SCMV (Sugarcane mosaic virus) is one of the most important virus diseases of maize in Europe and causes serious yield losses in susceptible cultivars. It is readily transmitted by aphids in a non-persistent manner. Thus, chemical control is not efficient for control of SCMV. Cultivation of resistant maize varieties is the most efficient and environmentally sound approach to limit yield-loss caused by SCMV. So far, the molecular mechanisms underlying the development and progression of SCMV infection in maize are poorly understood. Furthermore, the characterization of genes underlying QTL by positional cloning is very laborious and time consuming. Recently, the methods becoming available within the functional genomics toolbox provide an alternative for pinpointing genes underlying SCMV resistance, especially in view of the planned sequencing of major parts of maize genome. Two complementary approaches, SSH (Suppression subtractive hybridization) and microarray-based expression profiling, were used to isolate and identify candidate genes associated with SCMV resistance (1st and 2nd paper, respectively). Since current maize microarrays include less than 30% of all maize genes, SSH was conducted to identify rare transcripts associated with SCMV resistance. Expression profiling has become the predominant high-throughput transcript profiling method in understanding host-pathogen interaction. In the 1st paper, SSH was combined with macroarray hybridization to identify genes differently expressed in NILs (Near isogenic lines) F7+ (SCMV resistant, carrying Scmv1 and Scmv2 regions from FAP1360A) and F7 (SCMV susceptible). Altogether, 302 differentially expressed SSH-ESTs were identified in four comparisons addressing constitutive genetic discrepancy, inducible genetic discrepancy, compatible reaction, and incompatible reaction. Except for genes related to metabolism, most of the classified genes belonged to the three pathogenesis-related categories, cell rescue, defense, cell death and ageing, signal transduction and transcription, which accounted for 56-66% of the classified genes. In total, 19% (60 of 302) of the identified SSH-ESTs have previously been assigned to 29 bins distributed over all the 10 maize chromosomes. Among the mapped SSH-ESTs, 30% (18 of 60) were located within the Scmv2 and Scmv1 genome regions on chromosomes 3 and 6, respectively, conferring resistance to SCMV. Promising candidate genes have been identified, such as AA661457 (receptor-like kinase Xa21-binding protein 3) for Scmv1. In the 2nd paper, genes associated with SCMV resistance in the NIL pair F7+ and F7 were identified by transcript profiling based on maize unigene-microarrays. Altogether, 497 differentially expressed genes were identified in the same four comparisons as in the SSH approach, addressing constitutive genetic discrepancy, inducible discrepancy, compatible reaction, and incompatible reaction. Compared to the SSH approach, expression patterns of microarray-ESTs and SSH-ESTs were consistent for the same comparisons despite of technical discrepancies. Since pathogen-induced transcripts were underrepresented on the unigene-microarray, fewer microarray-ESTs (45.8%) were classified into pathogenesis-related categories than by using SSH-ESTs (60.5%). Moreover, fewer microarray-ESTs (4) co-segregated with Scmv QTL than SSH-ESTs (18). Therefore, our results demonstrate that SSH-macroarray complements incomprehensive microarrays. The candidate genes isolated and described in the 1st paper were employed to investigate their association with SCMV resistance across seven resistant or susceptible inbreds in the 3rd paper. The number of differentially expressed genes (SCMV infected versus non-infected) in individual lines was 177, 163, 165, 62, 47, 37, and 93, for FAP1360A, D21, D32, Pa405, F7, D145, and D408, respectively. All inbreds were divided into two groups by hierarchical cluster analysis: D32, D21, FAP1360A and D408 formed one group; Pa405, D145 and F7 another group. Due to the genetic structure among the seven inbreds, genetic background and resistance response are confounded. With or without the resistant U.S. inbred line Pa405, 22 and 112 genes were identified by t tests between resistant (D21, D32, and FAP1360A) and susceptible (D145, D408, and F7) inbred lines, respectively. The 112 candidate genes were divided into three clusters by K-means clustering and analyzed in more detail. These candidate genes identified from present analysis can be further investigated in a segregating population by genetical genomics approach. In conclusion, this thesis demonstrates the usefulness of expression profiling to study SCMV resistance and to identify candidate genes potentially affecting the signal transduction pathway or even for previously identified SCMV QTL. This information is relevant for plant breeders in view of development of functional markers. Due to oligogenic inheritance of SCMV resistance, marker-assisted selection (MAS) programs with functional markers would increase the breeding efficiency.
«