Structural Insight into Integrin Recognition and Anticancer Activity of Echistatin

Echistatin (Ech) is a short disintegrin with a long 42NPHKGPAT C-terminal tail. We determined the 3-D structure of Ech by X-ray crystallography. Superimposition of the structures of chains A and B showed conformational differences in their RGD loops and C-termini. The chain A structure is consistent with our NMR analysis that the GPAT residues of the C-terminus cannot be observed due to high flexibility. The hydrogen bond patterns of the RGD loop and between the RGD loop and C-terminus in Ech were the same as those of the corresponding residues in medium disintegrins. The mutant with C-terminal HKGPAT truncation caused 6.4-, 7.0-, 11.7-, and 18.6-fold decreases in inhibiting integrins αvβ3, αIIbβ3, αvβ5, and α5β1. Mutagenesis of the C-terminus showed that the H44A mutant caused 2.5- and 4.4-fold increases in inhibiting αIIbβ3 and α5β1, and the K45A mutant caused a 2.6-fold decrease in inhibiting αIIbβ3. We found that Ech inhibited VEGF-induced HUVEC proliferation with an IC50 value of 103.2 nM and inhibited the migration of A375, U373MG, and Panc-1 tumor cells with IC50 values of 1.5, 5.7, and 154.5 nM. These findings suggest that Ech is a potential anticancer agent, and its C-terminal region can be optimized to improve its anticancer activity.


Introduction
Echistatin (Ech) is a short disintegrin isolated from Echis carinatus that contains 49 residues with 4 disulfide bonds. Ech is the most extensively characterized disintegrin and has a conserved RGD motif and a long C-terminal NPHKGPAT tail [1][2][3][4][5][6]. Several studies support a functional role for the C-terminal residues of Ech in modulating its binding affinity and selectivity towards integrin αIIbβ3 [6][7][8]. For example, a synthetic C-terminal peptide of Ech inhibited the binding of integrin αIIbβ3 to Ech, activated the binding of integrin to immobilized ligands, induced the expression of activation-dependent conformational epitopes of integrin, and increased the binding of fibrinogen to peptide-treated platelets [6,7,9,10]. An RGD peptide fused with the C-terminal KGPAT residues of Ech selectively binds αvβ3 integrin but not αvβ5 integrin [11]. Replacement of contortrostatin's

The Overall Structure of Ech
The diffraction data of Ech were obtained to a resolution of 1.80 Å from a single frozen crystal, which contained two molecules per asymmetric unit. Figure 1A-D show a representative electron-density fit including the RGD loop from residues K21 to D30 and the C-terminus from residues P40 to T49. Ramachandran plot analysis of the final structure, obtained from the program MolProbity [34], showed 95.7% of the residues in the favored region, 4.3% in the additionally allowed region, and no residues in the outlier region ( Table 1). The data collection and structure refinement statistics are summarized in Table 1. The overall structure of chains A and B contained four disulfide bonds (C2-C11, C7-C32, C8-C37, and C20-C39) and one pair of antiparallel β-strands (residues 16-20 and 29-31) (Figure 2A,B). (B) Electron density map of the RGD loop with residues K21 to D30 (chain B); (C) Electron density map of the C-terminus with residues P40 to K45 (chain A); (D) Electron density map of the C-terminus with residues P40 to T49 (chain B).   Superimposition of the backbones of chains A and B with an RMSD value of 0.226 Å demonstrated that their overall structures were very similar, except for the RGD loop and C-terminal loop ( Figure 2C-E). Superimposition of the RGD loop (residues K21-D30) and C-terminus (residues P40-K45) resulted in RMSD values of 0.162 Å and 0.581 Å, respectively ( Figure 2D,E). Structural analysis showed that the RGD loop of chain B was squeezed by residues K21, R41, and K45 in contact with sulfate ions and residue S4 of chain A through crystal packing ( Figure S1). Therefore, the R and D side chains of the chain B structure interacted with each other and formed hydrogen bonds ( Figure 2D). In contrast to chain B, the R and D side chains in the RGD loop of chain A are far from each other ( Figure 2D). Furthermore, the backbones of the C-terminal GPAT residues of chain A showed a weaker electron density map ( Figure 1C). This was consistent with our unpublished NMR data and NMR nuclear Overhauser effect (NOE) analysis, which showed that the GPAT residues of the C-terminus exhibited low or negative NOE values ( Figure S2). The average NOE value of Ech is 0.605. In addition, residues R24-D27 in the RGD loop, E1-E3 in the N-terminus, and K45-T49 in the C-terminus had NOE values less than the average NOE value of Ech, suggesting that these regions of Ech were flexible. In contrast, the conformation of the C-terminal tail of Ech chain B was well defined ( Figure 1D), possibly because the C-terminal residue K45 can interact with the O atoms of SO 4 2-( Figure S1, Table S1). Moreover, the C-terminal residue T49 can interact with water molecules ( Figure S3). Although the tertiary fold of Ech was similar to the reported NMR structure (2Ech), superimposition of the backbones of Ech chain A and the NMR structure (2Ech) resulted in RMSD values of 1.688 Å ( Figure 2F,G) [13]. Nevertheless, the analyses of their structures showed that the RGD motif is located at the apex of the integrin-binding loop, which is structurally close to the C-terminus. These findings were consistent with the reported 3-D structures of medium disintegrins [14][15][16][17][18]. The resulting high-resolution crystal structure can provide an improved molecular basis to design anticancer agents using Ech as a scaffold.

Structural Comparison of the RGD Loop and the Interactions between the RGD Loop and C-Termini
Hydrogen bond connectivity and atom distances of the RGD loop were compared among Ech chains A and B and trimestatin, a medium disintegrin with an ARGDFP motif, to identify their structural differences in detail ( Figure 3). We found that the Cα to Cα distance between residues R and D in chain A, chain B, and trimestatin of the RGD motif was 6.5, 5.8, and 6.7 Å, and the Cζ to Cγ distance between residues R and D in chain A, chain B, and trimestatin was 12.4, 4.8, and 11.1 Å ( Figure 3A-C, Table 2). These results showed that Ech chain A and trimestatin exhibited similarity in the conformation of the RGD motif and in the sidechain orientation of the R and D residues, which were pointed away from each other. Moreover, comparison of the Cα(Ri)-Cα(Di+2), Cβ(Ri)-Cβ(Di+2), Cζ(Ri)-Cγ(Di+2), and Cα(Ri)-Cα(Xi+3) distances (Å) of Ech chain A and Ech's NMR structure (2Ech) revealed that the RGD motif in our X-ray structure (Ech chain A) was similar to that in the solution NMR structure (2Ech) ( Table 2). Interestingly, we found a conserved and extensively connected hydrogen bond network in the RGD loop. The residues located on both sides of the RGD motif in Ech formed hydrogen bonds, including K21 with D30, A23 with D29, and R24 with D27 ( Figure 3D,E, Tables S2 and S3). The corresponding residues of trimestatin also formed hydrogen bonds, including R46 with D55 and A48 with D54 ( Figure 3F, Table S4). These results showed that short and medium disintegrins have similar RGD motif and loop conformations, suggesting that these conserved interactions may be important to integrin recognition because they maintain the conformation of the RGD loop.   Similar to the conformation of the RGD loop, the interactions between the RGD loop and C-terminus were conserved in short and medium disintegrins ( Figure 4A-C, Tables S2-S4). The structural analysis of Ech chains A and B showed that the side chain of the D30 residue in the RGD loop and the amide proton of the N42 residue in the C-terminus formed hydrogen bonds ( Figure 4A,B). The corresponding residues of trimestatin also formed hydrogen bonds between the side chain of residue D55 in the RGD loop and the amide proton of residue W67 ( Figure 4C). In contrast to Ech chain A and trimestatin, Ech chain B formed additional hydrogen bonds, including residue M28 with residue K45 and residue D30 with residue K45 ( Figure 4B). These results implied that the longer C-terminal tail, which forms the loop-like structure, may play an important role in the interaction with integrins. The hydrogen bond between the side chain of the D30 residue in the RGD loop and the amide proton of the N42 residue in the C-terminus is shown; (B) Hydrogen bonds between the RGD loop and C-terminus (chain B). Three hydrogen bonds between the carboxyl group of residue M28 and the sidechain Nζ proton of residue K45, between the side chain of residue D30 and the amide proton of residue N42, and between the side chain of residue D30 and the sidechain Nζ proton of residue K45; (C) Hydrogen bond between the RGD loop and C-terminus in trimestatin. The hydrogen bond between the side chain of the D55 residue in the RGD loop and the amide proton of the W67 residue in the C-terminus is shown; (D) Interaction in Ech C-terminus (chain A). The interaction between the side chain of the N42 residue and the mainchain amide proton of residues P43 and H44; (E) Interaction in Ech C-terminus (chain B). The interaction between the side chain of the N42 residue and the mainchain amide proton of residues P43 and H44; (F) Hydrogen bond in the C-terminus of trimestatin. The hydrogen bond between the side chain of residue W67 and the backbone carboxyl group of residue N68.

Structural Comparison of the C-Terminal Regions
In addition to the similar conformation of the RGD loop and the interactions between the RGD loop and C-terminus, we also found that short and medium disintegrins shared a similar interaction in their C-termini ( Figure 4D-F, Tables S2-S4). The N42 residue can interact with the mainchain amide proton of the P43 residue in the Ech chain A ( Figure 4D) and chain B ( Figure 4E). The corresponding residues of trimestatin also formed hydrogen bonds between the sidechain amide proton of residue W67 and the carboxyl group of residue N68 ( Figure 4F). Thus, we know that these conserved regions of the C-terminus are essential for its structure. Furthermore, in contrast to trimestatin, the extended C-terminal tail residue H44 can interact with N42 in Ech chain A ( Figure 4D) and chain B ( Figure 4E). These results showed that the conformation of the C-terminal regions of Ech may be important in integrin recognition.

Structural Differences between the Integrin Complexes of Ech Chain A and Chain B
The docking of Ech chain A and chain B into integrin αvβ3 was used to identify the integrin-interacting residues. The analysis showed that the docking of Ech chain A and chain B into integrin αvβ3 resulted in different binding modes, especially in the C-terminus. (Figure 5A,B). The docking of Ech chain A into integrin αvβ3 showed that the side chain of the R22 residue of Ech chain A may interact with the main chain of N313 in the β3 subunit and that the main chain of the G25 residue of Ech chain A can also interact with the side chain of D218 in the αv subunit (Table 3 and Table S5). In contrast, Ech chain B did not show these kinds of interactions but instead interaction between the side chain of M28 and the main chain of D126 in the β3 subunit (Table 4 and  Table S6). On the other hand, the main chain of C-terminal residue H44 of Ech chain A can interact with Y122 of the β3 subunit, and this contact was not found in Ech chain B. However, the side chain of C-terminal residue K45 of Ech chain B can interact with the side chain of M180 of the β3 subunit. These results showed that the C-terminal tails of Ech chains A and B were involved in Ech and β3 subunit interactions. These results suggested that Ech with a longer C-terminal tail may increase its recognition site and binding affinity to αvβ3 integrin.

The C-Terminal Residues of Ech are Involved in Mediating Integrin Recognition
The lengths of the C-terminus and the K45 residue of Ech were found to be important for interacting with integrin αIIbβ3 [13]. In this study, the inhibition of the binding of cell-expressed integrins αvβ3, αIIbβ3, αvβ5, and α5β1 to their ligands by Ech and its mutants was used to determine their activity and selectivity (Tables 5 and 6). Compared to wild-type Ech, the mutant with the truncation of the C-terminal HKGPAT tail caused 6.4-, 7.0-, 11.7-, and 18.6-fold decreases in the inhibition of integrins αvβ3, αIIbβ3, αvβ5, and α5β1, respectively ( Table 5). The effect of the C-terminal region on the change in the relative binding affinity to integrins was α5β1 > αvβ5 > αIIbβ3 > αvβ3. These results indicated that the C-terminal tail of Ech played an important role in recognizing not only integrin αIIbβ3 but also integrins αvβ3, αvβ5, and α5β1 [7,10,13]. Alanine scanning mutagenesis of the C-terminal tail of Ech showed that wild-type, P43A, H44A, and K45A mutants exhibited similar activity in inhibiting the adhesion of CHO cells that expressed integrin αvβ3 to immobilized fibrinogen with IC 50 values of 20.7, 13.8, 16.4, and 14.7 nM (Table 6). In contrast to integrin αvβ3, some differences were found in their activities in inhibiting integrins αIIbβ3 and α5β1. They inhibited the adhesion of CHO cells that expressed integrin αIIbβ3 to immobilized fibrinogen with IC 50 values of 51.5, 67.7, 20.7, and 132.5 nM and inhibited K562 cell adhesion to immobilized fibronectin with IC 50 values of 132.6, 58.5. 30.0, and 84.9 nM. Interestingly, the H44A mutant caused 2.5-and 4.4-fold increases in the inhibition of αIIbβ3 and α5β1, and the K45A mutant caused a 2.6-fold decrease in the inhibition of integrin αIIbβ3. These results suggest that residues H44 and K45 may play different roles in integrin recognition. Indeed, in our docking structure of integrin αvβ3 with Ech chain A, we found that the main chain of C-terminal residue H44 in Ech chain A formed a hydrogen bond with the side chain of residue Y122 in the β3 subunit ( Figure 5A, Table 3 and Table S5). Moreover, in the docking structure of integrin αvβ3 with Ech chain B, we found that the sidechain of C-terminal residue K45 in Ech chain B formed a hydrogen bond with the side chain of residue M180 in the β3 subunit ( Figure 5B, Table 4 and Table S6). Our findings suggest that a longer C-terminal tail may improve the recognition site of Ech and increase its binding affinity to αvβ3 integrin, endowing Ech with inhibitory activity against αvβ3 integrin-related cancer. In the future, we can apply this property in Ech-based drug design for cancer therapy.

Inhibition of VEGF-Induced HUVEC Proliferation by Ech
Angiogenesis is a critical process for cancer cell growth and metastasis [35]. Endothelial cell proliferation is the key step in angiogenesis progression. Many studies have shown that disintegrins are potent angiogenic inhibitors [25][26][27][28][29][30][31][32]. The results of our analysis showed that Ech could inhibit HUVEC proliferation with an IC 50 value of 103.2 nM ( Figure 6, Table 7). In comparison with the reported disintegrin with IC 50 values of 100-400 nM [17,[30][31][32], Ech exhibited similar activity in inhibiting HUVEC proliferation (Table 7). These results suggest that Ech could be a good scaffold to design potent antiangiogenic agents.  Many studies have shown that disintegrins inhibit the adhesion and migration of tumor cells [24][25][26]29]. In our previous study, we also demonstrated that recombinant Ech effectively inhibited human A375 melanoma with an IC 50 value of 1.5 nM (Table 8) [13]. To study the inhibitory effect of Ech on the migration of other tumor cells, we examined its antimigration activities in human glioblastoma U373MG and pancreatic Panc-1 cells. Similar to the results of melanoma cell inhibition by Ech, it inhibited the migration of U373MG and Panc-1 tumor cells with IC 50 values of 5.7 and 154.5 nM, respectively (Figures 7 and 8, Table 8). We found that the K45E mutant exhibited 8.9-, 4.8-, and 10.4-fold decreases in inhibiting the migration of A375, U373MG, and Panc-1 tumor cells (Figures 7 and 8, Table 8), showing the importance of the K45 residue in inhibiting tumor cell migration. According to our X-ray structure, the conformation of Ech's C-terminus was stabilized by the carboxyl group of residue M28 and the side-chain Nζ proton of residue K45 and between the side chain of residue D30 and the side-chain Nζ proton of residue K45 ( Figure 4B). Furthermore, our docking structure of integrin αvβ3 with Ech chain B also showed that the side chain of C-terminal residue K45 in Ech chain B can interact with the side chain of residue M180 in the β3 subunit ( Figure 5B, Table 4 and Table S6). These findings suggested that the conformation of the C-terminus plays an important role in αvβ3 integrin recognition, resulting in the inhibition of tumor cell migration.

Discussion
RGD-containing disintegrins have potent antiangiogenic and antimetastatic activities [23][24][25][26][27][28][29][30][31][32]. To elucidate the structure-function relationships of Ech, we determined the crystal structure of Ech and its anti-integrin, antiangiogenic, and antimigration activities. We found that the conformation of the RGD loop and the interactions between the RGD loop and C-terminus were similar to those in the reported medium disintegrin trimestatin [15]. In contrast, significant differences between short and medium disintegrins were found in their C-terminal conformations. We also showed that the C-terminus of Ech plays important roles in integrin recognition and the inhibition of tumor cell migration. This structural and functional study of Ech can serve as a basis for insight into its structure-function relationships and for the design of anticancer drugs.
Ech is the most studied disintegrin, and Aggrastat (Tirofilban), an antiplatelet drug, was designed based on its RGD loop structure [33]. It is known that the RGD loop, linker, and C-terminal regions of disintegrins are involved in integrin recognition [18,25,36]. The incorporation of C-terminal structural information into current drugs may improve its potency and selectivity [37]. Previous studies also showed that the C-terminal residues of Ech possess a functional role in modulating the binding affinity to integrin αIIbβ3 [6][7][8][9][10]13]. The analysis of Ech's crystal structure showed that the sidechain of the D30 residue in the RGD loop and the amide proton of the N42 residue in the C-terminus formed hydrogen bonds. This is consistent with the report that the C-terminal tail of Ech acts in synergy with the RGD loop to interact with integrins [6,7,10,13]. According to our X-ray crystal structure and NMR NOE analysis, the residues from G46 to T49 of the C-terminus are flexible and cannot be observed in the X-ray structure. Disordered regions of proteins frequently have important biological functions. Interestingly, the carboxyl group of residue M28 interacts with the side-chain Nζ proton of residue K45, and the side chain of residue D30 interacts with the side chain Nζ proton of residue K45, resulting in a unique turn structure for Ech. These results showed that the K45 residue stabilizes the conformation of Ech's C-terminus and is important for its integrin recognition. The unique conformation of Ech's C-terminus can be used to design integrin-specific drugs.
Alanine scanning mutagenesis of the C-terminal tail of Ech showed that the H44A mutant caused 2.5-and 4.4-fold increases in the inhibition of αIIbβ3 and α5β1, and the K45A mutant caused a 2.6-fold decrease in the inhibition of integrin αIIbβ3. These results suggested that residues H44 and K45 may play different roles in integrin recognition. Moreover, we found a correlation between our docking results and these functional studies. In Ech chain A, the main chain of C-terminal residue H44 formed a hydrogen bond with the side chain of residue Y122 in the β3 subunit ( Figure 5A, Table 3 and Table S5). On the other hand, in Ech chain B, the sidechain of C-terminal residue K45 interacts with the sidechain of residue M180 in the β3 subunit ( Figure 5B, Table 4 and Table S6). The different interaction modes between H44/K45 and αvβ3 integrin may provide information on how H44/K45 binds to other integrins.
Integrins are highly expressed in many cancers, including melanoma, pancreatic cancer, glioblastoma, breast cancer, prostate cancer, ovarian cancer, and cervical cancer, and are involved in cancer cell progression [20][21][22][23]. A number of studies have reported that Ech is a potent antagonist of integrins αvβ3, αIIbβ3, and α5β1 for cancer therapy and inhibits HUVEC adhesion to immobilized fibronectin and vitronectin, as well as cell proliferation, migration, invasion, and adhesion of metastatic human osteosarcoma [6,10,[24][25][26]38]. T cells genetically engrafted with Ech-containing chimeric antigen receptor could also efficiently inhibit integrin-expressing solid tumors [39]. We showed that Ech could inhibit VEGF-induced HUVEC proliferation with an IC 50 value of 103.2 nM and inhibit the migration of human melanoma A375, glioblastoma U373MG, and pancreatic Panc-1 tumor cells with IC 50 values of 1.5, 5.7, and 154.5 nM. These findings indicate that Ech is a potent antiangiogenic and antimetastatic agent. The resulting structure-function relationships found in this study can provide a molecular basis for the use of Ech as a molecular scaffold to design anticancer drugs in the future.

Conclusions
In conclusion, this is the first report to determine a high-resolution crystal structure of a short disintegrin with a resolution of 1.80 Å. In comparison with the crystal structure of medium disintegrin, they have similar conformations for the RGD loop and the same hydrogen bond connectivity between the RGD loop and C-terminus. The major difference was found from the conformation of their C-termini. According to the results of the functional analysis, Ech may be a potential candidate for anticancer therapeutics with antiangiogenic and antimigration properties, and the C-terminal region of Ech can be optimized to improve its anticancer activity. The structure-function relationships of Ech can provide a molecular basis to design anticancer drugs using Ech as a molecular scaffold.

Expression, Purification, and Characterization of Ech and Its Mutants
The expression and purification of Ech and its mutants in P. pastoris were performed following previously described protocols [13,14,18,[40][41][42]. In brief, Ech and its mutants were expressed in the P. pastoris X33 strain with a 6xHis-tag, and purified to apparent homogeneity by Ni 2+ -chelating chromatography and HPLC. Wild-type Ech without a 6xHis-tag was expressed for crystallization and purified to homogeneity by CaptoMMC chromatography and HPLC. The yields of Ech and its mutants produced in P. pastoris were 2-7 mg/L. Mass spectrometry analysis showed that the experimental molecular weights of all recombinant proteins were in excellent agreement with theoretical values (less than 1 Da), which were calculated by assuming that all cysteines form disulfide bonds. Thus, our results indicated the formation of four disulfide bonds in Ech and its mutants.

Mass Spectrometric Measurements
The molecular weights of Ech and its mutants were confirmed using an LTQ Orbitrap XL mass spectrometer equipped with an electrospray ionization source (Thermo Fisher Scientific, Waltham, MA, USA) as previously described [18].

Cell Adhesion Assay
The cell adhesion assay was performed using previously described protocols [14]. In brief, Chinese hamster ovary (CHO) cells expressing integrins αvβ3 (CHO-αvβ3) and αIIbβ3 (CHO-αIIbβ3) were kindly provided by Dr Y. Takada (Scripps Research Institute) and maintained in Dulbecco's Modified Eagle's medium (DMEM). Human erythroleukemia K562 cells and the human colon adenocarcinoma cell line HT-29 were purchased from ATCC. K562 cells were cultured in RPMI-1640 medium containing 5% fetal calf serum, and HT29 cells were maintained in DMEM. Ninety-six-well microtiter plates (Costar, Corning, NY, USA) were coated with 100 µL of PBS buffer containing 200 µg/mL fibrinogen (for integrins αvβ3 and αIIbβ3), 50 µg/mL fibronectin (for integrins α5β1), and 5 µg/mL vitronectin (for integrins αvβ5) at 4 • C overnight. Wells were blocked by incubating with 200 µL of heat-denatured 1% bovine serum albumin (BSA) (Calbiochem, Darmstadt, Germany) at room temperature for 40 min. The heat-denatured BSA was discarded and was washed twice with 200 µL of PBS. CHO-αvβ3, CHO-αIIbβ3, and HT29 cells were detached by trypsinization and diluted to 3 × 10 5 cells/mL; K562 cells were diluted to 2.5 × 10 5 cells/mL; and 105 µL of cells was used for the assay. Cells were premixed with or without Ech and its mutants for 15-30 min at 37 • C in a 5% CO 2 atmosphere prior to dispensation into microtiter wells. The cells were then seeded to the coated plate and incubated at 37 • C (5% CO 2 ) for 1 h. The non-adhered cells were removed by washing twice with 200 µL of PBS. The adhered cells were fixed with 100 µL of 3.7% paraformaldehyde and quantified using a crystal violet assay. Colorization was performed by adding 150 µL of colorizing solution containing 50% alcohol and 0.1% acetic acid. The absorbance was read at 600 nm by an automatic microplate reader. Inhibition was defined as inhibition = 100 -[OD 600 (Rho protein-treated sample)/OD 600 (untreated sample)] × 100%. The reported IC 50 values are the average of at least three separate experiments.

Crystallization of Ech
Ech was dialyzed twice against 1 L ddH 2 O at 4 • C, lyophilized, and then dissolved in 50 mM Tris-HCl buffer (pH 8.0) at a concentration of~10 mg/mL. Ech was initially screened using commercial kits (The JCSG Core Suites, QIAGEN, Hilden, Germany) under 384 conditions. The CombiClover Junior sitting plate (24-well, Emerald BioSystems, Bainbridge Island, WA, USA) was used for crystallization. The crystals of Ech were obtained by mixing 0.8 µL protein solution and the reservoir solution containing 2.4 M (NH 4 ) 2 SO 4 (pH 9.0) and 0.1 M Bicine at pH 9.0 by the sitting-drop vapor diffusion method.

Diffraction Data Collection and Processing
All crystals were flash cooled to 100 K in a stream of cold nitrogen prior to data collection. X-ray diffraction data were collected on beamline TPS 05A at the National Synchrotron Radiation Research Center (NSRRC, Taiwan). The diffraction images were processed using the XDS program package [43]. The crystal data and data collection statistics of Ech are summarized in Table 1.

Structure Determination and Refinement
The initial phase of Ech was determined by molecular replacement, and the known structure of Rhodostomin (PDB ID 4RQG) in the PDB was used as our model protein. All structure determinations were carried out using alternate cycles of model building in Coot [44] and refinement in REFMAC5 [45]. The stereochemistry and structures of the final models were analyzed by MolProbity [34]. Ech belonged to the P3 2 21 space group with parameters a = 33.79, b = 33.79, and c = 120.90 Å. This structure was subsequently refined at 1.80 Å resolution using REFMAC5 to an R-factor of 0.212 and R-free of 0.255. The crystal structure of Ech was deposited in PDB ID 6LSQ.

NOE Analysis
The NOE experiment was conducted according to previously described protocols [14]. In brief, the backbone dynamics of Ech were determined by two-dimensional proton-detected heteronuclear NMR spectroscopy. NMR experiments were performed at 27 • C on a Bruker Avance 600 MHz spectrometer. The steady-state 1 H-15 N NOEs were collected from 1 H-detected 1 H-15 N correlation spectra recorded with sensitivity-enhanced pulse sequences. In the NOE experiment, two spectra, one with the saturated NOE and one without, were collected. The steady-state NOEs were the ratios that were calculated between the crosspeak intensities with and without saturation of the proton resonances (i.e., NOE = I sat /I nonsat ). The reported NOE value was the average value of three experiments.

Molecular Docking
The HADDOCK webserver was used for the docking of Ech into integrin αvβ3 [46]. The starting structure for docking was the X-ray structure of Echistatin. The interaction restraints were derived from the X-ray structure of integrin αvβ3 in complex with the 10th type III RGD domain of wild-type fibronectin (PDB ID 4MMX) [47] by using CCP4i software [48]. The defined distance threshold was 4 Å, and the interaction restraints between the RGD motif and integrin were used for the calculation. The input restraints between the R49, G50, and D51 residues of Ech and integrin αvβ3 were 27, 11, and 40, respectively. They were the contacts between the R49 residue and residues D150, Y178, Q180, and D218 of integrin αv; between residues G50 and Y178 of αv and residues R216 and A218 of β3; and between residue D51 and residues S121, Y122, S123, R214, N215, R216, D217, and A218 and Mn 2+ of the MIDAS of β3. An additional 0.5 Å distance was defined as the minimum and maximum limits of the interaction restraints. These restraints were used to perform the standard HADDOCK protocol for protein docking with minor modifications. This protocol combines three stages of molecular dynamics calculations, including heating and cooling, with a progressive increase in flexibility at the binding interface. In the first stage, 1000 conformations were calculated using a rigid-body docking protocol. The 200 structures with the lowest intermolecular energies were refined using semiflexible simulated annealing in the second stage. Semiflexible and fully flexible regions were automatically defined by HADDOCK. The 200 structures with the lowest intermolecular energy values were refined using explicit water molecules in the last stage. The clusters of structures were identified by fitting them with an average root mean square deviation (RMSD) value of 0.5-1.5 Å for the backbone atoms of all the amino acids. The selected structure cluster for the structural analysis was based on the lowest Z-score without any restraint violations. Hydrogen bonds and salt bridges interactions were analyzed with PISA software [49].

VEGF-Induced HUVEC Proliferation Assay
Human umbilical vein endothelial cells (HUVECs) were purchased from Cambrex-Lonza (East Rutherford, NJ, USA). They were maintained and expanded in endothelial cell growth medium-2 (EGM-2) and gradually replaced by 20% CCS/M199. For the proliferation assay, cells were seeded in 0.2% gelatin-coated 96-well plates at 1.5 × 10 4 cells/well, and then the cells were cultured in M199 serum-free medium at 37 • C overnight. The following day, 20 ng/mL recombinant human VEGF (R&D Systems, Minneapolis, MN, USA) was added, and the cells were treated with different concentrations of Ech for 48 h. After removing the medium, the cells of the plate were fixed with 120 µL of 3.7% paraformaldehyde and stained with 50 µL of 0.05% crystal violet solution for 20 min. Each well was then washed four times with 200 µL of distilled water and dried. Colorization was performed by adding 150 µL of colorizing solution containing 50% alcohol and 0.1% acetic acid. The absorbance was read at 600 nm by an automatic microplate reader. The reported values are the average of at least three separate experiments.

Cell Migration Assay
The transwell migration assay was conducted according to previously described protocols [13]. In brief, human A375 melanoma cells, human pancreatic Panc-1 cells, and U373 MG human glioblastoma cells were used to determine the inhibitory activities of recombinant Ech and its mutant. Then, 1 × 10 5 cells were added to the transwell inserts and incubated in 24-well plates. Cells in the inserts were cultured in serum free DMEM medium, whereas the lower wells were filled with DMEM medium containing 10% FBS as an attractant. After 6 h of incubation, the non-migrated cells on the upper side of the inserts were removed using a cotton swab. The insert wells were fixed with 3.7% paraformaldehyde and stained with 0.05% crystal violet. Cells that had transmigrated to the lower surface were imaged at 100× magnification, for five random fields by using OLYMPUS-IX71 microscopy. Migrated cells were counted using NIH ImageJ software.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6651/12/11/709/s1, Figure S1: Crystal packing of Ech, Figure S2: 1 H-15 N heteronuclear NOE values measured for individual amide sites along the backbones, Figure S3: The interactions of Ech with water molecules and sulfate, Table S1: Summary of the interactions between Ech (chain A and chain B) and sulfate ion, Table S2: Summary of the interactions in the RGD loop, C-terminus, and between the RGD loop and C-terminus in Ech chain A, Table S3: Summary of the interactions in the RGD loop, C-terminus, and between the RGD loop and C-terminus in Ech chain B, Table S4: Summary of the interactions in the RGD loop, C-terminus, and between the RGD loop and C-terminus in trimestatin, Table S5: Summary of the interactions between Ech chain A and αvβ3 integrin, Table S6: Summary of the interactions between Ech chain B and αvβ3 integrin.