Thursday, March 25, 2010

Compensatory gene amplification restores fitness after inter-species gene replacement.

Lind P.A., C. Tobin, O.G. Berg, C.G. Kurland, and D.I. Andersson (2010) Mol. Microbiol. 75: 1078-1089.

Transfer of genes to organisms can occur in various ways especially in microorganisms. Introduction of foreign genes follows one of the three fates: insertion to replacement of an existing homologous gene locus, uncertain locations on chromosome, or inactivation. If the transferred genes are neutral or deleterious to bacteria, they are likely to be lost over time (Berg and Kurland, 2002). Because of this reason, there have been few examples of experimental evidence for evolution of horizontally transferred genes.

To examine whether and how horizontally transferred genes evolve, the authors replaced the ribosomal protein genes of Salmonella typhimuriums with homologous genes of foreign origin and evolved the strain by repeating serial batch culture transfer. Since the ribosomal protein gene is essential in bacteria, the transferred gene will not be lost and the gene needs to adapt to the new host to allow the cell to grow more efficiently (to increase fitness). Low fitness of the six constructed strains was observed as expected, because the transferred gene and its product do not initially fit the new host for many reasons; for example, difference in codon usage causes translational problem. However, within 25-200 generation of growth, adaptive mutations did overcome the fitness defects of the strains.

An interesting finding was that all genetic changes observed in the six evolved strains were not directly related to the changes in the replaced protein coding sequence, but resulted in increased expression of the introduced gene product. It is known that protein concentration imbalance can cause fitness problems for several reasons (Papp and Pai et al. 2003). The increase in protein expression was required because the introduced alien protein was inefficiently expressed due to the differences in codon usage, or because the alien protein did not have sufficient affinities to the partner molecules (RNA and proteins) to reconstitute an effective ribosome complex. All observed mutations were duplication of DNA segments containing the introduced ribosomal protein gene. This indicates that the rate of beneficial mutations in the protein coding sequence, which can change codon usage or the function of the alien protein, is much lower than the rate of recombination event that results in increase in protein expression level. The former mutation can happen, and could eventually be fixed in the population if you kept evolving the strains for long time, but such beneficial mutations were not fixed in the population within 250 generations of growth of this bacterium.

This study support the hypothesis that gene paralogs and orthologs arise upon horizontal gene transfer in the presence of selection for the gene. Although gene duplication under selection is not a rare event (Reams and Neidle, 2004; Kugelberg and Kofoid et al, 2006), the hypothesis is attractive because bioinformatics analysis revealed that duplication is more common in laterally transferred genes than in indigenous genes (Hooper and Berg, 2003).
By the way, how long does it take for the changes in the coding sequence to be fixed? It depends on selection and population size. Ask mathematicians!

Lind P.A., C. Tobin, O.G. Berg, C.G. Kurland, and D.I. Andersson (2010) Compensatory gene amplification restores fitness after inter-species gene replacement. Mol. Microbiol. 75: 1078-1089

Papp B., C. Pai, and L.D. Hurst (2003)
Dosage sensitivity and the evolution of gene families in yeast. Nature 424: 194-197.

Berg O.D. and C.G. Kurland. (2002)
Evolution of microbial genomes: Sequence acquisition and loss. Mol. Biol. Evol. 19:2265–2276

Hooper S.D. and O.D. Berg. (2003)
Duplication is more common among laterally transferred genes than among indigenous genes. Genome Biol. 4: R48

Reams A.B. and E.L. Neidle (2004)
Gene amplification involves site-specific short homology-independent illegitimate recombination in Acinetobacter sp. strain ADP1. J. Mol. Biol. 338:643-656

Kugelberg E, E. Kofoid, A.B. Reams, D.I. Andersson, J.R. Roth (2006)
Multiple pathways of selected gene amplification during adaptive mutation. Proc. Natl. Acad. Sci. USA 103:17319-17324

H. Yano, University of Idaho

Monday, March 15, 2010

Exploring the evolutionary dynamics of plasmids: the Acinetobacter pan-plasmidome

Marco Fondi, Giovanni Bacci, Matteo Brilli, Cristiana M Papaleo, Alessio Mengoni, Mario Vaneechoutte, Lenie Dijkshoorn, Renato Fani

Bacteria belonging to the genus Acinetobacter are found in diverse ecosystems such as, soil, water and even animals. Some like A. baumannii are well-known opportunistic human pathogens while others can be useful in bioremediation because of their ability to degrade toxic hydrocarbons. Many of these bacteria have been found to have plasmids of different sizes that probably encode genes that help Acinetobacter sp. survive in the different ecosystems. The goal of this study was two-fold: i) reconstruct the evolutionary dynamics of plasmids of Acinetobacter sp. and ii) investigate the evolutionary cross-talk between plasmid and chromosome. A total of 29 plasmids and seven Acinetobacter genomes from NCBI were included in this study. A computation tool called Blast2Network was used for visualzing plasmid and chromosome relationships. For goal one, the authors retrieved 493 protein sequences form all 29 plasmids and used them as input for the Blast2Network program. This resulted in a network where each plasmid is represented by a ring of balls, each ball being a single protein. In addition there are lines connecting homologous proteins. Changing the degree of identity between proteins generates different networks, such that at 50% sequence identity more proteins are linked together than at 100% sequence identity. Overall the pattern remains the same with three distinct clusters. The first represents a group of plasmids called the pKLH-group that were isolated from different species/strains of Acinetobacter. The second cluster includes plasmids form A. baumannii strains and cluster three has plasmids form other Acinetobacter species. This shows that the pKLH plasmids are the most closely related since they have interlinks among all members even at 100% protein sequence identity, while interlinks decrease for the other two clusters. Thus, although the pKLH plasmids were isolated from different hosts, they have a high relatedness. This is an interesting result, since many of these plasmids are tra- and mob- making conjugation an impossible mechanism of gene transfer between bacteria. Another interesting observation is that plasmids from the same strain often have no connections at 100% identity meaning that no recent genetic exchange probably took place between them. This is surprising because transposition and recombination are common means of gene exchange between plasmids residing in the same host. On the other hand, some proteins, were found to have 100% identity between homologs on plasmids from clusters two and three, meaning that some gene exchange did take place between plasmids from different hosts. Thus, overall these networks are an easy way to visualize complex data. Next they analyzed the functional classes of proteins with the most interlinks. Not surprisingly they found transposition related proteins and mercury resistance related proteins to have high connectivity. It is also interesting that out of 493 proteins, 280 did not have any connections, suggesting that plasmids from Acinetobacter encode a high number of unknown functions. For the second goal, the authors included the genomes of seven completely sequenced Acinetobacter sp. and generated a network of plasmid and chromosome encoded proteins, similar to the network generated previously. The networks show the existence of interconnections between all chromosomes and most of the plasmids. The only exception to this is, four plasmids belonging to cluster three which have no connections to any of the chromosomes. The pKLH group on the other hand was found to be strongly interconnected to two A. baumannii strains. These connections were mostly with mercury resistance related proteins and transposases. Interestingly enough, some plasmids belonging to the same strain such as p1ABAYE, p2ABAYE and p4ABAYE were not found to have any connections to proteins of their host chromosome. This means that gene exchange did not take place between plasmid and chromosome in this case. There are many plasmid-encoded proteins that do not have sufficient identity to chromosome-encoded proteins, suggesting that these may have been acquired from other species/strains of Acinetobacter or other genera. Thus, this study provides an interesting visualization of plasmids of Acinetobacter and their relationships to each other and to some Acinetobacter genomes. I think the most surprising and interesting fact I learned is that tra- and mob- plasmids are promiscuous too and their evolutionary history is often independent of their hosts. There were a couple of areas that were unclear to me, such as, their identity thresholds, which seem to be absolute values instead of a range of values. Also, they do not say why they did not use all twenty-nine Acinetobacter genome sequences and restricted their study to only seven genomes. An interesting feature from figure 4, that they seemed to have overlooked is the fact that single plasmid-bearing proteins have numerous, multiple hits on the same chromosome (A. baumannii SDF) at the 90% threshold. To summarise, their illustrations are really pretty and show the different types of plasmid-chromosome relationships. With the availability of more sequences, such figures may get more difficult to create.

Diya Sen
Graduate student
University of Idaho