Genome-wide expression dynamics of a marine virus and host reveal features of co-evolution.

Interactions between bacterial hosts and their viruses (phages) lead to reciprocal genome evolution through a dynamic co-evolutionary process. Phage-mediated transfer of host genes--often located in genome islands--has had a major impact on microbial evolution. Furthermore, phage genomes have clearly been shaped by the acquisition of genes from their hosts. Here we investigate whole-genome expression of a host and phage, the marine cyanobacterium Prochlorococcus MED4 and the T7-like cyanophage P-SSP7, during lytic infection, to gain insight into these co-evolutionary processes. Although most of the phage genome was linearly transcribed over the course of infection, four phage-encoded bacterial metabolism genes formed part of the same expression cluster, even though they are physically separated on the genome. These genes--encoding photosystem II D1 (psbA), high-light inducible protein (hli), transaldolase (talC) and ribonucleotide reductase (nrd)--are transcribed together with phage DNA replication genes and seem to make up a functional unit involved in energy and deoxynucleotide production for phage replication in resource-poor oceans. Also unique to this system was the upregulation of numerous genes in the host during infection. These may be host stress response genes and/or genes induced by the phage. Many of these host genes are located in genome islands and have homologues in cyanophage genomes. We hypothesize that phage have evolved to use upregulated host genes, leading to their stable incorporation into phage genomes and their subsequent transfer back to hosts in genome islands. Thus activation of host genes during infection may be directing the co-evolution of gene content in both host and phage genomes.

Patterns and implications of gene gain and loss in the evolution of Prochlorococcus.

Prochlorococcus is a marine cyanobacterium that numerically dominates the mid-latitude oceans and is the smallest known oxygenic phototroph. Numerous isolates from diverse areas of the world's oceans have been studied and shown to be physiologically and genetically distinct. All isolates described thus far can be assigned to either a tightly clustered high-light (HL)-adapted clade, or a more divergent low-light (LL)-adapted group. The 16S rRNA sequences of the entire Prochlorococcus group differ by at most 3%, and the four initially published genomes revealed patterns of genetic differentiation that help explain physiological differences among the isolates. Here we describe the genomes of eight newly sequenced isolates and combine them with the first four genomes for a comprehensive analysis of the core (shared by all isolates) and flexible genes of the Prochlorococcus group, and the patterns of loss and gain of the flexible genes over the course of evolution. There are 1,273 genes that represent the core shared by all 12 genomes. They are apparently sufficient, according to metabolic reconstruction, to encode a functional cell. We describe a phylogeny for all 12 isolates by subjecting their complete proteomes to three different phylogenetic analyses. For each non-core gene, we used a maximum parsimony method to estimate which ancestor likely first acquired or lost each gene. Many of the genetic differences among isolates, especially for genes involved in outer membrane synthesis and nutrient transport, are found within the same clade. Nevertheless, we identified some genes defining HL and LL ecotypes, and clades within these broad ecotypes, helping to demonstrate the basis of HL and LL adaptations in Prochlorococcus. Furthermore, our estimates of gene gain events allow us to identify highly variable genomic islands that are not apparent through simple pairwise comparisons. These results emphasize the functional roles, especially those connected to outer membrane synthesis and transport that dominate the flexible genome and set it apart from the core. Besides identifying islands and demonstrating their role throughout the history of Prochlorococcus, reconstruction of past gene gains and losses shows that much of the variability exists at the "leaves of the tree," between the most closely related strains. Finally, the identification of core and flexible genes from this 12-genome comparison is largely consistent with the relative frequency of Prochlorococcus genes found in global ocean metagenomic databases, further closing the gap between our understanding of these organisms in the lab and the wild.

A structured interface to the object-oriented genomics unified schema for XML-formatted data.

Data management systems are fast becoming required components in many biology laboratories as the role of computer-based information grows. Although the need for data management systems is on the rise, their inherent complexities can deter the full and routine use of their computational capabilities. The significant undertaking to implement a capable production system can be reduced in part by adapting an established data management system. In such a way, we are leveraging the Genomics Unified Schema (GUS) developed at the Computational Biology and Informatics Laboratory at the University of Pennsylvania as a foundation for managing and analysing DNA sequence data in centromere research projects around Arabidopsis thaliana and related species. Because GUS provides a core schema that includes support for genome sequences, mRNA and its expression, and annotated chromosomes, it is ideal for synthesising a variety of parameters to analyse these repetitive and highly dynamic portions of the genome. Despite this, production-strength data management frameworks are complex, requiring dedicated efforts to adapt and maintain. The work reported in this article addresses one component of such an effort, namely the pivotal task of marshalling data from various sources into GUS. In order to harness GUS for our project, and motivated by efficiency needs, we developed a structured framework for transferring data into GUS from outside sources. This technology is embodied in a GUS object-layer processor, XMLGUS. XMLGUS facilitates incorporating data into GUS by (i) formulating an XML interface that includes relational database key constraint definitions, (ii) regularising traversal through that XML, (iii) realising automatic processing of the XML with database key constraints and (iv) allowing for special processing of input data within the framework for automated processing. The application of XMLGUS to production pipeline processing for a sequencing project and inputting the Arabidopsis genome into GUS is discussed. XMLGUS is available from the Flora website (http://flora.ittc.ku.edu/).

Temporal dynamics of Prochlorococcus ecotypes in the Atlantic and Pacific oceans.

To better understand the temporal and spatial dynamics of Prochlorococcus populations, and how these populations co-vary with the physical environment, we followed monthly changes in the abundance of five ecotypes-two high-light adapted and three low-light adapted-over a 5-year period in coordination with the Bermuda Atlantic Time Series (BATS) and Hawaii Ocean Time-series (HOT) programs. Ecotype abundance displayed weak seasonal fluctuations at HOT and strong seasonal fluctuations at BATS. Furthermore, stable 'layered' depth distributions, where different Prochlorococcus ecotypes reached maximum abundance at different depths, were maintained consistently for 5 years at HOT. Layered distributions were also observed at BATS, although winter deep mixing events disrupted these patterns each year and produced large variations in ecotype abundance. Interestingly, the layered ecotype distributions were regularly reestablished each year after deep mixing subsided at BATS. In addition, Prochlorococcus ecotypes each responded differently to the strong seasonal changes in light, temperature and mixing at BATS, resulting in a reproducible annual succession of ecotype blooms. Patterns of ecotype abundance, in combination with physiological assays of cultured isolates, confirmed that the low-light adapted eNATL could be distinguished from other low-light adapted ecotypes based on its ability to withstand temporary exposure to high-intensity light, a characteristic stress of the surface mixed layer. Finally, total Prochlorococcus and Synechococcus dynamics were compared with similar time series data collected a decade earlier at each location. The two data sets were remarkably similar-testimony to the resilience of these complex dynamic systems on decadal time scales.

Dynamic evolution at pericentromeres.

Pericentromeres are exceptional genomic regions: in animals they contain extensive segmental duplications implicated in gene creation, and in plants they sustain rearrangements and insertions uncommon in euchromatin. To examine the mechanisms and patterns of plant pericentromere evolution, we compared pericentromere sequence from four Brassicaceae species separated by <15 million years (Myr). This flowering plant family is ideal for studying relationships between genome reorganization and pericentromere evolution-its members have undergone recent polyploidization and hybridization, with close relatives changing in genome size and chromosome number. Through sequence and hybridization analyses, we examined regions from Arabidopsis arenosa, Capsella rubella, and Olimarabidopsis pumila that are homologous to Arabidopsis thaliana pericentromeres (peri-CENs) III and V, and used FISH to demonstrate they have been maintained near centromere satellite arrays in each species. Sequence analysis revealed a set of highly conserved genes, yet we discovered substantial differences in intergenic length and species-specific changes in sequence content and gene density. We discovered that A. thaliana has undergone recent, significant expansions within its pericentromeres, in some cases measuring hundreds of kilobases; these findings are in marked contrast to euchromatic segments in these species that exhibit only minor length changes. While plant pericentromeres do contain some duplications, we did not find evidence of extensive segmental duplications, as has been documented in primates. Our data support a model in which plant pericentromeres may experience selective pressures distinct from euchromatin, tolerating rapid, dynamic changes in structure and sequence content, including large insertions of mobile elements, 5S rDNA arrays and pseudogenes.

Centromere satellites from Arabidopsis populations: maintenance of conserved and variable domains.

The rapid evolution of centromere sequences between species has led to a debate over whether centromere activity is sequence-dependent. The Arabidopsis thaliana centromere regions contain approximately 20,000 copies of a 178-bp satellite repeat. Here, we analyzed satellites from 41 Arabidopsis ecotypes, providing the first broad population survey of satellite variation within a species. We found highly conserved segments and consistent sequence lengths in the Arabidopsis satellites and in the published collection of human alpha-satellites, supporting models for a functional role. Despite this conservation, polymorphisms are significantly enriched at some sites, yielding variation that could restrict binding proteins to a subset of repeat monomers. Some satellite regions vary considerably; at certain bases, consensus sequences derived from each ecotype diverge significantly from the Arabidopsis consensus, indicating substitutions sweep through a genome in less than 5 million years. Such rapid changes generate more variation within the set of Arabidopsis satellites than in genes from the chromosome arms or from the recombinationally suppressed centromere regions. These studies highlight a balance between the mechanisms that maintain particular satellite domains and the forces that disperse sequence changes throughout the satellite repeats in the genome.

A functional screen for the type III (Hrp) secretome of the plant pathogen Pseudomonas syringae.

Type III secreted "effector" proteins of bacterial pathogens play central roles in virulence, yet are notoriously difficult to identify. We used an in vivo genetic screen to identify 13 effectors secreted by the type III apparatus (called Hrp, for "hypersensitive response and pathogenicity") of the plant pathogen Pseudomonas syringae. Although sharing little overall homology, the amino-terminal regions of these effectors had strikingly similar amino acid compositions. This feature facilitated the bioinformatic prediction of 38 P. syringae effectors, including 15 previously unknown proteins. The secretion of two of these putative effectors was shown to be type III--dependent. Effectors showed high interstrain variation, supporting a role for some effectors in adaptation to different hosts.