|dc.description.abstracteng||Life can be understood as the systematic collaborative action of several entities on
molecular level reducing locally the entropy at the cost of the exterior environment.
These entities are in great part macromolecular complexes composed of different proteins and ribonucleic acids. In a bottom-up approach, structural studies on macromolecular complexes can help to understand the principles of life.
In this work, the structures of two very different macromolecular complexes, the human cohesin and the E. coli 70S ribosome, are studied by single particle electron microscopy. Structural studies on the 70S ribosome are well established, and very sophisticated questions regarding small binding partners or specific dynamics can be addressed today. In comparison, structural studies on cohesin are rather young and started less than 20 years ago. Only the structural definition of individual domains or subunits was possible so far. A complete model of cohesin has not been defined prior to the present work.
Single particle electron microscopy has proved to be superior to other structure defining methods in terms of studying large macromolecular assemblies (>200 kDa). Moreover, it is competent in resolving dynamics of these entities to a higher extent than any other method. In this work, it was aimed to study the influence of viomycin on the dynamics of the 70S ribosome. Moreover, it was aimed to generate the first structural model of cohesin. Single particle electron microscopy was the method of choice to study both, the 70S ribosome and cohesin.
Cohesin, a key regulator of sister chromatid cohesion, is mandatory for maintaining the topological arrangement of chromatin. Especially mitosis and meiosis require the correct and equal allocation of sister chromatids into daughter cells. An engineered cohesin version that largely lacks the long and flexible coiled-coil domains of Smc1 and Smc3 was shown to be an active ATPase and to interact with regulatory cohesin subunits (Pds5, Wapl). Due to the lower flexibility, this “bonsai cohesin” is more suitable to be studied by electron microscopy than its full-length counterpart.
In this study, the low-resolution structure of a bonsai cohesin tetramer (bonsai Smc1/3, Scc1, SA1), Pds5·bonsai cohesin tetramer (bonsai Smc1/3, Scc1, SA1, Pds5) and Pds5·bonsai cohesin trimer (bonsai Smc1/3, Scc1, Pds5) have been solved ab initio. By comparison of these three independently obtained EM models, specific densities, which were not present in all three models, have been attributed to Pds5 and SA1. Pds5 was shown to be located at the bottom of the Smc1/3 nucleotide binding domains (NBDs), and it seems to bend around both Smc NBDs. The size of the obtained Pds5 density corresponds to the expected size of the protein and shows similarity to a generated homology model. For Pds5, no crystal structure is available, and the binding site of Pds5 had not been known prior to this work. Computational sorting allowed to resolve a compositional heterogeneity of Pds5 occupancy within the sample of Pds5·bonsai cohesin tetramer. SA1 appears to be located at the back of the junction of the Smc1/3 NBDs and the Smc1/3 hinge domain. A 2xRFP tag was introduced to distinguish the Smc1 NBD from Smc3 NBD. An additional density is proposed to correspond to this tag. Accordingly, the correct attribution of Smc1 NBD and Smc3 NBD seems to be likely. The
fitting of the available crystal structures is in accordance to the topology, defined solely by single particle electron microscopy without any remaining undefined density segments. Accordingly, a complete topology map of the crystal structures (Smc1/3 hinge domain, Scc1(C)-Smc1 NBD, Scc1(N)-Smc3 NBD and SA1-Scc1(middle) within the EM model of bonsai cohesin was possible. The most detailed model (bonsai cohesin tetramer) reaches a resolution of roughly 27 Å. It appears that the low resolution is related to cohesin-specific heterogeneity due to the flexibility of the shortened coiled-coils. Based on this work, future structural studies may be enabled to obtain higher resolved 3D models of cohesin.
In the second part of this study, the influence of viomycin on the 70S ribosome dynamics was studied. Viomycin is a translocation-inhibiting antibiotic that was shown to stabilize A-site tRNA and to be located between h44 (30S subunit) and H69 (50S subunit). The focus of this work is to understand the inhibitory mechanism with respect to a change in the dynamical landscape. Prior to this work, the effect of viomycin on these dynamics had been discussed controversially. Specifically, the effect on the 30S body rotation and the tRNA dynamics had been unclear.
In the present work, it has been shown that viomycin induces higher 30S body rotation states. High concentrations of polyamines (spermidine and putrescine) can promote a similar but smaller effect on 30S body rotation. Regarding the 30S body rotation, the effect of one viomycin molecule can be mimicked by the collaborative action of several polyamine molecules. As high degrees of 30S body rotation are loosely coupled to more frequent tRNA hybrid states, it was examined if hybrid states are possibly induced by viomycin. Preliminary results suggest favoured hybrid states when viomycin is present. Further results are awaited to draw distinct conclusions. The effect of viomycin on 70S ribosome dynamics can be understood as a proof of principle that single particle electron cryo-microscopy is an appropriate way to address the inhibitory effect of a small molecule with respect to changing macromolecular complex dynamics.||de