Localization predictions for gene products involved in non-host resistance responses in a model plant/fungal pathogen interaction

The technologies for researching the interactions between plants and pathogens have evolved through small molecule extractions, enzyme analyses, gene cloning, mutational genetics, anti-sera-specific labeling, and beyond. This position paper demonstrates the use of primarily two predictive programs (SignalP and LOCtree) of the more than 24 existing computer programs, in deriving useful information from the N-terminal or total sequences of both host and pathogen gene products. The genes selected are known to be of importance in a non-host resistance model. In this model, pea is the non-host and expresses resistance to an “inappropriate” pathogen, a pathogen of bean. The usefulness of this technology increases as more sequences are known and the programs are re-defined. The predictive programs are not infallible, but are a useful supplement to other localization determinations. They can also uncover some unforeseen functions or properties of gene products related to their inherent potential for secretion or cellular localization. A selected sampling of gene products involved in race-specific resistance was also evaluated. Some disclaimers related to the N-terminal-based prediction programs and some factors that might alter their accuracy are also discussed. A position presented herein is that there has been an over estimation of the communication barriers observed cytologically and an under appreciation for how the inherent sequences of the protein product of a gene can determine its ultimate destination.