Virus-host dynamics in the Baltic Sea redoxcline

When comparing the whole genome sequence obtained from a single strain with environmental metagenomic data, differences in the genomes of the same prokaryotic species where found. Many of the genes showing highest levels of plasticity encode for potential virus recognition sites that are also important for nutrient uptake. Thus, without actual data on the strength of viral pressure on a prokaryotic community it is impossible to decide if genomic plasticity truly develops as a resistance mechanism against the infection of specific viruses or due to other mechanisms such as competition for nutrients. The Baltic Sea is characterized by a permanent halocline at 60-80 m depth, separating the mixed and oxygenated surface layer from the stratified and oxygen depleted waters below. Two prokaryotic key players in the Baltic Sea redoxcline have been identified, together accounting for 35-55% of total prokaryotic abundance. In this project we will relate virus pressure, that increases in anoxic waters due to low protist abundances, to the development of genomic plasticity. In experiments, we will challenge cultures of the two identified key players with viruses obtained from the Baltic Sea. Additionally we will also take samples from different depth zones of the Baltic Sea and estimate in situ virus pressure on prokaryotic communities. The project uses the newest high-throughput sequencing technology to identify genomic plasticity that will be related to virus pressure to decide whether viruses truly drive prokaryotic diversity.

The most significant result of this project is that viruses of bacteria are not the main cause of changes in genes encoding for potential virus recognition sites. Viruses infecting bacteria use specific sites on the outside of a bacterial cell to recognise and attach to their host. This recognition mechanism works akin to a key and lock and is generally highly specific in the sense that a specific virus type is only able to infect a specific population of its host bacterium. Given that viruses are incapable of active movements, viruses and their hosts meet per chance, i.e., high abundance of a specific host population increases the likelihood for viruses being able to infect this specific population. At the end of the infection cycle, many viruses kill there host cell, which eventually bursts open and releases a number of progeny viruses. One possible way for bacteria to evade virus infection are changes in genes that encode for the structures on the cell surface that are recognised by specific viruses and used for virus attachment. Even small changes in such cell surface structures may lead to the inability of a specific virus type to attach to the host cell (the key does not fit the lock anymore), making the host cell immune to infection by this specific virus type. It was found that genes encoding for such potential virus recognition sites of a specific bacterial taxon show an unusually high degree of diversity in aquatic environments and it was proposed thatviruses are the main cause for this finding. However, this hypothesis has a significant weak point: cell surface structures are essential in taking-up nutrients. Small changes in such structures have the downside of negatively affecting the ability of this particular cell to take-up nutrients. Thus, being immune to virus infection comes at a fitness cost compared to vulnerable cells of the same bacterial population. In this project, we have tested this hypothesis using a culture of a bacterial taxon called Sulfurimonas gotlandica GD1, that is found in high abundances in the Baltic Sea. We have added nutrients and a diverse community of viruses obtained from its natural habitat to cultures of GD1. At the end of the experiments we have obtained the genetic code of all the genes of GD1 found in each experimental treatment and compared the data with the reference genome ofGD1. We found that the number of genes encoding for potential virus recognition sites with a high degree of changes in the genetic code compared to the reference genome is equal or slightly higher when adding nutrients as compared to adding viruses. Thus, from this data we conclude that the tested hypothesis is not correct and that viruses are not the main cause of changes in genes encoding for virus recognition sites.