How We Got to Where We Are: the Ribosome in the 21st Century

PETER B. MOORE

[SECTION EDITOR: MICHAEL O'CONNOR]

Posted September 12, 2007

Departments of Chemistry and of Molecular Biophysics and Biochemistry, Yale University, New Haven, CT 06520

Mailing address: Dept. of Chemistry, Yale University, P.O. Box 208107, New Haven, CT 06520-8107. Phone: (203) 432-3995, Fax: (203) 432-5781, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

Scientific papers generally start with an introduction that summarizes what was known before the work to be described began and often a brief account of how that information was uncovered, i.e., with some history. The chapters that this essay introduces deal with the ribosome, and scientifically, the ribosome has had an unusually rich and interesting history. It has attracted the attention of many scientists of the first rank, and in its early years, ribosome research was an important component of the larger field we now call molecular biology. The history that follows is an informal one; those interested in a more scholarly scientific history of the ribosome should consult Rheinberger’s recent article on the subject (26).

HISTORICAL OVERVIEW

The first experiments relevant to the discovery of the ribosome are generally agreed to have been executed in the 1930s by scientists such as Brachet, Caspersson, and Claude, who were interested in characterizing biochemically the structures that can be seen in eukaryotic cells with the light microscope. By the early 1940s, several important generalizations had emerged. Like chromosomes, DNA is found exclusively in the nucleus, while RNA tends to localize in the cytoplasm, and the more RNA there is in a cell, the more active that cell is likely to be in protein synthesis (2, 6). Claude, who performed a series of important experiments that depended on centrifugal fractionation of cell lysates, identified a cellular fraction he dubbed microsomes (10), which included virus-like nucleoprotein particles that Brachet suggested might be involved in protein synthesis (3, 9).

In the years following World War II, experimental biology benefited enormously from technical advances in electronics and in many other areas that had been stimulated by the war. Two important types of instruments that had been developed in the 1920s and 1930s, ultracentrifuges and electron microscopes, were commercialized immediately after the war. In 1949, Spinco, which is now owned by Beckman-Coulter, produced the model L preparative ultracentrifuge, which was the first such instrument to reach the market, and by 1946, RCA was selling electron microscopes that were user friendly enough so that biologists could utilize them. The Manhattan Project, too, made a contribution. Prior to the war, biological experiments that used radioisotopes had been hard to execute because the isotopes required had to be produced by using cyclotrons, and instruments for detecting radiation were mostly handmade. After the war, radioisotopes were produced in abundance by using nuclear reactors, and an entire industry grew up to supply biochemists with radiolabeled compounds as well as the instruments to detect them. Thus, by 1950 or so, several powerful research tools that had been available previously only to those willing to make a career with them had become accessible to the larger scientific community, and the stage was set for what was to become a golden age of biology.

Starting in the 1950s, ribosome-related developments came thick and fast; only a few of the highlights will be noted here. We should start by remembering the work done by Palade, Porter, and Siekevitz, who used the electron microscope to revolutionize our understanding of the internal organization of the cell (23, 24, 25). Among their many discoveries was the endoplasmic reticulum, a structure that pervades the cytoplasm of most eukaryotic cells (Fig. 1). They quickly established that Claude’s microsomes were fragments of the endoplasmic reticulum and that Claude’s nucleoprotein particles were abundant in microsome preparations, some associated with reticulum membranes and others free in the cytoplasm. In 1954, Keller et al. published the results of an elegant series of experiments with radiolabeled amino acids that proved that Claude’s particles are indeed the sites where proteins are made in the cell (19).

Fig. 1Electron micrograph of the endoplasmic reticulum found in pancreatic cells. This micrograph shows what the endoplasmic reticulum looks like under an electron microscope when fixed, stained, and thin-sectioned specimens of pancreatic tissue are examined. The specks abundant in this view are ribosomes.

Source: Reprinted from NobelLecturesinMedicineorPhysiology,1971−1980 (22), with permission of the publisher.

The single most important fact that one needs to know about the structure of microsomal particles or ribosomes emerged shortly thereafter from studies of the particles’ centrifugal behavior. Ribosomes from all species are 1:1 complexes of two nonequivalent ribonucleoprotein particles, the larger being about twice the molecular weight of the smaller (7, 8, 29). Furthermore, the dissociation constant of this two-subunit complex is very sensitive to the Mg²⁺ ion concentration; high concentrations favor association.

At about the same time, a remarkable convergence between biological theory and biochemical discovery took place. Crick had realized on chemical grounds that protein synthesis was unlikely to involve direct interactions between amino acids and nucleic acids and, consequently, had proposed that amino acids must be attached to adaptor molecules, which he presumed would be nucleic acids, before they enter protein synthesis (11). Independently, Hoagland and coworkers demonstrated that amino acids are in fact delivered to the protein synthetic system esterified to low-molecular-weight RNAs of the class they called soluble RNAs (16). These RNAs are now called tRNAs (transfer RNAs), and they are, of course, Crick’s adaptor molecules.

mRNA (messenger RNA) was the last big piece of the protein synthesis system to be discovered. There was a period in the mid-1950s when many thought that the sequences of rRNAs encode information about protein sequences and that the ribosome was what resulted when the proteins responsible for synthesizing proteins from aminoacyl-tRNAs get to work on these bits of informational RNA. This view of the ribosome is incompatible with experimental observations that began accumulating almost as soon as it was articulated, and in 1959 and 1960, it was finally put to rest by the discovery of a hitherto unappreciated class of unstable RNA molecules, mRNAs, that carry sequence information from chromosomes to ribosomes (4, 14).

By 1964, Watson could discuss protein synthesis using cartoons like the one shown in Fig. 2, which the reader will instantly recognize to be remarkably similar to figures that have adorned molecular biology and biochemistry textbooks ever since (34). Figure 2 summarizes what was known about the ribosome by 1964 far more accurately than a casual observer may suspect. As the figure shows, the large subunit is rounder than the small subunit, but both have about the same maximum linear dimensions. mRNA does interact exclusively with the small subunit, which is where decoding occurs, and the large subunit is the site of peptide bond formation. In addition, there are two sites for tRNA binding on the ribosome, one occupied by peptidyl-tRNAs just before peptide bond formation takes place, the so-called P site, and the other, the A site, occupied at that same instant by aminoacyl-tRNAs. Finally, following peptide bond formation, the peptidyl-tRNA in the A site translocates to the P site before the next cycle of chain elongation occurs, as shown. From the perspective of 2007, the only things wrong with Fig. 2 are that there are three sites for tRNA on the ribosome, not two, the third one, the E site, being less interesting functionally than the A or P site, and the progression of tRNAs through these sites is more complicated than Fig. 2 implies.

Fig. 2Ribosome function as it was understood in 1964. This cartoon appears in the article Watson published in 1964, which did much to fix the idea in everyone’s mind that the ribosome has two principal sites for tRNA binding, one that interacts preferentially with peptidyl-tRNAs, the P site, and a second at which aminoacyl-tRNAs bind as they enter the peptide bond formation process, the A site (34). The left panel shows a piece of mRNA attached to the 30S subunit, forming base pairs with two tRNAs that both also interact with the large subunit, one carrying a nascent peptide and the other an aminoacyl-tRNA. In the middle panel, we see the state of affairs that exists after the nascent peptide has been transferred to the amino group of the amino acid of what used to be the aminoacyl-tRNA. The right panel shows the situation that prevails after the peptidyl-tRNA moves from the site where it forms initially to the site it must occupy if another amino acid is to be added to the growing chain. AA, amino acid; n, amino acid in nascent peptide.

Source: Reprinted from BulletindelaSociétédeChimieBiologique (34) with permission of the publisher.

These details notwithstanding, the most fundamental conclusion about the ribosome to emerge prior to 1964 was the realization that the ribosome is an enzyme. Aminoacyl-tRNAs are its substrates, mRNAs are its templates, proteins are its products, and ribosome-mRNA-tRNA interactions determine the sequences of the proteins that emerge. Thus, the ribosome is no less a polymerase than RNA or DNA polymerase, and it would probably be referred to today as polypeptide polymerase if its function had been properly understood when Roberts gave it its name in 1958 (27).

The rate at which new information about protein synthesis accumulated increased during the 1960s. The code got decoded, and the enzymology done on protein synthesis in that era did much to flesh out our understanding of the mechanism of the ribosomal phase of protein synthesis, even if it did not reveal every last detail. Thus, even though the story was still incomplete, by the late 1960s it was clear that the field of ribosome research was coming to a parting of the ways. The fundamental question that had motivated most of those who had worked with these particles, which was how the information contained in DNA sequences get translated into amino acid sequences, had been answered at a level satisfactory to most. The emphasis in the ribosome research field was rapidly shifting from figuring out what ribosomes do to figuring out how they do it, and it was clear that answers to these "how" question were unlikely to appear until the structure of the ribosome was understood. Given the size and complexity of the particle, that was a daunting prospect. Many left the field in search of greener pastures.

Logically, the first step in determining the structure of a macromolecular complex should be the identification of the components of which it is made. By the late 1960s, that work had begun, and the ribosome from Escherichia coli had been chosen as the paradigm particle for such studies. The proteins associated with that ribosome were being fractionated and characterized by Tissières’s group in Geneva and by Nomura, Kurland, and Craven in Madison, WI. Ribosomal protein turned out to be a complex mixture of molecules (31), and with one exception, each protein in the mixture is present in a single copy per ribosomal particle (15). This finding was entirely consistent with conclusions drawn by Waller several years earlier (33), but subsequent reports, which ultimately turned out to be spurious, had called Waller’s findings into question. Wittmann’s group in Berlin soon joined the fray, and its members ultimately sequenced all the ribosomal proteins in E. coli and those in many other species as well. The primary structure of the rRNAs was another kettle of fish entirely, and except for that of 5S rRNA, no sequence of any large rRNA was available for many years.

In 1968, one’s sense that the field of ribosome research would henceforth advance at a steady, if unspectacular, pace was upset by a paper published by Traub and Nomura in which it was reported that the small ribosomal subunit from E. coli can be reconstituted from its constituent proteins and RNA in vitro (30). This paper was not the first to deal with ribosome assembly in vitro, but it was the first to show complete self-assembly. The key to this success was the willingness of its perpetrators to incubate their (ugly) mixtures of partially precipitated, denatured ribosomal protein and naked 16S rRNA at elevated temperatures for hours. For the well-reared biochemist of the day, taught from birth that high temperatures destroy macromolecules, this method was insane, but it worked, and it inspired a host of interesting experiments which used reconstitution either as a tool to study ribosome structure and function or as a means of exploring the way ribosomes assemble. (It should be noted that it has long been understood that while bacterial ribosomes assemble in vitro in a way that is related to their assembly in vivo, the two processes are not identical.) The bacterial ribosome remains the most complex biological object for which self-assembly has been demonstrated. The reconstitution of any eukaryotic ribosome has yet to be shown.

About a decade later, the rRNA part of the picture began to come into focus. In 1978, Brosius et al. published the first complete sequence of a high-molecular-weight rRNA, that of the 16S rRNA from E. coli. Over a decade earlier, an ambitious effort to sequence large rRNAs by using technology Sanger had developed for sequencing small RNAs, like tRNAs, had been launched. These techniques were not powerful enough to sequence the large rRNAs, and Brosius and his colleagues solved the problem by doing what any graduate student would do today. They sequenced the gene, not its RNA product (5). (It was by no means as easy to do then as it is today!) The value of this sequence was multiplied manifold by the analysis Noller and Woese subsequently performed that led to a proposal for the secondary structure of 16S rRNA (21). We now know that the first version of secondary structure offered by Noller and Woese was incomplete, as they would certainly have told you at the time it was bound to be, but it had two powerful virtues. It correctly identified the three major domains of 16S rRNA, and every helix it proposed actually exists in the ribosome. 16S rRNA sequences became the "gold standard" for those who use sequences to study phylogenetic relationships, and this advance forced workers in the field of ribosome research to focus on the functional role of rRNAs in ribosomes with an intensity that they had not exhibited before. A sequence and a secondary-structure model of 23S rRNA followed shortly thereafter.

It took another 20 years for atomic resolution, three-dimensional structures of ribosomes to emerge. The two techniques that have contributed the most to our understanding of the three-dimensional ribosome are electron microscopy and X-ray crystallography, and electron microscopy came first. The first micrographs of ribosomes were obtained in the 1950s by using samples of tissue that had been fixed, stained, and thin sectioned. As Fig. 1 shows, in preparations of this sort, ribosomes appear to be dark specks devoid of internal structure. The first images of ribosomes with resolutions high enough to provide an indication of how the ribosomes are organized internally were published in 1960 by Huxley and Zubay, who used the then-novel technique of negative staining to prepare their samples (18) (Fig. 3). The images they published in 1960 were about as revealing as any obtained prior to 1990! The essence of the structural information in such images was extracted by Lake in the mid-1970s (20) (Fig. 4), and for years thereafter, Lake’s ribosome shapes, which were well known because he had distributed plaster of Paris replicas of his models to groups all over the world, were the shapes everyone used to think about ribosomes. They are remarkably accurate. By the early 1990s, however, advances in both the technology for producing electron micrographs of macromolecules and the algorithms for deriving three-dimensional models from (two-dimensional) electron micrographs led to the replacement of Lake’s ribosome models with models accurate to far higher resolutions (e.g., see reference 13). The resolution of electron density maps of ribosomes derived from electron micrographs had improved so much that some of the features could be interpreted in molecular terms by fitting higher-resolution structures of ribosomal components into them. However, the resolutions of even the best of these "second-generation" electron microscopy images still fall short of what would be required to interpret them in molecular terms independent of other, higher-resolution structural information. (This situation may change; the electron microscopy field continues its impressive advances.)

Fig. 3Electron micrograph of 70S ribosomes from E. coli embedded in a negative stain. This micrograph is a portion of Plate IIIb from an article on ribosome structure published by Huxley and Zubay in 1960 (18). The particles are embedded in phosphotungstic acid, which is a negative stain, and the division of the particles into two, unequal, subunits is plainly visible.

Source: Reprinted from the JournalofMolecularBiology (18) with permission of the publisher.

Fig. 4Lake’s models for the large (50S) and small (30S) ribosomal subunits from E. coli. The image shown is a photograph of the plaster models of the large subunit (left) and the small subunit (right) given by James Lake circa 1977. They correspond to the shapes he reported for these objects in 1976 (20). The maximum linear dimension of the ribosome is about 250 Å, and at this resolution, only a specialist can distinguish the ribosomes of one prokaryotic species from those of another. The ribosomal subunits of eukaryotes are similar in shape but somewhat larger, i.e., about 300 Å in maximum linear dimension.

X-ray crystallography provided the final denouement, but it took a while to happen. The first crystals of ribosomes large enough to work with crystallographically were prepared in the late 1970s in Wittmann’s laboratory in Berlin (35). These first ribosome crystals diffracted poorly, as did the crystals of ribosomes and ribosomal subunits prepared in Berlin and elsewhere for many years thereafter. But as experience deepened, resolution slowly improved (17, 32). By the early 1990s, the resolution of the diffraction patterns being obtained from ribosome crystals had improved so much that it was clear that atomic-resolution structures of ribosomes would result if reliable strategies could be found for phasing them. This last hurdle was overcome in 1998 (1), and in the summer of 2000, the first atomic-resolution structures appeared.

THE POST-STRUCTURAL AGE OF THE RIBOSOME

We are now happily living in the post-structural age of the ribosome. Rather than try to catalog all the ribosome crystal structures reported since 2000, or to summarize the rich harvest of information that has been reaped from them and from the far larger number of intermediate-resolution electron micrograph structures of ribosomes that have appeared, I close this essay by directing the reader’s attention to Fig. 5, which is an image of the 70S ribosome from Thermus thermophilus at the subunit interface from the perspective of the side where aminoacyl-tRNAs enter the ribosome during protein synthesis (28). For someone who began his career when little more was known about the structure of the ribosome than what is evident in one of Palade’s images, it is a continuing source of wonder that we know now where almost every atom is located in the ribosomes of several different species. It has been fun to have been along for the ride and a privilege to have helped make it happen.

Fig. 570S ribosome at atomic resolution. Selmer and coworkers recently published a 2.8-Å-resolution structure of the 70S ribosome from Thermus thermophilus (28), the highest-resolution structure of a complete ribosome that is presently available. The particle is oriented so that the viewer looks into the interface between the two subunits from the direction that aminoacyl-tRNAs are delivered to the particle by elongation factor EF-Tu. The RNA components of the large subunit are dark blue, and its protein components are light blue. The RNA components of the small subunit are dark green, while its proteins are light brown. A tRNA (dark red) can be seen to be bound at the P site, while beyond it, in magenta, a tRNA bound at the E site can just be seen. The pink fragment seen at the small-subunit end of the P site-bound tRNA is a fragment of mRNA. All components are shown in a stick representation in which only non-hydrogen atoms are taken into account. This image was prepared using PyMol (12).

Source: Reprinted from Science (28) with permission of the publisher.