Identification of potential interaction networks using sequence-based searches for conserved protein-protein interactions or "interologs"

Abstract

Protein interaction maps have provided insight into the relationships among the predicted proteins of model organisms for which a genome sequence is available. These maps have been useful in generating potential interaction networks, which have confirmed the existence of known complexes and pathways and have suggested the existence of new complexes and or crosstalk between previously unlinked pathways. However, the generation of such maps is costly and labor intensive. Here, we investigate the extent to which a protein interaction map generated in one species can be used to predict interactions in another species.

Figure 1

Experimentally verified interactions between Saccharomyces cerevisiae and Caenorhabditis elegans. (A–D Yeast diploid cells expressing each of 35 C. elegans potential interologs. Pairs are arranged in the order described in Table 2. The five patches at the bottom are controls (negative control on the left side and controls of increasing interaction strength towards the right side). See Vidal (1997) for a detailed description of these controls. (B) β-Galactosidase assay to detect the expression of GAL1::lacZ. (C) Growth assay on SC-Leu-Trp-His, +20 mM 3AT plates to detect the expression of GAL1::HIS3. (D) Growth assay on SC-Leu-Trp-Ura plates to detect the expression of SPAL10::URA3. (E) Conservation of interactions. Each C. elegans protein pair tested was plotted according to two E-values. The first E-value corresponds to the conservation between the X (from yeast) and X‘ (from C. elegans) proteins while the second E-value corresponds to the conservation between the Y (from yeast) and Y‘ (from C. elegans) proteins. The smaller of the two E-values was plotted on the X-axis and the greater on the Y-axis. The C. elegans protein pairs that tested positive in the two-hybrid system are labeled in black.

Figure 1

Experimentally verified interactions between Saccharomyces cerevisiae and Caenorhabditis elegans. (A–D Yeast diploid cells expressing each of 35 C. elegans potential interologs. Pairs are arranged in the order described in Table 2. The five patches at the bottom are controls (negative control on the left side and controls of increasing interaction strength towards the right side). See Vidal (1997) for a detailed description of these controls. (B) β-Galactosidase assay to detect the expression of GAL1::lacZ. (C) Growth assay on SC-Leu-Trp-His, +20 mM 3AT plates to detect the expression of GAL1::HIS3. (D) Growth assay on SC-Leu-Trp-Ura plates to detect the expression of SPAL10::URA3. (E) Conservation of interactions. Each C. elegans protein pair tested was plotted according to two E-values. The first E-value corresponds to the conservation between the X (from yeast) and X‘ (from C. elegans) proteins while the second E-value corresponds to the conservation between the Y (from yeast) and Y‘ (from C. elegans) proteins. The smaller of the two E-values was plotted on the X-axis and the greater on the Y-axis. The C. elegans protein pairs that tested positive in the two-hybrid system are labeled in black.

Figure 2

A predicted network of Caenorhabditis elegans protein interactions involving a subset of the 216 pairs of protein interactions tested in our analysis. Corresponding yeast proteins names are shown in parentheses. Nodes represent proteins and edges represent an interaction between two yeast proteins. Interactions also detected in C. elegans are represented by red arrows. Interactions detected in both orientations are represented by double-headed red arrows. Interactions involving proteins 1–8 have been implicated in RNA processing in yeast and the interaction cluster (Nos. 9–11) involves three proteins whose potential yeast orthologs have been found in TRAPP complexes.

Figure 3

Assessment of the reverse two-hybrid system as a method to further study Caenorhabditis elegans protein interactions. Reverse two-hybrid system. Each yeast diploid expressing one of the 35 conserved interacting protein pairs from C. elegans was tested to identify those that are sensitive to 5-Fluoroorotic acid (5-FOA). Pairs are arranged in the order described in Table 2. Diploids were replica plated onto both SC-Leu-Trp (A) and SC -Leu, -Trp + 0.2% 5-FOA (B). Twenty-seven of thirty-five pairs (77%) were 5-FOA sensitive.