MEETING BRIEF

Making Ribosomes
Fourth International Symposium on Ribosome Synthesis and Nucleolar Function

by Craig S. Pikaard

(Posted September 19, 1997 ? Issue 16; archived October 3, 1997)

Abstract

Ribosomes, the subcellular machines that carry out protein synthesis, are composed of scores of proteins and multiple RNAs, all of which must be synthesized and fitted together in proper stoichiometry. Remarkable progress is being made in understanding how these components are made, modified, transported, assembled, and coordinately regulated with one another and other cellular processes. By coming to understand the ribosome, we come closer to understanding the bustling world within living cells.

This past August, scientists from around the world converged on the coastal town of Noordwijkerhout, The Netherlands, not to join the throngs of vacationers on the beach but to do a little mental frolicking at a conference entitled "The Ribosome: Its Nucleolar Synthesis and Structure." This meeting, organized by H.A. (Dick) Raue and Jan van't Riet (both at Vrije Universiteit, Amsterdam) and C.W.A. Pleij (Leiden University) was the fourth incarnation of a conference normally held at Cold Spring Harbor every three years. The Netherlands gathering was to honor Professor Rudi Planta, a longtime leader in the field, in the year of his retirement as chairman of the Vrije Universiteit Department of Biochemistry.

The amazing ribosome catalyzes protein synthesis and is indispensable to life as we know it. The synthesis of ribosomes is no less amazing. In eukaryotes, ribosome assembly occurs in the nucleolus, where hundreds of genes encoding large ribosomal RNA (rRNA) precursors are transcribed by RNA polymerase I. These transcripts are cleaved in well-defined steps to produce 18S, 5.8S and 25-28S rRNAs. Chemical modification of specific nucleotides in rRNA by methylation or pseudo-uridylation occurs post-transcriptionally. A fourth RNA (5S) is transcribed elsewhere by RNA polymerase III and finds its way to the nucleolus. Meanwhile, over 80 ribosomal protein (r-protein) genes are transcribed by RNA polymerase II. Their mRNAs are exported to the cytoplasm, translated, and the resulting proteins shipped back the nucleus (and nucleolus) to bind the rRNAs and each other to produce the two ribosomal subunits. A multitude of additional proteins and small RNAs participate in these steps. In his talk, Jonathan R. Warner (Albert Einstein College of Medicine) pointed out the magnitude of the task in yeast, which have approximately 200,000 ribosomes in an actively growing cell. Warner calculated that 16 million ribosomal proteins must be synthesized in every cell cycle and that one assembled ribosomal subunit must exit each nuclear pore every three seconds. Overall, approximately 60% of all nuclear transcription is devoted to ribosome synthesis. As growth status changes, synthesis of ribosomal components changes accordingly, essentially ceasing in nongrowing or mitotic cells. Imagine building automobiles at a comparable pace, always in synch with consumer demand, and one gets an appreciation for what clever managers the cell employs.

How is all this activity coordinated? Many approaches have been taken, both biochemical and genetic, and a few answers are beginning to emerge. Larry Rothblum (Geisinger Clinic, Danville, Pennsylvania) and Ingrid Grummt (German Cancer Research Center, Heidelberg) presented data suggesting two ways pol I transcription can be downregulated. One involves the retinoblastoma (Rb) tumor-suppressor gene product, which binds to and blocks the activity of a key vertebrate transcription factor, UBF. A second mechanism involves phosphorylation of another transcription factor complex, SL1/TIF-IB, by the cell-cycle regulatory complex cyclin B/cdc2. Interestingly, the activity of the Rb protein is regulated by cell-cycle dependent phosphorylation, suggesting a way in which common signals ultimately impact upon multiple targets in the transcription machinery.

The mechanisms responsible for growth regulation in non-dividing cells remain a puzzle. Warner described a genetic approach to seek genes acting as master regulators of r-protein gene expression. His group identified multiple temperature-sensitive yeast mutants defective for transcription of r-protein genes, but unaffected for actin or enolase expression. The rRNA precursor is made at the non-permissive temperature in the mutants, but processing is blocked. Unexpectedly, the mutants are defective in genes encoding proteins of the secretory system. Chemical inhibitors of the secretory system, such as tunicamycin, mimic the mutant phenotype. Why secretory pathway defects should shut down r-protein transcription and pre-rRNA processing is not obvious. Nonetheless, an understanding of signaling pathways between the cytoplasm and nucleus may yet come from these perplexing initial results.

Another tale of the unexpected was delivered by Masayasu Nomura (University of California at Irvine), whose group is rapidly identifying novel pol I transcription factors in yeast. At least seven have been defined genetically, comprising two protein complexes, one of which is UAF (upstream activation factor). Purified UAF also includes several additional proteins not identified genetically. Two turn out to be histones H3 and H4! Interestingly, structural homologues of H3 and H4 were recently found to be part of the pol II transcription factor TFIID, but the presence of histones themselves in transcription factor complexes is unprecedented. Masayasu speculated that H3 and H4 may bend the yeast rRNA gene promoter the way UBF bends and wraps vertebrate promoters (UBF is absent in the yeast genome). Acetylation of these histones could also have regulatory significance. In this regard, I presented work of a postdoc, Z. Jeffrey Chen (Washington University, St. Louis, Missouri) studying nucleolar dominance in plants, an epigenetic phenomenon in which only one parental set of rRNA genes is expressed in interspecific hybrids. Silent rRNA genes appear to be repressed, in part, by histone deacetylation. Preventing deacetylation with butyrate or trichostatin derepresses these normally silent genes. We've assumed that changes in acetylation occur only on nucleosomal histones, but the yeast results suggest that histones in a UAF-like complex might also be targets.

The more than 80 ribosomal proteins are produced in roughly equal stoichiometry by a variety of mechanisms. In prokaryotes, most r-protein operons are subjected to feedback control if one or more protein products begin to accumulate in the cell. Such regulatory logic has been apparent for some time now, yet the precise mechanisms are still to be worked out. Janice Zengel (University of Maryland Baltimore County) presented the case of the S10 operon in E. coli, which is unusual in that it is subject to feedback control by ribosomal protein L4 at both transcriptional and translational levels. She showed that transcriptional control requires two stem-loop structures in the nascent RNA where L4, RNA polymerase, and the regulatory factor (NusA) apparently interact. Translational control requires an overlapping region of the RNA transcript. However, the combination of structures and/or sequences required for translational control have eluded identification, highlighting the difficulty of understanding the rules of regulation. Other interesting examples of regulation involving multiple stem-loops (Sheara Fewell; Carnegie Mellon University) or pseudoknots formed by sequences separated by hundreds of nucleotides (M.H. de Smit; Institut de Biologie Physico-Chimique, Paris) further illustrated the complexity of regulatory signals encoded in RNA.

Two sessions inspired a growing conviction that pre-rRNA processing is even more complicated than pre-mRNA splicing to remove introns. Scores of small RNAs and proteins are involved, but the small nucleolar RNAs (snoRNAs) stole the show. Maurille "Skip" Fournier (University of Massachusetts at Amherst) and Susan Gerbi (Brown University) reminded us that cells contain large populations of snoRNAs, perhaps more than 100, thus rivaling tRNAs as the largest class of

Hypothetical pairing of a box H/ACA guide snoRNA and rRNA site of pseudouridine formation. Figure provided by Jingwei Ni and Maurille J Fournier (Univ. Mass.)
small stable RNAs. Last year's big news was the discovery that members of one subset of the snoRNAs, the box C/D family (characterized by two conserved sequence motifs C and D) specify sites for ribose methylation in rRNA through base pairing interactions at the site of modification. There are ~115 such sites in vertebrates and ~65 in yeast. Other C/D box snoRNAs participate in pre-rRNA cleavage reactions. This year's hot news was that the conversion of uridines to pseudouridines in rRNA (~95 such sites in vertebrates, ~45 in yeast) is also specified by base pairing interactions with snoRNAs representing the other major family, the box H/ACA snoRNAs that associate with the protein Gar1p. C. Bousquet-Antonelli (Laboratoire de Biologie Moleculaire Eucaryote du CNRS, Toulouse, France) showed that depletion of Gar1p in yeast resulted in loss of pseudouridylation on 25S rRNA without affecting ribose methylation. Direct evidence for snoRNA involvement was presented by Jingwei Ni, in Fournier's lab, who knocked out numerous box H/ACA snoRNA genes in yeast, resulting in the loss of specific rRNA pseudouridylations. Likewise, insertions/deletions in snoRNAs shifted the sites of pseudouridylation accordingly. The biological significance of pseudouridines is not yet clear. Many are clustered near the peptidyl transferase center where protein synthesis is catalyzed. The Fournier lab prevented five of these modifications with their knockouts, yet protein synthesis was unaffected. However, it is possible that at least some modifications are needed for processing. D. Lafontaine in David Tollervey's group (Institute of Cell and Molecular Biology, Edinburgh) presented work on a yeast protein, CBF5p, that is an excellent candidate for the yeast pseudouridine synthase. It shares significant similarity to an E. coli tRNA pseudouridine synthetase, it associates with numerous box H/ACA snoRNAs, and knocking out CBF5p function results in the loss of all detectable rRNA pseudouridylation. Interestingly, the latter mutants fail to process 18S rRNA from pre-rRNA. Whether this defect is directly due to the missing pseudouridines remains to be shown.

E. Stuart Maxwell (North Carolina State University) and John Brown (Scottish Crop Research Institute, Invergowrie, Dundee) reminded us that many snoRNAs are encoded within introns of protein-coding genes, while others are encoded by genes devoted to their synthesis. No mechanistic connections between splicing of pre-mRNA and processing of the intronic snoRNAs have yet been described, but it is tempting to speculate that intronic snoRNAs and their host genes (often r-protein genes or translation factors) might comprise novel regulatory circuits in the cell.

A number of excellent talks focused on the roles of snoRNAs and other proteins in specific cleavage reactions in rRNA maturation. One remarkable protein complex needed for 5.8S rRNA processing was discussed by Tollervey. This complex consists of five proteins in yeast, all essential for cell viability. Biochemical studies show that most or all are potent nucleases that may play a role in mRNA turnover as well as 5.8S rRNA processing. How this awesome nuclease holoenzyme (likened to Star Wars' "Death Star") is kept from running amok, indiscriminately degrading all RNAs, remains an open, but important question.

Other progress reported at the meeting included the improved localization within the nucleolus of rRNA genes and associated transcription and processing factors, using microscopy and fluorescent antibodies. The impressive micrographs of W. Mosgoeller (Institute of Histology and Embryology, University of Vienna), Mark Olson (University of Mississippi Medical Center) and Peter Shaw (John Innes Centre, Norwich, England) were particularly striking. U. Thomas Meier (Albert Einstein College of Medicine) discussed the localization in mammalian cells of two interacting proteins, Nopp140p and Nap57p, found in structures known as coiled bodies and important for their formation. Interestingly, Nap57p is the homologue of the yeast CBF5p protein, the putative pseudouridine synthase, leading Meier to propose that coiled bodies might be the sites where pseudouridylation occurs. One is hopeful that a comprehensive view of the nucleolus may soon be available relating its structure to its various functions.

One cannot leave the subject of ribosome synthesis without admiring the finished product. Recent advances in cryoelectron microscopy and computer imaging have provided remarkable three-dimensional images of the E. coli ribosome, allowing the surface topology and locations of active sites associated with tRNAs to be seen or deduced. Richard Brimacombe (Max-Planck-Institute for Molecular Genetics, Berlin) described the use of this structural information to refine models of rRNA folding to better fit the visible structure. Also reporting the results of cryoelectron studies, Rajendra Agrawal in Joachim Frank's lab (Wadsworth Center, New York State Department of Health, Albany) showed tRNA-induced conformational changes in the ribosome deduced by changes in the position of the L1 ribosomal protein. The assignment of other ribosomal proteins to the 3-D structure and a better appreciation of the ribosome as a machine with moving parts is sure to follow. Many new surprises are sure to await us when we meet again in the year 2000.

Craig Pikaard is an associate professor in the Department of Biology at Washington University in St. Louis.

The ribosome image used above is from the Wadsworth Center Microscope 3D Database, � Dr. Joachim Frank.

Send us your comments and ideas for future articles.

Endlinks

Electron micrograph of a hepatocyte - shows, in a liver cell, how prominent and complex an organelle the nucleolus is. The image is a slow-to-download 373K file. This figure is from the National University of Singapore Histonet, a World Wide Web database for histology slides.

The RNA World - the principal site for RNA resource links. Links to software sites, meeting announcements, and databases and tools for structures. Maintained by the Institute of Molecular Biotechnology, Jena, Germany.

Structural Classification of Proteins (SCOP) - database (with several mirror sites worldwide) is a "must see" site for researchers interested in protein or RNA structure. Provides detailed and comprehensive descriptions of the structural and evolutionary relationships between all proteins whose structure is known.

RNA Modification Database - a comprehensive listing of post-transcriptionally modified nucleosides from RNA.

Favorite Software - HMS Beagle's capsule reviews of, and links to, sites for software such as Mfold, for RNA secondary structure prediction, and LoopDloop, freeware for drawing RNA secondary structures. See also the RNA Folding Software page on Harry Mangalam's home page at the University of California at Irving.

Robertus Group Web Page - the J. D. Robertus laboratory at the University of Texas researches ribosome inactivating proteins.