WayBackMachine Last Archived Version: December 21, 2005

Note: updates are marked in red (May 5, 2009)


Communicating Information about Virus Structure and Biology Via the World Wide Web

Stephan M. Spencer, Research Assistant (1)
Max L. Nibert, Assistant Professor (2)
Jean-Yves Sgro, Assistant Scientist

Institute for Molecular Virology
University of Wisconsin-Madison 53706

Update:
(1) www.stephanspencer.com
(2) Harvard Medical School

Abstract

In modern biology, there is an increasing emphasis on the relevance of the structure of a protein or other macromolecule to its biological functions. Similarly, the structure of a virus particle holds much information about its biology, and can be applied to such tasks as the interpretation of genetic data or the design of antiviral drugs. Recent technological advances in virus structure determination have allowed the structures of many viruses to be solved by x-ray crystallography and/or cryo-electron microscopy combined with image reconstruction. To facilitate the extraction of biological information from these structures, they should be made widely available and easily accessible in a meaningful format to specialists and non-specialists alike. In addition, more information can be extracted from such structures when they are animated or displayed interactively. Development of a resource that meets all of the above criteria has been made feasible with the advent of Mosaic and the World Wide Web Initiative.

The World Wide Web Server for Virology at the Institute for Molecular Virology, University of Wisconsin-Madison has been designed to disseminate virology-related information to scientists, students, and the general public. Resources include computer-generated images and animations of virus structures, topographical maps of virus surfaces, digitized electron micrographs, the ICTV Taxonomy of Viruses 5th Report with a search interface, an explanation of hardware and software requirements for visualizing virus structures, and tutorials on some selected topics in general virology.

The server is promoted not only as a reference resource, but also as a forum for teaching. Virologists are encouraged to announce themselves as being Internet-accessible, and to ask or answer questions in the bionet.virology USENET newsgroup. Virology instructors can submit their course syllabi to share with instructors with varying backgrounds in virology. Instructors can even customize the server for use by their own students. A fill-in form interface allows an instructor to create a "hypersyllabus", a course syllabus embedded with hyperlinks to documents on our server or elsewhere that serve as supplemental exercises.

Introduction

A virus is a submicroscopic parasite that must infect a host cell in order to replicate, i.e. make copies of itself. The genetic information - the viral genome - is encoded by nucleic acid, either DNA or RNA. The genome is enclosed in one or more layers of protein and, if the virus is enveloped, lipid as well. In many cases, the protein layers are highly symmetrical and are composed of many copies of a few viral proteins, arranged in shells or "capsids," in which repetitious protein-protein interactions are found. By employing the strategy of symmetry, viruses can reduce the number of different viral proteins required to form the capsid, and thus, the size of the viral genome. Virus symmetry comes in two general forms: icosahedral (twenty-sided) or helical.

Each virus encodes its own collection of viral proteins, acting in concert, each with specific roles that enable or enhance viral replication. The function of each viral protein is inherent in its tertiary structure (three-dimensional conformation). Some viral proteins function as components of the virus capsid. Others act as enzyme catalysts of chemical reactions that are essential to viral replication, such as RNA synthesis or proteolysis. Some viral proteins may actually do both, participating in forming the capsid and acting as an enzyme catalyst. Both processes are important targets for computer-based antiviral drug design: either affecting the capsid structure or inhibiting a key catalytic reaction can substantially decrease viral infectivity. Thus, studying the virus structure can reveal clues to its many biological functions.

The Virology Server as a Research Tool

The World Wide Web Server for Virology is an important tool for conveying structural information to virologists throughout the world. Until recently, the only structural data to which most virologists had access (except for researchers who actually collected the data and their collaborators) was in the form of static images, usually from journal articles. Now, with the advent of high speed data transfers, increasing global connectivity, and the easy-to-use Mosaic browser program, it is simple for researchers to obtain structure data from a variety of sources and view animated or interactive visualizations of these data on their own computers. Ease of access to rendered structure data is expected to catalyze speculation and collaboration among virologists. Most importantly, it facilitates the task of relating structural details to biological function, one of the primary goals of virological investigations. Virologists are encouraged to submit additional virus visualizations to the Virology Server on an ongoing basis.

Forms of Virus Structure Data

Several types of visual data are presented on the server, each containing biological information at different levels of complexity. Some types of data need not be processed (although processing may reveal additional structural details); others require processing and rendering.

Unprocessed Data Types

Digitized electron micrographs contain the least amount of detail. Electron micrographs are two-dimensional, showing viral particles spread out on a support surface in various orientations and positions relative to an electron beam (see Figure 1) [1]. No computer manipulation or averaging of images need be involved. The noise level is high and image quality is low due to staining and drying artifacts.

FIGURE 1: Digitized negative stain electron micrograph of brome mosaic virus, print magnification at 2600x [2].

Data Types Requiring Processing and Rendering

Cryo-electron microscopy in combination with image reconstruction can yield a three-dimensional structure that includes both surface and internal features and contains a large degree of structural detail (currently at resolutions as low as approximately 20 Angstroms) (see Figure 2). Image reconstruction is a computationally demanding process based on the assumption that each virus image from the raw data - the two-dimensional electron micrographs - is a two-dimensional projection of the same three-dimensional object [3]. The reconstruction process takes advantage of the icosahedral symmetry of viruses in assigning an orientation to each projection. The projections are subsequently reconstructed into a three-dimensional array of variable electron density values. The array is then rendered using the visualization techniques described below. Because of the assumption of icosahedral symmetry, icosahedrally symmetrical structures are reinforced and any non-icosahedrally symmetrical (e.g. flexible) structures present on or within the particle are averaged out. The gain in overall detail is marked, despite a lack of detail in non-icosahedrally symmetrical structures.

Virus structures solved by X-ray crystallography are three-dimensional and the resolution is very high (often 2 to 8 Angstroms) (see Figure 3). In this technique, viruses are crystallized and the crystals are subjected to an X-ray beam. Some of the X-rays in the beam collide with atoms of one of the many virus particles in the crystal and thus diffract. The regular arrangement of the crystalline lattice in which the virus particles are embedded causes some diffractions to be reinforced. Using involved mathematical procedures, the diffraction pattern (often captured on X-ray film) is interpreted and the virus particle's atomic coordinates (i.e. positions) are eventually determined [4]. This technique also utilizes icosahedral symmetry averaging, thus non-icosahedrally symmetrical structures are averaged out and lost, such as most or all of the nucleic acid content. X-ray crystallography is the only technique which currently provides enough detail to distinguish the individual atoms of a virus particle. A major drawback is that solving a virus structure by X-ray crystallography is time-consuming, often taking years, compared to weeks for cryo-electron microscopy combined with image reconstruction.

Rendering of the Virus Structure Data

Three-dimensional cryo-electron microscopy datasets are often visualized by surface rendering techniques, most commonly isosurfaces and cutting planes. Isosurfaces are used to portray the three-dimensional structure of the virus, whereas cutting planes portray a two-dimensional cross-section through the three-dimensional dataset (Generally, if the dataset is in Cartesian coordinates, slices are planar, and if in spherical coordinates, radial.) (see Figure 2) [5]. An isosurface is obtained by converting the subset of electron density values in the three-dimensional array that contours a chosen threshold value (isovalue) into a surface representation approximated by polygons. Cutting planes display a two-dimensional slice of the virus dataset, with data points represented by pixels that can be colored according to electron density value. Cutting planes display a greater amount of electron density data, i.e. a larger range of electron density values, but cannot portray the surface topography of a virus as an isosurface can.

FIGURE 2: Mammalian reovirus core as solved by cryo-electron microscopy and image reconstruction [6]. Panel A) Rendered as an isosurface using Iris Explorer [7] on a Silicon Graphics workstation. Panel B) An orthogonal slice through Cartesian space, rendered using Iris Explorer on a Silicon Graphics workstation.

Three-dimensional X-ray crystallographic coordinates for the entire virus particle are usually represented either as a molecular surface, a wireframe atomic model, or a space-filling atomic model (see Figure 3). Atoms in a space-filling model are represented as spheres with van der Waals radii. In a wireframe model, chemical bonds are represented as lines that connect adjacent atoms. A molecular surface is the surface containing the contact points between a water-sized spherical probe and the van der Waals atomic surfaces of a protein or oligomer of proteins (e.g. the capsid of a virus particle) [8]. For visualizing large structures such as a virus, a molecular surface representation is often preferable, as it gives a contiguous surface that more easily lends itself to the visual interpretation of the surface topography. Unfortunately, calculation of molecular surfaces is CPU- and memory-intensive and requires sophisticated programs like Grasp [9] for the Silicon Graphics workstation. Even visualization of the atomic surfaces of all component atoms of the virus can be too calculation-intensive to be highly interactive. When computing power is limited, the load on the CPU can be further alleviated by eliminating from the plot all atoms on the amino acid side chains of the viral proteins, leaving only the alpha-carbon backbone (also known as a C-alpha trace) (see Figure 3F). This type of representation has the additional benefit of visual simplification.

FIGURE 3: Rhinovirus 14 (a common cold virus) as solved by X-ray crystallography [10]. Panel A) Atoms are rendered as spheres with van der Waals radii using the program Grasp [9]. Panel B) Each alpha-carbon atom is rendered as a 4 Angstrom sphere using srf [11]. Panel C) Alpha-carbon backbone rendered as wireframe using MIDAS [12]. Panel D) Rendered as a molecular surface with radial depth cueing (surface topography colored according to distance from the center of the virus) [13] using Grasp. In Panels A, B and C, VP1 is in blue, VP2 is in green and VP3 is in red (VP4 is inside and not visible). All images were rendered on a Silicon Graphics workstation.

A reduction in computational requirements and image simplification can also be achieved by excluding from the visualizaton all of the structural redundancy due to symmetry, or more specifically, by displaying one icosahedral asymmetric unit. The icosahedral asymmetric unit is the smallest non-repetitive unit of an icosahedral virus particle and represents one-sixtieth of the entire virus particle. The icosahedral asymmetric unit can be displayed as a "topographical map," which depicts the virus surface as a landscape akin to a geographical terrain with mountains and valleys (see Figure 4).

FIGURE 4: Structure of rhinovirus 14. Panel A) Schematic diagram showing isosahedral asymmetric unit in relation to the complete virus particle. Panel B) Topographical map of rhinovirus 14 (as solved by X-ray crystallography), with radial depth cueing, rendered using Grasp on a Silicon Graphics workstation.

The structure of an individual viral protein (versus the whole virus particle) can be displayed. The atomic coordinates can be obtained by isolating the individual viral protein and subjecting it to X-ray crystallography, or by isolating the atomic coordinates of one protein from the coordinates of a complete virus particle. Protein structures can be displayed in a number of representations, including molecular surface, space filling, wireframe, ball-and-stick, or ribbon diagram (see Figure 5). The space-filling representation plots the positions of atoms; wireframe and ball-and-stick representations identify chemical bonds as well as atom positions. The protein's secondary structure (i.e. local structural elements such as alpha-helix or beta-sheet [14]) can be examined with a plot of the alpha-carbon backbone, or preferably, with a ribbon diagram [15].

FIGURE 5: VP1 (viral protein 1) of rhinovirus 14, one of the four polypeptides that form the virus capsid (entire particle solved by X-ray crystallography [9]). Panel A) rendered as a molecular surface using Grasp. Panel B) rendered as space-filling using MIDAS. Panel C) rendered as wireframe using SYBYL [16]. Panel D) rendered as ball-and-stick using SYBYL. Panel E) rendered as a ribbon diagram using Ribbons [17]. Panel F) alpha-carbon backbone, rendered as wireframe using SYBYL. All images were rendered on a Silicon Graphics workstation.

A researcher connecting to the Virology Server can not only view previously rendered images of virus crystal structures, but can also create his own images interactively using the Molecules R Us [18] service at the National Institutes of Health. Atomic coordinates stored in the Protein Data Bank (PDB) [19] can be either rendered on the NIH server by Raster3D [20] or downloaded to the local computer with the concurrent launching of a helper visualization program (which is possible thanks to the newly established "chemical" MIME type [21]). Hyperlinks to the Molecules R Us fill-in form page are integrated within the Virology Server when the PDB coordinates of viral particles or proteins are referenced within our documents.

At this time, interactive viewing of virus particles with Molecules R Us is limited to an icosahedral asymmetric unit, since removing the redundancy due to symmetry greatly reduces the PDB file size. The complete virus particle can be reconstructed by multiplying the coordinates sixty times by rotation matrices. The resulting atomic coordinates can be saved as a new PDB file. Reconstructed PDB files have been created for a number of solved virus structures and are available on our server in a compressed format, allowing local rendering of the complete virus structure. (However, the visualization program may need to be launched manually, as visualizing a compressed PDB file from within Mosaic requires the spawning of two applications consecutively.)

Presentation of the Virus Structure Data

Presentation of structural data comes in two forms: static images and animations. Static images are used to display all types of virus structure data. Animations (both MPEG [22] and Quicktime [23]) are used extensively, as they enhance the perception of depth and surface topography. Animation is only appropriate for displaying three-dimensional datasets, i.e. those obtained from X-ray crystallography or cryo-electron microscopy combined with image reconstruction.

The Virology Server as a Teaching Resource

Students, as researchers in training, can benefit greatly from the visualization of animated virus structures. They can have hands-on experience interpreting structural details of viruses. In addition, via hyperlinks to the Molecules R Us service at NIH, students can interactively probe the structures of individual viral proteins.

Although the visualization of virus structures is the key component of the Virology Server, other significant teaching resources are present on the server to aid in the instruction of students. The server maintains some tutorials on selected topics in general virology, a database of Internet-accessible virologists with search script and fill-in submission form, and reference to the bionet.virology USENET newsgroup. Virologists are encouraged to submit images or animations of virus structures, tutorials, course exams, and other cutting-edge materials to our FTP server (rhino.bocklabs.wisc.edu) by anonymous FTP for addition to the World Wide Web server (www.bocklabs.wisc.edu).

The server provides a place for virologists to share course materials, and for inexperienced virology instructors to take advantage of these contributed resources. An archive of virology course syllabi is maintained which will likely be useful for virologists to compare notes and get new ideas, and thus may help to standardize formats for virology courses. The syllabus archive may be even more useful for non-virologists who must teach some virology at the high school or college level and can discover what practicing virologists elsewhere are currently teaching. The syllabus archive and other collaborative efforts among virology instructors via the Internet are still in their infancy; it is hoped that in the future one will see the submission of course notes, figures, and other lecture material.

An additional archive is maintained to facilitate interactivity in learning, an archive which takes advantage of the World Wide Web's hypertext format. Instructors not only can contribute to the content of the server, but also can customize the server for their own students' use, through the use of hypersyllabi. A hypersyllabus refers to a course syllabus with lecture topics and corresponding supplementary exercises, the exercises being in the form of hyperlinks to documents on the Virology Server or elsewhere on the Internet. Hypersyllabi are created online with a fill-in form interface (see Figure 6) and can later be edited online as well. Instructors can choose to maintain their hypersyllabi locally (either on a local server or as a local file on disk), or on the Virology Server. In the latter case, the lectures are made available not only to the instructors' classes, but to anyone with access to the World Wide Web.

FIGURE 6: Creating a hypersyllabus involves multiple fill-in forms. Panel A) Fill-in form page for entering header information for the hypersyllabus. Panel B) Each lecture topic with up to four supplementary exercises is entered one at a time with an auxiliary fill-in form.

Summary

An array of resources for virologists and students is now available from the Institute for Molecular Virology at the University of Wisconsin-Madison, via the World Wide Web. The availability of new resources relating to the visualization of virus structures will likely have a great impact on the study of viruses, facilitating the interpretation of structural details and the exchange of ideas, thus helping shape the development of the field of virology. In addition, the teaching resources should aid in enhancing virology instruction.

Acknowledgements

We gratefully acknowledge Timothy S. Baker and colleagues at the Structural Biology Group at Purdue University for supplying us with the three-dimensional cryo-electron microscopy datasets which were used to generate Figure 2.

The World Wide Web Server for Virology at the Institute for Molecular Virology, University of Wisconsin-Madison, is supported by a grant from the Lucille P. Markey Charitable Trust.

References

1 Finch, J.T., and A. Klug. 1965. The Structure of viruses of the papilloma-polyoma type. J. Mol. Biol. 13:1-12.

2 Krol, M.A., and P.G. Ahlquist. 1994. unpublished results.

3 Martino, R.L., C.A. Johnson, E.B. Suh, B.L. Trus, and T.K. Yap. 1994. Parallel computing in biomedical research. Science 265:902-908.

4 Ladd, M.F.C., and R.A. Palmer. 1993. Structure determination by X-ray crystallography, 3rd ed. New York: Plenum Press.

5 Kaufman, A. 1990. 3D Volume Visualization. Eurographics Tutorial Notes 12.

6 Dryden, K.A., G. Wang, M. Yeager, M.L. Nibert, K.M. Coombs, D.B. Furlong, B.N. Fields, and T.S. Baker. 1993. Early steps in reovirus infection are associated with dramatic changes in supramolecular structure and protein conformation: analysis of virions and subviral particles by cryoelectron microscopy and image reconstruction. J. Cell Biol. 122:1023-1041.

7 Iris Explorer. Silicon Graphics, Inc., Mountain View, CA

8 Richards, F.M. 1977. Areas, volumes, packing, and protein structure. Ann. Rev. Biophys. Bioeng. 6:151-176.

9 Nicholls, A., K.A. Sharp, and B. Honig. 1991. Protein Folding and Association: Insights From the Interfacial and Thermodynamic Properties of Hydrocarbons. PROTEINS: Structure, Function, and Genetics 11:281-296.

10 Rossman, M.G., E. Arnold, J.W. Erickson, E.A. Frankenberger, J.P. Griffith, H.-J. Hecht, J.E. Johnson, G. Kamer, M. Luo, A.G. Mosser, R.R. Rueckert, B. Sherry, and G. Vriend. 1985. Structure of a human common cold virus and functional relationship to other picornaviruses. Nature 317:145-153. (PDB entry # 4RHV)

11 Connolly, M.L. 1993. The molecular surface package. J. Mol. Graphics 11:139-141

12 Ferrin, T.E., C.C. Huang, L.E. Jarvis, and R. Langridge. 1988. The MIDAS display system. J. Mol. Graphics 6:13-27.

13 Grant, R.A., S. Cranic, and J.M. Hogle. 1992. Radial depth provides the cue. Current Biology 2:86-87.

14 Chothia, C., M. Levitt, D. Richardson. 1977. Structure of proteins: packing of alpha-helices and pleated sheets. Proc. Natl. Acad. Sci. USA 74:4130-4134.

15 Richardson, J.S. 1985. Schematic drawings of protein structures. Methods Enzymol. 115:359-380.

16 SYBYL Molecular Modeling Software. TRIPOS Associates, Inc., St. Louis, MO.

17 Carson, M. Ribbon Models of Macromolecules. 1987. J. Mol. Graphics 5:103-106.

18 FitzGerald, P.C. Molecules R Us: A WWW forms interface to facilitate access (browsing, searching and viewing) of the molecular structure data contained within the Brookhaven Protein Data Bank (PDB). Chemistry Workshop, First International Conference on the World Wide Web.

19 Bernstein, F.C. 1977. The protein data bank: a computer archive. J. Mol. Biol. 112:535-542.

20 Bacon, D.J., and W.F. Anderson. 1988. A fast algorithm for rendering space-filling molecule pictures. J. Mol. Graphics 6:219-220.

21 Rzepa, H.S., B.J. Whitaker and M.J. Winter. 1994. Chemical Application of the World-Wide-Web J. Chem. Soc, Chem. Commun. 1907

22 Moving Pictures Experts Group, International Standards Organization. IEC-11172.

23 QuickTime. Apple Computer, Inc. Cupertino, CA.

Authors' Biographies

Stephan M. Spencer is a graduate student in the Department of Biochemistry at the University of Wisconsin-Madison, working in the laboratory of Dr. Max L. Nibert. He obtained his Bachelors of Science Degree with Honors in Cellular and Molecular Biology at the University of Michigan-Ann Arbor in 1993. His research focus is on structure-function relationships in mammalian reoviruses, and he is currently developing novel methods for the visualization of virus datasets solved by cryo-electron microscopy and image reconstruction. In addition to his research pursuits, he teaches a non-credit evening course on the Internet at the University of Wisconsin-Madison.
Stephan M. Spencer, M.Sc., is President and CEO of natural search marketing firm Netconcepts founded by Stephan in 1995. More at: http://www.netconcepts.com/who-we-are/executive-team/stephan-spencer/

Max L. Nibert is an Assistant Professor at the University of Wisconsin-Madison, with joint appointments in the Institute for Molecular Virology and the Department of Biochemistry. He received his Ph.D. in Microbiology and Molecular Genetics from Harvard University, where he studied various aspects of the natural history of reoviruses. He also received his M.D. from Harvard Medical School and completed a residency in Clinical Pathology at the Brigham and Womens Hospital in Boston, MA. Work in his new lab in Madison is focused on understanding how the structure of virus particles and proteins relates to the mechanisms of virus entry into cells.

Max Nibert is Faculty at the Department of Microbiology and Molecular Genetics at the Harvard Medical School
Current research interests.

Jean-Yves Sgro obtained his Ph.D. in Cellular and Molecular Biology from the University of Grenoble, France, where he studied RNA-proteins interactions in ssRNA viruses. In 1986 he moved to the Institute for Molecular Virology in Madison, Wisconsin, USA, where he sequenced the coat proteins of nodaviruses. He has developed techniques to visualize large virus structures from Protein Data Bank coordinates. Such visualizations provide insight into virus-cell interactions and are widely used in current scientific literature. They are also used to explain virus structures to non-scientists. Many of his images and computer movies are available on the Web.

Senior Scientist with joint appoitments at the Gene Expression Center, Biotechnology Center and the Institute for Molecular Virology.

Email address of contact author

sspencer@rhino.bocklabs.wisc.edu sspencer@netconcepts.com