Richard Cooke

While plant rRNA gene organization and expression have been studied for several decades, the repetitive nature and high sequence identity between the tandemly-repeated units has prevented precise studies to determine which units are active and elucidate mechanisms of spatial and temporal transcriptional control. We have detected four variants among rRNA genes of Arabidopsis thaliana by analysis of the 3' external spacer region. Surprisingly, the most abundant variant, representing ~50% of the genes, is not expressed in wild-type plants but is transcribed in lines mutated in one of the two nucleolin genes. Analysis of Arabidopsis ecotypes indicated considerable variability in numbers and presence/absence of variants, although more closely-related ecotypes show similar profiles. Sequence analysis showed that one of the variants is not only unexpectedly located within a 5S RNA gene cluster in the pericentromeric region of chromosome 3 but is also potentially highly expressed in wild-type plants. We present a model to explain how transcription of this this unusual variant, maintained in an active state by nucleolin binding, could be involved in control of expression of the major variant, while absence of nucleolin leads to silencing of the minor variant and consequent expression of the major variant.

Whole genome duplication (WGD) and subsequent evolution of gene pairs have been shown to have shaped the present day genomes of most, if not all, plants and to have played an essential role in the evolution of many eukaryotic genomes. Analysis of the rice (Oryza sativa ssp. japonica) genome sequence suggested an ancestral WGD ∼50–70 Ma common to all cereals and a segmental duplication between chromosomes 11 and 12 as recently as 5 Ma. More recent studies based on coding sequences have demonstrated that gene conversion is responsible for the high sequence conservation which suggested such a recent duplication. We previously showed that gene conversion has been a recurrent process throughout the Oryza genus and in closely related species and that orthologous duplicated regions are also highly conserved in other cereal genomes. We have extended these studies to compare megabase regions of genomic (coding and noncoding) sequences between two cultivated (O. sativa, Oryza glaberrima) and one wild (Oryza brachyantha) rice species using a novel approach of topological incongruency. The high levels of intraspecies conservation of both gene and nongene sequences, particularly in O. brachyantha, indicate long-range conversion events less than 4 Ma in all three species. These observations demonstrate megabase-scale conversion initiated within a highly rearranged region located at ∼2.1 Mb from the chromosome termini and emphasize the importance of gene conversion in cereal genome evolution.

Although marine picophytoplankton are at the base of the global food chain, accounting for half of the planetary primary production, they are outnumbered 10 to 1 and are largely controlled by hugely diverse populations of viruses. Eukaryotic microalgae form a ubiquitous and particularly dynamic fraction of such plankton, with environmental clone libraries from coastal regions sometimes being dominated by one or more of the three genera Bathycoccus , Micromonas , and Ostreococcus (class Prasinophyceae). The complete sequences of two double-stranded (dsDNA) Bathycoccus , one dsDNA Micromonas , and one new dsDNA Ostreococcus virus genomes are described. Genome comparison of these giant viruses revealed a high degree of conservation, both for orthologous genes and for synteny, except for one 36-kb inversion in the Ostreococcus lucimarinus virus and two very large predicted proteins in Bathycoccus prasinos viruses. These viruses encode a gene repertoire of certain amino acid biosynthesis pathways never previously observed in viruses that are likely to have been acquired from lateral gene transfer from their host or from bacteria. Pairwise comparisons of whole genomes using all coding sequences with homologous counterparts, either between viruses or between their corresponding hosts, revealed that the evolutionary divergences between viruses are lower than those between their hosts, suggesting either multiple recent host transfers or lower viral evolution rates.

In eukaryotes, 45S rRNA genes are arranged in tandem arrays in copy numbers ranging from several hundred to several thousand in plants. Although it is clear that not all copies are transcribed under normal growth conditions, the molecular basis controlling the expression of specific sets of rRNA genes remains unclear. Here, we report four major rRNA gene variants in Arabidopsis thaliana. Interestingly, while transcription of one of these rRNA variants is induced, the others are either repressed or remain unaltered in A. thaliana plants with a disrupted nucleolin-like protein gene (Atnuc-L1). Remarkably, the most highly represented rRNA gene variant, which is inactive in WT plants, is reactivated in Atnuc-L1 mutants. We show that accumulated pre–rRNAs originate from RNA Pol I transcription and are processed accurately. Moreover, we show that disruption of the AtNUC-L1 gene induces loss of symmetrical DNA methylation without affecting histone epigenetic marks at rRNA genes. Collectively, these data reveal a novel mechanism for rRNA gene transcriptional regulation in which the nucleolin protein plays a major role in controlling active and repressed rRNA gene variants in Arabidopsis. Chromatin remodeling plays a central role in controlling gene expression in all eukaryotic organisms. Chromatin can be found in a repressive or transcriptionally inactive state (heterochromatin) or in a more permissive or transcriptionally active state (euchromatin). The building block of chromatin is the nucleosome, which consists of four histones, H2A, H2B, H3, and H4, surrounded by 147 base pairs of DNA. In addition, a linker histone H1 directs the path of DNA between adjacent nucleosomes to form the chromatin fiber. Chromatin compaction also depends on DNA methylation and on a number of histone modifications, including methylation and acetylation of histone tails. However, other non-histone proteins are required to direct chromatin structure and remodeling. Nucleolin is a major nucleolar protein involved not only in rRNA transcription and processing of 45S pre–rRNA transcribed by RNA Pol I, but also in the control of RNA pol II transcription in the nucleoplasm. Through genetic, molecular, and immunocytological approaches, we studied the role of this protein in vivo in controlling rRNA chromatin structure and the expression of hundreds of clustered rRNA genes using the model plant Arabidopsis thaliana.

Domains in Arabidopsis proteins NRPE1 and SPT5-like, composed almost exclusively of repeated motifs in which only WG or GW sequences and an overall amino-acid preference are conserved, have been experimentally shown to bind multiple molecules of Argonaute (AGO) protein(s). Domain swapping between the WG/GW domains of NRPE1 and the human protein GW182 showed a conserved function. As classical sequence alignment methods are poorly-adapted to detect such weakly-conserved motifs, we have developed a tool to carry out a systematic analysis to identify genes potentially encoding AGO-binding GW/WG proteins. Here, we describe exhaustive analysis of the Arabidopsis genome for all regions potentially encoding proteins bearing WG/GW motifs and consider the possible role of some of them in AGO-dependent mechanisms. We identified 20 different candidate WG/GW genes, encoding proteins in which the predicted domains range from 92aa to 654aa. These mostly correspond to a limited number of families: RNA-binding proteins, transcription factors, glycine-rich proteins, translation initiation factors and known silencing-associated proteins such as SDE3. Recent studies have argued that the interaction between WG/GW-rich domains and AGO proteins is evolutionarily conserved. Here, we demonstrate by an in silico domain-swapping simulation between plant and mammalian WG/GW proteins that the biased amino-acid composition of the AGO-binding sites is conserved.

Summary: We report on the structural characterization of a functional U3 snoRNA ribonucleoprotein complex isolated from Brassica oleracea. The BoU3 snoRNP complex (formerly NF D) binds ribosomal DNA (rDNA), specifically cleaves pre‐rRNA at the primary cleavage site in vitro and probably links transcription to early pre‐rRNA processing in vivo. Using a proteomic approach we have identified 62 proteins in the purified BoU3 snoRNP fraction, including small RNA associated proteins (Fibrillarin, NOP5/Nop58p, Diskerin/Cbf5p, SUS2/PRP8 and CLO/GFA1/sn114p) and 40S ribosomal associated proteins (22 RPS and four ARCA‐like proteins). Another major protein group is composed of chaperones/chaperonins (HSP81/TCP‐1) and at least one proteasome subunit (RPN1a). Remarkably, RNA‐dependent RNA polymerase (RdRP) and Tudor staphylococcal nuclease (TSN) proteins, which have RNA‐ and/or DNA‐associated activities, were also revealed in the complex. Furthermore, three U3 snoRNA variants were identified in the BoU3 snoRNP fraction, notably an evolutionarily conserved and variable stem loop structure located just downstream from the C‐box domain of the U3 sequence structures. We conclude that the BoU3 snoRNP complex is mainly required for 40S pre‐ribosome synthesis. It is also expected that U3 snoRNA variants and interacting proteins might play a major role in BoU3 snoRNP complex assembly and/or function. This study provides a basis for further investigation of these novel ribonucleoprotein factors and their role in plant ribosome biogenesis.

The role of gene duplication in the structural and functional evolution of genomes has been well documented. Analysis of complete rice (Oryza sativa) genome sequences suggested an ancient whole genome duplication, common to all the grasses, some 50-70 million years ago and a more conserved segmental duplication between the distal regions of the short arms of chromosomes 11 and 12, whose evolutionary history is controversial.