Cn3D 2.50 is a viewing and analysis program that provides linked sequence and three-dimensional displays of the macromolecularl structures and VAST alignments of protein domains accessible through NCBI's MMDB (Molecular Modeling DataBase). Cn3D can be used in a standalone mode or configured to work as a helper application for Entrez where it facilitates rapid 3D visualization and comparison of macromolecular structures. Cn3D functions on PC's, Macs, and most UNIX platforms
Cn3D 2.50 Offers Several Enhancements over Previous Versions of Cn3D
Residue groups may
be defined as Features and assigned unique rendering and labeling schemes.
Installation is now a single-click operation for Netscape and Internet Explorer on PC platforms.
User-friendly control panels are provided for selecting residue rendering and visibility settings.
Residue selection in the Sequence and Structure Windows has been improved.
Support for viewing models based on sequence-structure alignments has been improved.
Cn3D 2.50 Displays Structural Alignments and the Implied Sequence Alignments in Linked Windows
Cn3D 2.50 can display a VAST structural alignment of multiple protein structures. Figure 1 shows both the structure window and linked Sequence Window used by Cn3D to display the VAST alignment of two 5'-3' exonucleases, Protein Data Bank (PDB) codes 1TFR (1) and 1EXN (2) , which are responsible for the removal of the RNA primers formed during the replication of DNA in bacteriophages T4 and T5, respectively. By default, the trace for 1TFR is colored green and that for 1EXN, blue with well-aligned, structurally conserved, portions of the traces colored red in both structures, so that regions of alignment are easily visible. This pre-defined color scheme is referred to as coloring by Neighbor and may be selected using the Color menu option. The proteins are depicted using alpha-carbon traces while the two magnesium ions of 1TFR are depicted as magenta-colored spheres. Three dimensional objects, cylinders and planks, are used to delineate alpha helices and beta strands respectively.
The alignment of Figure 1 may be created by searching
for 1TFR, the PDB code for the bacteriophage T4 RnaseH, on the MMDB
Home Page. Following the link to the Structural
Neighbors of 1TFR leads to the VAST Neighbors report featuring
a table listing the closest structural neighbors of 1TFR. The closest structural
match to 1TFR is seen at the top of the table and is 1EXN, a 5'-3' exonuclease
from bacteriophage T5. Checking the box next to 1EXN and adjusting the
radio buttons at the top of the form to select Launch
Viewer, Use Cn3D,
Chains only, and All
Atoms, followed by a click on View/Save
Alignments, invokes Cn3D with a display of the 1TFR/1EXN
Figure 1. Structure
and linked Sequence Windows showing a VAST alignment of two 5'-3' exonucleases,
colored by Neighbor.
The rendering shown in Figure 1 can be created using
one of four tabbed panels located in a Viewer
Controls window which
is new in Cn3D 2.50. The panel found under the Viewer
tab is shown in Table 1 and the settings indicated
correspond to those used to produce the image of Figure
1. The Viewer
Controls window is reached via the Cn3D
Controls option under View as
indicated below. For the rendering of Figure 1,
the default Structure Window background color of black has also
been changed to ivory using the Options/Color
Settings/Background menu item, also shown below.
The sequence window in Figure 1, compliments the Structure Window and represents the sequence implied by the VAST structural alignment. This sequence alignment window is linked to the three-dimensional Structure Window and shares its color scheme. Amino acid residues that are structurally aligned are shown in capital red lettering. Selecting a residue or group of residues in either the Structure Window or the Sequence Window highlights the appropriate residues in both, as can be seen in the Figure. It is therefore a simple matter to highlight corresponding residues in the two windows. Two regions of interest in the aligned exonuclease structures are highlighted in yellow. These two regions are discussed latter in the context of their biology in order to illustrate the sort of information that may be obtained easily using Cn3D.
Rotating the structural alignment using the mouse and zooming in from the top, as in Figure 2A, it is possible to see that the alignment of these two exonucleases extends to the level of four catalytic residues in the vicinity of one of the magnesium atoms of 1TFR. To generate this view, the alpha carbon traces have been turned off, and two groups of amino acids have been defined as Features and independently rendered as "fat tubes", revealing the excellent alignment of the corresponding side chains in the two structures. Features may be defined and given independent rendering and visibility settings using the Viewer Controls window of Cn3D 2.50. To define the feature "t4_cat", for instance, 1TFR residues D19, D71, D132, and D155 were selected in a discontinuous manner in the Sequence Window depressing the control key while using the mouse. The Viewer Controls Feature panel shown in Table 4 was then used to give the selected residues a Feature name and define their display and visibility properties.
In Figure 2B, the magnification has been increased,
the three-dimensional object representations have been turned off, and
the background color has been set to grey using the Options/Color
Settings/Background menu. The Viewer
panel has been used to label residues in "t4_cat" with three-letter amino
acid codes and to label the residues in t5_cat with one-letter codes .
The alignment shown, the rendering settings, and user-defined Features,
"t4_cat" and "t5_cat", can be saved in a single file for latter retrieval
using the File/Save/All
Figure 2A. Magnified view from the "top" of the aligned strucures.
Figure 2B. Magnified
view restricted to the two magnesium ions and two user-defined Features,
"t4_cat" and "t5_cat."
Cn3D 2.50 is a Versatile Alignment Analysis Tool
The alignment of Figure 1 highlights a small "tail" in 1TFR comprised of two alpha helices, which is absent in 1EXN. This "tail" is the leftmost highlighted area in the Structure Window and the bottommost highlighted area in the Sequence Window. One might hypothesize, on the basis of its location outside the region of core alignment, that this domain serves a function unique to 1TFR and not vital to the exonucleolytic activity which 1TFR shares with 1EXN. The C-terminal portion might represent an area of interface between 1TFR and other T4 proteins that facilitates cooperation between them.
Other evidence is consistent with this hypothesis. The C-terminal "tail" of 1TFR is homologous to a small domain which falls between the 5'-3' exonuclease domain and polymerase domain of the DNA polymerase of Thermus aquaticus (PDB code 1TAQ) (3). Hence, the homologous domain is positioned to serve as an interface between exonuclease and polymerase functionalities in the Thermus aquaticus enzyme. The interfacial location of the homologous domain in the Thermus polymerase may parallel the role of the C-terminal domain of 1TFR in the T4 system.
The second highlighted region of the alignment demonstrates an instance in which one crystal structure complements a second by providing a model for regions which are disordered in the second structure. This region is the uppermost highlighted region in the Sequence Windows of Figure 1. These highlighted residues form a helical arch in the 1EXN structure which passes over the central cleft of the enzyme core. It has been suggested that single-stranded DNA passes through this arch as part of the enzyme's catalytic mechanism (2). Most of the amino acids comprising the analogous region of 1TFR are disordered and, therefore, invisible in the 1TFR structure. From the alignment, however, it may be hypothesized that these disordered amino acids also form an arch in 1TFR that is similar to the arch present in 1EXN. The sequence display of Figure 1 reveals that both the C-terminal and N-terminal foundations of the helical arch of 1EXN are aligned with homologous residues of 1TFR. At the N-terminus of the arch, the 1EXN sequence "YKGNR" aligns with "YKKNR" in 1TFR. At the C-terminus of the arch the 1EXN sequence "FFE" is similar to the 1TFR sequence "YFE". The sequence alignment of Figure 1 indicates that there are four fewer amino acids in 1TFR available to form the arch than are used by 1EXN. A model for this disordered region of 1TFR may therefore be imagined in which the cleft is spanned by a helical arch akin to that of 1EXN but with shorter helical segments. It remains to be seen whether this proves to be the case.
Cn3D Displays MMDB-Derived Protein Domain Assignments
Another way to analyze the 1TFR/1EXN alignment is demonstrated in Figure
3. Here the two exonuclease structures have been colored by Domain.
Several other pre-defined color schemes are also available under the Color
menu as indicated below.
The domains highlighted are those defined, on the basis of geometrical compactness, in the MMDB structure file which Cn3D reads. From the figure it is apparent that both 1TFR and 1EXN are comprised of two large domains. An "extra domain", seen in the 1TFR structure and shown uncolored, is the C-terminal portion discussed earlier. The two helices comprising this portion of 1TFR are left uncolored (white in the Structure Window) to indicate that they are not assigned to a domain in the MMDB source file. Of the two large domains, the larger, shown in blue and cyan, is very similar in the two structures and aligns well. The smaller domain does not align well between the two structures. The division of these enzymes into similar and dissimilar domains may be helpful in disecting structure-function relationships.
Figure 3. Vast alignment of two 5'-3' exonucleases, colored by domain.
Cn3D Highlights Domains Formed from Discontinuous Sequences
An interesting feature is revealed in theSequence
Window of Figure 3. The Sequence Window reveals that the
major, highly similar domain (blue and cyan) is interrupted by the dissimilar
domain (red and green). This fact suggests the possibility that the
dissimilar domains which, from the Sequence Window, are seen to
be of similar lengths in the two structures, might converge upon similar
structures upon the binding of the nucleotide substrate.
Cn3D Can Map a Sequence onto an Appropriate Model Structure
Cn3D can also be used to map a subject protein sequence onto a structural template if sufficient sequence similarity between the sequences of the template and subject. An example is given in Figure 4 below.
The figure shows a Cn3D rendering of the structure of the soluble domain of oxidized cytochrome B5 from Bos taurus (4). Within the Structure Window, the protein is rendered as a Fat Tubes model, whereas the prosthetic heme group is rendered as a Space Fill model. The sequence window contains two cytochrome B5 sequences. The first sequence is the master sequence and corresponds to the protein structure visible in the structure window. The second sequence is that of cytochrome B5 of the distant eukaryote, Saccharomyces cerevisiae (5). Although the structure of the yeast protein is unknown, its sequence has been imported into the sequence window of Cn3D and automatically aligned with the Bos taurus sequence. Because the BLAST-produced alignment is good, it is possible for Cn3D to "map" the yeast sequence onto the Bos taurus structure. This has been depicted by coloring the structure by Neighbor to produce a solid dark tubular structure and then choosing Show Substitutions from the Alignment menu in the Sequence Window. The resulting rendering in the Structure Window is green where the two sequences are identical and yellow where they differ. This highlighting is reflected in the Sequence Window by the light shading of nonconserved residues in the 1CYO sequence. In this manner, the sequence differences between the two cytochrome B5 proteins have been mapped onto the structure of the Bos taurus protein.
The mapping reveals that a conspicuous run of conserved residues in
the sequence alignment, "EHPGG" in the two sequences, forms a loop in the
structure of the Bos taurus cytochrome. This loop contains the leftmost
of two conserved histidines, which are seen "kissing" the central iron
of the heme prosthetic group from opposite sides. These essential
histidines, His-39 and His-63, have been defined as Features,
colored red, and labeled using three letter amino acid codes using the Viewer
Controls Feature panel.
Note that these histidines are also colored red in the Sequence Window.
Figure 4. Close-up of
the heme-binding pocket of bovine cytochrome B (PDB code 1CYO) showing
the interaction of conserved residues with the heme (space-filled structure) group. The alignment of
the 1CYO sequence with a homologous sequence from S. cerevisiae is shown below.
The ability of Cn3D to map a protein sequence onto an appropriate template
structure will form the basis of a powerful new way to link sequences to
structures within the Entrez Genomes database. Genomic protein tables
are under development at NCBI which will provide links between individual
protein sequences within a genome and any known structures of proteins
with similar sequences. A click of the mouse will invoke Cn3D with
the genomic protein sequence of interest already mapped onto a template
protein structure. Hence, the loading of a template/sequence pair
from a genomic protein table will become automatic.
The following is a summary of the capabilities of Cn3D 2.50.
Data Input and Saving: The Save/File option on the MMDB web pages can be used to save the sequence and structure data as a local file. The same option on the VAST web pages will save the sequence, structure, and alignment. These files, as well as local ASN.1 biostruc files, can be loaded into Cn3D with the File/Open/Biostruc Local command. The data files saved from the web can also be loaded in Cn3D as an argument on the command line. The Export option from the File menu indied Cn3D will export PDB and Kinemage files for the master structure; GIF files can be exporte here as well. An alignment, complete with rendering settings and Feature definitions may be saved locally in an ASN.1 format for latter retrieval using the File/Save/All menu option. Saved alignments may be loaded using the File/Open/Local File menu option.
Feature Definition: Features, consisting of arbritrary groups of amino acid residues, may be defined and given unique rendering and visibility settings using the Render panel shown in Table 4.
Export Options: The master structure, that structure to which others are aligned (1TFR in the example above), may be exported as either a PDB format file or a Kinemage file. The image appearing in the Structure Window may be exported as a GIF file or copied to the clipboard.
Rendering: Structures may be rendered in standard modes such as the "cylinder and plank" secondary structural representation, the wire frame, spacefill, ball and stick, or alpha-carbon trace representations.
Color Schemes: Molecule coloring schemes include coloring by secondary structure, domain, molecule, residue, hydrophobicity, CPK color, and temperature factor.
Residue Labeling:Residues may be labeled at several predefined intervals with either one letter or two letter codes. Chain termini may also be labeled.
Alignment Display Modes: Structural alignments may be displayed in three modes. In the first, aligned residues are painted red while the others retain the color of their parent molecule as in Figure 1. Alternatively, the display may be limited to either non-aligned residues or aligned residues.
Animation: Animation using the View/Animation feature is a powerful way to visualize structure comparisons. Core regions of alignment appear stationary while non-aligned regions appear to "move" when an alignment is animated. Animation by quickly cycling through multiple structures, as in the case of models consisting of many NMR structures, is also supported.
Zooming and Translating: The displayed structures may be translated within the Structure Window by holding down the shift key and moving the mouse. Zoom levels are controlled through the use of a "zoom box" created by moving the mouse while depressing the control key.
Rotation Modes: Molecular rotations may be performed about each of the three Cartesian axes using the mouse. To change the rotation mode, use the Option/Mouse Settings menu option.
Residue Selection: In the Structure Window, double clicking on any atom in a residue highlights and selects the residue; a second double-click unhighlights and deselects the residue. In the Sequence Window, a double-click selects a residue, while a control-drag allows discontinuous selection of multiple residues. The view in the Structure Window may be restricted to selected residues using a check-box in the Viewer Controls Show/Hide panel. Both amino acid variations and substitutions between sequences may be highlighted using a pull-down menu in the Sequence Window. Selections in one window are mirrored in the other.
Highlight Color Selection: The color used to highlight residues may be changed using the Option/Color Settings/Highlight/User Defined menu option.
Help: An online manual for Cn3D is available at http://www.ncbi.nlm.nih.gov/Structure/cn3dhelp.html
Cn3D is Available for a Variety of Operating Systems
Cn3D is available in versions compiled for Windows, Mac-OS, and Linux.
Binaries are also available for the following UNIX systems: SunOS 4, Solaris, SGI Irix5.3 and higher, Linux, and DEC Alpha OSF1.
Cn3D 2.50 may be obtained from http://www.ncbi.nlm.nih.gov/Structure/cn3d.htm
Table 2. The Cn3D Label panel.
Table 3. The Cn3D Show/Hide panel.
Table 4. The Cn3D Feature panel. Shown are the rendering setting used for the four catalytic acidic residues comprising the "t4_cat" group of Figures 2A and 2B.
1. Structure of bacteriophage T4 RNase H, a 5' to 3' RNA-DNA and DNA-DNA exonuclease with sequence similarity to the RAD2 family of eukaryotic proteins. Mueser TC, Nossal NG, Hyde CC. Cell 1996 Jun 28;85(7):1101-1112.
2. A helical arch allowing single-stranded DNA to thread through T5 5'-exonuclease. Ceska TA, Sayers JR, Stier G, Suck D. Nature 1996 Jul 4;382(6586):90-93.
3. Crystal structure of Thermus aquaticus DNA polymerase. Kim Y, Eom SH, Wang J, Lee DS, Suh SW, Steitz TA. Nature 1995 Aug 17;376(6541):612-616.
4. Mathews, FS, P Argos, and M Levine. The structure of cytochrome b 5 at 2.0 Angstrom resolution. Cold Spring Harb Symp Quant Biol 36:387-95, 1972.
5. Truan, G, JC Epinat, C Rougeulle, C Cullin,
and D Pompon. Cloning and characterization of a yeast cytochrome b5-encoding
gene which suppresses ketoconazole hypersensitivity in a NADPH-P-450 reductase-deficient
strain. Gene 149:123-7, 1994.