Fig 1.
Separases from all species share a similar topology with highly conserved regions, which includes a caspase-like domain in their C-terminal region.
(A) Consolidated secondary structure prediction of separase from C. elegans using both PsiPred and JPred predicts a largely helical N-terminus (dark grey) with an unstructured region around residue 400 and a region of three β-strands from residues 720 to 750 (light grey). The conserved C-terminal half harbours the caspase-like domain (black), residues 900 to 1140. The catalytic dyad is indicated as white lines. (B) Multiple sequence alignment of the caspase-like domain of separase from C. elegans in comparison with human caspase 3, gingipain-R and MALT-1. Alignment was manually adjusted using Jalview to match secondary structure elements from predictions (PsiPred) of separases to structural elements as observed in caspase 3 (3EDQ), MALT-1 (3UO8) and gingipain R (1CVR). α-helices are shown in dark grey and β-strands in light grey. The conserved catalytic dyad (C, H) is shown in bold letters.
Fig 2.
Comparison of the structures of human caspase 3, human MALT-1 and gingipain-R.
(A) Overlaid structures of human caspase 3 (cyan, PDB code 3EDQ), human MALT-1 (green, PDB code 3UO8) and gingipain-R (magenta, PDB code 1CVR) show a similar overall fold consisting of a central six-stranded β -sheet flanked by α-helices. In gingipain-R, a helix connects the first four with the last two β-strands and replaces the inter-subunit linker present in caspases. In caspases the first four β -strands belong to the larger subunit p20 and the remaining two β-strands to the smaller subunit p10. The C-terminal Ig domain of MALT-1, subdomain A and the C-terminal IgSF domain of gingipain-R were omitted for clarity. (B) Comparison of the substrate binding pockets reveal a very similar binding mode of peptide inhibitors ace-LDESD-cho (for caspase 3), z-VRPR-fmk (MALT-1) and FFR-cmk (gingipain-R). Peptide inhibitors (magenta) and catalytic residues His and Cys (green) are shown as sticks. Residues mainly responsible for the recognition of the respective peptide P1 are shown as cyan sticks: Arg64 (for caspase 3), Asp163 (for gingipain-R) and Asp365 (for MALT-1). Figures were prepared with PyMol.
Fig 3.
Homology modelling of the caspase-like domain of C. elegans separase using caspase 3 as template.
(A) The homology model of the caspase-like domain of separase from C. elegans is shown in cartoon view with the proposed substrate peptide MEVER as sticks (magenta). The additional helices α2’ and α3’ that were introduced into the caspase 3 template (cyan) as well as the short helix modelled into the loop between β1 and α1 (red) are highlighted. Core model and loops were generated using MODELLER and energy minimized using GROMACS. (B) Surface electrostatics of the caspase-like domain in separases (shown here without substrate peptide) shows several positive (blue) and negative (red) patches that may be important for interactions with modulators such as securin or other domains within separase itself. Left: front view, same as view in (A). Right: view from back of molecule via vertical rotation by 180°. (C) Proposed interaction of the separase active site with substrate peptide MEVER (in sticks, magenta). Residues interacting with the peptide are labelled and shown in orange. Hydrogen bonds formed between peptide and protein are shown as black dashes. The catalytic residues H1014 and C1040 are indicated as green sticks. (D) Electrostatic properties of the substrate pocket reveal a deep negative pocket that receives the P1-Arg residue and a large positive patch that locks the side chains of P2-Glu and P4-Glu in place. All figures were prepared using PyMol.
Table 1.
Putative and confirmed separase substrates and their cleavage sites for different organisms.
Fig 4.
Interactions made between the predicted substrate peptide MEVER and the homology model of the C. elegans separase caspase domain.
(A) Residues in separases that interact with the substrate peptide are highlighted in an alignment. P1-Arg is anchored via hydrogen bonds to Glu917 (~) and Asp1082 towards the N-terminus of helix α4 (+). The main chain amide-NH of residue P1 interacts with main chain oxygen atom of Thr1075 (#). The side chain of P2-Glu forms hydrogen bonds to Arg1120 (which is located in a well-conserved region preceding β6) and Arg1044 (which is located in a loop region between helix α3’ and β4) which are both highlighted with * in the alignment. Aside from hydrophobic interactions, P3-Val interacts with both the main chain and side chain oxygen atoms of Thr1077 (^). P4-Glu forms hydrogen bonds to the main chain amide of Lys1118 and guanidino group of Arg1116 (–). α-helices are shown in dark grey and β-strands in light grey. The catalytic dyad (C, H) is shown in bold letters. (B) Schematic representation of interactions between C. elegans separase and the proposed substrate peptide EVER. Corresponding interacting residues in separases from other species are taken from Fig 4A. Core recognition sites of Scc1 proteins are shown in the boxed inset.
Fig 5.
The central region of separases may be similar to death domains and harbours a conserved WWxxRxxLD motif.
(A) The region N-terminal to the catalytically active caspase-like domain (black) is made up of six α-helices (grey) and may be structurally similar to death domains. A novel WWxxRxxLD-motif was found in the second helix of this domain whose function remains to be elucidated. Helices are numbered and indicated as grey bars, and their boundaries in C. elegans separase annotated. The region encompassing three β-strands is shown in light grey. Catalytic residues are marked with white bars. (B) Three-dimensional model of the proposed death domain in separase from C. elegans using the prodomain of human procaspase-9 as template. The six-helix bundle is shown in cartoon view with amino acids belonging to the proposed the WR motif shown as sticks (orange). A surface-exposed cysteine, C866 is indicated in magenta. Figure prepared with PyMol. (C) Surface representation and electrostatics of the proposed CARD domain show a large electropositive patch where the WR motif is located. Left: front view, same as view in (B), Right: view from back of molecule via vertical rotation by 180°. Figure prepared with PyMol. (D) Sequence alignment of the novel WR motif shows their high conservation within the central region of separase proteins. Sequences from mammals (Homo sapiens, Mus musculus), Caenorhabditis elegans, insects (Spodoptera frugiperda, Drosophila melanogaster, Drosophila virilis, Drosophila willistoni), fungi (Schizosaccharomyces pombe, Saccharomyces cerevisiae, Exophiala dermatitidis, Blumeria graminis), microsporidia (Encephalitozoon hellem, Encephalitozoon intestinalis, Encephalitozoon romaleae), protozoa (Giardia lamblia, Giardia intestinalis, Trypanosoma brucei, Trypanosoma cruzi, Leishmania major, Perkinsus marinus, Cryptosporidium muris, Cryptosporidium hominis, Cryptosporidium parvum), plants (Arabidopsis thaliana, Medicago truncatula, Ricinus communis) green algae (Ostreococcus tauri, Chlamydomonas reinhardtii, Volvox carteri) and the diatom Phaeodactylum tricornutum were aligned. Highlighted residues have 80% or more sequence identity (white letters on black background), 60–80% sequence identity (grey), or 40–60% (light grey). ‘Conservation’ indicates the degree of conservation of physico-chemical properties in each column of the alignment and is represented by numbers from 0 to 10. (E) Weblogo representation of the second predicted helix of the proposed CARD domain. The overall height of the stack indicates the sequence conservation at that position, while the height of symbols within the stack indicates the relative frequency of each amino acid at that position.