Crystal structure of tarocystatin-papain complex: implications for the inhibition property of group-2 phytocystatins

Abstract

Tarocystatin (CeCPI) from taro (Colocasia esculenta cv. Kaohsiung no. 1), a group-2 phytocystatin, shares a conserved N-terminal cystatin domain (NtD) with other phytocystatins but contains a C-terminal cystatin-like extension (CtE). The structure of the tarocystatin-papain complex and the domain interaction between NtD and CtE in tarocystatin have not been determined. We resolved the crystal structure of the phytocystatin-papain complex at resolution 2.03 Å. Surprisingly, the structure of the NtD-papain complex in a stoichiometry of 1:1 could be built, with no CtE observed. Only two remnant residues of CtE could be built in the structure of the CtE-papain complex. Therefore, CtE is easily digested by papain. To further characterize the interaction between NtD and CtE, three segments of tarocystatin, including the full-length (FL), NtD and CtE, were used to analyze the domain-domain interaction and the inhibition ability. The results from glutaraldehyde cross-linking and yeast two-hybrid assay indicated the existence of an intrinsic flexibility in the region linking NtD and CtE for most tarocystatin molecules. In the inhibition activity assay, the glutathione-S-transferase (GST)-fused FL showed the highest inhibition ability without residual peptidase activity, and GST-NtD and FL showed almost the same inhibition ability, which was higher than with NtD alone. On the basis of the structures, the linker flexibility and inhibition activity of tarocystatins, we propose that the overhangs from the cystatin domain may enhance the inhibition ability of the cystatin domain against papain.

Fig. 1

Three groups of phytocystatins. Phytocystatins can be divided into three groups on the basis of molecular mass: Group 1 is about 12–16 kDa and has one cystatin (CY) domain; group 2 is about 23 kDa and has one cystatin domain and an extended cystatin-like (CY-L) domain at the C terminus; group 3 is about 85 kDa and has several repetitive cystatin domains. The resolved cystatin domain structures are shaded in the gray box

Fig. 2

Crystal structures of tarocystatin–papain complex. a Stereo view of the structure of the tarocystatin–papain complex. The remaining NtD of tarocystatin is in lime green and papain is in warm pink. b The remnant residues of CtE are in lime green and papain is in warm pink. c Hydrogen bonding networks between tarocystatin and papain. Nine hydrogen bonds are highlighted by yellow dashed lines between NtD and papain (chain B to chain A): Met4 (N) to Gly66 (O), Met4 (O) to Gly66 (N), Met4 (SD) to His159 (N), Gly5 (N) to Asp158 (O), Gly5 (N) to OCS25 (OD2), Gln49 (NE2) to Cys63 (O), Val51(O) to Trp177 (NE1), Ser52 (N) to Gly20 (O), Ser52 (OG) to Asn18 (OD1). One water-mediated hydrogen bond is Trp80 (NE1) to Trp177 (O) and one salt bridge is Glu18 (OE2) to Lys139 (NZ). The residues from 10 to 15 of tarocystatin are not observed and are represented by black dots. d The 1σ of 2Fo-Fc electron density map for the oxidized sulfhydryl group of Cys25 of papain. e The dipeptide, Ser-Asn, of CtE is contoured by 1σ of the 2Fo-Fc electron density map

Fig. 3

Structure comparison and sequence alignment of NtD of tarocystatin with OC-1 and PMC-2. a NtD of tarocystatin (CeCPI-NtD), OC-1, and PMC-2 are represented by cyan, orange, and magenta, respectively. The superimposition of the Cα atoms of NtD of tarocystatin, OC-1 and PMC2 gives an average RMS difference of 1.35 and 0.87 Å, respectively. The structure numbers indicate the residue number of NtD, and the conserved motifs are labeled by Trunk, L1 and L2. b The sequence alignment of CeCPI-NtD, OC-1, and PMC-2 reveals the similarity of the cystatin domain, except the loop of residues 10–15 of NtD is a highly flexible region. The residues of NtD interacting with papain by hydrogen bonding are labeled by asterisks. The secondary structure located above the sequence alignment is extracted from the structure of NtD, and the unresolved region is represented by dashed lines

Fig. 4

Protease activity assay of papain. FL, NtD and CtE represent the full-length, N-terminal domain, and C-terminal extension of tarocystatin, respectively. The proteins from tarocystatin mixed with/without nonactivated papain were analyzed on 15% SDS-PAGE. Lane 1 FL only, lane 2 FL mixed with papain and remained about a band of 9 kDa, lane 3 NtD mixed with papain and remained a 9 kDa band, lane 4 CtE mixed with papain and no proteins left, lane 5 NtD only, lane 6 CtE only, lane 7 papain only, lane M standard protein marker

Fig. 5

Glutaraldehyde (GA) cross-linking analysis of NtD and CtE. The purified proteins FL, NtD and CtE were cross-linking by glutaraldehyde and resolved on 10% SDS-PAGE. The shifts of protein band are indicated by the arrow. Lane M standard protein marker. a NtD and CtE were cross-linked by >0.002% GA. b NtD and NtD were cross-linked by >0.002% GA. c CtE and CtE could not be cross-linked by GA. d FL and FL were cross-linked by >0.002% GA. e The interaction between GST-fused CtE (GCtE) and NtD was further confirmed. The lower band indicates the interaction of NtD–NtD, and the higher band indicates the interaction of GCtE–NtD. f The results confirm the interaction between NtD and NtD but not NtD and GST

Fig. 6

Domain–domain interaction by yeast two-hybrid assay. The bait represents the vector containing the GAL4 DNA binding domain, and the prey represents the vector containing the GAL4 activation domain. a The protein–protein interaction of NtD and NtD could be identified in the pGBT9-NtD and pACT-NtD constructs, whereas the constructs of pGBT9-NtD/pACT and pGBT9/pACT-NtD confirmed that the interaction of NtD and NtD was not caused by autoactivation. Lack of yeast grown with the construction of pGBT9-CtE and pACT-CtE showed that no interactions could be found between CtE and CtE. However, a few yeast colonies in pGBT9-NtD and pACT-CtE showed a weak interaction between NtD and CtE. b The constructs of NtD, CtE and FL were subcloned into pGBT9 and pACT vectors. An interaction could be observed in FL and FL, and NtD and FL. The constructs of pGBT9-FL/pACT and pGBT9/pACT-FL also confirmed that the interaction of FL was not caused by autoactivation. pGBKT7-p53/pGADT7-T and pGBT9/pACT served as the positive and negative control, respectively. Yeast was incubated in the SD/-Trp/-Leu/-His selection medium containing 5 mM 3-AT

Fig. 7

In-gel inhibitory activity assay with different segments from tarocystatin. The brightness of the band indicates the protease activity of papain to digest the substrate gelatin and the brighter band represents the higher residual activity of protease. All inhibitory activity assays entailed 3 pmol papain. Lane P the positive control indicating the papain activity only. a With increased concentration, GCtE shows the more digestive ability of papain from 64 pmol. CtE and GST show only a little elevated protease activity. b The inhibition ability of different combinations of tarocystatin. The inhibition ability was GFL > FL > GNtD > NtD. GCtE shows the highest enhanced capacity, whereas CtE and GST show little enhanced capacity. Lane P represents the recovered activity of papain of 100%. Percentages indicate the recovered activity of papain relative to lane P