In order to be able to develop new drugs against infectious diseases, researchers need to understand their molecular basis. How does the proliferation of a pathogen proceed, what interactions take place between it and the host cell and how are these processes regulated? For this purpose, protein-RNA complexes are examined. They play an important role at different times in the lifecycle of the pathogen or the host cell and are thus involved in infection processes. Together with colleagues, Prof. Alexandre Bonvin has now developed an integrated structure analysis platform that is able to very simply and effectively calculate the structure of large protein-RNA complexes on the basis of diverse experimental data. The so-called “M3 Framework”, which is an extension of the HADDOCK software developed in the Bonvin lab,  is available free of charge for non-profit researchers. The researchers published their results in the scientific journal Nature Methods.

Example of a complex protein structure calculated with the M3 Framework: the box C/D enzyme for RNA methylation.


Like tiny machines, proteins in our body do hard work. Virtually all processes are performed or controlled by these highly specialised protein molecules. They transmit signals, convert energy, initiate chemical reactions or provide growth and movement. These partially very complex protein machines, such as RNA polymerases, are not easy to decode in their structure and function. Usually, scientists use methods of protein crystallography or electron microscopy for this purpose. However, these methods can have the disadvantage that they can impair the natural form and function of the proteins and nucleic acids.

Protein machines

In the current project, researchers follow a different approach, studying large protein-RNA complexes in solution with nuclear magnetic resonance spectroscopy (NMR). This is a method for examining the electronic environment of individual atoms and the interactions with the neighbouring atoms. “The great advantage of the method is that we can experience complex protein machines as active enzymes at work, with their natural dynamic folding and shape,” says research leader Teresa Carlomagno (Helmholtz Centre for Infection Research).

Experimental data

The researchers went a step further and have now developed a modern platform that is capable of integrating various experimental data. “The data can, for example, come from mutational analyses, NMR spectroscopy, electron microscopy, fluorescence spectroscopy, or modelling,” explains Carlomagno. “By combining all methods, a calculation of the protein complex is possible, which gives us a first idea about what we are faced with. At the same time, the M3 platform also provides scientists with information when data are not yet sufficient and further experiments are necessary to accurately describe the protein machine.”

M3 Framework

The M3 Framework is a broadly applicable and user-friendly platform, which effectively integrates single and very mixed structural and biochemical data of different origins and calculates an atomic structure of large complex molecular structures. The protocol of the platform is based on the HADDOCK software developed by Prof. Alexandre Bonvin from Utrecht University. The additional files required to run the M3 Framework in combination with HADDOCK are publicly available for download.


Ezgi Karaca*, João P.G.L.M. Rodrigues*, Andrea Graziadei, Alexandre M.J.J. Bonvin**, Teresa Carlomagno
M3: an integrative framework for structure determination of molecular machines.
Nature Methods, 2017, DOI: 10.1038/nmeth.4392

* Formerly affiliated with Utrecht University
** Affiliated with Utrecht University