Viral genome sequencing by random priming methods

Abstract

Background: Most emerging health threats are of zoonotic origin. For the overwhelming majority, their causative agents are RNA viruses which include but are not limited to HIV, Influenza, SARS, Ebola, Dengue, and Hantavirus. Of increasing importance therefore is a better understanding of global viral diversity to enable better surveillance and prediction of pandemic threats; this will require rapid and flexible methods for complete viral genome sequencing.

Results: We have adapted the SISPA methodology 123 to genome sequencing of RNA and DNA viruses. We have demonstrated the utility of the method on various types and sources of viruses, obtaining near complete genome sequence of viruses ranging in size from 3,000-15,000 kb with a median depth of coverage of 14.33. We used this technique to generate full viral genome sequence in the presence of host contaminants, using viral preparations from cell culture supernatant, allantoic fluid and fecal matter.

Conclusion: The method described is of great utility in generating whole genome assemblies for viruses with little or no available sequence information, viruses from greatly divergent families, previously uncharacterized viruses, or to more fully describe mixed viral infections.

Figure 1

Overview of the strategy. Viral particles are separated from host contaminants using centrifugation and filtration. Viral particles are treated with DNAse I to remove contaminated nucleic acids. Random priming is used to generate 500–1000 bp amplicons which are size-selected and cloned. Colonies are picked and sequenced. Sequence is trimmed and assembled. Contigs are closed using sequence-specific primers.

Figure 2

Outline of the SISPA method. A. Viral RNA is converted to cDNA using random-tagged and poly-A tagged primers (FR26RV-N and FR40RV-T). B. Second strand DNA is synthesized using Klenow exo-DNA polymerase, in the presence of random tagged and virus specific 5' end oligo primers. C. Double stranded DNA is amplified by PCR using the primer tag (FR20RV). D. Amplicons are separated by electrophoresis and products ranging from 500–1000 nucleotides are cloned into the TOPO vector. 96–288 colonies are picked, plasmid DNA is purified and the inserts are sequenced.

Figure 3

Representative assemblies of viruses described in this study. Images shown were generated using DNASTAR Seqman program. A. Enterobacteriophage MS2 (3569 bp). B. Human Rhinovirus 16 (7124 bp). C. Newcastle disease virus Lasota (15186 bp). All assemblies have been aligned with their reference genomes. Gaps and low coverage areas which require closure are circled.

Figure 4

A. Depth of coverage of viruses. Depth of coverage statistics were generated for each contig (using the output of DNAStar Seqman program). Average coverage is the summed length of all sequence reads in a contig, including gaps divided by the contig length. The average and standard deviation for each virus was determined. B. Correlation of genome coverage with colonies picked. The SISPA method was performed for enterobacteriophage M13 (6407 bp), Newcastle disease virus Lasota (15,186 bp) and enterobacteriophage lambda (48502 bp). One, two or three 96-well blocks of clones were sequenced, trimmed and assembled. The sum of the total lengths of edited contigs for each condition was calculated as percent of the total reference genome length.

Figure 5

Relationship between initial virus particle number, genome coverage and percent non-specific sequences generated by SISPA. MS2 viruses were diluted to 10⁸, 10⁶, 10⁴, and 10² particles per SISPA DNAse I reaction. The sum of the total lengths of edited contigs for each dilution was calculated as percent of the total reference genome length. Non-specific sequences were determined as those sequences which did not match reference genome with a cutoff value less than 10^-25.