Comparative structural analysis of the caspase family with other clan CD cysteine peptidases

Abstract

Clan CD forms a structural group of cysteine peptidases, containing seven individual families and two subfamilies of structurally related enzymes. Historically, it is most notable for containing the mammalian caspases, on which the structures of the clan were founded. Interestingly, the caspase family is split into two subfamilies: the caspases, and a second subfamily containing both the paracaspases and the metacaspases. Structural data are now available for both the paracaspases and the metacaspases, allowing a comprehensive structural analysis of the entire caspase family. In addition, a relative plethora of structural data has recently become available for many of the other families in the clan, allowing both the structures and the structure-function relationships of clan CD to be fully explored. The present review compares the enzymes in the caspase subfamilies with each other, together with a comprehensive comparison of all the structural families in clan CD. This reveals a diverse group of structures with highly conserved structural elements that provide the peptidases with a variety of substrate specificities and activation mechanisms. It also reveals conserved structural elements involved in substrate binding, and potential autoinhibitory functions, throughout the clan, and confirms that the metacaspases are structurally diverse from the caspases (and paracaspases), suggesting that they should form a distinct family of clan CD peptidases.

Figure 1. The topology and structure of the caspase dimer

Caspase-7 (PDB ID 1F1J, see Supplementary Table S1) is used to represent a typical caspase with all β-strands coloured blue and α-helices grey. Structural elements from the second monomer in the dimer are coloured paler than the first. (A) The topology and simplified nomenclature of the caspases. The central β-strands and major α-helices are named from the N-terminus (β1–β6 and α1–α5, respectively); the important loop regions (L) are named according to the strands that they follow (L1–L5); and the small sections of β-turns on L3 and L5 are named β¹–β³ and β¹–β², respectively. The position of the catalytic dyad histidine (H) and cysteine (C) is highlighted in red on L3 and L4, respectively, and the position of the conserved aspartate cleavage site (D) is shown on L4 (/). A vertical dashed line represents the dimer interface and the C-terminus of the second monomer is shown (′). The CHF SSEs are highlighted (*). (B) Ribbon diagram of a caspase dimer. Topdraw [90] and PyMOL (http://pymol.sourceforge.net, Schrodinger) were used for topology diagrams and molecular images throughout the present review, respectively.

Figure 2. The S₁-binding pockets of the clan CD family members

The catalytic dyad is shown in red and conserved aromatic residues are shown in green. With the exception of TbMCA2, residues that form hydrogen bonds to the P₁ residue of a bound inhibitor are shown in blue (the darker shade of blue represents interactions through functional groups, whereas the lighter blue shows interactions from main chain atoms). Residues and SSEs involved in P₁ binding are labelled and SSEs structurally homologous (but topologically diverse) to those found in the caspases are highlighted (^H). (A) Caspase-7 PDB (ID 1F1J). (B) Inhibitor-free TbMCA2 (PDB ID 4AFR) in which residues shown to be important in substrate binding are highlighted in blue, with those responsible for specificity in P₁ [11] shown in navy blue. (C) MALT1 paracaspase domain (MALT1-P) (PDB IB 3UOA). (D) Legumain (PDB ID 4AW9). (E) Gingipain R (PDB ID 1CVR). (F) MARTX-CPD (PDB ID 3GCD). Inhibitors used in complex structures are shown in Supplementary Table S1.

Figure 3. The structural topologies of the clan CD enzymes

(A) Caspase-7; (B) TbMCA2; (C) MALT1-P; (D); legumain; (E) gingipain R; and (F) MARTX-CPD. The S₁-binding pockets are highlighted as in Figure 2 and the topologies are based on the PDB codes described in the same Figure. Strands in the central β-sheet are numbered from the N-terminus in black. Black numbering is also used for the five major α-helices and important S₁-binding loops (L) when they are located in the structure in the same order as they are in the caspases. SSEs that are structurally homologous to those found in the caspases, but appear in the structure in a different order, are highlighted with an (^H), followed by the caspase numbering, and shown in purple (α and β have been omitted as a result of space constraints but are used in the text). The position of the catalytic dyad (H/C) is shown in red on loops L3 and L4 (or ^HL3 and ^HL4), respectively.

Figure 4. The structural diversity in the central β-sheet of the clan CD enzymes

(A) The β-sheet topologies exhibited by the families in the clan. The β-strands described by the CHF [23] are shown in grey whereas the other strands are shown in blue; the N- and C-terminal ends of the enzymes are labelled accordingly. (B) A phylogenetic tree based on structure, in which Q^H [85] is a measure of structural homology. This Figure was produced using a STAMP [87] structural alignment and VMD [87].