Assessment of pleiotropic transcriptome perturbations in Arabidopsis engineered for indirect insect defence

Abstract

Background: Molecular characterization is an essential step of risk/safety assessment of genetically modified (GM) crops. Holistic approaches for molecular characterization using omics platforms can be used to confirm the intended impact of the genetic engineering, but can also reveal the unintended changes at the omics level as a first assessment of potential risks. The potential of omics platforms for risk assessment of GM crops has rarely been used for this purpose because of the lack of a consensus reference and statistical methods to judge the significance or importance of the pleiotropic changes in GM plants. Here we propose a meta data analysis approach to the analysis of GM plants, by measuring the transcriptome distance to untransformed wild-types.

Results: In the statistical analysis of the transcriptome distance between GM and wild-type plants, values are compared with naturally occurring transcriptome distances in non-GM counterparts obtained from a database. Using this approach we show that the pleiotropic effect of genes involved in indirect insect defence traits is substantially equivalent to the variation in gene expression occurring naturally in Arabidopsis.

Conclusion: Transcriptome distance is a useful screening method to obtain insight in the pleiotropic effects of genetic modification.

Figure 1

Constructs used for transgene introduction in Arabidopsis. (a) Genes stacked in COX + and COX++ lines by crossing: HMGR1S, short (cytosolic) isoform of 3-hydroxy-3-methylglutaryl coenzyme A reductase 1; FPS1L, farnesyl diphosphate synthase 1 long isoform; FaNES1, nerolidol synthase 1 from Fragaria x ananassa; COX, mitochondrial signal peptide; P35S, CaMV 35S promoter; Tnos, terminator of the Agrobacterium tumefaciens nopaline synthase. (b) Average signal intensities of the overexpressed (HMGR1S and FPS1L) and housekeeping (GADPH and ACT2) genes in Col-3 (WT), COX + and COX++ lines. Bars indicate the 95% confidence interval.

Figure 2

Potential genetic sources for transcriptome variation. (a) genomic differences between accessions. (b) genomic differences between population individuals. (c) genomic differences between a genetically modified individual and the corresponding wild type. Each bar represents the whole genome of an individual line. WT, wild type. GM, genetically modified.

Figure 3

XY scatter plots of transcriptome data. The correlation constant (R²) represents the variability of the global gene expression profile of two samples or groups. A small R² indicates large variation, a large R² indicates small variation. (a). XY scatter plots of the Col-3 replicates (WT) with each other (1st row) and with a replicate of COX + and COX++ with the smallest R² value (2nd and 3rd row, respectively). (b). XY scatter plot for the averages of the groups.

Figure 4

PCA plots using RMA normalized gene expression data of Col-3 plants and COX + and COX++ lines. a, PCA plot with PC1 (X) and PC2 (Y). b, PCA plot with PC3 (X) and PC2 (Y). c, PCA plot using genes with CV < 20% across wild type samples with PC1 (X) and PC2 (Y). d, PCA plot using ANOVA insignificant genes (α = 5%) with PC1 (X) and PC2 (Y). Percentages are the variation explained by the corresponding PC. Dark grey circles: wild type samples, light grey: COX + and black: COX++ samples.

Figure 5

Transcriptome distances between Arabidopsis genotypes. Values above the diagonal are the distances based on intact (un-weighted) PC scores and values below the diagonal are distances based on the weighted scores on the first 9 PCs of a PCA on gene expression data of the wild type and transgenic lines of this study (a), wild type accessions of public databases (b) and groups of the Cvi/Ler RIL population (c). Significant values are shown in bold (permutation test, P-value = <0.05). a, b and c correspond to the source of the data which is delimited by blocks. Accessions shaded by the same color belong to the same experiment.

Figure 6

A schematic overview of the possible position and pleiotropic effects of inserted construct and/or gene(s).