APPROACHING THE LOWER LIMITS OF TRANSGENE VARIABILITY

The inclusion of chicken lysozyme matrix-associated regions (MARs) in T-DNA has been demonstrated to reduce the variation in beta-glucuronid ase (GUS) gene expression among first-generation transformed plants. T he residual variation observed between transgenic plant lines with MAR s at the T-DNA borders was investigated. By definition, any phenotypic variance between or within genetically identical plants is caused by random or environmental variation. This variation therefore sets a low er limit to the variation in GUS activities. The variance of GUS activ ity in offspring plant populations of genetically identical individual s was used as an estimate of environmental variation, For transgenic p lants with MARs at the T-DNA borders, the variation between independen t transformants could not be distinguished from the environmental vari ation. The variation could be attributed mainly to the variation in th e GUS activity measurement. Therefore, the MAR element approached the maximal possible reduction of transgene variability given current tech nology and sample sizes. The role of MARs in offspring plants was eval uated by comparing such populations of transgenic plants for the magni tude of and variation in GUS activity. Pairwise comparisons showed tha t the presence of MARs reduced variation in offspring generations in t he same manner as demonstrated for primary transformants. The populati ons carrying a doubled cauliflower mosaic virus 35S promoter-GUS gene tended to be more variable than the Lhca3.St,1 promoter-GUS gene-carry ing populations. This tendency indicated an intrinsic susceptibility o f the doubled cauliflower mosaic virus 35S promoter to variation. Homo zygous plants were approximately twice as active as the corresponding hemizygous plants and tended to be more variable than the hemizygous p lants. We hypothesized that the magnitude of environmental variations is related to a higher susceptibility to transgene silencing.