Identification of an RNA-Protein Bridge Spanning the Ribosomal Subunit Interface

See allHide authors and affiliations

Science  24 Sep 1999:
Vol. 285, Issue 5436, pp. 2133-2135
DOI: 10.1126/science.285.5436.2133


The 7.8 angstrom crystal structure of the 70Sribosome reveals a discrete double-helical bridge (B4) that projects from the 50S subunit, making contact with the 30Ssubunit. Preliminary modeling studies localized its contact site, near the bottom of the platform, to the binding site for ribosomal protein S15. Directed hydroxyl radical probing from iron(II) tethered to S15 specifically cleaved nucleotides in the 715 loop of domain II of 23S ribosomal RNA, one of the known sites in 23Sribosomal RNA that are footprinted by the 30S subunit. Reconstitution studies show that protection of the 715 loop, but none of the other 30S-dependent protections, is correlated with the presence of S15 in the 30S subunit. The 715 loop is specifically protected by binding free S15 to 50S subunits. Moreover, the previously determined structure of a homologous stem-loop from U2 small nuclear RNA fits closely to the electron density of the bridge.

Ribosomes are large ribonucleoprotein complexes that are responsible for the fundamental process of protein synthesis. They are composed of two asymmetric subunits, each of which contributes to specific functions during translation. The interface between these subunits allows for the coordination of these discrete functions and also provides the binding surfaces for many substrates and ligands. Thus, the identification of specific molecular interactions between the two subunits is of great importance. Numerous experiments have identified RNA and protein elements that potentially contribute to this subunit-subunit interface (1–3). However, in the absence of high-resolution structural information, identification of the molecular components comprising specific subunit-subunit interactions has been difficult.

The 7.8 Å x-ray crystal structure of the Thermus thermophilus 70S ribosome (4) shows that the two ribosomal subunits are connected by a complex network of molecular interactions. One of these (bridge B4) can be identified as a double-stranded RNA stem-loop that is continuous with the 50S subunit and makes contact with the bottom of the platform of the 30S subunit (Fig. 1). Immunoelectron microscopy and preliminary modeling studies of the 30S subunit based on extensive biochemical, biophysical, and phylogenetic evidence localize the binding site for protein S15 to this region of the 30Ssubunit (5). Additionally, evidence for the placement of S15 at the subunit interface has come from intersubunit cross-linking studies (2) and a temperature-sensitive S15 mutant that is defective in subunit association (6).

Figure 1

Electron density from the 7.8 Å crystal structure of the T. thermophilus 70S ribosome (4) showing interaction of a discrete RNA feature of the 50S subunit (white) with the bottom of the platform of the 30S subunit (blue). Electron density is contoured at 1.1σ.

To test the possible proximity of S15 to 23S ribosomal RNA (rRNA), we performed directed hydroxyl radical probing (7). Iron(II) was tethered by a linker, 1-(p-bromoacetamidobenzyl)-EDTA (BABE) (8), to unique cysteine (C) residues on the surface of S15 at amino acid positions 12, 36, 46, and 70 by directed mutagenesis (9), using the published solution and crystal structures of S15 as a guide (10). The Fe(II)-derivatized proteins were incorporated into 30S subunits by in vitro reconstitution (11) and associated with 50S subunits to form 70Sribosomes, which were then purified by sucrose gradient centrifugation. The 30S subunits containing S15 derivatized at position 70 failed to associate with 50S subunits, although they appeared to be normally assembled according to other criteria (12). This result is also consistent with contact between S15 and the 50S subunit. Localized hydroxyl radical production was initiated, and the entire 23S rRNA chain was scanned for sites of rRNA cleavage by primer extension (13). A single region of 23S rRNA, the 715 stem-loop of domain II, was targeted by hydroxyl radicals generated from positions 12 and 46 of S15 (Fig. 2) (14). Overlapping cleavage patterns were observed; nucleotides 707 to 711 and 715 to 717 were targeted by Fe(II)-C12-S15 (Fig. 2), and nucleotides 706 and 707 and 716 to 722 were cleaved by Fe(II)-C46-S15 (Fig. 2). No other cleavage was observed in 23S or 5S rRNA from Fe(II)-S15 (12). Placement of the 715 stem-loop at the subunit interface is supported by hydroxyl radical footprinting studies using free Fe(II)-EDTA (3), which showed that this same region of domain II became protected upon 70S ribosome formation (Fig. 2).

Figure 2

(A) Directed hydroxyl radical cleavage of 23S rRNA in 70S ribosomes containing Fe(II)-derivatized ribosomal protein S15 in reconstituted 30S subunits. Sites of cleavage were localized by primer extension with reverse transcriptase. A and G, sequencing lanes; wt, mock Fe(II)-BABE–treated (cysteine-free) wild-type S15; C12, Fe(II)-C12-S15; C36, Fe(II)-C36-S15; C46, Fe(II)-C46-S15. Vertical lines at the right indicate regions of cleavage. (B) Positions of directed hydroxyl radical cleavage from Fe(II)-C12-S15 and Fe(II)-C46-S15. Nucleotides within this region of 23S rRNA that are protected by subunit association from free hydroxyl radicals (3) are also shown. Intensities of cleavages are indicated by large, medium, or small solid circles (high through low intensity, according to size).

The possibility of direct interaction of S15 with the 715 stem-loop was first tested using footprinting analysis of in vitro reconstituted 30S subunits lacking S15 (15,16). Solution hydroxyl radical footprinting (17) of the 70S particles showed that the 30S-dependent protection of the 715 region of 23SrRNA was abolished as a result of S15 omission (Fig. 3A), although 30S-dependent footprints were unaffected elsewhere in 23S rRNA (Fig. 3, B through D). Protection of 23S rRNA by 30Ssubunits reconstituted with S15 was indistinguishable from that of natural 30S subunits (Fig. 3, A through D).

Figure 3

(A through D) Free Fe(II)-EDTA–generated hydroxyl radical footprinting of 23S rRNA in 50S subunits by 30Ssubunits lacking S15. Protection of 23S rRNA by S15-deficient 30S subunits is identical to that of S15-containing subunits for the (B) 900, (C) 1700, and (D) 1940 regions, but is abolished in the 715 region (A). A and G, sequencing lanes; K, unmodified 50S subunits; K′, unmodified 70S ribosomes. Samples in all other lanes are modified with hydroxyl radicals generated from free Fe(II)-EDTA. Dash, 50S subunits; n30S: natural 30S subunits; r30S-S15, in vitro reconstituted 30S subunits lacking S15; r30S+S15, in vitro reconstituted 30S subunits containing S15. (E) Free Fe(II)-EDTA–generated hydroxyl radical footprinting of 23S rRNA in 50S subunits by free protein S15. All lanes contain 20 pmol of 50Ssubunits. Dash, no S15; 3×, 60 pmol of S15; 9×, 180 pmol of S15. Vertical lines at the right indicate regions of cleavage.

To further test the possible interaction of S15 with the 715 loop, we bound free S15 protein to 50S subunits, and we assessed its interaction with 23S rRNA by hydroxyl radical footprinting (18). The 715 loop of 50S subunits was protected in a concentration-dependent manner by S15 in a way that is similar to that observed with intact 30S subunits (Fig. 3E). No change in the reactivity of any other 23S rRNA nucleotides was observed upon S15 binding, and no change in reactivity of any 23S rRNA nucleotides was observed upon binding a mixture of small subunit ribosomal proteins lacking S15 (12). These results suggest a specific interaction between ribosomal protein S15 and the 715 stem-loop of 23S rRNA. Initial efforts to footprint S15 on naked 23S rRNA or on an RNA fragment corresponding to the 715 loop have proven unsuccessful. We conclude that an intersubunit contact is formed by the interaction of protein S15 with the 715 stem-loop of 23S rRNA.

There is a striking similarity between the 715 stem-loop of 23S rRNA and stem-loop IIA of U2 small nuclear RNA (snRNA), whose solution structure has been determined by nuclear magnetic resonance spectroscopy (19). The loop sequences of theEscherichia coli 715 loop and U2 loop IIA are identical, whereas that of T. thermophilus differs from them in only two positions (Fig. 4A). The close fit of the stem-loop IIA structure to the electron density of the observed bridge (Fig. 4B) provides further confirmation of the identification of this molecular feature of the ribosome interface as the 715 stem-loop of 23S rRNA. The arrangement of the loop suggests that nucleotides around the conserved purines at positions 715 and 716 (E. coli numbering) make specific interactions with the 30S subunit (Fig. 4B). At lower contour levels, electron density from the minor groove side of the 715 stem-loop merges with that of the 30S subunit, consistent with the observed hydroxyl radical protection pattern. At the current resolution of the 70S ribosome structure, the regions of S15 that participate in this intersubunit contact cannot be unambiguously identified and will require further structural and biochemical studies. It should be possible to use similar combinations of x-ray structural information and biochemical probing analysis to identify additional molecular features of the ribosome and its functional complexes.

Figure 4

(A) Sequences of the 715 stem-loop (E. coli numbering) from E. coli and T. thermophilus and the U2 snRNA stem-loop IIA. (B) Stereo image of the fit between the solution structure of U2 snRNA stem-loop IIA (19) and the electron density of the intersubunit bridge (4), contoured as in Fig. 1.

  • * Present address: Department of Biochemistry, Biophysics, and Molecular Biology, Iowa State University, Ames, IA 50011, USA.

  • Present address: Whitehead Institute for Biomedical Research and Massachusetts Institute of Technology, Cambridge, MA 02142, USA.


View Abstract

Navigate This Article