Rapid cDNA synthesis and sequencing techniques for the genetic study of bluetongue and other dsRNA viruses

Abstract

The genetic study of double-stranded (ds) RNA viruses by sequence analyses of full-length genome segments, or entire viral genomes, has been restricted by the technical difficulties involved in analyses of dsRNA templates. This paper describes improved methods for sequence-independent synthesis of full-length cDNA copies of dsRNA genes and associated sequencing strategies. These methods include an improved version of the ‘Single Primer Amplification Technique’ (SPAT – [Attoui, H., Billoir, F., Cantaloube, J.F., Biagini, P., de Micco, P. and de Lamballerie, X., 2000. Strategies for the sequence determination of viral dsRNA genomes. J. Virol. Methods 89, 147–158]), which is described here as ‘Full-Length Amplification of cDNAs’ (FLAC). They also include the development of direct sequencing methods (without cloning) for the resulting full-length cDNAs. These techniques, which are applicable to any viruses with segmented dsRNA genomes and conserved RNA termini, make it possible to generate sequence data rapidly from multiple isolates for molecular epidemiology studies.

Introduction

The family Reoviridae is one of eight families of dsRNA viruses (Birnaviridae, Picobirnaviridae, Cystoviridae, Hypoviridae, Partitiviridae, Chrysoviridae, and Totiviridae [Mertens, 2004]). The reoviruses (a term used here to describe any member of the family Reoviridae) have 10–12 dsRNA genome segments, which range in size from ∼ 0.2 to ∼ 4.5 kbp (Mertens et al., 2005a). Most of the methods previously used for sequencing individual genome segments are technically demanding, time consuming and require relatively large amounts of purified and undamaged RNA. Although some sequence data have been accumulated for a few of the reoviruses (e.g. rotavirus SA11 – Mitchell and Both, 1990) mostly by a number of laboratories working in collaboration, studies of their evolutionary relationships have been hampered by the lack of rapid and efficient sequencing methods.

Direct sequencing of RNA molecules by enzymatic techniques is possible, although it usually generates only relatively short terminal sequences and the ubiquitous nature of ribonucleases makes it technically demanding (Donis-Keller et al., 1977, Mertens et al., 1984). Makeyev and Bamford (2001) have recently reported use of the highly processive Phi 6 RNA dependent RNA polymerase in RNA sequencing reactions, which may provide a basis for direct RNA sequencing methods in the future. However, the current absence of such methods has made it essential to develop strategies for the synthesis of cDNA copies from RNA templates, which can then be analysed by established DNA sequencing techniques (Attoui et al., 2000, Bigot et al., 1995, Lambden et al., 1992, Potgieter et al., 2002, Vreede et al., 1998).

Individual reovirus genome segments need to be separated, or selected in some way prior to sequencing, to avoid generating data from several different RNAs simultaneously, making the data unreadable. Separation can be achieved by electrophoresis, by bacterial cloning of cDNAs (although this requires extra work/time and can generate ‘cloning artifacts’) or by synthesis of cDNAs using primers that are specific for a single genome segment. However, the design of such primers is difficult for any segment that has not already been sequenced. Even then, mis-priming (especially during reverse transcription), RNA secondary structure, and high GC content can result in a failure to generate exclusively full-length cDNA products in the RT-PCR reactions. Inefficient transfection or cloning of vectors in host bacteria are also factors that can cause problems, particularly with larger dsRNA segments.

The synthesis of cDNAs from previously uncharacterised dsRNA templates can be achieved by ligating a defined oligonucleotide to both 3′ termini, followed by reverse transcription using a primer complementary to the oligonucleotide (Attoui et al., 2000, Lambden et al., 1992). However, this does not prevent mispriming and does not provide a method for direct sequencing of the cDNA products.

This paper describes improved and optimised methods for the synthesis, amplification and sequencing of full-length cDNAs from bluetongue virus (BTV) genome segment 2 (∼2950 bp). This includes an improved version of the ‘Single Primer Amplification Technique’ (SPAT – originally described by Lambden et al. (1992) as modified by Attoui et al. (2000)). The new method, which is described here as ‘Full-Length Amplification of cDNAs’ (FLAC), prevents mis-priming during cDNA synthesis by reverse transcriptase. FLAC is not dependent on previous sequence data for primer design, can therefore be used to amplify any large and previously uncharacterised dsRNA virus gene, and has been used successfully by Shapiro et al. (2005) to characterise the genome segments of a novel cypovirus.

A new strategy is also described for direct sequencing of the terminal regions of the full-length cDNAs generated by FLAC. This method was used to analyse bluetongue virus (BTV) genome segment 2 (Seg-2), without a cloning step. It is also applicable to other uncharacterised dsRNA virus genes that have conserved RNA termini (e.g. those of the other reoviruses), as demonstrated by its successful use with Indian sheep orthoreovirus (ISOV). Sequencing of the central portion of the viral genes is then possible, using the initial data generated to design sequencing primers. Because these methods do not involve a cloning step, they generate a consensus sequence for the original RNA gene population, avoiding cloning ‘artifacts’ and increasing the speed of analysis.

Using these methods, sequence data can be obtained from a novel virus isolate within 48 h.