2.3 Methods
2.3.3 Building a structure-guided sequence alignment (Alternately, use 2.3
1. Network Load your target structure in Cn3D by selecting the Network Load option from theFile menu and entering either the PDB or MMDB identifier. This will render the protein structure in the open Structure Viewer window and open the corresponding sequence in the
Alignment Viewerwindow. See Note: 9.
2. Adjust the appearance of your target structure by selecting Style → Rendering short- cuts→Tubes, which aids in comparing structural alignment, andStyle→Coloring shortcuts
→ Sequence Conservation→ Fit, which colours the sequence and structure according to the position specific scoring matrix (PSSM) of the sequence alignment. See Note: 10.
3. Open the Import Viewer window by selectingImports→ Show Imports. Arrange your windows so that the Structure, Alignment, and Import windows are all visible. Import new structures into the Import Viewer window by selectingEdit→ Import Structure and thenVia
Networkto enter the PDB or MMDB identifier, orFrom a Filefollowed by a .pdb file (Figure 3.1). See Note: 11.
Figure 2.1: Screenshot of the three windows comprising the Cn3D workspace showing an analysis of the LAGLIDADG Homing Endonuclease family. The Structure Viewer window (top left) shows a protein structure alignment of the protein family. The Sequence Viewer window (bottom) shows the sequence alignment which corresponds to the structure alignment with additional sequences from the protein family. The Import Viewer window (rop right) shows sequences which have have been imported into Cn3D, but have not been added to the sequence alignment.
4. Attempt to Merge All the new structures from the Import Viewer into the Alignment Viewer by selecting Alignments →Merge All. There may be conflicts when your new struc- tures are imported (residues highlighted in pink in the Import Viewer); conflicts will prevent structures from being added to theAlignment Windowvia theMerge Allcommand.
blocks located above the alignments in theAlignment Viewer. View the blocks in theAlignment Viewerby selectingEdit→Enable Editor. Each block defines a structurally conserved segment of the alignment which cannot accept insertions or deletions. Adjust the each block in the
Alignment Viewer so that the new structures in the Import Viewer no longer conflict; Split, Merge, Create, Delete (under theEdit menu), and Horizontal Drag (under theMouse Mode
menu) the blocks to resolve all pink conflicts in theImport Viewerwindow. See Note: 12. 6. Transfer the structures from the Import Viewer into the Alignment Viewer by selecting
Alignments→Merge All in theImport Viewerwindow. Save your work to a file by selecting
Savefrom theFilemenu in theStructure Viewerwindow. After saving, re-open the file to view the imported structures. Cn3D will not render the newly imported structures until the structures are merged into the Alignment Viewer and the file is saved and re-opened.
7. Select File → Realign Structures to superimpose the newly merged structures in the Structure Viewer as defined by the sequence relationships in theAlignment Viewer. Repeat this step whenever a change is made to the blocks or any structure in the Alignment Window, to update the structure alignment. See Note: 13.
8. Now that all structures are visible, revise the blocks to reflect the structural diversity of the protein family. Each block defines a conserved and critical structural feature that cannot ac- cept insertions or deletions. Adjust the blocks in the Alignment Viewer so that they correspond to the structurally conserved core of the protein. Create gaps between the blocks in regions of structural divergence where insertions or deletions would be tolerated. As in step 5, use the options in theEditandMouse Modemenus to make changes. See Note: 14.
9. Structure alignment algorithms are vulnerable to shift errors, where major structural elements are superimposed, but individual residues are not (Figure 3.2); for example, two β- sheets may be superimposed, but in a configuration where aligned residues sidechains point in opposite directions. Search the alignment for shift errors by highlighting aligned residues (Mouse Mode → Select Column in the Alignment Viewer) and examining the corresponding structural alignment in theStructure Viewer; highlighted residues are highlighted in yellow in
both windows. Select alternate residues in theStructure Viewer by double clicking. Use both sequence homology and structural evidence to revise the structure alignment. Be thorough, as errors at this step will be propagated through the rest of the alignment creation process.
Figure 2.2: An example of a shift error added to a segment of a LAGLIDADG alignment. A vertical column in the sequence alignment was selected which highlights the corresponding residues in yellow in the Structure Viewer. One of the structures helices is out of alignment by 1 residue turn. This should be corrected by shifting the corresponding sequence in the sequence alignment.
10. Import the sequences collected by PSI-BLAST by selectingEdit→Import Sequences
fasta file. These imported sequences will behave as the structures did in this window.
11. Align the new sequences to the target structure by selectingAlgorithms→Block Align N; this procedure aligns sequences in theImport Viewerto the PSSM of the sequence alignment in theAlignment Viewer. After the block align procedure is complete, merge the sequences that fit with the existing block model. See Note: 15.
12. Sort the sequences by selecting Edit→ Sort Rows→ By Score and thenEdit→ Sort Rows→Float PDBsin the Alignment Viewer window. Inspect the alignment in theAlignment Viewer. The Fit colouring scheme will colour residues red that fit will with the PSSM for that position and blue for residues that fit poorly. Each sequence in the Alignment Viewer
contributes to the PSSM, so poorly aligned sequences should be moved back to the Import Viewerby selectingImports→Realign Rows from List.
13. In the Import Viewer perform aBlock Align N to realign the sequences to to the ex- panded PSSM defined by theAlignment Viewer. Once again Merge Allthe sequences into the
Alignment Viewer.
14. Repeat steps 12 and 13 iteratively until the alignment in theAlignment Viewercontains as many homologous sequences that will fit with both the block model and the sequence fit. See Note: 16.