B BENCHARKI¹, J MUTTERER², M EL YAMANI³, D ZAOUI⁴ J V ZEIGLER-GRAFF² and G JONARD²

¹Faculte des Sciences et Techniques, PO Box 577 Settat, Morocco; ²lnstitut de Biologie Moleculaire des Plantes, 12 rue General Zimmer, Strasbourg Cedex 87084, France; ³INRA de Settat- Morocco; and⁴ Faculte des Sciences, El Jadida, Morocco

Background and objectives
Barley yellow dwarf virus (BYDV), the type member of the luteovirus group, occurs worldwide and reduces the productivity of small grains and pasture grasses. Luteoviruses are phloem limited and obligately transmitted by aphids in a circulative persistent manner. Five variants of BYDV have been identified, and named originally on the basis of their aphid transmission specificity. In Morocco, PAV-type is the most common variant virus in cereals. In the present work, we investigated the biological and the coat protein coding sequence variabilities of the Moroccan PAV-type isolates.

Materials and methods
A total of 125 virus-infected plant samples collected in surveys were characterized through aphid transmission on Clintland 64 oat and using DAS-ELISA techniques. Among those that tested positive for PAV-type presence, 12 BYDV isolates were maintained on the basis of source plant (barley, durum wheat, oat and Bromus spp.) and origin. The biological variability of each one of 12 isolates was investigated through inoculation of the isolates to a set of eight barley and one oat cultivars. The vector used consistently was Rhopalosiphum padi at a rate of 10 to 12 aphids per plant. The experiment was carried out in the greenhouse at 18±2°C.

Results and conclusions
Significant biological variability was observed among the Moroccan PAV-type isolates studied. The results differentiated between two clusters (1 and 2). Cluster 1 grouped the isolates MA9501, MA9502, MA9504, MA9512 and MA9513, and cluster 2 the MA9415, MA9505, MA9508, MA951 1, MA9514, MA9516 and MA9517. Based on symptom severity expressed by the inoculated differential plants, isolates of cluster 2 were more severe and stunting than those of cluster 1.

In parallel, nucleic acid sequences were studied for the 12 isolates and focused on the coat protein (CP) gene using reverse transcriptase-polymerase chain reaction and sequencing. Sequence analyses showed that all the isolates were identical in size: 603 nt and 200 amino acid residues. Sequence homology grouping yielded the same clusters of the isolates as found earlier. The CP of the isolates in cluster 1 showed 98.5-100% nucleotide and 97.5-100% amino acid identity, while the CP of the isolates in cluster 2 showed 95.9-99.7% nucleotide and 97.5-99.5% amino acid identity. In contrast, the CP between the two clusters ranged from 88.1 to 91.1% nucleotide identity and 83.0 to 87.5% amino acid identity. The comparison of the CP Moroccan isolates and those already established reveals that the difference occurs in 32 amino acid residues. The CPs of cluster 1 were closely related to the Australian (Vic-PAV), North American (NY-PAV, P-PAV), and Japanese (JPN-PAV) PAV-type isolates, while the CPs of cluster 2 were closely related to the North American PAV-129 isolate. Clusters obtained on the basis of genetic variability of the CP gene correlated well with those yielded by the biological variation.