1Plant Protection Unit, Department of Primary Industries, Indooroopilly 4068, Queensland, Australia; 2Centre for Wet Tropics Agriculture, South Johnstone Road, South Johnstone 4859, Queensland, Australia

Background and objectives
Banana streak disease, caused by banana streak badnavirus (BSV), is found in almost every banana-growing region of the world. BSV was first detected in Australia in 1992 in Musa AAB Group cv. Mysore. Since 1992, there have been sporadic occurrences of BSV in other Musa cultivars such as Williams (AAA Group, Cavendish Subgroup) and Lady Finger (AAB Group, Pome Subgroup), in both tropical and sub-tropical production areas. Detection of BSV in bananas is problematic, due partly to the great variability of the virus [1]. In this report, experiments to characterize and develop diagnostic tests for two strains of BSV (BSV-Cav and BSV-Mys), isolated from Williams and Mysore plants, respectively, are described.

Materials and methods
BSV-Cav and BSV-Mys were collected near Innisfail, northern Queensland. BSV and sugarcane bacilliform virus (SCBV) ELISA detection kits (Agdia) were used according to manufacturer's instructions, except that the virus extraction buffer was 0.05 M Tris-HCl, pH 7.4, containing 0.5% sodium sulphite and 5% skim milk. Virus mini-preps were done using a modified method of Ahlawat et al. [2]. BSV-Mys was purified from leaf material using a modified method of Lot et al. [3]. BSV-Cav sequence was amplified from a virus miniprep by immunocapture PCR, utilizing SCBV antibodies for virus capture, and Badna-T/Badna-3 primers for PCR [1]. XhoII-digested genomic DNA of BSV-Mys was cloned into pBluescript SK+ vector (Stratagene). Additional BSV-Mys clones were generated by PCR, using a virus-specific primer and the Badna-T primer. PCR products were cloned using a TA cloning kit (Invitrogen).

Results and conclusions
ELISA: BSV-Cav could be detected using both the BSV and SCBV ELISA systems, but neither system was suitable for the detection of BSV-Mys. For BSV-Cav, greater sensitivity of detection was obtained using the SCBV ELISA and all subsequent indexing was done with this system. To assess the reliability of ELISA, a field plot of 120 BSV-Cav-infected Williams plants was tested at 4-12-week intervals over two cropping cycles. At the beginning of the cropping cycle (sucker stage), detection efficiency was as little as 3%. However, the number of plants with detectable virus increased as the crop developed, reaching a maximum of about 80% at the time of bunch emergence. Plant developmental stage, as well as climatic conditions favouring a rapid rate of plant growth, appeared to be factors favouring virus expression.

Cloning and sequencing of BSV: Amplicons of BSV-Cav and BSV-Mys, with sizes of 1.3 and 2.0 kb, respectively, were cloned and sequenced. GenBank database searches were done and no exact matches obtained, but significant similarity was found with other badnaviruses. The sequences were translated and aligned with those of other badnaviruses and an RNase H motif found in both sequences, as well as a reverse transcriptase motif in the BSV-Mys sequence. Phylogenetic analyses were done, and BSV-Cav and BSV-Mys found to be most closely related to each other and to Commelina yellow mottle badnavirus. The magnitude of divergence of sequences suggests that BSV-Cav and BSV-Mys are distinct viruses.

PCR primers have been designed which hybridize to BSV-Cav and not BSV-Mys, and vice versa, allowing rapid, sensitive and specific detection of the two viruses.

1. Lockhart BEL, Olszewski NE, 1993. In Ganry J, ed., Breeding bananas for resistance to diseases and pests. CIRAD/INIBAP, Montpellier, pp. 105-113.
2. Ahlawat YS, Pant RP, Lockhart BEL et al., 1996. Plant Disease 80, 590-592.
3. Lot H, Djiekpor E, Jacquemond M, 1991. Journal of General Virology 72, 1735-1739.