Identifying Unknown Herbicide Targets
While the strategy just discussed continues to be exploited for herbicide discovery, there are still large gaps in our knowledge of other potential, non-enzyme herbicide targets. For example, the functions of a large proportion of Arabidopsis genes are either partially (about 50%) or completely (more than 30%) unknown. Many of these genes likely encode proteins that are components of signal transduction cascades, transcription factors, or have other regulatory functions. All of these have potential as herbicide targets, since inhibition of their function could be lethal or injurious to a plant. However, in most cases the genes must be identified and the functions of their encoded proteins determined before they can be pursued. There are several strategies for accomplishing this, including silencing or interfering with gene expression, predicting protein function, and comparing profiles of RNAs, proteins, or metabolites.
|The overall purpose of these strategies is to knock out or reduce the expression level of individual genes, choose those that create a negative growth phenotype, and then work ‘backwards’ towards identifying their function.|
These techniques are carried out for large numbers of genes simultaneously in a HTS system. Detailed information on the molecular structure and interactions of the target protein with other proteins must then be obtained, once the function is known.
Interference with Gene Expression: RNA interference (RNAi) is a technique to partially reduce the expression levels of genes by introducing double-stranded RNA segments into plants through transformation. The plant cell has endogenous systems that degrade double-stranded RNAs, and thus simultaneously destroy the normal, single-stranded transcripts. Similar methodologies are antisense RNA and virus-induced gene silencing (VIGS). VIGS has the advantage of being used in plant species that are recalcitrant to high-throughput transformation.
Predicting Protein Function: DNA sequences obtained from genome sequencing projects and other sources can be translated into putative protein sequences. Once obtained, the protein sequences can be compared with other, known sequences in databases, and similarities can help predict possible functions. Recent advances in computer-based homology searches, high-speed protein crystallization and structure determination, and the development of large databases of three-dimensional protein structures will help tremendously in narrowing possible functions for unknown genes.
RNA, Protein, and Metabolite Profiling: Analysis of profiles or arrays of mRNAs (transcriptomics), proteins (proteomics), and metabolites (metabolomics) can provide a global indication of the physiological state of a cell, tissue, or whole plant. Comparisons between treated (with a potential herbicide) and untreated samples can highlight particular pathways or processes that are inhibited, and thus provide clues to the mechanism of toxicity. mRNA arrays (Figure 2) are commonly known as microarrays, and are designed to monitor and quantify changes in gene expression in organisms or tissues in response to a treatment, in this case a potential herbicide. Briefly, individual known cDNAs, gene fragments, or Expressed Sequence Tags (ESTs) are spotted in duplicate onto a solid support and individually hybridized with labeled RNA extracted from treated and untreated tissues or cells. The relative strengths of the resulting hybridization signals are a direct measure of the amount of each particular mRNA. These signals thus reflect the degree of up- or down-regulation of gene expression in response to the treatment. Such assays have been used to compare herbicide-treated to untreated plant tissues in attempts to determine unknown mechanisms of action.
Comparison of protein arrays from treated and untreated plants gives similar information. When combined with RNAi or another gene expression interference technique, these techniques can quickly lead to identification of a target gene.
In an application of metabolomics, single proton nuclear magnetic resonance (NMR) profiling of metabolites coupled with principal component analysis/cluster mapping was used to compare metabolite profiles of untreated plants with those treated with known and unknown compounds. Using a HTS system, this approach successfully confirmed known herbicides as well as accurately predicted the herbicidal activity of several unknown compounds. The technique was subsequently further refined and used to assign metabolite profiles to 400 plant extracts for future testing of candidate herbicides.