Structure of the chloroplast ribosome using cryo-EM and single-particle reconstruction

To better understand chloroplast translation we have identified the protein components of chloroplast ribosomes and ribosome associated proteins using a proteomic approach (Yamaguchi et al., 2002, Yamaguchi et al., 2003). Many orthologues of E. coli 70S ribosomal proteins were identified in chloroplast ribosomes. In addition, a novel S1 domain-containing protein was identified, as well as novel plastid specific ribosomal proteins (PSRPs). Several ribosomal proteins that show good homology with their bacterial orthologs also contain large additional domains, including S2 (57 kDa), S3 (76 kDa), and S5 (84 kDa). These chloroplast ribosomal proteins are much larger than their E. coli counterparts, containing N-terminal extensions (S2 and S5) or insertion sequences (S3). Structural predictions based on the crystal structure of the bacterial 30S subunit suggest that the additional domains of the chloroplast S2, S3 and S5 are located adjacent to each other on the solvent side of the subunit near the binding site for the S1 protein. These additional domains may interact with the novel S1 proteins in aspects of mRNA recognition or translation initiation that are unique to chloroplast.

Since our proteomic analyses revealed that chloroplast ribosomes were significantly different in protein composition from bacterial ribosomes, we determined that it was important and informative to calculate the structure of the chloroplast ribosome and determine the locations of chloroplast-unique proteins and domains. The small subunit of the C. reinhardtii ribosome has 25% more mass than the bacterial small subunit, due to additional chloroplast-unique proteins and protein domains as introduced above. The rRNA components of the chloroplast ribosome and bacterial ribosomes are quite conserved, and overall 92% of the bacterial small subunit and 94% of the large subunit proteins have clearly identifiable orthologs in the chloroplast ribosome, demonstrating that the chloroplast ribosome has a bacterial core with additional domains.

In collaboration with the National Resource for Automated Molecular Microscopy (NRAMM) at TSRI, we first solved the structure of the more abundant cytosolic ribosome from C. reinhardtii using cryoelectron microscopy and single particle reconstruction. We accompanied this structural work with a comprehensive proteomic analysis of the cytosolic 80S ribosome (Manuell et al., 2005). These studies indicated that though separated by hundreds of millions of years of evolution, 80S ribosomes from mammals, yeast, and photosynthetic algae are quite similar in structure and composition, and by deduction, function. In part we did this to verify the feasibility of such structural studies, and in part we needed to confirm that C. reinhardtii is not a non-model organism in terms of ribosome composition and structure.

We have currently calculated the structure of the chloroplast ribosome to 15.5 angstroms resolution. We are able to computationally fit bacterial ribosome crystal structure data to our chloroplast ribosome map and calculate two different types of difference map. The first of these differences, bacterial ribosome structures that are lacking from the chloroplast ribosome, allow us to validate the quality of our ribosome structure. The small regions of density falling into this difference category are nearly all explained as helices that are lacking from the chloroplast rRNAs, or bacterial proteins that were not identify in the chloroplast ribosome. Regions of the chloroplast ribosome that are heavily conserved with the bacterial ribosome structure are the subunit interface, the factor-binding region, and most of the large subunit, which houses the peptidyl transferase reaction active site.

The second type of difference density we can calculate is structures found on the chloroplast ribosome and not on a bacterial ribosome. Even at this moderate resolution large chloroplast-unique (cu) structures are easily visualized, mostly on the small subunit. Computational modeling using a bacterial ribosome crystal structure can assist in the potential assignment of identity to these cu structures, although further experiments will be necessary to validate this hypothesis. For example, cu density in the beak and head region is closely associated with the location of protein S3, hence is likely composed of the chloroplast-unique domain we have identified on S3.

Continued refinement of this structure coupled with genetic and biochemical analyses of the unique domains should allow us to unambiguously assign specific proteins to the chloroplast-unique structures on the small subunit to specific functions within chloroplast translation.