The principal goals of the group of Dr. Marek Mutwil are elucidation of gene function, discovery of master switches controlling cell wall polysaccharide biosynthesis, and development of new methods for bioinformatics and biology. Our group’s unique feature is integration of bioinformatics (dry lab) and molecular biology (wet lab) approaches.
Arabidopsis thaliana has been studied by thousands of scientists for decades; still, approximately 40% of its genes have no described function. Co-expression analysis uses the fact that functionally related genes are often transcriptionally co-regulated, and has been invaluable in uncovering novel participants of biological processes (for review see Usadel et al. 2011). Nevertheless, the analysis often returns large amounts of false negatives (due to genes missing from microarrays) or false positives (low amount of expression data). An improvement of the analysis, comparative co-expression analysis, utilizes the fact that co-expression relationships are conserved across species and can further highlight biologically relevant associations (Mutwil et al. 2011). By utilizing comparative co-expression analysis we want to:
- Elucidate function of unknown gene products. We are developing novel algorithms to compare and visualize conserved (and therefore important) co-expression relationships. Interesting gene candidates found by the analysis are subsequently analyzed in the lab by knock-out/over-expression analyses, as described below. Currently, our analysis pipeline includes eight plant species but we plan to include data from other model organisms, such as human, E.coli, yeast, Drosophila and C. elegans, which should further augment the predictive power of our analyses (Figure 1).
- Apply knowledge transfer between species. While Arabidopsis thaliana is a great plant for basic research, it has no industrial/agricultural value. Sequence comparison is currently used to find the most likely functional correspondent between species, but it’s becoming apparent that this is not so simple (Patel et al. 2012). A true functional homolog should have the same molecular function (i.e. sequence) and same spatio/temporal expression (i.e. similar co-expression relationships), which can be uncovered by comparative transcriptomics.
- Study the evolution of transcriptome. Since transcriptional networks are conserved across species, it enables us to study the age and method of creation of different biological modules.
- RNASeq co-expression networks. Microarrays are still the most accessible source of transcriptomic data, but are slowly being phased out in favor of RNA sequencing technology. RNASeq data promises a more complete transcriptomic snapshot and will be used in future research.
- Develop web-tools to analyze, visualize and interpret biological networks. Any method is as useful as its accessibility, and therefore we strive to provide user-friendly web-tools that can be used with minimal understanding of bioinformatics. We have constructed several popular web-tools, such as GeneCAT and PlaNet.
Figure 1. Interactive family association network centered around COBRA family involved in cellulose biosynthesis. A node represents a gene family, and edge represents significant association between families. Many of the families associated with COBRA are also essential for cellulose biosynthesis, such as Cellulose_synthase, Glycosyl_hydrolase_9 and others. Interestingly, several DUF (domains of unknown function) proteins are also present, making them prime candidates for functional characterization for involvement in cellulose biosynthesis. Right click on the picture to open menu.
The group uses established and novel methods to tackle several aspects of biology, with focus on gene function and plant cell walls. More specifically, the group utilizes molecular biology, microscopy and biochemistry to address the following subjects:
- Gene function confirmation. Gene candidates found by our bioinformatics pipeline are further characterized in the lab (e.g. Mutwil et al. 2010) by knock-out, over-expression, localization and interaction studies.
- Regulators of cell wall polysaccharide biosynthesis. Polysaccharides contained in plant cell walls are essential resources in nutrition, industry and next generation biofuels. Utilizing novel high-throughput screening methods, we want to uncover transcriptional networks that can induce, or repress, biosynthesis of the cell wall polysaccharides in Arabidopsis. The ultimate goal of this project is to be able to modulate abundance of polysaccharides in cell walls of Arabidopsis and agriculturally important plants.
- Turnover of cell wall associated proteins. While biosynthesis of proteins and polysaccharides contributing to cell wall integrity is the major focus of many researchers, no information regarding the degradation of these components inside the cell wall is available. Our goal is to uncover how degradation machinery in cytoplasm (possibly via ubiquitination and subsequent proteasomal degradation) and cell wall (proteases) influence the abundance and activity of cell wall producing and modifying enzymes.
- New methods. Scientists are constrained by available methods to answer biological questions. Cell wall related proteins are especially difficult to analyze, since they are often lowly expressed, membrane bound and/or glycosylated. To overcome this, we are continuously developing new, transferable methods:
- Proteins involved in cell wall homeostasis are often working in multi-protein complexes that still remain to be characterized. Conventional interaction detection methods are difficult to implement, due to the localization (golgi, plasma membrane, cell wall) and properties (glycosylated, low abundance) of the cell wall proteins. We are developing a new method for protein-protein interaction detection based on promiscuous proximity biotinylation. This method will then be used to identify and further characterize novel components of complexes involved in cell wall biosynthesis (Figure 2).
- Gene knockouts in plants often result in no visible knock-out phenotypes, due to functional redundancy caused by large gene families, which in turn complicates functional characterization of gene function. To overcome this, we are attempting to target whole protein complexes for proteasomal degradation. This new method, dubbed “proximity degradation”, will be used to uncover function of gene of interest and its interactors.
Figure 2. Biotin ligase (birA) can biotinylate GST. To investigate whether birA can biotinylate another protein, we have purified birA and GST, and incubated both proteins with ATP and biotin over time of 24 hours. The reaction was then run on SDS gel and probed with streptavidin-HRP (streptavidin binds to biotin). As seen on the picture, birA rapidly biotinylates itself. GST becomes increasingly biotinylated after 1 hour.[less]
Figure 2. Biotin ligase (birA) can biotinylate GST. To investigate whether birA can biotinylate another protein, we have purified birA and GST, and incubated both proteins with ATP and biotin over time of 24 hours. The reaction was then run on SDS gel and probed with streptavidin-HRP (streptavidin binds to biotin). As seen on the picture, birA rapidly biotinylates itself. GST becomes increasingly biotinylated after 1 hour.