Peptide-Alkoxyamine Drugs: An Innovative Approach to Fight Schistosomiasis: “Digging Their Graves with Their Forks”

The expansion of drug resistant parasites sheds a serious concern on several neglected parasitic diseases. Our recent results on cancer led us to envision the use of peptide-alkoxyamines as a highly selective and efficient new drug against schistosome adult worms, the etiological agents of schistosomiasis. Indeed, the peptide tag of the hybrid compounds can be hydrolyzed by worm’s digestive enzymes to afford a highly labile alkoxyamine which homolyzes spontaneously and instantaneously into radicals—which are then used as a drug against Schistosome adult parasites. This approach is nicely summarized as digging their graves with their forks. Several hybrid peptide-alkoxyamines were prepared and clearly showed an activity: two of the tested compounds kill 50% of the parasites in two hours at a concentration of 100 µg/mL. Importantly, the peptide and alkoxyamine fragments that are unable to generate alkyl radicals display no activity. This strong evidence validates the proposed mechanism: a specific activation of the prodrugs by the parasite proteases leading to parasite death through in situ alkyl radical generation.


Introduction
Schistosomiasis (or bilharziasis) is a major parasitic disease in terms of mortality and morbidity, as almost 200 million people are infected, and at least 200,000 deaths per year are associated with this disease.This disease is caused by worms of the Schistosoma genus present in freshwater bodies of subtropical regions [1].Consequences of chronic infection include fibrosis and calcification of the urinary tract, renal failure or bladder cancer and acute hepatitis, liver and intestine fibrosis, and portal hypertension.On children, it may cause poor growth and learning difficulty.Up to now, a single drug against schistosomiasis, praziquantel (PZQ), has been used for more than 50 years, and many cases of resistance or decrease in sensitivity have been evidenced [2,3].Moreover, praziquantel demonstrates activity against adult worms (after 5-7 weeks post-infection) but lacks efficacy against early human stages (i.e., schistosomula).Because individuals living in endemic areas are continually re-infected, the strategy for reducing the worm burden in each person necessitates the development of novel drugs capable of efficiently targeting all life stages of schistosomes in humans.Consequently, research on new drugs able to circumvent the praziquantel™ (PZQ) resistance and to kill all life stages of the parasite is an absolute priority.Schistosome parasites ingest host red blood cells and digest hemoglobin to obtain the amino-acids they need for the synthesis of their own proteins [4].A molecule targeting the parasite's mandatory hemoglobin digestion activity could therefore be efficient at several stages of its development [5].Recently, we proposed an approach based on the use of a mandatory enzymatic activity of cancer cells to deliver specifically to the intended target a drug known to have a non-selective activity [6].Here we propose to test such an approach on Schistosoma mansoni adult worms, an approach which we could not resist the pun of naming digging their graves with their forks (Scheme 1).Indeed, any parasite has specific enzymatic activities different from the host's.Then, hybrid peptide-alkoxyamines are specifically hydrolyzed by enzymes (here proteases, due to their peptide tags) into free peptides and free alkoxyamines.The latter homolyzed spontaneously and instantaneously at room temperature into nitroxyl (blue part) and alkyl (red part) radicals.The latter are potentially cytotoxic for their environment, leading to the death of the parasites, i.e., here schistosomes.
schistosomiasis, praziquantel (PZQ), has been used for more than 50 years, and ma cases of resistance or decrease in sensitivity have been evidenced [2,3].Moreover, p ziquantel demonstrates activity against adult worms (after 5-7 weeks post-infection) b lacks efficacy against early human stages (i.e., schistosomula).Because individuals livi in endemic areas are continually re-infected, the strategy for reducing the worm burd in each person necessitates the development of novel drugs capable of efficiently targeti all life stages of schistosomes in humans.Consequently, research on new drugs able circumvent the praziquantel™ (PZQ) resistance and to kill all life stages of the parasite an absolute priority.Schistosome parasites ingest host red blood cells and digest hem globin to obtain the amino-acids they need for the synthesis of their own proteins [4] molecule targeting the parasite's mandatory hemoglobin digestion activity could the fore be efficient at several stages of its development [5].Recently, we proposed an a proach based on the use of a mandatory enzymatic activity of cancer cells to deliver sp cifically to the intended target a drug known to have a non-selective activity [6].Here propose to test such an approach on Schistosoma mansoni adult worms, an approach wh we could not resist the pun of naming digging their graves with their forks (Scheme 1).deed, any parasite has specific enzymatic activities different from the host's.Then, hyb peptide-alkoxyamines are specifically hydrolyzed by enzymes (here proteases, due their peptide tags) into free peptides and free alkoxyamines.The latter homolyzed spo taneously and instantaneously at room temperature into nitroxyl (blue part) and alkyl (r part) radicals.The latter are potentially cytotoxic for their environment, leading to death of the parasites, i.e., here schistosomes.Scheme 1. Scheme for the concept: digging their graves with their forks.
It is expected that the schistosomes' hemoglobin-digesting proteases, which are n naturally present in human cells, are specific parasite enzymes responsible for the re cling of amino acids from the host's hemoglobin.Moreover, it has been demonstrated th within the same protease family, such as cathepsin D, the differences in cleavage si between schistosomes and humans are so great that it is possible to develop specific hibitors targeting specifically the hemoglobin-ingesting parasite [7].In schistosome, serine protease has been identified in the gut of the parasite, suggesting that the protea involved in the digestion are totally different from those of the vertebrate host.Howev unlike Plasmodium's, the schistosome's hemoglobin digestion cascade has not been fu characterized.The hemoglobin digestion implies at least three classes of proteases pres in the parasite gut lumen: cathepsin (B, C and L), metallo-aminopeptidases and aspara nyl endopeptidases [8].Considering that both parasites produce the same waste prod (i.e., hemozoin) from the same nutrient (i.e., hemoglobin), we hypothesized that th might be a convergence between Plasmodium and Schistosoma in the enzymes involved the proteolytic activity.Indeed, the proteolytic cleavage profiles of crude hemoglobin pH around 5 are almost identical when obtained by Plasmodium digestive plasmepsins by schistosome digestive cathepsins [9].This convergence is probably due mostly to structural resistance to proteolysis of blood proteins, which allows them to be less sen tive to microbial infections.Thus, proteolysis can only occur on a few less structured si as exemplified by the main proteolysis site 33-34 at the C-terminal extremity of a he Scheme 1. Scheme for the concept: digging their graves with their forks.
It is expected that the schistosomes' hemoglobin-digesting proteases, which are not naturally present in human cells, are specific parasite enzymes responsible for the recycling of amino acids from the host's hemoglobin.Moreover, it has been demonstrated that within the same protease family, such as cathepsin D, the differences in cleavage sites between schistosomes and humans are so great that it is possible to develop specific inhibitors targeting specifically the hemoglobin-ingesting parasite [7].In schistosome, no serine protease has been identified in the gut of the parasite, suggesting that the proteases involved in the digestion are totally different from those of the vertebrate host.However, unlike Plasmodium's, the schistosome's hemoglobin digestion cascade has not been fully characterized.The hemoglobin digestion implies at least three classes of proteases present in the parasite gut lumen: cathepsin (B, C and L), metallo-aminopeptidases and asparaginyl endopeptidases [8].Considering that both parasites produce the same waste product (i.e., hemozoin) from the same nutrient (i.e., hemoglobin), we hypothesized that there might be a convergence between Plasmodium and Schistosoma in the enzymes involved in the proteolytic activity.Indeed, the proteolytic cleavage profiles of crude hemoglobin at pH around 5 are almost identical when obtained by Plasmodium digestive plasmepsins or by schistosome digestive cathepsins [9].This convergence is probably due mostly to the structural resistance to proteolysis of blood proteins, which allows them to be less sensitive to microbial infections.Thus, proteolysis can only occur on a few less structured sites as exemplified by the main proteolysis site 33-34 at the C-terminal extremity of a helix.Substrate specificity profiles for the digestive plasmepsins of Plasmodium are much better characterized than for schistosome digestive enzymes [10].Thus, we selected the short peptide Phe-Val-Phe extracted from the substrate specificity study of Plasmodium plasmepsin II [11] because the synthesis did not require further protection/deprotection steps.
Therefore, drawing on our experience with alkoxyamines' efficiency against tumors [6] and Plasmodium [12], alkoxyamines A3L-A7L (Figure 1) carrying the peptide sequence L-Phe-Val-Phe-OH are tested (More details are provided in [12]).For comparison, the series D made of unnatural peptides is also investigated (Figure 1).peptide Phe-Val-Phe extracted from the substrate specificity study of Plasmodium plasmepsin II [11] because the synthesis did not require further protection/deprotection steps.Therefore, drawing on our experience with alkoxyamines' efficiency against tumors [6] and Plasmodium [12], alkoxyamines A3L-A7L (Figure 1) carrying the peptide sequence L-Phe-Val-Phe-OH are tested (More details are provided in [12]).For comparison, the series D made of unnatural peptides is also investigated (Figure 1).

Materials and Methods
Preparation of alkoxyamines and reference samples.All reagents including anhydrous solvents used for the chemical synthesis were purchased from commercially available sources such as Merck, TCI, and Fluorochem, and were used without further purification.All air and/or water-sensitive reactions were carried out under an argon atmosphere with dry solvents using standard syringe-cannula/septa techniques.Routine monitoring of reactions was performed using Merck Silica gel 60 F254, aluminum-supported TLC plates; spots were visualized using UV light and ethanolic acidic para-anisaldehyde solution or ethanolic phosphomolybdic solution, followed by heating.Purifications by means of column chromatography were performed with Silica gel 60 (230-400 mesh). 1 H, 13 C, and 19 F NMR spectra were obtained in CDCl3, MeOD (with residual H2O), D2O, DMSO-d6, or C6D6 solutions on by using Bruker AC400 (400 MHz) and Bruker AC300 (300 MHz) spectrometers.High-resolution mass spectra (HRMS) have been performed using a SYNAPT G2 HDMS (Waters) mass spectrometer equipped with a pneumatically assisted atmospheric pressure ionization (API) source.The sample was ionized in positive electrospray mode under the following conditions: electrospray voltage: 2.8 kV; port voltage: 20 V; Nebulizing gas flow rate (nitrogen): 100 L/h or in negative electrospray mode under the following conditions: electrospray voltage: −2.27 kV; port voltage: −20 V; Nebulizing gas flow rate (nitrogen): 100 L/h.The mass spectra (MS) were obtained with a time-of-flight (TOF) analyzer.The exact mass measurements were carried out in triplicate with external calibration.Homolysis constants were measured by EPR on an EMX Bruker spectrometer from 10 −4 M samples in tertbutyl benzene or in a mixture of H2O-nPrOH (1:1; v:v).
General ethyl chloroformate/triethylamine procedure GP1.A solution of N-Bocphenylalanine (1.3 eq) in THF (150 mL) is cooled to −15 °C under an argon atmosphere.Triethylamine (1.3 eq) is added, and then the ethyl chloroformate (1.3 eq).The reaction mixture is stirred at −15 °C for 1 h.Then, alkylaniline (1.0 eq) dissolved in THF (20 mL) is Figure 1.Peptide-alkoxyamines investigated as anti-schistosomal drugs.L and D are for the configuration of the residues on the peptides as given by Fischer's rules.

Materials and Methods
Preparation of alkoxyamines and reference samples.All reagents including anhydrous solvents used for the chemical synthesis were purchased from commercially available sources such as Merck, TCI, and Fluorochem, and were used without further purification.All air and/or water-sensitive reactions were carried out under an argon atmosphere with dry solvents using standard syringe-cannula/septa techniques.Routine monitoring of reactions was performed using Merck Silica gel 60 F 254 , aluminum-supported TLC plates; spots were visualized using UV light and ethanolic acidic para-anisaldehyde solution or ethanolic phosphomolybdic solution, followed by heating.Purifications by means of column chromatography were performed with Silica gel 60 (230-400 mesh). 1 H, 13 C, and 19 F NMR spectra were obtained in CDCl 3 , MeOD (with residual H 2 O), D 2 O, DMSO-d 6 , or C 6 D 6 solutions on by using Bruker AC400 (400 MHz) and Bruker AC300 (300 MHz) spectrometers.High-resolution mass spectra (HRMS) have been performed using a SYNAPT G2 HDMS (Waters) mass spectrometer equipped with a pneumatically assisted atmospheric pressure ionization (API) source.The sample was ionized in positive electrospray mode under the following conditions: electrospray voltage: 2.8 kV; port voltage: 20 V; Nebulizing gas flow rate (nitrogen): 100 L/h or in negative electrospray mode under the following conditions: electrospray voltage: −2.27 kV; port voltage: −20 V; Nebulizing gas flow rate (nitrogen): 100 L/h.The mass spectra (MS) were obtained with a time-of-flight (TOF) analyzer.The exact mass measurements were carried out in triplicate with external calibration.Homolysis constants were measured by EPR on an EMX Bruker spectrometer from 10 −4 M samples in tertbutyl benzene or in a mixture of H 2 O-nPrOH (1:1; v:v).
General ethyl chloroformate/triethylamine procedure GP1.A solution of N-Bocphenylalanine (1.3 eq) in THF (150 mL) is cooled to −15 • C under an argon atmosphere.Triethylamine (1.3 eq) is added, and then the ethyl chloroformate (1.3 eq).The reaction mixture is stirred at −15 • C for 1 h.Then, alkylaniline (1.0 eq) dissolved in THF (20 mL) is added and the reaction is stirred for 1 h at −15 • C, then at room temperature overnight.The solvent is removed in vacuo, and the residue is taken up in DCM.The organic layer is washed sequentially with 1 M HCl twice, saturated NaHCO 3 solution twice and brine twice, dried over MgSO 4 , and evaporated under reduced pressure.The purification is performed by flash column chromatography (EtOAc/Petroleum ether).
General Salen/MnCl 2 alkoxyamine synthesis procedure GP2.To a stirred solution of N,N'-bis(salicylidene)ethylene diamine (salen ligand, 0.25 eq.) in THF (10 mL) is added MnCl 2 (0.25 eq) in an open flask.After 10 min of stirring at room temperature, a solution of TEMPO (1.0 eq) and 4-vinylanilido phenylalanine-Boc (1.2 eq) in THF (50 mL) is added, and then solid NaBH 4 is added (4.0 eq) in four portions, every 10 min.The mixture is stirred between 24 h and 48 h at room temperature.At the end, the solution is dissolved in EtOAc (100 mL) in an Erlenmeyer and carefully quenched at 0 • C by the addition of 1 M HCl until the solution becomes colorless or slightly orange.Then, NaHCO 3 is added up to neutral pH and the organic layer is washed with water (80 mL) and brine (80 mL).The organic layer is dried on MgSO 4 , filtered, and evaporated to dryness.The purification is performed by flash column chromatography (EtOAc/Petroleum ether).
General Boc deprotection procedure GP3.The Boc-residue (1.0 eq) is dissolved in DCM (45 mL, V DCM = 4V TFA ) and TFA (10.0 eq) is added under air.The reaction mixture is stirred at room temperature.Once the substrate is fully consumed, toluene (25 mL) is added, and the solvents are removed in vacuo.The co-evaporation is repeated twice.The purification by flash column chromatography (EtOAc/Petroleum ether) is performed.
General DCC/HOBt procedure GP4.Diisopropylethylamine (1.0 eq) is added to a stirred solution of TFA salt (1.0 eq) under argon.The mixture is stirred for 10 min at room temperature, then Boc-dipeptide (1.1 eq) and HOBt (1.1 eq) are added and stirred until dissolved.The mixture is cooled to 0 • C, and DCC or EDCI (1.1 eq) is added.The mixture is stirred overnight at room temperature.The mixture is filtered with cold DCM, and washed with 1 M HCl, NaHCO 3 (saturated solution), and brine.The purification is performed by flash column chromatography (EtOAc/Petroleum ether).
General TFA removing procedure GP5. 1 M NaOH solution (6 mL) is added to a stirred solution of TFA salt (1.0 eq) under air.The mixture is stirred for 10 min at room temperature, the mixture becomes cloudy.Then, a few drops of MeOH are added and the aqueous phase is extracted three times with DCM (5 mL).The organic layer is dried with MgSO 4 , filtered, and the solvent is evaporated to dryness.
General hydrogenation procedure GP6.To a solution of the benzylated compound (1.0 eq) in MeOH (20 mL) is added Pd/(C) (10% weight).Then, the atmosphere in the flask is replaced by the H 2 atmosphere and stirred.Once the substrate is fully consumed, the mixture is filtrated through a pad of celite, rinsed with MeOH, and the solvent is evaporated to dryness.
14 g, 4.185 mmol, 1.2 eq) in THF (50 mL) is cooled to −15 • C under argon atmosphere.Triethylamine (0.56 mL, 4.185 mmol, 1.2 eq) is added, and then ethyl chloroformate (0.37 mL, 4.185 mmol, 1.2 eq).The reaction mixture is stirred at −15 • C.After 1 h, 4-ethylaniline (0.43 g, 3.487 mmol, 1.0 eq) dissolved in THF (15 mL) was added and the reaction was allowed to stir for 1 h at −15 • C, then at room temperature overnight.The solvent is removed in vacuo, and the residue is taken up in DCM.The organic layer is washed sequentially with 1 M HCl twice, saturated NaHCO 3 solution twice and brine twice, dried over MgSO 4 , and evaporated under reduced pressure.The purification by flash column chromatography (EtOAc/Petroleum ether) affords P21D as a white solid (0.47 mg, 22%).
4-(Benzyloxy)-4-oxobutanoic acid.Succinic anhydride (8.00 g, 79.944 mmol, 1.0 eq) was dissolved in the mixture of anhydrous dichloromethane (10 mL) and Benzyl alcohol (12 mL).Triethylamine (3.2 mL, 23.983 mmol, 0.3 eq) and DMAP (9.76 g, 79.944 mmol, 1.0 eq) were added to this solution.The resulting clear solution was stirred at ambient temperature overnight.After which time, the mixture was dissolved in EtOAc (150 mL) and washed with a saturated solution of NaHCO 3 (3 × 50 mL).The aqueous extracts were carefully acidified to pH 2 with concentrated HCl and then extracted with DCM (3 × 50 mL) and subsequently dried (Mg 2 SO 4 ), filtered, and concentrated to afford the title compound (16.23 g, 97%) as a white solid with no further purification needed.
Kinetic experiments.Kinetics of A1-A7 (Scheme 2) for series L and D are performed by monitoring the growth of nitroxide by Electron Paramagnetic Resonance (X-band EMX Bruker machine) in tert-butylbenzene and a water/n-PrOH (for a good solubility of the alkoxyamine) as solvents and using O 2 as alkyl radical scavenger to suppress the back reaction (k c in Scheme 2) as already reported [18].Homolysis rate constants k d are given by Equation (1) ([nitroxide] ∞ = [alkoxyamine] 0 = 0.1 mM, see Table S1) and the subsequent activation energies E a given by Equation (2).
4-(Benzyloxy)-4-oxobutanoic acid.Succinic anhydride (8.00 g, 79.944 mmol, 1.0 eq) was dissolved in the mixture of anhydrous dichloromethane (10 mL) and Benzyl alcohol (12 mL).Triethylamine (3.2 mL, 23.983 mmol, 0.3 eq) and DMAP (9.76 g, 79.944 mmol, 1.0 eq) were added to this solution.The resulting clear solution was stirred at ambient temperature overnight.After which time, the mixture was dissolved in EtOAc (150 mL) and washed with a saturated solution of NaHCO3 (3 × 50 mL).The aqueous extracts were carefully acidified to pH 2 with concentrated HCl and then extracted with DCM (3 × 50 mL) and subsequently dried (Mg2SO4), filtered, and concentrated to afford the title compound (16.23 g, 97%) as a white solid with no further purification needed.
Kinetic experiments.Kinetics of A1-A7 (Scheme 2) for series L and D are performed by monitoring the growth of nitroxide by Electron Paramagnetic Resonance (X-band EMX Bruker machine) in tert-butylbenzene and a water/n-PrOH (for a good solubility of the alkoxyamine) as solvents and using O2 as alkyl radical scavenger to suppress the back reaction (kc in Scheme 2) as already reported [18].Homolysis rate constants kd are given by Equation (1) ([nitroxide]∞ = [alkoxyamine]0 = 0.1 mM, see Table S1) and the subsequent activation energies Ea given by Equation (2).
Scheme 2. Dynamic Covalent Bond in alkoxyamines: kd for the homolysis rate constant and kc for the re-formation reaction.
= − . ( Cytotoxicity experiments. A non-cancer line of Vero cells was used to determine the cytotoxicity of the compound against mammalian cells.They were evaluated at a final DMSO concentration of 0.5%.The culture medium was MEM (Dutscher, Bernolsheim, France) supplemented with 10% fetal bovine serum (Fisher Scientific, Illkirch, France), 1X non-essential amino acids (Fisher Scientific, Illkirch, France), 100 U/mL, 100 µg/mL penicillin/streptomycin (Fisher Scientific, Illkirch, France), and 2 mM L-glutamine (Fisher Scientific, Illkirch, France) at 37 °C in a humidified 5% CO2 atmosphere [19].Vero cells (100 µL of 10 5 cells/mL per well) were plated in 96-well plates for 24 h then treated with 100 µL of compound dilutions (in duplicate) during 48 h.All molecules were tested from 5 nM to 50 µM.Each well was then examined under the microscope to detect any precipitate formation before the supernatant was removed by flicking the plate.A volume of 100 µL of 0.5 mg/mL in MEM from a stock solution at 5 mg/mL of PBS-dissolved MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, Sigma Aldrich/Merck, Darmstadt, Germany) was then added to each well [20].
After incubation for 1 h at 37 °C and 5% CO2 the supernatant was removed and 100 µL of DMSO were added.Plates were gently shaken to dissolve formazan crystals resulting from the MTT reduction by living cells and read at 570 nm with the VICTOR Nivo plate reader (Perkin Elmer, Waltham, USA).CC50 values were determined using GraphPad Prism 7 software (San Diego, CA, USA).
Experiments on the viability of Schistosoma mansoni adult worms.In vitro tests were performed on Schistosoma mansoni NMRI strains.This parasite is lab maintained on

Cytotoxicity experiments.
A non-cancer line of Vero cells was used to determine the cytotoxicity of the compound against mammalian cells.They were evaluated at a final DMSO concentration of 0.5%.The culture medium was MEM (Dutscher, Bernolsheim, France) supplemented with 10% fetal bovine serum (Fisher Scientific, Illkirch, France), 1X non-essential amino acids (Fisher Scientific, Illkirch, France), 100 U/mL, 100 µg/mL penicillin/streptomycin (Fisher Scientific, Illkirch, France), and 2 mM L-glutamine (Fisher Scientific, Illkirch, France) at 37 • C in a humidified 5% CO 2 atmosphere [19].Vero cells (100 µL of 10 5 cells/mL per well) were plated in 96-well plates for 24 h then treated with 100 µL of compound dilutions (in duplicate) during 48 h.All molecules were tested from 5 nM to 50 µM.Each well was then examined under the microscope to detect any precipitate formation before the supernatant was removed by flicking the plate.A volume of 100 µL of 0.5 mg/mL in MEM from a stock solution at 5 mg/mL of PBS-dissolved MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, Sigma Aldrich/Merck, Darmstadt, Germany) was then added to each well [20].
After incubation for 1 h at 37 • C and 5% CO 2 the supernatant was removed and 100 µL of DMSO were added.Plates were gently shaken to dissolve formazan crystals resulting from the MTT reduction by living cells and read at 570 nm with the VICTOR Nivo plate reader (Perkin Elmer, Waltham, MA, USA).CC 50 values were determined using GraphPad Prism 7 software (San Diego, CA, USA).
Experiments on the viability of Schistosoma mansoni adult worms.In vitro tests were performed on Schistosoma mansoni NMRI strains.This parasite is lab maintained on Biomphalaria glabrata (BRE strain) as intermediate mollusc host and Golden Hamster (Janvier Labs, Le Genest-Saint-Isle, France) as definitive vertebrate host.Methods, for molluscs' and hamsters' infection and for parasite recovery, were previously described [21].Adult worms were recovered by hepatic perfusion technique between 42 and 47 days post hamster exposition to parasite larvae.Worms were carefully collected and disposed in 12-well Falcon ® plate containing 2 mL of RPMI 1640 (supplemented with L-glutamine and Hepes 25 mM).The plates containing a minimum of 15 worms by well were then placed in an incubator chamber at 37 • C and 5% CO 2 .The worms sex ratio was almost balanced in each well.The drug was first dissolved in DMSO (Sigma-Aldrich) to give a 100 mg/mL mother solution.Then the stock solution was complemented with Tween 80 and RPMI to obtain the following final ratio dilution: RPMI 1640/Tween 80/DMSO and 1000/0.95/3.8,v/v/v.
All molecules were tested at a final concentration of 100 µg/mL and 10 µg/mL.A negative control consisted in adding the same RPMI 1640/Tween 80/DMSO solution, but without any drug and a positive control consisted in adding praziquantel [22].All molecules and control were performed in duplicates of well.The worm viability was checked after 1, 2, 3, 4, 5, 6, 7, 8, 24, 48, and 72 h after adding the molecule.Parasites exhibiting no body contractions during a 30 s observation were considered as dead.

A3L-A7L.
Alkoxyamines A3L-A7L were prepared in 4 steps for A3L and a few more for A4L-A7L (Scheme 3) starting from the amino acid P1L transformed into amide P3L.The latter was coupled to TEMPO using the conventional Mn/salen procedure (see experimental section and references [23,24]) to afford alkoxyamine A1L which was hydrolyzed as salt A2L.Then, it was coupled to dipeptide P5L (Scheme 4) affording peptide-alkoxyamine A3L.The latter was deprotected as salt A4L which was, in turn, either neutralized as A5L or homologated in succinate and succinic derivatives A6L and A7L, respectively.The same procedure (see Supplementary Materials) was applied for the preparation of the unnatural series A3D-A5D in similar yields for each step.To investigate the potential effect of peptides or peptide analogues potentially observed during the decomposition of A3L-A7L, derivatives P6L-P12L, P17L, P18L, and P21L-P25L were prepared as depicted in Schemes 4-6.The same procedure (see Supplementary Materials) was applied for the preparation of the unnatural series D in similar yields for each step.

Kinetic Analysis
All homolysis rate constants kd are collected in Table 1.Activation energies Ea (given by Equation ( 2)) are around 130 kJ/mol as expected [25] for para-substituted amino-grouparyl-TEMPO-based alkoxyamines meaning that alkoxyamines are stable under our experimental conditions (t1/2 > 170 day at 37 °C).As expected, Ea does not depend on the configuration of the peptide chain.Moreover, no difference in Ea is observed between diastereoisomers.
Table 1.Homolysis rate constants kd for 3-9 (0.1 mM) of series L and D in t-BuPh and a water/n-PrOH mixture as solvents at a temperature T using O2 as alkyl radical scavenger, and the subsequent activation energies Ea.   2) and an averaged frequency factor is used A = 2.4 10 14 s -1 , R = 8.314 J•K -1 •mol -1 , kd and T are given in columns 2, 3, 5, and 6, see reference [17].e In water/n-PrOH, v/v: 1:1.f Not soluble.g Not estimated.

Kinetic Analysis
All homolysis rate constants k d are collected in Table 1.Activation energies E a (given by Equation ( 2)) are around 130 kJ/mol as expected [25] for para-substituted amino-group-aryl-TEMPO-based alkoxyamines meaning that alkoxyamines are stable under our experimental conditions (t 1/2 > 170 day at 37 • C).As expected, E a does not depend on the configuration of the peptide chain.Moreover, no difference in E a is observed between diastereoisomers.
Table 1.Homolysis rate constants k d for 3-9 (0.1 mM) of series L and D in t-BuPh and a water/n-PrOH mixture as solvents at a temperature T using O 2 as alkyl radical scavenger, and the subsequent activation energies E a .2) and an averaged frequency factor is used A = 2.4 10 14 s −1 , R = 8.314 J•K −1 •mol −1 , k d and T are given in columns 2, 3, 5, and 6, see reference [17].e In water/n-PrOH, v/v: 1:1.f Not soluble.g Not estimated.

Cytotoxicity Activity
Due to technical problems, it was not possible to perform cytotoxicity in the same concentrations as those used for toxicity experiments on schistosomes.Nevertheless, the cytotoxicity (CC 50 > 50 µM) obtained on VERO cells after 48 h of treatment and reported for Plasmodium [12] shows that these molecules are rather inactive in mammalian cells [26].
Then, the approach developed in Scheme 1 may have involved different parts of the alkoxyamines in the inhibition processes both for toxicity and cytotoxicity.Hence, activity might be due to (a) the activity proposed in Scheme 1; (b) the toxicity of the released peptide (Scheme 7a); or (c) a non-radical activity of alkoxyamines which is mimicked with structures displayed in Scheme 7b.To address items (b) and (c), peptides P7L, P10L-P12L, and P18L and P21L-P25L were then prepared, respectively, as described in Schemes 4-6.

Cytotoxicity Activity
Due to technical problems, it was not possible to perform cytotoxicity in the s concentrations as those used for toxicity experiments on schistosomes.Nevertheless cytotoxicity (CC50 > 50 µM) obtained on VERO cells after 48 h of treatment and repo for Plasmodium [12] shows that these molecules are rather inactive in mammalian cells Then, the approach developed in Scheme 1 may have involved different parts of the al yamines in the inhibition processes both for toxicity and cytotoxicity.Hence, act might be due to (a) the activity proposed in Scheme 1; (b) the toxicity of the released tide (Scheme 7a); or (c) a non-radical activity of alkoxyamines which is mimicked structures displayed in Scheme 7b.To address items (b) and (c), peptides P7L, P10L-P and P18L and P21L-P25L were then prepared, respectively, as described in Scheme 4 Interestingly, all models from the L series, whether peptides alone or hybrid c pounds, are non-cytotoxic (cytotoxicity concentration CC50 > 50 µM) [12].Importa peptide-alkoxyamine hybrids in the natural series L are stable in our experimental co tions and not cytotoxic at 50 µM, confirming their specificity on schistosome.How peptides P8L and P9L, which display the benzylated end-carboxy function, show an expected toxicity contrary to P6L (Table S1).The latter is tentatively ascribed to an in tion effect due to a binding constant enhanced by the presence of the benzyl group on carboxy-terminal function.
Peptide-D models P6D, P8D, P21D, and P22D are unexpectedly toxic (Table whereas only P6D and P22D are also cytotoxic (Table S1).This activity is tentativel cribed to an inhibition effect due to the wrong D configuration.Alkoxyamines A4D, A5D (Table S1), which have been developed because eucaryote cells are not able to hy lyze peptides of D-series, are surprisingly both cytotoxic for VERO cells and toxic to s tosomes.We tentatively ascribed these unexpected results either to the inhibition o D-configuration of peptides (vide supra) or to the activity of bacteria in the digestive of schistosomes.Recent studies have evidenced the presence of bacteria interacting schistosomes which could lead to the development of new control strategies targeting crobiomes.[27] However, at this stage, no better reasons can be provided and these al yamines and peptides are not further discussed as they are not the focus of this articl

Schistosomicidal Activity
By analogy with plasmepsins of Plasmodium, which is also a hemoglobin-inges parasite, the peptide sequence Boc-Phe-Val-Phe can be considered as a potential subs for the schistosome's digestive enzymes.A3L-A7L should be hydrolyzed by schistos Interestingly, all models from the L series, whether peptides alone or hybrid compounds, are non-cytotoxic (cytotoxicity concentration CC 50 > 50 µM) [12].Importantly, peptide-alkoxyamine hybrids in the natural series L are stable in our experimental conditions and not cytotoxic at 50 µM, confirming their specificity on schistosome.However, peptides P8L and P9L, which display the benzylated end-carboxy function, show an unexpected toxicity contrary to P6L (Table S1).The latter is tentatively ascribed to an inhibition effect due to a binding constant enhanced by the presence of the benzyl group on the carboxy-terminal function.
Peptide-D models P6D, P8D, P21D, and P22D are unexpectedly toxic (Table S1) whereas only P6D and P22D are also cytotoxic (Table S1).This activity is tentatively ascribed to an inhibition effect due to the wrong D configuration.Alkoxyamines A4D, and A5D (Table S1), which have been developed because eucaryote cells are not able to hydrolyze peptides of D-series, are surprisingly both cytotoxic for VERO cells and toxic to schistosomes.We tentatively ascribed these unexpected results either to the inhibition of the D-configuration of peptides (vide supra) or to the activity of bacteria in the digestive tube of schistosomes.Recent studies have evidenced the presence of bacteria interacting with schistosomes which could lead to the development of new control strategies targeting microbiomes [27].However, at this stage, no better reasons can be provided and these alkoxyamines and peptides are not further discussed as they are not the focus of this article.

Schistosomicidal Activity
By analogy with plasmepsins of Plasmodium, which is also a hemoglobin-ingesting parasite, the peptide sequence Boc-Phe-Val-Phe can be considered as a potential substrate for the schistosome's digestive enzymes.A3L-A7L should be hydrolyzed by schistosome and should exhibit activities, while they should not display any activity in mammalian cells because they do not have any chymotrypsin-like secreted activity.Table 2 shows that A4L and A5L are the most efficient molecules, A3L is inactive, and A6L and A7L are less active.The inactivity of A3L can be explained by its poor solubility.In the context of the schistosome digestive tube at pH 5.5, A6L is not protonated until the Boc group is enzymatically removed and A7L is less protonated than at pH 7, thus also poorly soluble.A4L and A5L however are fully protonated at pH 5.5, thus more soluble.As mentioned above, peptide models are not cytotoxic for VERO cells.Surprisingly, peptide P10L displays some activity towards schistosomes which cannot be understood at this stage of our investigations.However, the cytotoxicity of P10L impacts only the discussion concerning AL7.On the other hand, and as expected, peptide models P21L-P25L carrying the phenylethyl moiety display no activity (closest non-reactive models of alkoxyamines A3L-A7L, that cannot produce alkyl radicals, see Scheme 7b and Table 2), which suggests that non-activated alkoxyamine are likely to be non-cytotoxic.Importantly, all alkoxyamines are thermally stable under our experimental conditions (see Table 1) meaning that the activities observed are likely due to their activation by schistosome enzymes.
None of the peptide-alkoxyamine hybrids displays any activity at concentrations of 10 µg mL.

Discussion
As mentioned above, peptide models are not cytotoxic for VERO cells.Then, matching the activities of peptide-alkoxyamine hybrids with the activities of peptides P11L and P22L for A4L, P18L and P25L for A5L, and P23L for A6L shows that all these peptides display no activity.Therefore, the activity observed is due to the alkoxyamine moiety.
The lower activity of A6L and the non-activity of A3L highlight the importance of the substituent at the N-terminal position since, namely, a Boc-group cancels the activity (see A3L in Table 2), Benzylated succinyl groups lower the activity, and a succinyl moiety affords a peptide 10L displaying some activity as in A7L.Nevertheless, A6L is able to kill schistosomes by releasing alkyl radicals as expected.
In sharp contrast to A6L, alkoxyamines A4L and A5L display a higher activity which is similar, i.e., the time to kill 50% of parasites is 2 h for A4L and A5L, vs. 24 h for A6L (Table 2).
Importantly, the free-amino-peptide-based alkoxyamine A5L and the TFE-salt-peptidebased alkoxyamine A4L display the same structure in water solution, i.e., the protonated amino-terminal group, and as expected both alkoxyamines display the same activity (Table 2) (For technical reasons, it was not possible to record the same after 8 h for A4L.Moreover, the discussion of the expected reactivity is spoiled by the unexpected activity of P8L).Therefore, these results highlight the need for positively charged NH 2 -terminal groups (Table 2) to observe the activity displayed in Scheme 1, which makes sense at pH 5.5.Consequently, alkoxyamines A4L-A6L nicely support the proposed strategy, as displayed in Schemes 1 and 8.
The lower activity of A6L and the non-activity of A3L highlight the importance of the substituent at the N-terminal position since, namely, a Boc-group cancels the activity (see A3L in Table 2), Benzylated succinyl groups lower the activity, and a succinyl moiety affords a peptide 10L displaying some activity as in A7L.Nevertheless, A6L is able to kill schistosomes by releasing alkyl radicals as expected.
In sharp contrast to A6L, alkoxyamines A4L and A5L display a higher activity which is similar, i.e., the time to kill 50% of parasites is 2 h for A4L and A5L, vs. 24 h for A6L (Table 2).
Importantly, the free-amino-peptide-based alkoxyamine A5L and the TFE-salt-peptide-based alkoxyamine A4L display the same structure in water solution, i.e., the protonated amino-terminal group, and as expected both alkoxyamines display the same activity (Table 2) (For technical reasons, it was not possible to record the same after 8 hours for A4L.Moreover, the discussion of the expected reactivity is spoiled by the unexpected activity of P8L).Therefore, these results highlight the need for positively charged NH2-terminal groups (Table 2) to observe the activity displayed in Scheme 1, which makes sense at pH 5.5.Consequently, alkoxyamines A4L-A6L nicely support the proposed strategy, as displayed in Schemes 1 and 8. Peptide-alkoxyamine hybrid activity targets the parasite's mandatory digestion cascade.This strategy has already been used with other molecules such as trioxaquines, which have shown both in vitro and in vivo activities [28].Our strategy considered that since there was a convergence with Plasmodium in the degradation chain from hemoglobin to hemozoin production, there could be a convergence in the actors involved in this proteolysis.This strategy proved effective, as our molecules are effective against schistosomes even though the plasmepsins are not present in their genome.This does not prevent convergent activity from a different family of proteases from occurring.Indeed, in a study comparing the genomic structures of the aspartic proteinase family genes, including the cathepsin D gene of schistosome, it was shown that there were exons that were common to both plasmepsins I, II, and IV from Plasmodium falciparum and cathepsin D from schistosome.Both plasmepsins I, II, and IV and cathepsin D have been evidenced to be major actors in hemoglobin degradation in Plasmodium [10] and schistosome [29], respectively.In vitro, the parasite's digestive tract probably contains only a few globules, which could lead to low protease secretion in the digestive tract and hence slow activation of the alkoxyamine prodrugs.We would expect better activity by using parasites fed with red blood cells in in vitro or in in vivo experiments.

Conclusions
Although the activity of A5L and A4L (time to kill 50% = 2 h) is lower than that of praziquantel (time to kill 50% < 1 h), our results highlight the potential of our approach (Scheme 1 and Scheme 8), i.e., that the Schistosoma parasite could dig its grave with its fork.That is, the mandatory feeding enzymatic activity of worms can be used to activate a prodrug able to release specifically in the desired target, due to the specific peptide, a non-selective and highly toxic free radical drug leading to the death of parasites.This original approach could lead to the development of antiparasitic compounds that are particularly safe for patients, with no or limited side effects.Peptide-alkoxyamine hybrid activity targets the parasite's mandatory digestion cascade.This strategy has already been used with other molecules such as trioxaquines, which have shown both in vitro and in vivo activities [28].Our strategy considered that since there was a convergence with Plasmodium in the degradation chain from hemoglobin to hemozoin production, there could be a convergence in the actors involved in this proteolysis.This strategy proved effective, as our molecules are effective against schistosomes even though the plasmepsins are not present in their genome.This does not prevent convergent activity from a different family of proteases from occurring.Indeed, in a study comparing the genomic structures of the aspartic proteinase family genes, including the cathepsin D gene of schistosome, it was shown that there were exons that were common to both plasmepsins I, II, and IV from Plasmodium falciparum and cathepsin D from schistosome.Both plasmepsins I, II, and IV and cathepsin D have been evidenced to be major actors in hemoglobin degradation in Plasmodium [10] and schistosome [29], respectively.In vitro, the parasite's digestive tract probably contains only a few globules, which could lead to low protease secretion in the digestive tract and hence slow activation of the alkoxyamine prodrugs.We would expect better activity by using parasites fed with red blood cells in in vitro or in in vivo experiments.

Conclusions
Although the activity of A5L and A4L (time to kill 50% = 2 h) is lower than that of praziquantel (time to kill 50% < 1 h), our results highlight the potential of our approach (Schemes 1 and 8), i.e., that the Schistosoma parasite could dig its grave with its fork.That is, the mandatory feeding enzymatic activity of worms can be used to activate a prodrug able to release specifically in the desired target, due to the specific peptide, a non-selective and highly toxic free radical drug leading to the death of parasites.This original approach could lead to the development of antiparasitic compounds that are particularly safe for patients, with no or limited side effects.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/pathogens13060482/s1,Table S1: Doses of alkoxyamines provided to Schistosoma at 37 • C, time to reach the death of half of the population of worms, time to kill all parasites, and their cytotoxicity on Vero cells (CC 50 in µM).Table S2: Biological activity/ raw data.Analatical data, NMR and HRMS data of all chemiscals prepared in the article are provided in SM.

Figure 1 .
Figure 1.Peptide-alkoxyamines investigated as anti-schistosomal drugs.L and D are for the configuration of the residues on the peptides as given by Fischer's rules.

Scheme 2 .
Scheme 2. Dynamic Covalent Bond in alkoxyamines: k d for the homolysis rate constant and k c for the re-formation reaction.

Scheme 8 .
Scheme 8. Scheme for the activation of alkoxyamine A5L by enzyme.

Scheme 8 .
Scheme 8. Scheme for the activation of alkoxyamine A5L by enzyme.

Table 2 .
Alkoxyamines and peptides are provided to schistosome at 37 • C at massic concentration (100 µg/mL) a , time to reach the death of half of the population of worms, and time to kill all parasites.