Identification and pharmaceutical evaluation of novel frog skin-derived serine proteinase inhibitor peptide–PE-BBI (Pelophylax esculentus Bowman-Birk inhibitor) for the potential treatment of cancer

Amphibian venom-derived peptides have high potential in the field of anticancer drug discovery. We have isolated a novel Bowman-Birk proteinase inhibitor (BBI)-type peptide from the skin secretion of Pelophylax esculentus (PE) named PE-BBI, and evaluated its bio-functions and anti-cancer activity in vitro. PE-BBI is a heptadecapeptide with C-terminal amidation. The mRNA sequence and primary structure of PE-BBI were identified using RT-PCR and LC/MS, respectively. A trypsin inhibitory assay was used to characterize the serine proteinase inhibitory activity of synthetic PE-BBI. PE-BBI’s myotropic activity was analyzed using isolated rat bladder and rat-tail artery smooth muscle tissues, and the anti-cancer ability of PE-BBI using human colorectal cancer cells. PE-BBI’s mechanism of action was investigated using Discovery studio software. PE-BBI showed trypsin inhibitory activity (Ki = 310 ± 72 nM), strong myotropic activity, and cytotoxicity that were specific to cancer cells, and no side effect to normal epithelial cells. The docking stimulation showed that PE-BBI had high affinity to several members of human kallikrein related peptidase (KLK) family. This finding helps to enrich our understanding of BBI peptides’ mode of action. Moreover, the data presented here validates frog secretions as sources of potential novel proteinase inhibitors for cancer treatment.

The serine protease inhibitor family is one of the most widely distributed natural proteinase inhibitors. Serine protease inhibitors are able to regulate certain proteases participated in many human physiological and pathological processes, such as peptide hormone release, digestion, host defence, blood coagulation, disease progression, hypertension, muscular dystrophy, cancer, acquired immunodeficiency syndrome (AIDS), and microbial infections [1][2][3][4] . Serine proteinase inhibitors are also crucial in proteolytic activity disorder processes, which may cause a disruption in physiological processes, homeostatic balance, the expanding of tumour cells and degenerative diseases [5][6][7] . Understanding the roles serine protease inhibitors play in diseases may offer opportunities for development of therapeutic agents 8 . Proteinase inhibitors have been classified into more than 10 families, based on their structural features, including Kunitz, Kazal, and Bowman-Birk proteinase inhibitor (BBI) motifs 9 .
Trypsin Inhibitory Activity of PE-BBI. Synthetic replicates of PE-BBI was serially quantified for trypsin inhibitory activity in the range of 0.1-100 µM. PE-BBI showed a competitive reversible trypsin inhibitory activity. The dose-dependent hydrolysis of the fluorogenic substrates curve of PE-BBI were shown in Fig. 2A. The initial rates (v i ) were calculated from Fig. 2A, and the initial rates were used to calculate the inhibition constants (K i ) value by using 'Morrison plots' (Fig. 2B) 26 . Compared with some other BBI-type peptides such as HV-BBI and pLR-HL, PE-BBI was found to behave as a moderately potent trypsin inhibitor with a K i value of 310 ± 72 nM. To further confirm the binding model between PE-BBI and bovine trypsin protease, we applied ZDock to simulate the binding model. From the simulation results, we can find that the active site (Lys 9 ) of PE-BBI inserts into the reaction site (His 58 , Asp 102 , and Ser 195 ) of bovine trypsin proteinase, thus blocking trypsin's proteolytic activity. The simulation result shows PE-BBI might affect trypsin proteinase in the same way as reported BBI inhibitor peptide(s)/protein(s) (Fig. 2C). Myotropic Effects of PE-BBI. Synthetic PE-BBI was found to be active in the rat tail artery and the bladder smooth muscle preparations (Fig. 3A,B). In the preparations of rat bladder smooth muscle, PE-BBI showed a dose-dependent contraction with an EC 50 value of 14.25 nM (Fig. 3A). In preparations of rat artery, PE-BBI evoked a dose-dependent relaxation with an EC 50 value of 8.79 nM (Fig. 3B).
Antimicrobial Effects of PE-BBI. Antimicrobial activity of synthetic PE-BBI was assessed by determination of MIC assays using E. coli, S. aureus, and C. albicans. In this assay, PE-BBI didn't show any obviously antibacterial activity even in the highest concentration (512 μM). The results illustrated that PE-BBI did not have antibacterial activity against bacterial and fungi (Data not showed).
Anti-cancer Activity of PE-BBI. Previous studies supported the idea of the potential therapeutic use of BBI-type peptides as the anticancer therapy against human colon cancer. Thus, four human cancer cell lines including DLD-1, DKS8, HCT116, and HKE3 were used to examine the anti-proliferative effects of PE-BBI. Human colon endothelial cell line (FHC) was used to evaluate the inherent cytotoxicity of PE-BBI against normal human cells. PE-BBI showed selective cytotoxicity against the four different cancer cell lines and almost no cytotoxicity against FHC cells (Fig. 4A-E  simulation was shown in Fig. 5. ZDock module of Discovery Studio was used for docking and scoring; then RDock was applied to optimize the binding results. The models optimized by RDock module were performed in Fig. 5A-D, respectively. The docking analysis revealed that PE-BBI is highly likely binding to the active site of KLKs. These results further confirmed our prediction of peptide binding mechanism. PE-BBI had high affinity to certain members of human KLKs protein, which could induce the inhibition of KLKs. Consequently, the inhibition of the KLKS would in turn also had implications for related cancer pathology processes.

Discussion
Amphibian skin secretions have long been known as complex cocktails of bioactive peptides. Several different classes of serine protease inhibitor peptides have been isolated and identified from frog skin secretions, including Bowman-Birk-like trypsin inhibitor peptides, which are classified according to their primary structure 9 . Skin secretion proteinase inhibitors are believed to be widespread in Rana frog species and their primary structure-based modifications can produce a series of analogues (shown in Fig. 6), such as HV-BBI and pLR-HL, two BBI trypsin inhibitor peptides with only trypsin inhibitory activity but no antimicrobial activity or anti-cancer activity 27,28 . Besides these peptides, some structurally similar peptides such as Nigroain-A and Ranacyclin-T, show broad antimicrobial activity but lack any trypsin inhibitory activity 29,30 . Interestingly, these peptides have highly similar primary structure but distinct bio-functions. Although many structurally similar trypsin inhibitor peptides have been found in the past years, until now, none of them showed any anticancer  activity 28,30-32 . Leguminous BBIs have been identified with anti-cancer activity against human prostate, colon, breast, and skin cancer cell lines 33 . According to previous studies, a leguminous BBI showed in vitro and in vivo anti-tumour ability and without cytotoxicity in the dimethylhydrazine (DMH) induced colorectal tumour rat model 4,34 . Besides, BG-4, a novel proteinase peptide from Momordica charantia (bitter melon), induced apoptosis in HCT-116 and HT-29 colorectal cancer cells 20 . Another BBI peptide BTCI could form a stable complex with the 20S proteasome. In the complex form, BTCI can inhibit the trypsin-like proteolytic activity of the 20S proteasome, thereby inducing cytostatic and cytotoxic effects in MCF-7 breast cancer cells by induction of apoptosis 35 .
In this study, a novel short cyclic peptide PE-BBI was identified from the skin secretions of Pelophylax esculentus. It is the first time a frog skin BBI peptide was found to have the anticancer ability. The primary structure of PE-BBI contains the canonical disulphide-bridged loop between Cys 6 and Cys 16 , which has been proved to be essential for their trypsin inhibitory capability 29 . Molecular docking analysis results also suggested that PE-BBI inactivates the trypsin proteinase through the combination between its cyclic part (Cys 6 ~ Cys 16 ) and the active site of trypsin proteinase. Synthetic PE-BBI demonstrated strong trypsin inhibitory activity (K i = 310 ± 72 nM). Besides the proteinase inhibitory activity, it is interesting to see that PE-BBI also has a myotropic activity on isolated rat bladder and rat-tail artery smooth muscle on nano-/subnano-molar concentration. Through MTT assay, we also found that PE-BBI has considerable anticancer ability against a range of human colorectal cancer cell lines. Meanwhile, PE-BBI did not show any cytotoxicity against the normal human epithelial cell, FHC. Comparing the primary structure between PE-BBI and other frog skin-derived proteinase inhibitor peptides (Fig. 6) 18,[28][29][30]32 , all of these peptides contained the same disulphide-bridged loop structure but differ entirely at the N-terminal and C-terminal extension amino acid residues. PE-BBI contains more lysine (K) and amidated C-terminal, which means PE-BBI has a more positive charge than pLR-HL and HV-BBI. From the previous research findings, we noticed that the cancer cell membrane contains more anionic lipids (e.g., phosphatidylserine) than noncancerous cells. Thus, the membrane of cancer cells is negative charged compared to the noncancerous cells. On the surface of noncancerous cells, due to zwitterionic phospholipids (e.g. phosphatidylcholine and sphingomyelin) occupies a large part, the surface of noncancerous cells is negligible or weak negative charged 36 . This hypothesis may explain why HV-BBI and pLR-HL peptide lack anti-cancer ability, due to the lack of the cell membrane binding capacity.
Regarding the mechanism of anti-cancer cytotoxicity of PE-BBI, several possibilities exist, based on some previous studies. Some researchers have suggested the anti-cancer ability of BBI is related to the suppression of the release of tumourigenesis promoting factors such as the superoxide anion radical and hydrogen peroxide 33,37-40 .  Some other researchers propose that BBIs are able to inhibit some crucial proteolytic processes, which play key roles in carcinogenesis 31,33,41 .
Prognosis/diagnosis value of human tissue kallikrein-related peptidase (KLK) family members have been proved in many studies [21][22][23][24] . For instance, KLK4 dysregulation was detected in colon cancer and induced PAR1 signalling in caner. This study implies KLK4 may represent a novel pathway via PAR1 in colon tumourigenesis 42 . KLK6 mRNA overexpression evokes a poor prognosis in the colon cancer patient cohort 43 . Moreover, elevated KLK8 mRNA and protein level are detected in CRC tissue, which leads to an aggressive phenotype and poor prognosis 44 . The expression level of KLK10 is significantly higher in ovarian, pancreatic and colon cancer compared to normal tissues 23,45,46 . Previous studies have identified that trypsin and KLK4 shared significant structural and sequence similarity, and were both trypsin-like serine proteases. Thus, trypsin inhibitor would be able to inhibit KLK4 in addition to trypsin 47 . Swedberg et al. identified that SFTI-1 had high affinity to KLK4 47 . Thus, PE-BBI contains the homology active site region (-TK P1 SIPP-) with SFTI-1, which means PE-BBI may represent similar affinity to KLK4. The docking results in the present study revealed the active regions of PE-BBI and KLK4 superimpose extremely well. Moreover, we also tested the binding between PE-BBI and other potential colon cancer biomarkers (KLK6, KLK8, and KLK10) 23,[43][44][45][46] and got the similar results as KLK4. The myotropic regulation effect of PE-BBI proved further evidence. KLK family is one of the key regulatory factors of tissue smooth muscles by regulating the kallikrein-kinin system 48 . The inhibition of KLKs would result in smooth muscle contraction/release. Thus, PE-BBI might inhibit cancer cell growth by blocking KLK proteins, which also affects the kallikrein-kinin system in vivo.
Overall, our study identified a novel multi-functional peptide PE-BBI and assigned its potential anticancer mechanism. Compared with leguminous derived BBIs, frog skin-derived BBI peptide PE-BBI has a more conserved primary structure with smaller molecule size, and is a potential anti-cancer lead compound. Further studies on the structure-function relationships and mechanisms of the anticancer ability of frog skin-derived BBI peptides may provide new candidates for peptide anti-cancer therapies. were kept for 4 months prior to secretion harvesting. Transdermal electrical stimulation was used to harvest the skin secretions. This technique was first mentioned by Tyler and colleagues in 1992 49 . Briefly, the dorsal skin of frogs was stimulated by platinum electrodes (6 V DC; 4 ms pulse-width; 50 Hz) for two times, 20s per duration, then chilled deionised water was used to wash and collect the skin secretion before submerging into liquid nitrogen for lyophilise. Lyophilisate was stored at −20 °C prior to analysis. All of the above protocols were performed under the UK Animal (Scientific Procedures) Act 1986, project licence PPL 2694, issued by the Department of Health, Social Services and Public Safety, Northern Ireland. Procedures had been vetted by the IACUC of Queen's University Belfast, and approved on 1 March 2011.

Fractionation of Lyophilised Pelophylax esculentus Skin Secretion by Reverse Phase
HPLC. The lyophilised skin secretion (5 mg) was dissolved in 0.5 mL of trifluoroacetic acid (TFA)/water (0.05/99.95, v/v). Then the solvent was centrifuged to collect the supernatant. The supernatant was analysed by using Thermoquest gradient system with an analytical column (Phenomenex C-5; 0.46 cm × 25 cm). The linear elution gradient (From 0.05/99.95 (v/v) TFA/water to 0.05/19.95/80.0 (v/v/v) TFA/water/acetonitrile) was applied to separate the peptides. The flow rate was 1 mL/min. The effluent absorbance was monitored at λ214 nm and 1 mL fractions were collected at minute intervals. After 240 minutes, all of the fractions were collected. The fractions (240 fractions in total) were separated, lyophilised and stored at −20 °C prior to analysis and screening.

Structural Characterisation of PE-BBI.
After trypsin inhibitory screening, fraction #94 which showed the highest trypsin inhibitory activity was further applied in the structural characterisation. The molecular weight of peptides in fraction #94 was detected by using MALDI-TOF mass spectrometry (Voyager DE instrument, Perspective Biosystems, MA, USA). The peptides with similar molecular weight with those mature peptides obtained from molecular cloning were subjected to LCQ Fleet electrospray ion-trap mass spectrometer (Thermo Scientific, San Jose, CA, USA) to clarify their primary structure. Then the mature peptide of PE-BBI was chemically-synthesized by solid phase peptide synthesis (SPPS) technology.

Cloning of PE-BBI Precursor-Encoding cDNA.
Lyophilised frog skin secretion (5 mg) was lysed by mRNA protection buffer by Dynal Biotec, UK. The polyadenylated mRNA was isolated by using magnetic oligo-dT beads as described by the manufacturer (Dynal Biotec, UK). To obtain the completed full-length trypsin inhibitor precursor encoding region, 5′-and 3′-rapid amplification of cDNA ends (RACE) procedures were applied using a SMART-RACE kit (Clontech, UK) followed manufacturer protocol 50 . Staphylococcus aureus (NCTC 10788) and pathogenic yeast Candida albicans (NCYC 1467) were used in minimum inhibitory concentration (MIC) assays to investigate the board antimicrobial activity ability of synthetic PE-BBI. Microorganisms involved in MIC assays were cultured in MHB for 16-18 h. Then sub-cultured for 2~4 h to allow the microorganism to enter the logarithmic growth phase. Subsequently, proper density of bacterial (1 × 10 6 colony forming units (cfu)/mL) and fungi (1 × 10 5 cfu/mL) were seeded into 96 well plate. Different concentration ranged from 1 to 512 μM of PE-BBI were added into each well (6 replicates per concentration). The absorbance value at 550 nm was detected using a Synergy HT plate reader (Biotech, Minneapolis, MN, USA) after 18 h incubation in 37 °C incubator 50 .
Myotropic Assay of PE-BBI. Female Wistar rats (250-300 g) were applied in this study and sacrificed followed the principle of institutional animal experimentation ethics and UK animal research guidelines. All of the procedures involved in this study was approved by IACUC of Queen's University Belfast. Rat-tail artery and bladder were isolated and located in pre-chilled Kreb's solution equilibrated with 95% O 2 , 5% CO 2 51 . Isolated muscle strips (2-10 mm) were vertically suspended in a 2 mL organ bath containing 37 °C Kreb's solution. The Kreb's solution was constant refreshing by constant flowing (2 mL/min) during the assay. Potassium chloride (KCL, 60 mM) was used to test for viability of muscle strips after 20 min equilibration. After the smooth muscle strip were prepared, the synthetic replicates of PE-BBI (10 −11 to 10 −5 M) was used to quantify the myogenic potency on rat-tail artery and bladder smooth muscle. Between each concentration, the smooth muscle strip was washed 5 mins and re-equilibration for another 5 min. Each data point was repeated on at least five independent muscle strips. The tension change was amplified and recorded by a PowerLab System (AD Instruments Pty Ltd.) 51 .
Trypsin Inhibitor Assay of PE-BBITrypsin (0.1 µM in 1 mM HCL) and trypsin substrate (50 µM) were prepared prior to the assay. Trypsin inhibitory assay included positive group (10 µL trypsin, 180 µL substrate, 20 µL PBS), experiment group (10 µL trypsin, 180 µL substrate, 20 µL PE-BBI ranging from 0.1-100 µM) and control group (180 µL substrate, 30 µL PBS). In each well, the final volume was 210 µL. After 8 mins incubation, the fluorescent strength released from substrate hydrolysis was monitored continuously at 460 nm (excitation at 360 nm) in a CYTOFLUOR ® multi-well plate reader Series 4000 spectrofluorometer. Each sample was carried out in triplicate and repeated at least three times. The fluorescent strength values were used to calculate the initial rate (vi) and inhibition constants (K i ) 32 .
MTT Assay of PE-BBI. Synthetic replicates of PE-BBI were screened on the human colon cancer cell lines, DLD-1, DKS8, HCT116, and HKE3. Human colon endothelial cell FHC was used to evaluate the cytotoxicity of synthetic peptides on normal human cell line. All the cell lines were seeded with 5000 cells/well in 96 well plate. The seeded cells were incubated for 24 h and further starved for 12 h in prior of experiment, a series of PE-BBI concentrations (10 −7 -10 −4 M) were prepared in related growth medium and added in the experiment group with 6 replicates. The same volume of PBS was added in the control group. After 24 h PE-BBI treatment, MTT assay was used followed the protocol described in detail previously 52 . Briefly, 10 µL of MTT (5 mg/mL) was added and incubated for 4 h. Then the supernatant was removed and 100 µL of DMSO was added to dissolve the purple formazan generated by live cells. The absorbance was measured at 570 nm by using EL808 TM Absorbance Microplate Reader (Biolise BioTek EL808, Winooski, VT, USA). The cell viability (%) of each concentrations were calculated by normalized to the control, respectively. Molecular Docking Simulation Analysis. Molecular docking simulation of PE-BBI and trypsin proteinase was carried out using BIOVIA Discovery Studio 2017 R2 (Biovia, San Diego, CA, USA). The purpose of molecular docking was simulated the predominant binding condition(s) between ligand peptide (PE-BBI) and receptor proteins (KLK family members) with known three-dimensional structure. The three-dimensional (3D) structure of bovine trypsin (PDB ID: 2O9Q), KLK4 (PDB ID: 4K8Y), KLK6 (PDB ID: 3VFE), KLK8 (PDB ID: 5MS4), and KLK10 (PDB ID: 5LPE) were obtained from the Protein Data Bank (https://www.rcsb.org/). Before docking, interior water molecules and ligands should be deleted to clear the docking candidates. The defaulted hided hydrogen atoms were added to both proteins. Subsequently, we used ZDock 53 , a rigid-body docking algorithm, to create the initial docking poses between PE-BBI and trypsin, KLK4, KLK6, KLK8, and KLK10. We applied the protocol with parameters as following: Angular Step Size as 15; RMSD cutoff as 6.0; Interface Cutoff as 9.0; Maximum Number of Clusters as 60. The ZDock protocol generated lots of docking poses ranked by the ZDock score. ZDock score was based on shape complementary to evaluate the reliability of docking poses. The top ranked poses were classified into different clusters 54,55 , and further refined by RDock 56 . RDock applied the Chemistry at Harvard Macromolecular Mechanics (CHARMm) force field to generate the energy minimized docking models. In this study, the top ranked ZDock poses were optimized by RDock then further applied in the molecular docking 57 .
Statistical Analysis. Statistical analysis was performed using GraphPad Prism 6 software. P values were calculated by Student's t-test from the mean values of the indicated data. Significant differences were demonstrated with asterisks (*P < 0.05; **P < 0,01; ***P < 0.001).