Deep Lake is a marine derived, hypersaline system in Antarctica that remains perennially ice-free with water temperatures dropping to -20°C. These harsh environmental conditions have led to a low complexity microbial community, completely dominated by members of the haloarchaea, including four isolated species (tADL, DL31, Hrr. lacusprofundi and DL1) that account for ~72% of the lakes cellular population. Genomic sequencing and analysis of the four isolated species combined with metagenomics have revealed an unprecedented level of inter-genera exchange of long (up to 35 kb) stretches of identical DNA. However, despite the rampant, promiscuous exchange of DNA, distinct haloarchaeal lineages appear to prevail in the lake by virtue of their unique capacities for niche adaptation (1, 2). With no apparent metazoan grazers present in the lake, viruses are hypothesised to play a dominant role in shaping the microbial community of Deep Lake. In this present study we applied metaproteomics for the first time on a hypersaline environment and combined it with in-depth genomic and metagenomic analysis of Deep Lake CRISPR (Clustered Regularly Interspaced Short Palindromic Repeat) and BREX (Bacteriophage Exclusion) (3) systems to elucidate host-virus interactions.
Shotgun metaproteomics was performed on Deep Lake biomass from 5 distinct depths, captured by sequential filtration through 3 µm, 0.8 µm and 0.1 µm filters during the Antarctic summer of 2008/2009. All identified proteins were manually annotated and grouped into taxonomic and functional categories. We characterized CRISPR systems of the four genomes and the Deep Lake metagenome and used CRISPR spacer and repeat sequences to identify sources of invading DNA.
The Deep Lake metaproteome comprised around 1100 detected proteins. A striking feature was the identification of multiple, highly abundant cell surface proteins with a high degree of sequence variation compared to the genomes of the isolate species (“variants”). E.g. we identified 6 distinct proteins all matching the main S-layer component of tADL. Furthermore we detected variants for archaella (archaeal flagella), pili and other cell surface proteins. Multiple viral proteins were detected with sequence similarity to other, mainly haloarchaeal viruses. Functional CRISPR loci could be identified in the genomes of all four isolated species and CRISPR-associated (Cas) proteins were detected for two of them. CRISPR spacers could be linked to different sources of invading DNA, with most, but not all spacers targeting viruses. We detected one BREX protein (PglX) for Hrr. lacusprofundi. Some detected proteins, including cell surface proteins, were encoded on metagenome contigs together with putative viral genes.
The detection of multiple protein variants for cell surface structures like S-layer and archaella is indicative of phylotypes that are present in the lake. Introducing variation in cell surface structures likely provides the haloarchaeal populations with a way of evading viral infection. Consistent with this is the presence of a diverse viral population in Deep Lake. We detected proteins from at least eight distinct haloarchaeal viruses (eight major capsid proteins), with some proteins confirming active viral life cycles (e.g. prohead protease). Furthermore, the CRISPR spacer analysis revealed that some viruses infect multiple species (broad host range). In addition to the acquired cell surface variation, haloarchaeal host cells have employed active CRISPR and BREX systems as defense against viral infection. The presence of cell surface genes on metagenomic contigs together with putative viral genes, and the high degree of sequence variation observed in many cell surface proteins, suggests that viruses are involved in the acquisition, mutation and distribution of cell surface variants within the haloarchaeal populations. Overall, we were able to identify and describe a complex network of virus-host interactions, revealing a pivotal role of viruses in shaping the microbial community in Deep Lake (4).