Using Situs for the integration of multi-resolution structures
- First Online:
- Cite this article as:
- Wriggers, W. Biophys Rev (2010) 2: 21. doi:10.1007/s12551-009-0026-3
- 545 Downloads
Situs is a modular and widely used software package for the integration of biophysical data across the spatial resolution scales. It has been developed over the last decade with a focus on bridging the resolution gap between atomic structures, coarse-grained models, and volumetric data from low-resolution biophysical origins, such as electron microscopy, tomography, or small-angle scattering. Structural models can be created and refined with various flexible and rigid body docking strategies. The software consists of multiple, stand-alone programs for the format conversion, analysis, visualization, manipulation, and assembly of 3D data sets. The programs have been ported to numerous platforms in both serial and shared memory parallel architectures and can be combined in various ways for specific modeling applications. The modular design facilitates the updating of individual programs and the development of novel application workflows. This review provides an overview of the Situs package as it exists today with an emphasis on functionality and workflows supported by version 2.5.
KeywordsStructural models3D data setsMulti-platformModeling
Scientific computing, including modeling and simulation, is crucial for solving biophysical research problems that are beyond the reach of traditional theoretical and experimental approaches (U.S. Department of Energy 2005). Originally confined to a supporting role with respect to experimental or theoretical approaches, modeling and simulation are increasingly seen as capable of creating new evidence in their own right (Lee et al. 2009). Computer-generated hypotheses can be confirmed or refuted, like their experimental or theoretical counterparts, even though the virtual (in silico) world is at best an imperfect mirror of the physical (in vivo) world.
In the late 1990s, funding agencies in the biological sciences took notice of this opportunity. In April 1998, a special Cell Biology and Biophysics Subcommittee of the U.S. National Advisory General Medical Sciences Council examined research trends in the areas of molecular cell biology, structural biology, and biophysics. Among the needs identified by the panel were better (computational) methods for structural analysis of large macromolecular assemblies and imaging macromolecules in cells. Based in part on these recommendations, the National Institutes of Health (NIH) issued a new program announcement that altered the more traditional biological hypothesis-driven review and award criteria in favor of method development (National Institutes of Health 2000). Instead of the traditional proposal style, biophysical scientists in the U.S. could for the first time submit applications based solely on the merit of computational techniques. This paradigm shift was important for the advancement of computational biology, because opportunities for funding computational research had hitherto existed mainly in the physical sciences (U.S. Department of Energy 2005).
Against the backdrop of the emerging research opportunities in computational biophysics, the Situs package was created for the modeling and simulation of large biomolecular assemblies at variable resolution scales. Situs was initially conceived as a platform for the dissemination of structural coarse-graining algorithms to the biophysical community. Powerful experimental techniques such as cryo-electron microscopy (Baker and Johnson 1996), tomography (Medalia et al. 2002), and small-angle scattering (Niemann et al. 2008), which routinely produced 3D structures at a reduced spatial resolution, had emerged. These methods were capable of yielding low-resolution density maps under a wide range of biochemical conditions that allow atomic structures of components to be fitted and docked (Baker and Johnson 1996; Wriggers and Chacón 2001b), and they were in need of software to help integrate the structural data.
The goal of Situs is to characterize the structure and functionally relevant motions of biomolecular systems by integrating experimental data across the resolution scales, using advanced algorithms from neurocomputing, image processing, and visualization. A decade has passed since the publication of the original Situs paper (Wriggers et al. 1999). This review will assess the software as it is used by scientists today. Naturally, the workflow has changed in many ways over the years as compatible molecular graphics programs have evolved and Situs tools have been enhanced, updated, or replaced.
The following sections highlight the current Situs workflow using published usage examples kindly provided by other laboratories. In the first section, a typical correlation-based docking approach in electron microscopy (EM) is described, using the recent model of the influenza virus ribonucleoprotein complex (Coloma et al. 2009) as an example. Next, the integration of structural data with small-angle X-ray scattering (SAXS) data is shown on models of the extracellular region of an EGF receptor family member, s-dEGFR (Alvarado et al. 2009). Finally, a flexible fitting approach is shown using coarse-grained resolution models of myosin. Personal comments and annotations by the author are provided in the electronic supplementary material.
Correlation-based docking in electron microscopy
Visualization and modeling of small-angle scattering data
Another specific problem in the interpretation of SAXS data is the visualization of the beads. We found it useful to render not the densely packed beads themselves, but rather an envelope that can be created by isocontouring a volumetric map that was created by convolution with a soft kernel such as a Gaussian (using pdb2vol).
An atomic model of F-actin (Holmes et al. 2003) was fitted to the 14 Å resolution actomyosin map (data kindly provided by Rasmus R. Schröder, now at University of Heidelberg, during his visit to Houston in 2003). The F-actin structure allowed us to create a mask for a single myosin S1 unit by low-pass filtering from the docked atomic structure using pdb2vol. As described by Wriggers and Chacón (2001b), the mask was needed by the tools voledit and voldiff to segment and subtract densities from actin and neighboring symmetry-related subunits and to obtain the density of a single myosin S1 from the helical 3D map. This single myosin S1 map was then compared to the atomic structure.
The longitudinal distance constraints in the MCN were assigned manually, as described by Wriggers et al. (2004), by following the connectivity of the polypeptide chain and to ensure robustness of the control points during the shape change. We found by trial and error that motion capture was best achieved through allocating more flexibility to the 50K regions (effectively allowing cleft closure) by eliminating all constraints on the motion of control points in this region. The final network used for the automated flexing is shown in Fig. 6.
We performed the flexing by adding a constraint energy function to the Hamiltonian of a molecular dynamics simulation that penalizes global shape differences between the data sets (Wriggers et al. 2004). In the molecular dynamics run, we added water molecules predicted by DOWSER (Zhang and Hermans 1996) to the system, which resulted in a total system size of 12,008 atoms.
One can expect that at 14 Å resolution the flexing faithfully reproduces conformational differences with a precision of 2 Å if atomic structures are locally conserved (Wriggers et al. 2004). Side chains are rearranged automatically to accommodate global conformational changes. Otherwise, the algorithm leaves the initial structure intact at the local level. Whether this assumption holds depends on the nature of the conformational difference between the two isoforms, which is not known a priori. However, it has been shown that only about 7% of protein domain rearrangements documented in the PDB are irregular motions where the tertiary structure is significantly perturbed (Gerstein and Krebs 1998). Therefore, it is plausible, at least for the predominantly hinge-type domain motions exhibited by myosin, that the low-resolution flexible fitting approach visualizes conformational changes with a precision of single amino acid residues. The final flexing-induced rms deviation in the atomic model was 5.3 Å.
To validate the precision and probe for systematic errors, we also performed a control flexing calculation on the structure of myosin 5 (Coureux et al. 2003). Myosin 5 is deemed to be in closer agreement with the 3D map of S1 in actomyosin (Holmes et al. 2004). Following the above protocols, we created a model of myosin 5, resulting in a total system size of 11,150 atoms. The observed flexing-induced rms deviation in the atomic model was 3.8 Å, which was indeed much lower than that observed in the myosin 2 case.
The above tests validate the Situs-based flexible fitting approach with a real EM data set. More detailed and systematic tests of flexible fitting were published in Rusu et al. (2008). In addition, the myosin fitting was recently extended to full thick filaments of tarantula muscle in collaboration with the group of Raúl Padrón in Venezuela (Alamo et al. 2008).
One key to the success of Situs over the years has been that the programs were ported to multiple platforms and their source code was freely available on the Internet (http://situs.biomachina.org). While we strive to teach at workshops and symposia, it seems that many researchers prefer to explore software in their own laboratories.
Our web-based tutorials have helped hundreds of electron microscopists and small- angle scattering experts to learn the use of the programs. For their dissemination, we obtained our own web domain (http://biomachina.org) and web server.
Another helpful aspect was the modular design of the programs. As mentioned above, the software consists of multiple, stand-alone tools that can be combined in various creative ways. The modular design allowed us to update individual programs over time (inevitably, it becomes necessary to update algorithms and to implement bug fixes for problems reported to us). We are managing an e-mail list to communicate with the more than 2,000 registered users who opted to receive information. Readers should feel free to send comments to firstname.lastname@example.org.
This brief review primarily focused on the scientific use of the software, but the development of Situs was also a personal journey for the author, with many memorable encounters along the way. A personal history of this work, as well as annotated references, can be found in the electronic supplementary material.
I would like to thank Jaime Martín-Benito Romero, Diego Alvarado, and Mark Lemmon for providing Figs. 2 and 4, and for critical reading of this manuscript. In addition, I thank Rasmus R. Schröder and Kenneth C. Holmes for discussions. This work was supported in part by grants from National Institutes of Health (R01GM62968), the Alfred P. Sloan Foundation (BR-4297), and the Human Frontier Science Program (RGP0026/2003).
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.