingdoms of life. The three principal conserved features of this domain, the GNRA tetraloop, the symmetric internal loop, and the asymmetric internal loop, are highlighted in cyan. Nucleotides in red in the E. coli SRP domain IV RNA are those that have been shown to be important for in vivo function (35) as well as Ffh binding (15). (D) Sequence alignment of SRP54/Ffh M domains from various organisms, with the numbering consistent with that of E. coli Ffh. The secondary structural assignments were based on the x-ray structure of the bound E. coli M domain. Boxed in cyan are invariant residues in the RNA binding region of the protein. The sequence of the E. coli M domain shown corresponds to the one that was used for crystallization. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

In bacteria, the SRP is essential for cell viability and efficient protein export (3) and consists of the 4.5S RNA and the Ffh protein. These components as well as the SRP receptor, FtsY, share sequence and functional homology with their eukaryotic counterparts 7SL RNA, SRP54, and SRα, respectively (4–8). The evolutionary conservation of this fundamental cellular component is demonstrated by the ability of human SRP54 to bind with high affinity to the Escherichia coli4.5S RNA and to rapidly hydrolyze GTP in the presence of the SRP receptor (5). Similarly, Ffh is able to replace SRP54 in a chimeric mammalian SRP that is capable of elongation arrest and signal sequence recognition (6). Thus, the Ffh-4.5S RNA complex appears to be a minimized structural and functional homolog of the eukaryotic SRP, which makes it attractive for detailed structural and mechanistic studies.

The Ffh/SRP54 proteins contain three domains: N, G, and M. The NH2-terminal N domain, a four-helix bundle, is closely associated with the adjacent G domain, a Ras-like guanosine triphosphatase (GTPase) (9) responsible for mediating the interaction of the SRP with its receptor and regulating SRP function through hydrolysis of GTP (7, 10). Structurally related N and G domains are also present in the SRP receptor (11). The methionine-rich M domain near the COOH-terminus of Ffh/SRP54 contains recognition sites for both the signal peptide and the SRP RNA (12–14).

The Ffh binding site on the 4.5S RNA is localized to domain IV, a ∼50-nucleotide (nt) region whose highly conserved secondary structure consists of two internal loops that include noncanonical base pairings and unpaired nucleotides (Fig. 1, B and C) (15–18). In bacterial and human SRP, several lines of evidence suggest that domain IV (Fig. 1, B and C) stabilizes the particle and its interaction with the signal peptide (19, 20). Outside of domain IV, the size, sequence, and secondary structure of SRP RNAs vary widely, even among bacterial species (21), and it is unclear if these regions of the RNA are essential for SRP function.

Here we present the 1.8 Å resolution crystal structure of the universally conserved ribonucleoprotein core of the SRP, a complex between domain IV of 4.5S RNA and the M domain of Ffh. Nucleotide analog interference mapping and genetic results show that the structure represents the functional complex in vivo. In the structure, unique RNA-protein interactions characterize the molecular interface, and a network of highly ordered waters and metal ions also mediate key contacts. The structure explains why phylogenetically conserved residues in the protein and RNA are required for SRP assembly, and it suggests a possible role for the RNA in molecular recognition of signal peptides.

Structure determination and overview. To obtain well-ordered crystals of the M domain in complex with SRP RNA, we screened a series of modified protein and RNA constructs (22). None of these modifications altered the apparent dissociation constant of the RNA-protein complex from that of the intact E. coli SRP, which is 40 pM under the assay conditions we used (23). Crystals of the M domain bound to a 49-nt SRP RNA were grown that diffract x-rays to 1.55 Å resolution. The structure was solved by multiwavelength anomalous dispersion (MAD) phasing from crystals containing selenomethionine-labeled protein (Table 1). Using diffraction data extending from 30.0 to 1.8 Å resolution, a model of the complex was built and refined to a crystallographic R factor of 19.9% and a freeR factor of 22.1% (Table 1).

Table 1

Crystal structure determination and refinement. The 49-nt SRP RNA was transcribed by T7 RNA polymerase and purified by standard techniques (44). Selenomethionine-labeled M domain was expressed in the E. coli methionine auxotrophic strain B834 (Novagen) and purified with a combination of affinity, reversed-phase, and ion-exchange chromatography. The RNA-protein complex was crystallized by the sitting drop vapor diffusion method by the addition of 2 μl of macromolecular solution to 2 μl of a precipitating solution containing 10% isopropanol, 50 mM Na-MES (pH 5.6), 200 mM KCl, 12.5 mM MgCl2, and 35 mM C-HEGA 10 (Anatrace). Over 2 to 3 days, the crystals grew to maximum cell dimensions of 0.2 mm by 0.1 mm by 0.1 mm, belonging to the C2 space group with unit cell dimensions of a = 136.0 Å, b = 77.9 Å, c = 32.9 Å, and β = 96.3°. For cryoprotection, the crystals were exchanged into mother liquor containing 30% 2,4-methyl pentanediol and flash-frozen in liquid propane. Diffraction data were measured at beamline X4A of the National Synchrotron Light Source and processed with the programs DENZO and SCALEPACK (45). Selenium sites were identified and refined with the Crystallography & NMR System (CNS) (46), and this phase information plus density modification yielded an electron density map into which most of the RNA and protein could be unambiguously built with the O graphics interface (47). Rounds of manual building, torsion-angle simulated annealing, and individual B-factor refinement with CNS generated the final model. Magnesium ions were identified by their octahedral coordination geometry as well as the distance (2.1 to 2.2 Å) of inner-sphere coordinated water molecules, whereas potassium ions were identified by their peak height in the electron density (>6.5 SDs above mean density level in each case) as well as inspection of the nature of the functional groups forming inner-sphere coordinations. PROCHECK analysis of the protein revealed that all residues lie within the allowed regions in the Ramachandran plot (92% in the most favorable region) with an overall G factor (2.8 bandwidths from the mean) better than that expected for a protein at this resolution (48). In the RNA, all sugar puckers are C3′-endo except for G58, which is C2′-exo, and A42, which was restrained as C2′-endo during refinement.

View this table:

The structure reveals minor groove recognition of a noncanonical RNA helix by a helix-turn-helix (HTH) motif contained within the five α helices of the M domain (Fig. 2, A and B). The HTH fold, corresponding to helices 2 to 4 (Figs. 1D and 2A), is virtually identical to that observed in the crystal structure of the Ffh protein alone as well as that of the M domain of the human SRP54 protein (24–26) (Fig. 2A). The HTH motif presents two α helices to the minor groove of the RNA, making a series of direct and solvent-mediated contacts between conserved amino acids and nucleotides. This mode of binding is distinct from that observed for the structurally similar HTH motifs of DNA binding proteins, which bind to DNA primarily through interactions of a single helix in the major groove (27). A 33–amino acid segment of the protein containing the proposed signal peptide recognition site is disordered in the crystals and was not observed in the electron density map (Figs. 1D and 2A). This observation is consistent with biochemical evidence for conformational flexibility of this region in the absence of a bound signal sequence (20).

Figure 2

(A) (Left) Stereo representation of the E. coli 4.5S RNA–M domain structure. The protein is shown as a cyan ribbon; superimposed on the structure is that of the unbound M domain from T. aquaticus, drawn as a magenta ribbon. The RNA backbone is represented as a dark blue ribbon, with nucleotides that are universally conserved within SRP RNAs colored yellow and those that are highly conserved colored green. This coloring scheme for the RNA is used in all subsequent figures. (Right) Secondary structure of the RNA used for crystallization drawn to reflect the three-dimensional architecture. The G54A mutation in the crystallized RNA is highlighted in cyan, and additional nucleotides that are not native to the E. coli 4.5S RNA are shown in outline. The double arrow indicates the tertiary contact observed between the two internal loops. (B) Molecular surface representation of the complex, oriented to show the proposed signal peptide recognition groove. RNA is dark blue, protein is magenta; hydrophobic residues in the groove are yellow, and adjacent phosphates in the RNA are red. (C) Ribbon representation of the free form of domain IV of the E. coli SRP RNA (16–18), shown in the same orientation as the RNA in (A). Changes observed in the RNA within the complex include exposure of base pairs in the minor groove of the symmetric internal loop, and extrusion of the four bases on the 5′ side of the asymmetric internal loop (see text for details). This and subsequent figures, except where stated, were generated with RIBBONS (49).

In contrast to the protein, the structure of the RNA changes significantly upon binding to the M domain (Fig. 2C). Solution nuclear magnetic resonance (NMR) structures of domain IV of the RNA showed the nucleotide bases of both internal loops stacked into the helix (16–18). In the complex, however, the four nucleotides of the asymmetric loop of the RNA are extruded from the helix. Three of these bases are stacked and wrap around the outside of the helix to form a unique surface that positions invariant nt A39 for contacts with the M domain.

Minor groove recognition mediated by two RNA internal loops. The internal loop nucleotides in domain IV of the 4.5S RNA create an unusual surface along the minor groove face of the helix that forms the recognition site for the M domain of Ffh. The nucleotides of the symmetric internal loop of the RNA form five noncanonical base pairs that fill the minor groove and create a unique surface recognized by the protein (Fig. 3A). A symmetric heteropurine G-A base pair is adjacent to a sheared G-G pair that pinches together the opposing strands of the RNA backbone. The narrowed helical cross section is maintained by two structurally distinct A-C base pairs, stacked beneath the G-G pair. Although the Watson-Crick face of the C in each case interacts with the Hoogsteen face of the A, the A47-C62 pair involves two hydrogen bonds between the bases, while the bases in the C46-A63 pair are bridged in part by a potassium ion (Fig. 3A). The phosphodiester backbone in this region interacts directly with the bases of G61, A63, and C46, stabilizing the unusual base pairs. A cross-strand stack between G61 and C46 displaces the A47-C62 base pair from the helical axis (Fig. 3A). Consequently, the A47-C62 pair is thrust into the minor groove, presenting the carbonyl and exocyclic amine groups of C62 and A47, respectively, for direct recognition by the M domain. As a result of the unique set of base pairs in the symmetric loop, the minor groove surface is flat compared with that of a normal A-form helix and contains a binding pocket for a second potassium ion situated at the G-G pair (Fig. 3B). Notably, there is a dense network of hydrogen bonds between the protein and the RNA in this region (Fig. 3A), most of which involve backbone carbonyl groups in the protein rather than specific side chains. This differs from most protein–nucleic acid complexes in which groove interactions are mediated by protein side chains (28, 29).

Figure 3

Structure of the symmetric internal loop of the 4.5S RNA. (A) Stereo representation of protein-RNA interactions at the symmetric internal loop. The RNA backbone is shown in blue, with 2′-OH groups represented as red spheres; orange spheres, potassium ions; red sphere and rods, hydrated magnesium ion. The G49-A60 pair is a Watson-Crick–type purine-purine pair with hydrogen bonds between the G49 O6 and A60 N6, and G49 N1 and A60 N1. The G48-G61 pair is a sheared type purine-purine pair with only one direct hydrogen bond between G48 O6 and N61 N2. G61 forms further hydrogen bonds between the N1 and N2 positions and a phosphate oxygen of A47, inducing a slight kink in the phosphodiester backbone at this position. The universally conserved A47-C62 form a reverse Hoogsteen pair between A47 N6 and C62 N1, and A47 N7 and C62 N2. The second mismatch contains a single direct hydrogen bond between C46 O2 and A63 N6. This pair has additional base-backbone hydrogen bonds between C46 N4 and a phosphate oxygen of C62, and A63 N6 and C46 O2′. Additionally, the N4 of C62 and O6 of C46 are within hydrogen bonding distance, indicating an additional interaction between these two cytosines to potentially form a “pair of pairs.” The extrusion of the reverse-Hoogsteen A47-C62 base pair toward the minor groove presents this pair for recognition by the protein and induces cross-strand stacking between G61 and C46. Ser-381 Oγ forms base-specific hydrogen bonds with the O2 of C62 and N6 of G48. Helix 2 of the M domain makes further base-specific contacts between Ala-377 O and N6 of G48 and Asn-380 and N2 of A47. At the end of helix 3, the conserved GSG sequence participates in RNA recognition through an interaction between Ser-406 O and O2′ of G49 and the coordination of the potassium ion by Gly-405 O. The close approach of the protein backbone to the RNA dictates the requirement of glycine at position 405, whereas Ser-406 and Gly-407 are necessary for maintenance of the turn between helices 3 and 4. All other contacts between the protein and RNA are mediated through hydrogen bonding to 2′-OH groups, while the phosphate oxygens of the symmetric loop are not used by the protein for recognition. (B) The molecular surface of the RNA, rotated ∼90° around a vertical axis from the view shown inFig. 2A; coloring is according to the curvature of the surface (green, convex; gray, concave). This representation highlights the relatively flat minor groove of the symmetric internal loop, in which there is a large pocket that binds a potassium ion. The close proximity of the phosphodiester backbones of the two internal loops at the site of a tertiary contact creates a groovelike feature that serves as a site for further RNA-protein interactions. This figure was created with GRASP (50).

In the asymmetric internal loop of the RNA, the single unpaired adenosine on the 3′ side remains stacked in the helix. The phosphodiester backbone of the four nucleotides on the 5′ side of the loop is inverted such that the phosphates are pointed toward the interior of the helix. The nucleotide bases are completely expelled from the helix, and nt A39, C40, and C41 stack perpendicular to the helical axis to form an arch that crosses the major groove (Figs. 2A and 3B). A42, a nonconserved residue, is unstacked and serves as a connector between the stack and the helix above it. This unique geometry positions the universally conserved nt A39 to contact invariant residues in both the symmetric internal loop of the RNA and in the protein (Figs. 2A and 3B).

The heart of the RNA-protein interface is a protein-mediated interaction between the symmetric and asymmetric internal loops. A dense network of contacts occurs between helices 2b and 3 in the M domain (Fig. 1D) and the A47-C62 pair and A39 in the RNA (Fig. 4A). The location of A39 on the outside of the RNA helix enables stacking with the first invariant arginine residue (Arg-398) and hydrogen bonding to the second (Arg-401) within the highly conserved amino acid sequence RXXRXXXGSG (residues 398 to 407) in helix 3 of the M domain. Furthermore, the 2′-OH of A39 is hydrogen-bonded to the phosphate of A63 between the two A-C base pairs of the symmetric loop, directly connecting the two internal loops of the RNA.

Figure 4

Asymmetric loop and protein contacts. (A) Within the asymmetric loop, nt A39, C40, and C41 form a continuous stack that is capped by Arg-398. Additional contacts are between A39 N3 and NH1 of Arg-401, and A39 N1 and Ser-397 Oγ. Also shown is the single tertiary contact observed in the RNA structure between A39 2′-OH and a phosphate oxygen of A63. (B) The heart of the 4.5S RNA–M domain interaction consists of a hydrogen bonding network between universally conserved residues. The salt bridge between Glu-386 and Arg-401 is cradled between the asymmetric and symmetric loops of the RNA.

Arg-398 and Arg-401 form part of a central salt bridge that includes the invariant Glu-386 and the polypeptide backbone of helix 2b (Fig. 4B). In the structure of the uncomplexed M domain of human SRP54, these side chains were proposed to be inaccessible for direct RNA recognition (25). However, with no rearrangement upon RNA binding, these residues form critical contacts to the RNA. Instead of extending into the RNA major or minor grooves, the arginines in the salt bridge present a surface to which the RNA binds. The bridge links the two internal loops of the RNA by way of hydrogen bonds between Arg-401, the N3 of A39 in the asymmetric loop, and the 2′-OH of C62 in the symmetric loop (Fig. 4B). Thus, unlike RNA binding proteins including L11, U1A, U2B" and sex-lethal (29), and L30 (30), the M domain of Ffh is preorganized for SRP RNA recognition.

A network of waters and metal ions facilitate the RNA-protein interaction. One of the most striking features of the M domain–RNA complex is the abundance of well-ordered solvent molecules and metal ions that are coordinated to the RNA and, in one case, bridge invariant residues at the RNA-protein interface (Fig. 3A). The extrusion of the asymmetric internal loop nucleotides from the helical stack creates a large cavity in the center of the RNA helix that is filled with cations and water molecules (Fig. 5, A and B). Two hydrated magnesium ions and 28 water molecules were unambiguously identified within this cavity. The waters have temperature factors similar to those of the surrounding atoms of the RNA, indicating that they are highly ordered and occupy specific binding sites (Fig. 5B). In the symmetric loop of the SRP RNA, two potassium ions are coordinated to the noncanonical base pairs. One potassium ion, bound to the carbonyl oxygen of G48 and the N3 and 2′-OH groups of G61 in the G-G base pair, also coordinates to the backbone carbonyl of Gly-405 (Fig. 3A). The functional groups that define this ion-binding pocket are universally conserved in the SRP, suggesting that the potassium ion is integral to their interaction. The other potassium ion is situated between the C46-A63 base pair and the U45-G64 pair in the major groove side of the symmetric internal loop. This ion is also coordinated to several water molecules, forming part of the core of the extensive solvent network in the major groove of the asymmetric loop (Fig. 5, A and B).

Figure 5

Metals and solvent in the protein-RNA complex. (A) Solvent-flattened experimental electron density map contoured at 1.5 SDs above the mean density level, superimposed on the refined atomic model, shows well-ordered solvent molecules (cyan), hydrated magnesium ions (red), and a potassium ion (orange sphere) located in the interior of the asymmetric internal loop. (B) Ribbon and stick diagram of the complex emphasizing the hydrogen bonding network (magenta dashes) between solvent molecules, magnesium ions, and a potassium ion in the asymmetric loop. The averageB factor for the 28 water molecules represented is 28.5, which is only slightly higher than that of RNA atoms surrounding these solvents, indicating that they are well ordered. (C) Histogram of interference values for the M domain binding region of the 4.5S RNA; nucleotide positions correspond to the numbering for the E. coli 4.5S RNA. The dashed line represents an interference value of 2, which is considered to be a site whose functional group substitution significantly impairs protein binding (32). (D) Substitution of the pro-Rp oxygen by sulfur or the N7 by carbon of A63 (shown as red spheres) results in a significant interference of protein binding. These two functional groups form hydrogen bonds with well-ordered water molecules and coordinate a potassium ion, respectively, indicating that these solvent molecules contribute to the stabilization of the protein-RNA complex.

To test whether these ions are functionally important for the formation of the SRP complex, we probed their coordination to the RNA using nucleotide analog interference mapping (31,32). Full-length E. coli SRP RNA (110 nt) was transcribed in vitro to contain randomly incorporated 7-deaza adenosine or 2′-deoxyguanosine nucleotide analogs tagged with α-phosphorothioate groups (7dAαS and dGαS, respectively). These analogs were chosen for study because N7 of adenosine and the 2′-OH of guanosine are involved in numerous structural contacts between the RNA and protein. From the resulting pools of RNA, molecules competent to bind to the M domain protein were selected, and positions of incorporated nucleotide analogs were identified by site-specific cleavage.

Six of the RNA functional groups tested in this way were found to make significant energetic contributions to the formation of the SRP complex (Fig. 5C). Strikingly, removal of one of the coordinating ligands to either of the potassium ions seen in the crystal structure (dGαS at G61 and 7dAαS at A63) was deleterious to protein binding (Fig. 5, C and D). This result is analogous to observations made in group I self-splicing introns, where potassium ions bind and stabilize RNA tertiary structure (33). Furthermore, phosphorothioate substitution alone at A63 interfered with complex formation. Rather than participating in a direct RNA-RNA or RNA-protein contact, this phosphate oxygen is hydrogen-bonded to two well-ordered water molecules that coordinate directly to one of the potassium ions (Fig. 5D). Because these two waters are part of the solvent network seen in the asymmetric loop, this network appears to substantially stabilize the RNA structure required for protein recognition.

Three additional sites where analogs interfere with protein binding are observed in the symmetric internal loop. Two of these involve direct protein-RNA contacts at the 2′-OH groups of G48 and G49 (Fig. 3A). The third, a 7-deaza substitution at A47, disrupts one of the two hydrogen bonds in the A-C base pair at the heart of the RNA-protein interface (Figs. 3A and 4B). Furthermore, results of detailed chemical probing of the 4.5S RNA-Ffh interaction (15) as well as additional analog interference experiments (34) are consistent with the crystal structure. Thus, interactions observed crystallographically are functionally important for formation of the SRP.

In vivo function of the SRP core RNA. Previous studies of the E. coli SRP suggested that domain IV of the RNA is necessary and sufficient for specific recognition by Ffh (15, 35) (Fig. 1, B and C). Its binding affinity for the M domain is virtually identical to that for the full-length Ffh protein and it also supports GTP hydrolysis by Ffh and FtsY in vitro (36). To test whether the 49-nt domain IV RNA used for crystallography is functional in vivo, we used a strain of E. coli (S1610) lysogenic for λ phage carrying the essential gene ffs encoding the 4.5S RNA (37). Under normal growth conditions, the phage is present in the chromosome, the 4.5S gene is expressed, and the cells can grow. Upon heat shock the λ lysogen is excised, removing the 4.5S gene and preventing bacterial cell growth. This strain is ideal for testing the ability of engineered constructs of the 4.5S gene, supplied on plasmids, to support cell growth at the nonpermissive temperature.

Bacterial cells were transformed with plasmids containing a series of constructs of the 4.5S RNA gene cloned behind its native promoter (Fig. 6A). When these constructs were expressed from low-copy plasmids, all but the smallest RNA (Fig. 6A, min.) supported growth at the nonpermissive temperature (38). The min RNA construct supported growth at the nonpermissive temperature when expressed from a high-copy vector. Analogous RNA constructs containing a point mutation (C62G), which disrupts the protein-RNA interface (Fig. 4B) and thus prevents binding to Ffh (35), failed to support growth at the higher temperature (38,39). As a control for proper RNA expression and processing, and to ensure that endogenous 4.5S RNA was absent in cells grown at the nonpermissive temperature, total cellular RNA was tested by Northern blot analysis (Fig. 6B). These results show that all of the SRP RNAs are stable and are correctly processed in vivo. Thus, the nonconserved portion of the SRP RNA is not essential in E. coli, though it may play a role in stabilizing the RNA and preventing premature degradation. This demonstrates that the RNA construct that has been crystallized contains all of the elements that are necessary and sufficient for the essential function of the 4.5S RNA in vivo.

Figure 6

In vivo function of truncated forms of the SRP RNA. (A) Sequences of the RNAs tested for in vivo function. The C62G mutation that was used as a control for proper RNA function is highlighted in red and the G54A tetraloop mutation used in the crystallized RNA is shown in cyan. (B) Northern blot analysis of SRP RNAs tested in vivo.E. coli S1610 was transformed with plasmids containing wild-type (wt) or minimized (E4 and min) SRP RNAs, followed by a brief heat shock at 45°C to denature the temperature-labile λ repressor and induce genes responsible for prophage excision (37). These cultures were allowed to further incubate at 30°C, followed by plating on LB agar. At the permissive temperature (30°C), the chromosomal copy of the 4.5S RNA gene is not lost, and thus all cells are viable. At the nonpermissive temperature (42°C), the sole source of SRP RNA is transcribed from a gene encoded on the plasmid. In each case, the RNA with the wild-type domain IV sequence supported cell growth, whereas incorporation of the C62G mutation was lethal (38). Total cellular RNA (5 μg) isolated from these transformants wasadded to each lane (2 to 7) alongside RNAs of known length transcribed in vitro (lane 1) and subjected to electrophoresis on a 12% denaturing polyacrylamide gel. The RNA was transferred to a positively charged nylon filter by electroblotting, followed by probing with a psoralen-biotin labeled oligonucleotide that is complementary to nt 32 to 74 of the E. coli 4.5SRNA; detection was by chemiluminescence. Lane 1, in vitro–transcribed SRP RNAs that are 110, 51, and 39 nt long serve as molecular weight markers; lane 2, RNA from S1610 cells grown at 30°C; lane 3, RNA from cells transformed with a plasmid encoding the wild-type RNA; lane 4, RNA from cells transformed with a plasmid encoding RNA E; lane 5, RNA from cells transformed with a plasmid encoding RNA E4; lane 6, RNA from cells transformed with a plasmid encoding RNA E4 (G54A); lane 7, RNA from cells transformed with a plasmid encoding RNA min. The E4 RNA (lane 5) reproducibly yielded a smaller degradation product, which was also observed in other highly minimized SRP RNAs (38).

Although the portion of the SRP RNA outside of the Ffh recognition site is dispensable in vivo, what is the role of the conserved RNA in SRP function? In the crystal structure, the signal peptide binding site proposed by Keenan et al.(24) lies alongside the RNA backbone adjacent to the symmetric internal loop (Fig. 2, A and B). This creates a continuous molecular surface in which two-thirds of the cleft is contributed by the protein and one-third by the RNA backbone. While signal peptides vary widely in sequence, they are characterized by a hydrophobic region of 6 to 15 amino acids flanked by 2 to 5 positively charged residues. Thus, a binding site composed of both protein and RNA would neatly accommodate such a sequence through a combination of hydrophobic interactions and electrostatic contacts.

The SRP RNA is required to elicit GTP hydrolysis during each cycle of signal sequence binding, protein translocation, and release by the SRP in both bacterial and eukaryotic systems (40,41). Thus, the RNA–M domain complex (the ribonucleoprotein domain) within the SRP may function analogously to GTPase activating proteins (GAPs) that have been characterized in other GTPase-regulated systems. While most of the universally conserved amino acids and nucleotides are buried in the M domain–RNA interface, A39 and Arg-398 are stacked on the surface of the complex (Fig. 4A). Although invariant within Ffh/SRP54 proteins, Arg-398 is not required for stable association of the SRP RNA with Ffh (42), suggesting another role for this region of the SRP core. An intriguing possibility is that the A39–Arg-398 stack modulates GTPase activity of Ffh and/or FtsY (10), perhaps by enabling the Arg-398 residue to facilitate GTP hydrolysis as has been observed for other GAPs (43).

The structure and analysis of the SRP core presented here provides a detailed understanding of one of the most conserved ribonucleoprotein elements in biology. The unique interactions of protein, RNA, cations, and ordered water molecules create the core of the SRP responsible for signal peptide recognition. Together with earlier crystallographic studies of the NG domain of Ffh (9) and the NG domain of the SRP receptor (11), structures of the all of the conserved functional centers required for protein translocation by the SRP are now in hand. These structures provide the foundation for understanding how these elements interact to effect protein targeting in all cells.

  • * To whom correspondence should be addressed. E-mail: doudna{at}


View Abstract

Stay Connected to Science

Navigate This Article