Featured Article - November 2009
Short description: A comprehensive atomic-resolution analysis of the centralmetabolic system of Thermotoga maritima sheds light on protein evolution.
For decades, metabolic pathways have been viewed as a series of chemical reactions, and scientists have worked away at cataloging substrates, reactions and products. Yet it is increasingly clear that this representation no longer adequately describes the mass of genomic and proteomic information we now have about a cell's metabolic system. Viewing this information as a biological network seems to be the way forward.
Now an entire network can be viewed at atomic resolution in three dimensions. All of the 478 proteins forming the central metabolic network of the marine bacterium Thermotoga maritima has been overlayed with structures. This remarkable feat was achieved by PSI JCMM, PSI JCSG, the Department of Bioengineering, University of California at San Diego and the Burnham Institute for Medical Research at La Jolla. Between them, they compiled and analyzed the structures of 120 proteins, solved by the JCSG, other PSI centers, and other structural biology groups, and modeled the remaining 358.
T. maritima is a thermophilic bacterium that thrives at around 80°C. It was first discovered in a geothermal vent and is of interest for two reasons. The first is that from an evolutionary point of view it represents the deepest known branch point in the bacterial domain and it is also one of the slowest-evolving bacterial lineages. The second is that T. maritima metabolizes many carbohydrates, including cellulose and xylan, producing hydrogen as a waste product, and so is a potential source of renewable energy.
The team first constructed a framework for the metabolic network using information gleaned from nearly 150 publications on T. maritima, from which they were able to annotate more than half of the metabolic reactions. They then worked out the remaining interactions using a variety of approaches, including comparison of reactions with those in other similar microorganisms.
The initial framework was tested using flux balance analysis, a mathematical technique that identifies gaps in a network. Eventually, the network consisted of 478 genes, 503 metabolites and 562 intracellular and 83 extracellular reactions. With this information, the team was able to simulate the metabolism of T. maritima at both the biochemical and the molecular level.
Three-dimensional structures of each of the 478 gene products were generated, either by experimental structural biology or through homology modeling. This structural genomic approach added considerable functional information to the network, as it supported the functional assignment of at least 181 proteins.
But to the team's surprise, what caught their eye from the structural information was how few different folds there were in the T. maritima proteins when compared with a random set of proteins. In the 478 proteins in the network, containing a total of 714 domains, there were only 182 different folds; about 300 different folds would be expected for a random protein set. Of the folds, the triosephosphate isomerase (TIM) barrel was most frequent, followed by the Rossmann folds.
The team explored further this non-random fold distribution by looking at just the core-essential group of proteins. This group has a surprisingly large number of folds (111 folds for 177 proteins) compared with the non-essential protein group (92 folds for 203 proteins). This suggests that core-essential proteins carry out unique functions that require specific folds and non-essential tasks can exploit and adapt existing folds.
Y. Zhang, I. Thiele, D. Weekes, Z. Li, L. Jaroszewski et al. Three-dimensional structural view of the central metabolic network of Thermotoga maritima.
Science 325, 1544-1549 (2009). doi:10.1126/science.1174671