The Arabidopsis Predicted Interactome

    Using orthologs (same gene/ different species), we have predicted what protein-protein interactions might be occurring in Arabidopsis. Over 20,000 PPi's are now predicted, which significantly overlap with the 700 known interactions. Predicted PPi's are significantly more co-expressed, than randomly chosen protein pairs, and they are often co-localized (found in the same organelle). This roadmap of interactions in Arabidopsis is a valuable tool in understanding how metabolic, signaling and regulatory pathways are constructed.

Click on the image (or here) to browse the interactome yourself, enter your favorite genes and see who they are predicted to interact with! The raw data is available at The Arabidopsis Information Resource (TAIR) website.

This work is done in collaboration with Dr. Jane Geisler-Lee (SIUC), Dr. Nick Provart (U of Toronto), and Dr. A. Harvey Millar (U. Western Australia).

Check out other resources at the Bio-Array Resource (many bioinformatic tools) and SUBA (the subcellular localization database).

      Looking at regulation from the CIS- side.

      Regulation of genes involves more than just transcription (TRANS-acting) factors. By refocusing on how promoter sequences relate to patterns of gene expression we can understand regulation from the inside looking out. Our lab has developed bioinformatic algorithms that correlate the presence of cis-element (a small sequence within the promoter) with patterns of gene expression across the Arabidopsis genome. This Regulatory Fingerprint (see table) shows not only what kinds of things regulate the cis-element, but also the strength (P-value) of this correlation. The ABRE element is regulated foremost by ABA (a plant hormone), but also by many other things, including time of day (circadian rhythm) and blue light. The DRE element is much more specific in its response to cold, wounding and abiotic stress. The ERE element responds to both biotic and abiotic stresses. Using this cis-side analysis and annotation in conjunction with what we know about trans-acting factors, we can begin to reconstruct complete regulatory pathways, and learn how decision making is wired inside a plant. We can also learn how elements act in combination to build regulatory modules, and create designer promoters which express their cargo gene in exactly the right tissue and under the right conditions.

Table 1 notes: 1. Enrichment factor is the number of genes with CRE in the proximal 500bp promoter divided by the number of genes expected to have this CRE by chance. 2. Number of genes with CRE on indicated microarray showing 2-fold or greater induction (blue) or at least 2-fold suppression (orange) or both (purple) and number of genes expected based on percent induction of whole genome. 3. Chi-squared probability that the differences between observed and expected (based on genomic average) induced or suppressed genes with CRE occur by chance. P <0.05 is considered significant correlation without correction for multiple hypothesis testing, p<1.3e-04 is considered significant after Bonferroni correction (red numbers). Genes with CREs were tested on all microarrays, but only those with significant correlation are listed here. Mannitol = 300mM mannitol, ABA = Abscisic Acid treatment (10 or 100 uM), Salt = NaCl 150mM, Cold = 4C, Ozone = 200 ppb 1 hour, Heat = 38C, Met. Viol. = methyl viologen, Wound = mechanical wounding of leaves, CK = t-Zeatin 20uM, 1-3hr, Brass. = Brassinolide 10nM, NO= Nitric Oxide, Peroxide = Hydrogen peroxide, GA= gibberilic Acid 1 uM.

Filling in the Gaps in Biosynthetic Pathways

Many of the biosynthetic pathways for secondary compounds like suberin, lignin or flavinoids still have gaps. Regulatory proteins and transcription factors controlling these branching pathways have not all been identified. Several gaps exist where metabolites are known, but the enzyme making that metabolite is unknown. Large families of similar enzymes are often found where the general function is known, but not which specific pathway, or under what conditions is that part of the pathway activated.

We use a combination of different bioinformatic approaches to help identify candidate genes. In the schema to the left, many of the enzymes (balls) are co-regulated (blue). Other enzymes (orange) are expressed in the same pattern. Are these enzymes under different regulational control? Is their activity coordinated through something other than transcription? Are these constitutive steps? Have we misidentified these enzymes? Still more enzymes (grey balls) are completely unknown. Using co-regulation pattern, predicted interactions, and the presence of common cis-regulatory elements, we will try to find candidate genes to fill in the gaps in this and other pathways.

This project is a collaborative effort with Dr. Geisler-Lee (SIUC plant biology) and Dr. Zhu (SIUC computer science).

Transcriptomics of Competition in Arabidopsis

Arabidopsis, like many other plants often finds itself adjacent to many of its sisters and sibling rivalry ensues. How does a plant know how many neighbors it has and how closely they are packed together? What resources do siblings fight over? We have sought answers to these questions by examining the transcriptome of Arabidopsis plants in different degrees of intra-specific competition. The image to the left shows the hybridization of RNA from Arabidopsis plants (above image) planted at different densities. We are using the 27k chip containing probes for 26,000 different Arabidopsis genes from Qiagen-Operon.

This project is a collaborative effort with Dr. David Gibson (SIUC Plant Biology) and Dr. Andrew Wood (SIUC Plant Biology)