Metagenomic analysis of RNA viruses in a fresh water lake

Abstract

Freshwater lakes and ponds present an ecological interface between humans and a variety of host organisms. They are a habitat for the larval stage of many insects and may serve as a medium for intraspecies and interspecies transmission of viruses such as avian influenza A virus. Furthermore, freshwater bodies are already known repositories for disease-causing viruses such as Norwalk Virus, Coxsackievirus, Echovirus, and Adenovirus. While RNA virus populations have been studied in marine environments, to this date there has been very limited analysis of the viral community in freshwater. Here we present a survey of RNA viruses in Lake Needwood, a freshwater lake in Maryland, USA. Our results indicate that just as in studies of other aquatic environments, the majority of nucleic acid sequences recovered did not show any significant similarity to known sequences. The remaining sequences are mainly from viral types with significant similarity to approximately 30 viral families. We speculate that these novel viruses may infect a variety of hosts including plants, insects, fish, domestic animals and humans. Among these viruses we have discovered a previously unknown dsRNA virus closely related to Banna Virus which is responsible for a febrile illness and is endemic to Southeast Asia. Moreover we found multiple viral sequences distantly related to Israeli Acute Paralysis virus which has been implicated in honeybee colony collapse disorder. Our data suggests that due to their direct contact with humans, domestic and wild animals, freshwater ecosystems might serve as repositories of a wide range of viruses (both pathogenic and non-pathogenic) and possibly be involved in the spread of emerging and pandemic diseases.

Figure 1. Taxonomic classification of assemblies.

Assemblies were classified based on comparison to the CAMERA database using the BLASTX algorithm and an e-value of 1×10e-5 or lower. Sequences in assemblies without significant matches to existing protein sequences (e-value>1E-5) were classified as “Unknown”. The remaining sequences were classified based on best BLASTX hits for their assemblies. Of the “known” sequences, 67% of the November sample and 70% of the June sample had homology to published viral sequences.

Figure 2. Composition of viral types.

Assemblies were assigned into one of five categories based on nearest BLASTX homology results. For both the November and June samples, approximately 87% of all viral sequence reads for each season were in assemblies matching RNA viruses.

Figure 3. Distribution of viral reads by viral types.

Names were assigned to assemblies based on the best BLASTX match. Assemblies with the same virus name were grouped together, and the numbers of reads comprising these assemblies were added to generate a reads per unique types value. Although RNA viruses were the target of this study and the best-represented in the data, DNA viruses have been included as well for comparison.

Figure 4. Distribution of potential hosts of RNA viruses.

BLASTX results were used to classify viral hits. The information obtained from the bibliographic description of the identified virus was then used to identify the most probably host.

Figure 5. Phylogenetic trees of Banna virus.

The entire vp1 and vp2 segments were chosen for sequence comparison analysis. Other sequences used for the analysis were downloaded from the NCBI database. All sequences were analyzed using ClustalX with default parameter settings as described in Materials and Methods. Consensus tree bootstrapping was performed with Geneious 4.0.4 using the neighbor-joining method and 1,000 samples.

Figure 6. Genomic organization of a novel paralysis virus.

(A) Genomic organization of cricket paralysis virus. (B) Genomic organization of the putative novel paralysis virus identified in Lake Needwood. Two contiguous sequences with sizes of 6000 and 1500 nucleotides assembled from combined June and November reads exhibited significant homology with the cricket paralysis virus. Using targeting amplification a DNA fragment of 700 nt was amplified, sequenced and used to link the two contigs thus generating a consensus sequence of 8086 nt. Using BLASTX we mapped the boundaries of the two (non structural and structural) polyproteins. (C) Phylogenetic analysis of the Lake Needwood virus assembly with homologous paralysis viruses. A region (containing ∼600 amino acids residues) of the replicase polyprotein was used for phylogenetic analysis after multiple sequences alignment using ClustalX with default parameter settings as described in Materials and Methods. Consensus tree bootstrapping was performed with Geneious 4.0.4 using the neighbor-joining method and 1,000 samples.

Figure 7. Phylogenetic tree of circoviruses.

A region of 90 amino at the 3′ end of the circoviruses polymerase gene was selected for sequence comparison analysis. Other sequences were downloaded from the GOS and the NCBI databases. Selected sequences were analyzed using ClustalX with default parameter settings as described in Materials and Methods. Consensus tree bootstrapping was performed with Geneious 4.0.4 using the neighbor-joining method and 1,000 samples.