Working from the particular to the general – the development of comparative genomics – A meeting report from the 4th Annual Conference on Microbial Genomes

Each genome sequenced contains a wealth of information on that particular strain. However, one of the central challenges of the post-genome era is to use the information from the sequenced example strains to investigate the bacterial population and the species as a whole. Developments in this field were prominent at the latest TIGR Microbial Genomes conference held in Chantilly, VA. Comparisons are now becoming possible within clonal populations over time, and between unrelated strains and species Pairs of genomes from a single species are now available for Helicobacter pylori, Mycobacterium tuberculosis, Neisseria meningitidis and soon Escherichia coli. The published comparison of the H. pylori genome sequences (1) showed many differences including reorganizations, polymorphisms in hypermutable loci and 6–7% differences in gene complement. R. Fleischman (TIGR) presented the results of a comparative analysis of the two sequenced M. tuberculosis genome strains, H37Rv and CDC1551, which showed a very different pattern of differences between two strains of the same species namely, little evidence of genome reorganization and very similar gene complements and sequences. CDC1551 has a moderately larger chromosomal size due to the presence of 15 or 16 IS6110 as opposed to the four copies present in H37Rv. The two sequences have been compared using suffix trees and this has identified a number of tandem repeats, insertions and deletions which distinguish the two strains. Some of the tandem repeats are associated with the surface located putative virulence determinants, PPE–PGRS proteins, the others are associated with proteins of unknown function. Insertion–deletion events include whole genes, one example being a region close to the phospholipase C gene, where interestingly the genes that are absent from H37Rv are present in both CDC1551 and the unpublished Mycobacterium leprae sequence from the Sanger Centre. One intriguing observation is that when the presence of insertions and deletions is investigated in multiple strains there appears to be a relationship between the number of insertion and deletions and the number of IS6110 copies that are present. Whilst there is no evidence of a physical relationship between the IS elements and the sites of the changes this suggests that there may be an association between these two processes of strain diversification. F. Blattner (Univerity of Wisconsin) presented ongoing work comparing the sequences of E. coli strains. His group have identified in the region of 0.5 Mb of DNA that is present in E. coli strain K-12 but is absent from the sequenced strain of O157, and 1.3 Mb of sequence that is present in O157 but absent from K-12. Rather than being exclusively associated with a small number of large regions they have identified 235 points scattered over the whole chromosomal map at which these insertions or substitutions can be identified. The smallest is an insertion of 6 amino acids in a leader sequence. More classical pathogenicity islands have also been detected and there is an example in which the two strains have different pathogenicity islands located at the same sequence location adjacent to the leuX tRNA gene. Blattner's group have made substantial progress in using these polymorphic sites in order to investigate the wider E. coli population using PCR primers to the flanking regions of each. The differences at these sites cluster into clearly definable groups which has the potential to form the basis of a subdivision of strains into biologically similar groups. This has revealed, amongst other things, that there are naturally occurring strains that are very similar to K-12, suggesting that this strain may indeed be representative of some naturally occurring isolates despite its long history of laboratory culture. B. Tummler (Klinisce Forschergruppe) presented comparative analyses of Pseudomonas aeruginosa strains which are characterized by rapid strain adaptation to new environmental niches and in cystic fibrosis are associated with a predictable series of changes affecting virulence that occur regardless of the initial colonizing clone. Comparison of strains considered to be part of a single clone were considered to display differences of up to 10% of their genomes. Detailed comparison of two members of one circulating clone revealed hot spots for insertions and an example of a hypervariable region with abnormal %G+C composition and codon usage in which 80% of the open reading frames (ORFs) were of unknown function but in which the ORF density was the same as the rest of the chromosome. The time frame for these changes is thought to be short and the origin of some is recognizable (for example one region was previously found in Pseudomonas putida). The sites of insertion were associated not only with tRNA genes but also sequences capable of forming Holliday junctions. In contrast to these large numbers of changes between isolates, when early and late colonizing isolates from cystic fibrosis cases were compared using signature tagged mutageneis to find the changes responsible for the familiar changes in phenotype, it was shown that gain of only a very small number of genes were responsible. Comparative analyses are also possible at the level of protein families. L. A. McCue (NY State Department of Health) presented just such a comparison in which the superfamiy of cAMP binding proteins were investigated in the completed genome sequences using the probe software. This revealed that M. tuberculosis differs from the other sequenced genomes in having a very large number of apparently cAMP regulated proteins which include cytoplasmic, internal membrane, and receptor-type proteins and was the only bacterial species to date to have translocases and esterases in this class of regulated proteins. This type of analysis shows how major differences in metabolism can be rapidly detected using genome sequences. The theme of pursuing metabolic function through comparative analysis was continued by S. G. E. Anderson (Uppsala University) who has been comparing the genomes of intracellular parasites, endosymbionts and mitochondria. These studies reveal that these species continue to loose genes and are in a continuous process of becoming progressively more dependent upon the metabolisms of the cells in which they reside. Whilst they are loosing genes through degeneration there is little evidence of gene acquisition from other sources, probably due to their relative isolation within the cells that they live. Their most striking observation comes from a comparison of the mitochondrial genes that are located in the chromosome and mitochondrial chromosomes in yeast. Investigation of the approximately 400 genes encoding mitochondrial functions and considering those that have homology to bacterial genes it seems likely that the original bacterial symbiont did not provide genes for ATP production, and perhaps could not produce ATP. This led to the suggestion that the acquisition of the endosymbiont ancestor of mitochondria was related to the control of toxic intracellular oxygen at a period in evolutionary history when oxygen concentrations were increasing and that only later did the mitochondrial ATP translocases evolve. The corresponding genes in the chromosome then developing in the chromosome de novo rather than by translocation from the ancestor symbiotic bacterium. These comparative approaches have revealed many differences between strains, but also reveal significant qualitative differences in the evolutionary biology and the processes of change in each of the different species groups that have been investigated. Horizontal transfer is clearly widespread, indeed W. F. Doolittle (Canadian Institute for Advanced Research) suggests that it is sufficient to undermine our current models of phylogenitics and evolutionary trees. However, the comparative analysis of microbial genomes reveals important differences in the extent, rate and characteristics of horizontal exchange, and other processes generating change within and between different bacterial species. The use of arrays, generated using genome sequence data, has added a further dimension to comparative analyses: comparisons of the same strain over time. In the first example, M. Laub (Stanford University) presented the use of arrays to study the life cycle of Caulobacter crescentus. This approach has great promise for use in other systems in which there are clearly separated phases of the cell cycle. They have been able to clearly demonstrate the expression of the genes associated with flagellar biosynthesis in the order in which they are assembled. They have also made many new observations including the identification of over 400 new cell cycle regulated genes and 13 operons that are under the control of CtrA. The second example, is a study of yeast evolution in chemostat cultures. F. Rosenzweig (University of Florida) discussed their array based analysis which demonstrated that there was a very rapid change in the pattern of mRNA production following strong selection (glucose limitation) and that there is a very rapid reorganization of the cellular metabolism and duplication of some genes that are adaptive to the changing environment. This emphasises the enormous flexibility that cellular systems have, presumably following only minor changes in the genome that can have a profound effect upon cell function. This comparative approach, although perhaps still in its infancy, is already providing new and important insights into cell biology, adaptation, evolution, and differences within and between strains. As the number of genomes that are available continues to grow, the power of comparative genomics is clearly becoming an increasingly important and pervasive field affecting many areas of microbial research and may ultimately become the prime reason for sequencing additional genomes. N.J.S. is supported by a Wellcome Trust Fellowship in Medical Microbiology.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.