Macromolecular assemblies

The approaches used by structural biologists evolve as they tackle the study of increasingly large, complex, and flexible assemblies. Such studies require synergistic combinations of methods over different length and timescales to provide a functional picture of the macromolecular machine under scrutiny. This themed issue on Macromolecular Assemblies comprises seven reviews covering recent conceptual and methodological developments to study complex structures. Two of the reviews analyze biological advances — obtained by X-ray crystallography — in understanding the viral molecular machines that are used for the invasion of animal cells. The other reviews focus on assemblies that have been approached by complementary methodologies. One of the first steps during the entry of a virus into an animal cell is specific recognition of a cellular receptor. In this issue, Stehle and Casasnovas review several examples of crystal structures of viral envelope proteins or whole virus particles in complex with cellular recepors, bringing forth important common features that govern this specific recognition. Virus–receptor interactions are indeed important in defining key aspects of virus pathogenicity and understanding receptor switching — which appears in many cases to be distressingly facile — is a major issue in contemporary virology discussed in this review. A key step in the infection of a cell by enveloped viruses is the fusion of the viral and cellular membrane, and the review by Backovic and Jardetzky covers the recently identified viral membrane fusion proteins belonging to structural class III. The membrane fusogenic glycoproteins of totally unrelated RNA and DNA viruses were found to share a common 3D fold, indicating homology in spite of the absence of sequence conservation. This may have arisen by horizontal gene transfer between viruses, or by the independent acquisition of the gene from the cellular host at different times during evolution. Like the previously characterized class I and II fusion proteins, class III proteins change conformation to induce fusion of the target membrane with the virus envelope. These conformational changes can be mapped onto the structural homologs, providing a striking example of the use of structural biology to identify links between otherwise totally unrelated organisms. The viral proteins analyzed in this issue were studied by X-ray crystallography, and additional studies by other techniques will certainly contribute in the understanding of their function. This is the case for the poorly understood, higher order complexes in membrane fusion. Indeed, the study of most complex macromolecular assemblies relies increasingly on hybrid methods, combining different methodologies such as electron microscopy, NMR spectroscopy, X-ray crystallography, solution studies such as small angle X-ray or neutron scattering (SAXS and SANS), computational modeling and mass spectrometry/proteomics. One example is the tumor suppressor molecule p53, which can adopt a number of different conformations because of its intrinsic flexibility, which relates to its high degree of promiscuity of binding partners. The study of such proteins involves X-ray crystallographic and/or NMR analyses of the individual domains, and then fitting the pieces together with the aid of other techniques, like electron microscopy and solution scattering. Other examples include the proteasome, as well as large virus particles that cannot be crystallized as a whole. The structure of the nuclear pore complex and of cellular organelles are other challenging examples. Such studies are the topic of the remaining reviews in this issue. The structure of p53 has been studied by X-ray crystallography, NMR, SAXS, and EM, but remains controversial. A prominent model of the human p53 tetramer is based on X-ray and NMR structures of the so-called tetramerization domain, which forms tetramers when this domain is expressed in isolation. In this model, four of these domains form a tetramer unit from which the other domains extend. However, a recent cryo-EM structure of the murine p53 tetramer shows a different, more compact organization, in which the tetramer is formed by pairwise contacts between N-terminal and C-terminal regions. Some properties of p53 implicate a coordinated action of N-terminal and C-terminal domains and are better explained by the recent cryo-EM structure. This topic is covered in the review by Okorokov and Orlova. Regulated proteolysis plays a key role in cellular function, and two of the reviews in this volume are concerned with the mechanisms of action of large proteolytic machines. Cheng is concerned with the structure and function of the 26S proteasome, the large machine responsible for regulated protein degradation in eukaryotic cells. The barrel-shaped core structure has been characterized by X-ray crystallography, but targeting and access to the proteolytic chamber are determined by elaborate regulatory complexes at the ends of the barrel. This is the case of the 19S complex containing the components involved in recognizing and processing ubiquitinated substrates, as well as the ATPases required for unfolding of these substrates so that they can be threaded into the proteolytic chamber. Cryo EM is being used to determine the structure of the full 26S complex including core and 19S complexes. In a related review, Striebel et al. discuss the ATPases of the AAA+ family, which use the energy of ATP hydrolysis to unfold proteins that are introduced into the proteolytic chamber of the proteasome. Their review focuses on their specificity, mechanisms of action, and regulation by cofactors to recognize and unfold substrates in the proteasome as well as in bacterial ATP-dependent proteases. For a full understanding of molecular machineries in the cell, it is necessary to relate atomic structures to larger scale information. The resolution of cryo-EM maps has been steadily improving, with the best single particle structures of icosahedral viruses around 4 Å, and increasing numbers of intermediate resolution maps in the 4–10 Å range. In order to extract the biological information from such maps, efforts must be put into interpretation of the densities. Given the large number of atomic structures available, it is often possible to dock atomic models of components of a large assembly into EM maps. In other cases, it is necessary to create de novo models and use the cryo-EM map to filter the models obtained. Atomistic models can be fitted into maps as rigid bodies or by flexible docking of domains or secondary structure elements. Such hybrid approaches involve computer modeling and density correlation, and ideally make use of all available structural and biochemical information in deriving structural models. The current methodologies that are being developed — and in particular the interpretation of electron density maps in the 4–10 Å resolution range — is covered in the review by Lindert et al. An example of a multiscale approach to a huge cellular assembly, the 100 MDa nuclear pore complex, is presented in the review by Elad et al. This large ring structure spanning the two nuclear membranes forms the portal for selective transport of cargo between nucleus and cytoplasm. A difficultly in defining the full structure of the complex is that the pore is lined with a large mass of natively unstructured chains, the FG repeats, which are directly implicated in the transport process. Owing to the size and complexity of this assembly, a wide variety of approaches are being combined to study its structure and function, including cryo-electron tomography, AFM, and bioinformatics to incorporate the full range of structural and biochemical information. Such a broad combination of approaches to understand the structural dynamics of macromolecular processes in the cell is also the topic of a review appearing in a parallel themed issue on ‘Cell Structure and Dynamics’ in Current Opinion in Cell Biology by Sali and colleagues [1]. Ultimately one would like to relate all this information to the in vivo context and extend the hybrid approaches to even larger scales. Methods for determining the 3D density of whole organelles or cell structures are now rapidly developing [2]. Cryo-electron tomography of vitrified samples has the capacity to reveal the 3D structure of cell and tissue sections or organelles in their native, hydrated state, in the 3–5 nm resolution range. In favorable cases, it is possible to relate molecular and cellular information, such as the ATP synthase complexes on mitochondrial membranes, which can be discerned in tomograms of isolated mitochondria (Figure 1a). The resolution of mitochondrial features in vitrified sections is lower, owing to the greater technical difficulty of imaging vitrified sections, but the membrane topology is clearly seen (Figure 1b). The 3D structure of mitochondria has been recently reviewed [3]. In a study of vitrified sections of skin desmosomes, it has been possible to interpret tomogram density in terms of the molecular packing of the cadherin proteins forming the inter-membrane junction [4]. Electron tomography has also been applied to the native 3D organization of bacterial polysomes, revealing a pseudo-helical arrangement of the ribosomes [5]