Technical Highlight - April 2014
Short description: A hybrid approach that integrates four mass spectrometry methods with structural modeling takes on tough protein complexes.
Classical structural biology approaches are increasingly complemented by emerging experimental techniques. Recently, hybrid methods have been rising to the challenge by leveraging complementary information from multiple strategies. Aebersold, Robinson and colleagues have developed a generic hybrid approach that integrates four different mass spectrometry (MS) methods with structural modeling to enable the characterization of diverse protein assemblies.
In the authors' workflow, purified protein complexes undergo bottom-up, label-free quantification using liquid chromatography–tandem MS to identify and quantify individual protein subunits, and chemical cross-linking prior to MS to detect interface proximities within and between subunits. Information on protein topology is then derived from collisional cross-sections measured with ion mobility–MS, and native MS is used to determine subunit connectivity and global and subcomplex stoichiometry.
Synthesis happens on the computational side. The MS data are used as restraints to refine and score thousands of models that are iteratively sampled with a Monte Carlo approach. The initial high-resolution models for the subunits used as input come from X-ray crystal, NMR or homology modeling data and ensemble analysis predicts the most likely structures after restraint-guided modeling.
The authors used well-characterized bacterial protein assemblies with differing topologies for training and validation of their approach: methane monooxygenase hydroxylase, toluene/o-xylene monooxygenase hydroxylase and urease. Statistical analysis allowed assessment of the relative contribution of each data type in structure prediction, leading to the assignment of a general weighing factor in the scoring function.
As a challenging application, the researchers chose the yeast proteasomal lid, the structure of which was recently characterized by high-resolution electron microscopy. Overall, the best scoring models from this MS-guided approach agreed well with the published electron density maps. In addition, this approach was able to place the smallest of the lid subunits, Sem1, which had been very challenging even in high-resolution EM structures. Finally, the authors structurally characterized an assembly intermediate of the proteasomal base. Thus, this approach holds the potential to tackle proteasomal assemblies previously challenging for classical approaches in structural biology.
A. Politis et al. A mass spectrometry–based hybrid method for structural modeling of protein complexes.
Nature Meth. (9 February 2014). doi:10.1038/nmeth.2841