ABSTRACT
The 80S ribosome from Saccharomyces cerevisiae has been reconstructed from cryo electron micrographs to a resolution of 35 Å. It is strikingly similar to the 70S ribosome from Escherichia coli, while displaying the characteristic eukaryotic features familiar from reconstructions of ribosomes from higher eukaryotes. Aside from the elaboration of a number of peripherally located features on the two subunits and greater overall size, the largest difference between the yeast and E.coli ribosomes is in a mass increase on one side of the large (60S) subunit. It thus appears more elliptical than the characteristically globular 50S subunit from E.coli. The interior of the 60S subunit reveals a variable diameter tunnel spanning the subunit between the interface canyon and a site on the lower back of the subunit, presumably the exit site through which the nascent polypeptide chain emerges from the ribosome.
The recent completion of the sequence of the genome of Saccharomyces cerevisiae (1), as well as determination of the entire set of mammalian ribosomal proteins (2; I.G.Wool, personal communication), has permitted us a look at the entire complement of the yeast ribosome. The yeast ribosome is composed of two subunits. The 60S subunit contains three RNA molecules: 25S RNA of 3392 nt, hydrogen bonded to the 5.8S RNA of 158 nt and associated with the 5S RNA of 121 nt. There are 42 proteins in the large subunit, plus two copies each of two acidic exchangeable proteins. Altogether there are 3671 nt and 7235 amino acids in the 60S subunit, giving it a ratio of ~61% RNA to 39% protein. The 40S subunit has a single RNA of 1798 nt and 32 proteins with a total of 4749 amino acids, giving it a ratio of 54% RNA to 46% protein. All identified yeast ribosomal proteins are represented by homologous proteins in the mammalian ribosome, which has a single protein, L28, that appears not to be present in yeast. It will be interesting to learn whether L28 is present in ribosomes of higher plants as well as those of intermediate species such as Caenorhabditis elegans and Drosophila. Mammalian ribosomes differ from fungal ribosomes mainly in the size of the large RNA of the large subunit: mammalian 28S RNAs may exceed 5000 nt in length, compared with roughly 3400 nt for the fungal (as well as higher plant) 25S RNAs.
Thus, except for the mammalian ribosome having longer RNA molecules and slightly larger proteins, there is a remarkable conservation of the elements of this master machine of translation.
Following our study of the three-dimensional (3D) structure of the wheatgerm ribosome (3) we felt it valuable to examine in detail the structure of another eukaryotic ribosome, one that provides the opportunity of exploration using genetic methods.
Ribosomes were isolated from S.cerevisiae strain W303 (4) essentially as described (5). Briefly, the cells were collected by centrifugation, washed in water and resuspended in TMN (50 mM Tris-acetate, pH 7.4, 50 mM NH4Cl, 12 mM MgC12, 1 mM DTT) and broken by vigorous agitation with glass beads. After two cycles of low speed centrifugation the supernatant was centrifuged for 3 h at 50 000 r.p.m. through a cushion of TMN containing 5% NH4(SO4)2 and 10% sucrose, in order to separate the ribosomes from any soluble proteins and to strip off any loosely bound proteins. Subunits derived from ribosomes prepared in a nearly identical way have proved to have activity in several protein synthesis assays (6,7). This was also the method used to prepare the ribosomes used in the definition of the set of ribosomal proteins present in yeast (5).
Grids were prepared for cryo microscopy according to standardized methods in our laboratory following Wagenknecht et al. (8) and Dubochet et al. (9). Micrographs were recorded using low dose protocols in a Philips EM420 at a magnification of 38 000×; the resulting pixel size was 5.26 Å.
Twelve micrographs at 0°, 35° and 50° tilts were analyzed. Analysis was of 7470 projections with a projection matching scheme (10). For this an initial reference was a low pass filtered structure obtained previously (3) for the wheatgerm ribosome. This wheatgerm structure had been obtained from tilt pair data according to a random conical reconstruction scheme, with the missing cone information filled in by projection onto convex sets (POCS; see for example 11). After the first iteration of reconstruction for the initial yeast data set, however, the reference was successively replaced by the current highest quality yeast ribosome reconstruction.
Initially 2348 projections from four 0° micrographs were processed, but multivariate statistical analysis and hierarchical ascendent classification revealed the paucity of certain view ranges, indicating that the 80S ribosome could not be reconstructed without tilt data, in contrast to the E.coli 70S ribosome. Subsequently tilt micrographs were added two by two. These data sets were processed independently, after incorporation of the original tilts allowed adequate infilling of the angular space. Analysis of subsequent data increments was based on the reference derived from this stage, but the projection sets used were completely independent. This allowed the quality and completeness of the 35° and 50° tilt sets to be gauged. Finally, all data sets were pooled and the SIRT back-projection algorithm (12) was iterated until the results stabilized.
The full data set was highly anisotropic in nature, with an extreme over-representation of only two or three preferred views, in striking contrast to the E.coli 70S ribosome, which assumes a sufficient number of orientations that tilt micrographs are not needed. For the eukaryotic ribosome it was found necessary to collect 35° and 50° tilt images to fill in the areas of missing information in Fourier space. The main consequence, beyond the necessity of recording tilt data, was that a large number of iterations of the SIRT back-projection algorithm were needed to overcome the weighting bias caused by the preponderance of [Phi] = 0, [theta] = 0 projections. For the final reconstruction 200 iterations of the SIRT algorithm were used. The resultant structure was examined and compared with existing structures for the eubacterial and eukaryotic ribosomes through surface representations. The structure was further analyzed using the AVS (Advanced Visual Systems Inc., Waltham, MA) and O (13) graphics softwares.
The 80S ribosome from S.cerevisiae has been reconstructed from 7470 individual ribosome images to a resolution of 35 Å according to the Fourier Shell Criterion measure, using a critical value of 0.5 (14), corresponding to a signal-to-noise ratio of ~2 (15). This value closely agreed with the value obtained by the 45° phase residual criterion used in previous studies. Low pass filtration to 35 Å and, for correction of the transfer function, Wiener filtration were then applied.
The yeast ribosome is a bipartite structure that varies from roughly equidimensional to somewhat elongate depending on viewing angle (Fig. 1). In the orientation shown in Figure 1A the two subunits are seen side by side, 40S subunit on the left and 60S subunit on the right, with their `heads' at the top. In this view the height of the ribosome is ~254 Å, the width ~278 Å and the thickness ~267 Å. These height and width dimensions are 11-14% smaller than those calculated for the ribosome from a higher eukaryote, wheatgerm (3), but greater than those for the 70S E.coli ribosome (Fig. 1). That the thickness dimension is significantly greater for the yeast structure than for wheatgerm could be a result of experimental differences (see Discussion).
Although a number of low density internal features can be recognized in the yeast ribosome; features continuous between the interface surface and the back of the large subunit are of particular interest. These represent candidates for the postulated tunnel through which the nascent polypeptide chain exits the ribosome (see Discussion). In the following we make use of surface representations (Fig. 2A and B) and a cutaway stereo representation (Fig. 3) to investigate two low density features that issue from a deep depression in the IC and traverse the subunit to debouche at sites on its back. The lower tunnel feature is strikingly goblet-shaped (Fig. 3), tapering rapidly to a narrow stem that ends as a small hole on the lower back of the subunit (Fig. 2B). The upper tunnel feature is broader and less variable in diameter and it is situated more to the stalkward side of the subunit, i.e. in a plane closer to the viewer in Figure 3. This latter broad tunnel is sectioned by the clipping plane in this figure, whereas the lower tunnel is further from the viewer (closer to the midline of the subunit), behind the sectioning plane. The broad tunnel ends as a very large cavity in the mid back of the subunit (Fig. 2B).
Comparison of large subunits from the eubacterium E.coli and the eukaryote S.cerevisiae (18,21; Malhotra et al., submitted for publication; Verschoor et al., work in progress) reveals that the lower tunnel is a highly conserved morphological feature and thus merits particular attention. It provides a strikingly direct conduit between the interface and back aspects of the subunit. As can be seen from Figure 3, it appears to represent the shortest possible connection, in geometric terms, between the interface and cytoplasmic surfaces of the subunit or between the classical translational and exit domains.
Overall the yeast ribosome is strikingly similar to the E.coli ribosome at a comparable resolution (Fig. 1). It is somewhat larger and more elaborated morphologically, but the degree of agreement in the shapes as the structures are rotated is remarkable. For many features the difference is a matter of degree; generally the E.coli ribosome has a counterpart feature to the yeast, but it is less strongly expressed. Particularly notable is the small subunit beak.
In most of the views shown in Figure 1 (except for panels B, F and G) the yeast ribosome strongly resembles the 70S E.coli ribosome. The principal difference is in the shape of the small subunit. In views where the small subunit is seen in lateral view (e.g. Fig. 1B and C) the characteristically eubacterial versus eukaryotic features of this subunit allow the two ribosomes to be identified unambiguously. Although the E.coli 30S subunit is now recognized as having a broad nose-like front portion to the head, with a very marked bend to the tip, the beak or bill of the eukaryotic subunit differs in its greater degree of extension. Likewise, the crest described for the eukaryote (22-24) remains a key identification element, even subsequent to recognition (e.g. Fig. 4E and F in 18) of a comb- or wattle-like feature trailing down the back of the neck of the E.coli subunit (see 19, fig. 4A and B). The eukaryotic crest is a feature of the head per se and has not been observed to lie along the neck.
The features that do differ tend to be peripherally located. In the 40S subunit we see no trace of the spur feature noted for the E.coli subunit (10,18-19), but rather a broad square base. The classical eukaryotic feet (3,25) are recognizable, though not highly elaborated in this yeast structure; a small conical front foot can be discerned (Fig. 1A and H-J). Also, the platform, while now recognized as a well-developed feature in eukaryotes (cf. 3), is nevertheless not as extended in the yeast structure as in the E.coli structure (Fig. 1D).
Recent E.coli ribosome reconstructions (16,18-19) have revealed some surprising features, which in fact serve to increase the resemblance to the 40S subunit, with all of its so-called eukaryotic additions. Among the most marked similarities are comparable morphologies of the head, platform and the groove that constitutes the interface surface of the small subunit neck. Exploded views (Fig. 1K) reveal virtually identical morphologies in this key region of the translational domain (see below).
For rotations in which there is a difference in overall form between the two ribosome structures this difference is usually attributable to the larger and more ellipsoidal form of the 60S subunit, as compared with the globular 50S subunit of the E.coli ribosome. The 60S subunit is characterized by its obliquity (relative to an axis in which the central protuberance is directed vertically), with a trapezoidal rather than circular outline. On the side of the ribosome bearing the so-called L1 analog arm the base of the subunit has a strongly diagonal facet that seems to represent a local concentration of ribonucleoprotein not seen in the 50S subunit.
This report characterizes the third medium resolution eukaryotic ribosome structure to be achieved by cryo electron microscopic methods, after the wheatgerm (3) and rabbit (40S subunit only; 24) ribosomes. In comparing the constituents of these three ribosomes we note that the yeast and plant rRNAs are quite similar in size, while the mammalian rRNAs are larger, by ~10% for the 40S subunit and by nearly 33% for the 60S subunit (26). The current catalog of ribosomal proteins is probably complete for yeast and mammals. The latter have only a single additional protein, although many mammalian ribosomal proteins are slightly larger than their yeast homologs. While we as yet lack sufficient data to fully evaluate the plant ribosomal proteins, there is no reason to believe, from the data described thus far, that they will be appreciably different.
The principal difference between the two 80S ribosome structures, wheatgerm and yeast, is a difference in proportions. Relative to the wheatgerm structure, the yeast ribosome appears more compact in height and width (Fig. 2C) and notably more globular. It is also greater in thickness (see below). Compared with the wheatgerm structure, resolution into two separate subunits is superior for the yeast structure in the lower or body portion of the ribosome. Except for the central bridge and the bridge between the bases, the two subunits are completely distinct, at least at the threshold used in this study.
The notable difference in thickness between the yeast and wheatgerm structures points up a cautionary note. Several differences in methods exist between the studies, which make a strict comparison problematical. The wheatgerm data had a missing cone of angular information, due to the random conical tilt geometry used, which was compensated for by the use of POCS techniques (see for example 11). In the yeast study, however, angular space was completely (albeit unevenly) infilled and POCS was not applied. Moreover, a different back-projection algorithm was used, in which a constraint similar to one of the POCS constraints was in effect. As well as this difference in data collection geometry and computational techniques between the studies, one preparational difference cannot be excluded as having some bearing on the comparison. The yeast ribosomes were high salt washed, in order to ensure that non-ribosomal components were not present. This could possibly give rise to an enhanced openness of the upper intersubunit space. However, the ability to visually separate the subunits is probably due more to higher effective resolution (due to the far larger number of projections in the image data set used for reconstruction) than to stripping off of factors and ligands.
One final comparison with a 40S ribosome structure from rabbit reticulocytes (24) is surprising. The yeast small subunit structure in both frontal and lateral views more strongly resembles this mammalian 40S subunit structure than it does the wheatgerm 40S structure. (Intermediate views of the rabbit structure were compromised by loss of information in certain angular ranges due to the missing cone.) Particularly for the crest, beak, enrolled front lobe (23), cupped platform and pointed front foot the agreement is unexpectedly good. Why the details of the morphology of the yeast 40S subunit should resemble those seen for a mammalian subunit (which has a slightly larger 18S RNA) more closely than those seen for a higher plant (with an 18S RNA very similar in size to that of yeast) is unclear. The data collection and methods of analysis of the two earlier structures, rabbit and wheatgerm, were closely similar to one another, whereas the methods for yeast ribosome analysis were in some respects closer to the methods used for the E.coli reconstructions (where data from a larger angular range could be used; see for example 18,19).
Since completion of the reconstruction reported here recent continued studies at improving resolution on the E.coli and yeast ribosomes (21; Malhotra et al., submitted for publication; Verschoor et al., work in progress) have led to the consistent finding of a low density `tunnel' feature extending from the IC to the lower back of the large subunit. Although other prominent low density features can be discerned in individual reconstructions, the siting and orientation of what is apparently a conserved feature is of primary interest.
Although we are interested in eukaryotic specializations related to, for example, membrane attachment, we must first identify features related to the most universal functions in the translation process. If we accept the experimental evidence that a tunnel traverses the large ribosomal subunit for the purpose of conducting the nascent chain from the peptidyltransferase center (PTC) then we should be able to identify this tunnel feature in eubacteria, Archaea and eukaryotes. The existence of such a tunnel through the large subunit is strongly implied by studies involving iodide ion quenching of nascent chain photoreactive probes (see for example 27), in which the nascent chain is shown not to be exposed to the cytoplasm in membrane-bound ribosomes.
It is surprising, at the moderate resolution of this study, that a feature as small as the exit hole should be so distinct. The 60S subunit can be sighted through from the IC side (Fig. 2A) to the back (Fig. 2) at an only slightly elevated threshold. There is a clear qualitative difference between this unique feature and the larger, smooth holes or cavities that appear as the structure pinches out in thin regions once a locally critical threshold is reached. Internally the effects of resolution limitation can be seen in the fact that the goblet structure, although open to and in communication with the IC, retracts away from it to form the wavy rim feature.
At higher resolution in both E.coli (18, Figs 2 and 4 ; 19, Fig. 2) and yeast (21; Verschoor et al., work in progress) the tunnel is seen to be more tubular, i.e. less variable in width than the rapid taper seen in the present yeast structure. However, not only is the trend of the feature identical, but the siting of the exit hole on the lower back of the large subunit also corresponds precisely. This hole has subsequently been demonstrated to be involved in post-translational processing of the nascent chain. In a study just completed on a 3D reconstruction of the yeast sec61p-ribosome complex (21) the central pore of the sec61p complex, which is involved in signal sequence recognition and binds directly to the ribosome, aligns precisely with the hole marking the debouchement of this tunnel feature, suggesting that the nascent chain is conducted successively through tunnel and pore after its formation at the PTC.
Thus we are able to use the strong structural conservation of the ribosome to delineate internal features that are strikingly alike in ribosomes from different taxa. Ligand experiments such as the one just described can then confirm that these well-characterized morphological features interact with ligands of known function, which demonstrates that we are indeed justified in assigning functional roles to such features.
The structure calculated for the yeast ribosome leaves no question that the architecture of the translational machinery of the cell has been strongly conserved from one kingdom to another. The degree of resemblance to the eubacterial (E.coli) ribosome is considerable, while the similarity of this lower eukaryotic ribosome to the ribosome of a higher eukaryote (wheatgerm) is also unambiguous. Thus the task of tracing fundamental translational processes, such as the itineraries of the mRNA and tRNAs into and out of the ribosome, is to some degree simplified if we are able to recognize the same structural landmarks on evolutionarily divergent ribosomes and to understand how they delineate functional domains that are strongly conserved. Then, conversely, we can also identify non-conserved features and begin to explore their possible functions in terms of known inter-kingdom differences or variations in the translational process.
The data presented above will serve as a baseline for future studies. Saccharomyces cerevisiae is perhaps the one eukaryote for which we have both the genetic data and the genetic tools to approach a structure-function study of the ribosome. There are a number of yeast ribosomal proteins which are not essential for life yet which are important for optimal assembly, stability and function of the ribosome (28). In addition, there are mutant ribosomal proteins that have a substantial effect on the accuracy of the ribosome's translational capabilities (29). There exist more general mutations, such as those that block N-acetylation of numerous ribosomal proteins (30) without having much apparent effect on ribosome function. As cryo electron microscopy continues to advance it is likely to become the method of choice for understanding the subtle effects of mutations such as these and for reconciling the differences that inevitably appear between prokaryotic and eukaryotic ribosomes.
Financial support to J.F. from the NIH (1R01 GM 29169) and NSF (BIR 9219043) and to J.R.W. (NIH 1R01 GM 35532) is gratefully acknowledged. We thank Arun Malhotra, Pawel Penczek and Amy Heagle for assistance with visualization in the preparation of the final manuscript.
Nucleic Acids Research
Pages
Introduction
Materials And Methods
Results
Morphology of the yeast ribosome
Large subunit tunnel
Discussion
Comparison with the E.coli ribosome
Comparison with ribosomes from higher eukaryotes
Internal structure of the yeast ribosome: the exit tunnel
Conclusions
Acknowledgements
References
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 6 Jan 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Wang, C. Li, E. Wang, and X. Wang Uncovering the rules for protein-protein interactions from yeast genomic data PNAS, March 10, 2009; 106(10): 3752 - 3757. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Champion, B. H. Lane, M. E. Jackrel, L. Regan, and S. J. Baserga A Direct Interaction between the Utp6 Half-a-Tetratricopeptide Repeat Domain and a Specific Peptide in Utp21 Is Essential for Efficient Pre-rRNA Processing Mol. Cell. Biol., November 1, 2008; 28(21): 6547 - 6556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. I. Moy and P. A. Silver Requirements for the nuclear export of the small ribosomal subunit J. Cell Sci., July 15, 2002; 115(14): 2985 - 2995. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-W. Lee, S. J. Berger, S. Martinovic', L. Pasa-Tolic', G. A. Anderson, Y. Shen, R. Zhao, and R. D. Smith Direct mass spectrometric analysis of intact proteins of the yeast large ribosomal subunit using capillary LC/FTICR PNAS, April 30, 2002; 99(9): 5942 - 5947. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. BECKMANN, C.M.T. SPAHN, J. FRANK, and G. BLOBEL The Active 80S Ribosome-Sec61 Complex Cold Spring Harb Symp Quant Biol, January 1, 2001; 66(0): 543 - 554. [Abstract] [PDF]
|
|
|
|
|
|
A.W. JOHNSON, J.H.-N. HO, G. KALLSTROM, C. TROTTA, E. LUND, L. KAHAN, J. DAHLBERG, and J. HEDGES Nuclear Export of the Large Ribosomal Subunit Cold Spring Harb Symp Quant Biol, January 1, 2001; 66(0): 599 - 606. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Nadano and T.-A. Sato Caspase-3-dependent and -independent Degradation of 28 S Ribosomal RNA May Be Involved in the Inhibition of Protein Synthesis during Apoptosis Initiated by Death Receptor Engagement J. Biol. Chem., April 28, 2000; 275(18): 13967 - 13973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. I. Moy and P. A. Silver Nuclear export of the small ribosomal subunit requires the Ran-GTPase cycle and certain nucleoporins Genes & Dev., August 15, 1999; 13(16): 2118 - 2133. [Abstract] [Full Text]
|
|
|
|
|
|
M. C. Justice, T. Ku, M.-J. Hsu, K. Carniol, D. Schmatz, and J. Nielsen Mutations in Ribosomal Protein L10e Confer Resistance to the Fungal-specific Eukaryotic Elongation Factor 2 Inhibitor Sordarin J. Biol. Chem., February 19, 1999; 274(8): 4869 - 4875. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. V. Rodnina and W. Wintermeyer Form follows function: Structure of an elongation factor G-ribosome complex PNAS, June 23, 1998; 95(13): 7237 - 7239. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||