Synthesis and Evaluation of Antimycobacterial and Antiplasmodial Activities of Hirsutellide A and Its Analogues.

Hirsutellide A is nature-derived cyclic hexadepsipeptide with reported antimycobacterial and antiplasmodial activities. To verify its structure, hirsutellide A was synthesized following a solution-phase peptide synthesis approach. A detailed analysis of the 1H and 13C NMR spectra of the synthesized compound revealed structural variation from what had been originally assigned for hirsutellide A, despite the use of identical building blocks. This variation occurred at the two allo-Ile moieties. To investigate the structure-activity relationship, the depsipeptide and peptide analogues of hirsutellide A were prepared and tested for antimycobacterial and antiplasmodial activities. The compounds displayed antiplasmodial potency against Plasmodium falciparum 3D7 while showing weak or no activity against Mycobacterium tuberculosis H37Rv. The drug-likeness of the series was assessed through in vitro absorption, distribution, metabolism, and excretion (ADME) profiling, revealing systematic differences between the pharmacokinetic properties of cyclic hexapeptides and hexadepsipeptides.

Introduction

Due to the growing concerns over the emergence and spread of antibiotic-resistant microbes, the search for new chemical entities is intensifying.[1] In this regard, small-sized cyclic peptides offer good promise owing to their potential to target previously nondruggable drug targets.[2,3] Accordingly, a number of small-sized peptides, nature-derived or synthetically modified, are being investigated for their antimicrobial activities.[4−6] Recently, we have reported the anti-TB potential of wollamide B and its analogues by synthesizing and testing their activities against Mycobacterium tuberculosis (Mtb).[7,8] The compounds demonstrated potent antimycobacterial activities and interesting druglike attributes such as water solubility, plasma stability, and microsomal stability, while the membrane permeability still needs further improvement. The poor membrane permeability was found to be associated with the overall low lipophilicity of wollamides. Optimizing the structures of other previously reported lipophilic antimicrobial cyclic peptides could result in more membrane-permeable bioactive peptides. Additionally, this could help us figure out the limit of lipophilicity needed to design membrane-permeable small-sized cyclic peptides. Accordingly, in this work, we chose another more lipophilic antimicrobial depsipeptide (hirsutellide A) as a lead for further structural optimization.

Hirsutellide A (Figure ) is an 18-membered cyclic hexadepsipeptide that was isolated from the entomopathogenic fungus Hirsutella kobayasii BCC 1660.[9] The compound showed in vitro antimycobacterial (M. tuberculosis H37Ra, minimum inhibitory concentration (MIC) = 6–12 μg/mL) and antiplasmodial (Plasmodium falciparum, IC50 = 2.8 μg/mL) activities.[9] The structure of hirsutellide A was elucidated by analyzing its NMR spectra and using Marfey’s method. Its structure was then assigned to be a symmetrical cyclic hexadepsipeptide containing allo-isoleucine (allo-Ile), (R)-2-hydroxy-3-phenylpropanoic acid (HPPA), and sarcosine (Sar) residues.[9]

Figure 1

Structures of hirsutellide A (I) and reported synthetic stereoisomer (II).

The first attempt to synthesize hirsutellide A was reported in 2005 by Xu et al.[10] The method involved a stepwise construction of the linear hexadepsipeptide precursor in the solution phase followed by macrocyclization. However, due to the use of a “wrong” building block (i.e., Ile instead of allo-Ile), this resulted in the synthesis of one of its stereoisomers (II).[11] This stereoisomer (II) showed very similar physical and spectral properties to hirsutellide A, but when tested against M. tuberculosis (Mtb) H37Rv, its antimycobacterial potency was lost.[12]

Apart from the aforementioned study, the structure and biological activities of hirsutellide A were not confirmed through total synthesis. Moreover, to the best of our knowledge, its structure–activity relationship (SAR) remained unexplored.

The aim of the present work was, therefore, to synthetically verify the structure of the natural product, to synthesize analogues, and to evaluate the in vitro antiplasmodial and antimycobacterial activities. Aiming to assess the drug-likeness of the class, in vitro pharmacokinetic (absorption, distribution, metabolism, and excretion (ADME)) profiling and cytotoxicity assays were conducted.

Results and Discussion

Synthesis of Hirsutellide A

The synthesis of hirsutellide A was initiated using a solution-phase peptide synthesis approach following a method described by Xu et al., with some modifications.[10] The first step involved diazotization hydrolysis of d-Phe to the corresponding α-hydroxyl carboxylic acid (1) (Scheme ). After protecting the carboxy group of 1 as benzyl ester (2), it was coupled to N-Boc-Sar-OH using EDC*HCl/dimethylaminopyridine (DMAP) to obtain the fully protected didepsipeptide 3. This reaction took only 2 h for completion with a yield of 88%. The same reaction with the coupling agent dicyclohexylcarbodiimide (DCC) took 24 h with a lower yield (72%).[10,13] The Boc protecting group of the didepsipeptide was then removed via treatment with trifluoroacetic acid (TFA) (6.5 equiv) to give compound 4.

Scheme 1

Synthesis of the Protected Tridepsipeptide Precursor of Hirsutellide A

Reagents and conditions: (a) 40% NaNO2, H2SO4, 0 °C to rt, 20 h; (b) BnOH, pTsOH, reflux, 4 h, toluene; (c) EDC·HCl, DMAP, dichloromethane (DCM), 0 °C to rt, 2 h; (d) TFA, DCM, 0 °C to rt, 4 h; (e) 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU), N,N-diisopropylethylamine (DIPEA), DCM, 0 °C to rt, 12 h.

Coupling the didepsipeptide 4 with N-Boc-allo-Ile-OH using HATU/DIPEA/DCM gave the protected tridepsipeptide (5). After purification, compound 5 was divided into two equal portions. The first portion was subjected to catalytic hydrogenation, using Pd/C (10%), to remove the benzyl protecting group. This afforded compound 6. The Boc protecting group of the second portion was removed using TFA and yielded compound 7.

The synthesis of the linear hexadepsipeptide (8) was then achieved through a reaction that involved coupling of 6 and 7 using BOP-Cl/DIPEA (Scheme ).[14] The use of other coupling reagents such as HATU/DIPEA and benzotriazol-1-yl-N-oxy-tris(pyrrolidino)phosphonium hexafluorophosphate (PyBOP)/DIPEA gave very little product. The low yield associated with the use of these coupling reagents was related to the higher incidence of diketopiperazine formation.[15,16] This was confirmed here by isolating the diketopiperazine side product.

Scheme 2

Synthesis of the Hexadepsipeptide Precursor of Hirsutellide A and Macrolactamization

Reagents and conditions: (a) 10% Pd/C, H2 (5 atm), EtOAc, rt, 12 h; (b) TFA, DCM, 0 °C to rt, 4 h; (c) bis(2-oxo-3-oxazolidinyl)phosphinic chloride (BOP-Cl), DIPEA, DCM, 0 °C to rt, 18 h; (d) 10% Pd/C, H2 (5 atm), EtOAc, rt, 12 h; (e) TFA, DCM, 0 °C to rt, 5 h; (f) HATU, hydroxybenzotriazole (HOBt), DIPEA, dimethylformamide (DMF), 0 °C to rt, 3 days.

The subsequent reductive removal of the benzyl group and cleavage of the Boc group from 8 gave the linear hexadepsipeptide precursor. Finally, the desired cyclic depsipeptide (9) was obtained via macrocyclization of this precursor using HATU/HOBt/DIPEA in DMF at a high dilution (1 mM) and isolated yield of 50%. This is a significant improvement compared to the previously reported yield of macrocyclization (22%).[10] The purity and identity of the final compound were ascertained using high-resolution mass spectrometry (HRMS), ultraperformance liquid chromatography–mass spectrometry (UPLC–MS), and NMR.

Compound 9 Exhibited Different Spectral Properties from Hirsutellide A (I)

While displaying identical HRMS values, a detailed analysis of the 1H and 13C NMR spectra of 9 revealed differences from that reported for hirsutellide A.[9] The 1H chemical shifts of hirsutellide A, its previously reported synthetic stereoisomer (II),[11] and the newly synthesized compound (9) are listed in Table . All three compounds displayed similar chemical shifts for protons that belong to the 2-hydroxy-3-phenylpropionic acid and sarcosine moieties. However, for the two allo-Ile (Ile) moieties, while the natural product (I) and its stereoisomer (II) showed similar 1H chemical shifts, the newly synthesized compound (9) was slightly different. This particularly pertains to the δH of the two ethenyl protons on the 4′ positions (H-4′). 1H NMR spectra for hirsutellide A showed signals of the two H-4′ at δH = 1.19 and 1.55 ppm. This value was also similar for its stereoisomer (δH = 1.20 and 1.54 ppm). For the newly synthesized compound, the two H-4′ protons appeared to be more shielded and occurred at a higher field (δH = 1.05 and 1.36 ppm).

Table 1

Comparison of the 1H Chemical Shifts of Hirsutellide A, Stereoisomer (II) in the Literature, and the Newly Synthesized Compound (9)

		δH, multiplicity (J in Hz)a
structural unit		hirsutellide A (I)b9	streoisomer (II)b10	compound 9
2-hydroxy-3-phenyl- propanoic acid	1
	2	5.63, dd (11.8, 2.9)	5.61, dd (2.8, 11.6)	5.60, dd (11.7, 3.0)
	3	2.74, dd (14.0, 11.9)	2.73, dd (11.6, 14.0)	2.73, dd (14.2, 11.8)
		3,68, dd (14.0, 2.8)	3.66, dd (2.8, 14.0)	3.65, dd (14.2, 3.0)
	4
	5, 9	7.16, br d (7.0)	7.29–7.14, m	7.30–7.09, m
	6, 8	7.28, dd (7.0, 7.0)
	7	7.23, m
l-allo-isoleucine	1’
	2’	4.93, dd (10.1, 9.7)	4.91, t (10.2)	4.87, t (10.1)
	3’	2.24, m	2.24, m	2.22, m
	4’	1.19, m	1.20, m	1.05, m
		1.55, m	1.54, m	1.36, m
	5’	0.91, t (7.4)	0.90, t (7.4)	0.89, m
	6’	0.87, d (6.7)	0.85, d (6.8)
	NH	7.57, d (9.7)	7.54, d (10.0)	7.53, d (9.7)
sarcosine	1’’
	2’’	3.20, d (17.1)	3.17, d (17.2)	3.17, d (17.1)
		4.46, d (17.2)	4.44, d (17.2)	4.40, d (17.0)
	NMe	3.27, s	3.25, s	3.26, s

For all of the three compounds, 1H NMR spectra were measured in CDCl3 using a 400 MHz instrument.

Previously reported 1H NMR data.

Similarly for hirsutellide A and the reported stereoisomer, while the two methyl protons at positions 5′ and 6′ appeared separately as triplets and doublets, they were merged and appeared as multiplets for the synthesized compound (Figure ). In the 13C NMR spectra, a similar discrepancy was observed for the chemical shift values of C-2′ to C-6′ (Table S1).

Figure 2

1H NMR spectra of (A) hirsutellide A[9] and (B) compound 9 both measured in CDCl3 using a 400 MHz instrument.

The facts that (1) hirsutellide A was shown to have almost identical physical and spectral properties with the reported stereoisomer (II)[11] and (2) differences observed in the 1H and 13C NMR between hirsutellide A and the compound synthesized by incorporating allo-Ile (9) lead us to propose that hirsutellide A contains Ile rather than allo-Ile.

To further test our assumption, we synthesized two peptide analogues of hirsutellide A incorporating Ile (14) and allo-Ile (16). The two compounds exhibited the same chemical shift patterns for the Ile and allo-Ile ethenyl protons as hirsutellide A and 9, corroborating our proposed structure (Figure S11).

Moreover, the 1H NMR spectra of the desmethyl analogue of hirsutellide A (13) synthesized through incorporating Ile showed very similar chemical shifts for the two ethenyl protons with hirsutellide A, which further strengthened our assumption (Figure S6).

Synthesis of Depsipeptide and Peptide Analogues of Hirsutellide A for SAR Study

Depsipeptide (10–13) and peptide (14–20) analogues of hirsutellide A were designed and synthesized for preliminary SAR studies (Table ). While the depsipeptide analogues were synthesized following the approach used for 9, a standard Fmoc-based solid-phase peptide synthesis (SPPS) approach on 2-chlorotrityl chloride resin was utilized to construct the linear hexapeptide precursors of the cyclic peptide analogues, followed by macrocyclization in the solution phase.[7,8,17] Details of the synthesis protocol are described in the Supporting Information (Section S1).

Table 2

List of the Synthesized Depsipeptide and Peptide Analogues of Hirsutellide A

cpd#	I	II	III	X	R
9	allo-Ile	d-Phe	Gly	–O–	–CH3
10	Leu	d-Phe	Gly	–O–	–CH3
11	Val	d-Phe	Gly	–O–	–CH3
12	Gly	d-Phe	Gly	–O–	–CH3
13	Ile	d-Phe	Gly	–O–	–H
14	Ile	d-Phe	Gly	–NH–	–CH3
15	Val	d-Phe	Gly	–NH–	–CH3
16	allo-Ile	d-Phe	Gly	–NH–	–CH3
17	allo-Ile	d-Phe (4-Cl)	Gly	–NH–	–CH3
18	allo-Ile	d-Phe (4-OCH3)	Gly	–NH–	–CH3
19	allo-Ile	Leu	Gly	–NH–	–CH3
20	allo-Ile	d-Phe	Pro	–NH–	–CH3

The depsipeptide analogues (10–12) were designed to assess the effect on the activity of replacing the aliphatic side chains of the two Ile moieties. Ile was replaced with other lipophilic amino acids including, allo-Ile, Leu, Val, and Gly.

The desmethyl analogue 13 was prepared to see the effect of removing an N-methyl group. This was done by replacing the two Sar moieties with Gly.

The peptide analogues (14–20) were synthesized for the purpose of comparing their metabolic stability and potency with the depsipeptides.[18,19] Compound 15 was the peptide analogue of hirsutellide A that was synthesized through the exchange of (R)-2-hydroxy-3-phenylpropanoic acid moieties with d-Phe. Compounds 17 and 18 were synthesized by introducing d-Phe analogues whose phenyl side chains were blocked with either chlorine (17) or a methoxy group (18) at the para position. This was done to slow down metabolic clearances. Similarly, to investigate the importance of the aromatic side chains for the observed antimycobacterial activities, the two d-Phe were exchanged by the aliphatic amino acid d-Leu (19). The two Sar units of hirsutellide A were also replaced with Pro (10) to evaluate the effect of a pronounced change of the conformation (Schemes and ).

Scheme 3

Structure of Hirsutellide A

Scheme 4

General Structural Feature of the Target Cyclic (Depsi)Peptides

Biological Activities and ADME Profiles of Hirsutellide A (9)

To compare with what was reported for hirsutellide A, in vitro bioactivity testing was conducted for compound 9. The antimycobacterial activity was assessed by determining the MIC against standard drug-susceptible Mtb H37Rv and Mycobacterium vaccae using a microplate alamar blue broth dilution assay.[20] The antiplasmodial activity was tested against the chloroquine (CQ)-sensitive P. falciparum strain 3D7,[21] and the result was expressed as IC50 values, representing the concentration required to inhibit the growth of the parasite by 50% as indicated by the in vitro uptake of [3H]-hypoxanthine by P. falciparum.[22] In addition, cytotoxicity was assessed using human hepatocellular carcinoma cells (Hep2G). The results of the tests are displayed in Table .

Table 3

Biological Activities and ADME Profile of Compound 9

bioactivity/ADME profile of compound 9	values
antimicrobial activities
MIC (Mtb H73Rv)	>80 μM
MIC (M. vaccae)	18.8 μM
IC50 (P. falciparum 3D7)	2.3 μM
cytotoxicity
IC50 (Hep2G cells)	>100 μM
membrane permeabilitya	280 nm/s
% human serum albumin binding	88%
intrinsic microsomal clearance (Clint)
human	46.53 mL/(min g) tissue
mouse	38.3 mL/(min g) tissue
plasma stability (% remain after 2 h incubation)b
human plasma	103%
mouse plasma	18%

Passive artificial membrane permeability.

Estimated by quantifying the amount of a compound remaining after incubation with human and mouse plasma for 2 h at 37 °C.

Parallel to bioactivity testing, in vitro ADME profiling was also conducted to evaluate its druglike properties.[23] The ADME profiling included (1) a study of physicochemical properties, viz., passive artificial membrane permeability, human serum albumin binding affinity, and lipophilicity; (2) determination of plasma stability after 2 h incubation of the compound with human and mouse plasma; and (3) metabolic stability, expressed as intrinsic clearance of the compound in both mouse and human liver microsomes. The results are also included in Table .

While being moderately active against the nonpathogenic Mycobacterium vaccae (MIC of 18.8 μM), compound 9 failed to show activity against Mtb H37Rv (MIC > 80 μM). This finding does not corroborate the anti-Mtb activity previously reported for hirsutellide A (MIC against Mtb H37Ra, 6–12 μg/mL). The stereoisomer (II) was also reported to be inactive against Mtb H37Rv.[12] We ascribe the discrepancy to the difference of the mycobacterial strains. While the natural product had been tested against an attenuated form of Mtb H37Ra, compound 9 and II were tested against the virulent strain, Mtb H37Rv. The virulent strain was either intrinsically unsusceptible or able to somehow inactivate or pump out hirsutellide A and analogues. This serves as a note of caution toward a recent study that found both strains to be equally susceptible to standard antitubercular agents, including the cyclopeptide capreomycin.[24]

In contrast, compound 9 displayed similar antiplasmodial potency (IC50 = 2.3 μM) to that reported for hirsutellide A (IC50 = 4.2 μM), reconfirming the (moderate) antiplasmodial activity of the natural product. The compound showed a dose-dependent inhibition of P. falciparum, as shown in Figure . However, the antiplasmodial potency was still weak compared to chloroquine, emphasizing the need for structural modification to perhaps turn it into a better antimalarial lead.

Figure 3

Plot showing dose-dependent inhibition of P. falciparum 3D7 by compound 9. The error bars represent standard error, n = 9.

Coassessing the toxicity and ADME profiles with bioactivity assay is important in early hit-to-lead optimization phase as it helps reduce the attrition rate in later stages of drug discovery.[25−27] Accordingly, while testing its activities, ADME profiling and cytotoxicity assay were conducted. An assay against Hep2G cells proved that compound 9 did not exhibit cytotoxicity (IC50 > 100 μM).

Similarly, ADME profiling revealed some desirable druglike attributes. Among them was its very good permeability through an artificial membrane (280 nm/s). This is quite interesting as poor membrane permeability is one of the major hurdles in the design of peptide-related drugs.[28,29] The structural features that seem to be responsible for the observed good permeability of 9 are the two depsipeptide (ester) bonds and the lipophilic amino acid side chains. The other relevant feature of compound 9 was its excellent stability in human plasma, as indicated by 100% recovery after incubation for 2 h. However, in mouse plasma, it showed poor stability (18% recovery) because of the different types of plasma esterases in the two species. Similarly, compound 9 showed moderate stability in human but low stability in mouse microsomes (Table ). These findings underscore the need to carefully interpret in vivo drug bioactivity assays when done in mice models as a lack of activity might have resulted from poor plasma and/or microsomal stability.

Biological Activities and ADME Profiles of Depsipeptide and Peptide Analogues of Hirsutellide A

The antimicrobial tests and cytotoxicity assay of the newly synthesized depsipeptide and peptide analogues of hirsutellide A were conducted in the same way as for compound hirsutellide A (9). For a comprehensive comparison, ADME profiling was also done for some selected depsipeptide and peptide analogues. The results of the tests are presented in Tables –7.

Table 4

Antimycobacterial and Antiplasmodial Activities and Cytotoxicities of Depsipeptide and Peptide Analogues of Hirsutellide Aa

cpd#	MIC(Mtb H37Rv) (μM)	IC50 (P. falciparum 3D7) (μM)	IC50 (Hep2G Cells) (μM)
11	>80	7.5	>100
12	>80	20.1	>100
13	>80	13.8	>100
15	>80	7.7	>100
16	>80	2.2	>100
17	40	1.8	26
18	>80	4.1	>100
20	>80	2.0	>100
INH	2–4	nd	nd
CQ	nd	0.005	nd

INH = isoniazide, CQ = chloroquine, nd = not determined.

Table 7

In Vitro Plasma Stabilities of Selected Peptide and Depsipeptide Analogues of Hirsutellide A

	plasma stability (% remaining)a
cpd#	human	mouse
12	67	0
13	82	0
19	90	88
20	110	100

Estimated by quantifying the amount of a compound remaining after incubation with plasma for 2 h at 37 °C.

Similar to hirsutellide A (9), its analogues also showed poor or no activity against Mtb H37Rv. Only compound 17 showed weak activity with MIC = 40 μM. The antiplasmodial activities, on the other hand, were retained moderately and comparable to 9 (Table ). Modifications of the exocyclic substituents of the depsipeptide ring such as replacement of allo-Ile with Val (11), Gly (12), and demethylation of the N-methyl sarcosine (13) reduced the antiplasmodial activity. Peptide analogues of hirsutellide A were generally more active against Plasmodium species (IC50 = 1.8–7.7 μM) than the depsipeptide analogues (IC50 = 7.5–20.1 μM). As observed for 9, the inhibition of P. falciparum growth occurred in a dose-dependent manner (Figure ).

Figure 4

Plots showing dose-dependent inhibition of P. falciparum 3D7 by compounds 16 (a), 17 (b), 18 (c), and 20 (d). The error bars represent standard error, n = 9.

The ADME profiling revealed some differences between depsipeptide and peptide analogues.

In general, compared to the peptide analogues, the depsipeptides exhibited improved microsomal stabilities in both human and mouse plasmas, though the overall microsomal stability of the class in general is still considered low (Table ). As noted for compound 9, the depsipetide bond reduced the aqueous solubility and plasma stabilities, especially against mouse plasma (Tables and 7). Within the depsipeptide class, exchange of the allo-Ile moiety (in compound 9) with Gly (12) significantly lowered plasma stability (67%) probably because the alkyl side chain of allo-Ile sterically shielded against plasma esterases. The removal of the N-methyl groups (13) reduced the plasma stability (82%), highlighting the importance of N-methylation of peptides to improve their metabolic stability.[17,30]

Table 6

In Vitro Microsomal Stabilities of Selected Peptide and Depsipeptide Analogues of Hirsutellide A

	Clint (mL/(min g) tissue)a
cpd#	mouse	human
12	52.00	34.00
13	65.90	32.80
15	7.52	3.60
16	47.71	18.90
18	13.25	8.18
19	5.79	6.07
20	8.83	7.14

Clint, in vitro intrinsic microsomal clearance.

Table 5

Physicochemical Properties of Selected Depsipeptide and Peptide Analogues of Hirsutellide A (9)

cpd#	solubilitya(μg/mL)	permeabilityb,c (nm/s)	%HSA
11	80	560	nd
12	81	150	54.7
13	6	470	91.02
17	34	330	95.24
18	124	290	77.2
19	≥184	380	37.15
20	≥254	330	84.22

Kinetic aqueous solubility.

Passive artificial membrane permeability; HSA, human serum albumin binding; nd, not determined.

GSK cutoff values for solubility:[31] >100 μg/mL, high solubility; 30–100 μg/mL, moderate solubility; <30 μg/mL, low solubility. *GSK cutoff values for permeability:[31] >200 nm/s, high permeability; 10–200 nm/s, intermediate permeability; <10 nm/s, low permeability.

The peptide analogues (9 and 10) generally exhibited higher aqueous solubility (>180 μg/mL) than the depsipeptide analogues. Exchanging the two ester groups for amides was found to enhance mouse plasma stabilities. In addition to their excellent human plasma stability (≥90%), the peptide analogues 19 and 20 also exhibited improved mouse plasma stabilities (88 and 100%, respectively).

In addition to the ester-to-amide substitution, the membrane permeability of hirsutellide A analogues seems also to be dependent on the nature of the amino acid substituents. The more lipophilic the side chains are, the better is their membrane permeability. Similarly, as the depsipeptides, the peptide analogues exhibited a wide range of binding affinities to the human serum albumin. Within the peptide analogues, while the introduction of para-chloro substituents to the two phenyl groups of d-Phe (17) increased their affinity (95%), the removal of the aromatic side chains (19) significantly lowered their albumin binding (37%).

Conclusions

Hirsutellide A was synthesized using a solution-phase peptide synthesis approach. A comparison of the 1H and 13C NMR spectra of the synthesized compound with the natural product revealed slight variations in the chemical shifts that belong to the two allo-Ile moieties. A thorough analysis of NMR spectra showed that hirsutellide A contains Ile rather than allo-Ile, as had been previously reported. Depsipeptide and peptide analogues of hirsutellide A were also synthesized. While the compounds retained the reported antiplasmodial activity of hirsutellide A, the antimycobacterial activity of the natural product was not reproduced. A discrepancy in the antimycobacterial activity of hirsutellide A and the synthesized analogues could be explained by the use of different mycobacterial strains (Mtb H37Ra vs Mtb H37Rv). In vitro ADME profiling proved the drug-likeness of the synthesized depsipeptide and peptide analogues and provided some rationale for differences. The findings revealed structural features, especially for membrane permeability and stability that are relevant for the future design of cyclic peptide-related drugs.

Experimental Section

Chemistry

General

All of the chemicals and reagents used for synthesis and purification were purchased from a commercial source and used without further purification. A polypropylene syringe (Inject, B.Braun Melsungen AG, Germany) fitted with a frits column plate was used for the solid-phase synthesis of linear hexapeptides. Whenever needed, analytical thin-layer chromatography (TLC) (Silica gel 60 F254, Merck, Germany) was used to monitor the reaction progress. The purification of the synthesized compounds was done on column chromatography using silica gel 60 (0.063–0.200 mm, Merck) as a stationary phase. Electrospray ionization mass spectra (ESIMS) were acquired using an SSQ 710 C mass spectrometer, Finningen MAT GmbH, Bremen, Germany. NMR spectra were recorded on a Bruker Varian Inova Unity 500 (operating frequency, 300 or 500 MHz for 1H NMR and 101 or 126 MHz for 13C NMR), Kolthoff GmbH, Filsum, Germany. Proton and carbon chemical shifts are reported in ppm. 1H NMR spectral data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad, ovlp = overlapping), coupling constant in hertz, and integration. The purities of the final target compounds were assessed by an independent quality control team at GlaxoSmithKline (GSK), Stevenage, using UPLC–MS with a UV diode array or an evaporative light scattering detector (ELSD), and MS was used to confirm identity. Generally, the purity of the synthesized final compounds was found to be >95%.

Synthesis of the Target Cyclic Hexadepsipeptides

Synthesis of Compound 9

(R)-2-Hydroxy-3-phenylpropanoic Acid (1)

NaNO2 (12 mL, 40%) was added to an ice-cooled solution of d-phenylalanine (6.6 g, 40 mmol) in H2SO4 (1 N, 45 mL) over a period of 1 h. The reaction mixture was continuously stirred at 0 °C for 8 h and kept at room temperature for an additional 12 h. To push the reaction to completion, additional H2SO4 (1 N, 10 mL) and 40% NaNO2 (6 mL) were added in similar conditions to above and were continuously stirred for a further 8 h. At the end of the reaction, the reaction mixture was saturated with NaCl and extracted with EtOAc (3 × 60 mL). The combined EtOAc extract was washed with brine, dried with MgSO4, filtered, and evaporated to give a white crude solid mass, which was recrystallized from diethyl ether to give the target compound. Yield (68%), R = 0.64 in MeOH, [α]D20 = +20.4 (c 1, ethanol), 1H NMR (400 MHz, MeOH-d4) δ 7.32–7.13 (m, 5H), 4.33 (dd, J = 8.0, 4.4 Hz, 1H), 3.09 (dd, J = 13.9, 4.4 Hz, 1H), 2.89 (dd, J = 13.9, 8.0 Hz, 1H). ESIMS m/z calcd for C9H10O3: 166.17; found: 165.22 [M – H]−.

(R)-Benzyl 2-Hydroxy-3-phenylpropanoate (2)

Compound 1 (1.66 g, 10 mmol), p-toluenesulfonic acid monohydrate (0.069 g, 0.40 mmol), benzyl alcohol (1.6 mL, 15 mmol), and toluene (40 mL) were placed into a 100 mL round-bottom flask equipped with a Stark and Dean distilling receiver. The reaction mixture was heated under reflux (250 °C) for 4 h until water was no longer distilled off. After the reaction was completed, the mixture was allowed to cool at room temperature and 30 mL of ether and 50 mL of saturated Na2CO3 were added to set the organic and aqueous phases. The organic layer was separated, and the aqueous layer was washed with ether (2 × 30 mL). The combined ether extract was again washed with 100 mL of brine, dried with MgSO4, and filtered, and the solvent was evaporated under reduced pressure to get an oily crude product. The crude product was then purified via column chromatography on silica gel using petroleum ether (40–60) and EtOAc (10:1) as a mobile phase to get a faint yellow oily product. Yield (77%), R = 0.11 in EtOAc /heptanes (1:1), 1H NMR (400 MHz, CDCl3-d) δ 7.47–7.11 (m, 10H), 5.20 (t, J = 1.9 Hz, 2H), 4.51 (ddd, J = 6.4, 4.7, 1.5 Hz, 1H), 3.15 (ddd, J = 14.0, 4.7, 1.6 Hz, 1H), 3.07–2.96 (m, 1H), 2.95–2.71 (bs, 1H). ESIMS m/z calcd for C16H16O3: 256.30; found: 279.03 [M + Na]+.

Synthesis of the Didepsipeptide (3)

N-Me-Boc-Gly-OH (0.84 g, 4.45 mmol), 2 (1.14 g, 4.45 mmol), and DMAP (0.05 g, 4.45 mmol) were dissolved in 25 mL of dry DCM and allowed to cool to 0 °C while stirring. EDC·HCl (1.02 g, 5.34 mmol) was added to the mixture, and the whole reaction mixture was continued to be stirred for 2 h at room temperature. After the completion of the reaction, the contents were transferred to a separatory funnel. Three layers were formed when water (63 mL) and EtOAc (84 mL) were added to the separatory funnel. Upon shaking, the dichloromethane portion that sank down to the aqueous layer was reextracted to the EtOAc layer to form an organic and aqueous phase. The organic phase was then separated, and the aqueous layer was reextracted twice with 50 mL of EtOAc. Finally, the combined organic extract was saturated with KHSO4, washed with brine, dried by MgSO4, and the solvent was removed under reduced pressure. The obtained oily mass was purified by column chromatography using silica gel as a stationary phase and petroleum ether/ EtOAc (1:1) as a mobile phase. Yield (88%), R = 0.51 in EtOAc/hexane (1:1), 1H NMR (400 MHz, CDCl3-d) δ 7.25 (m, 10H), 5.33 (dt, J = 9.0, 4.3 Hz, 1H), 5.21–5.04 (m, 2H), 4.14–3.96 (m, 1H), 3.93 (s, 1H), 3.15 (dtd, J = 22.4, 14.3, 8.9 Hz, 2H), 2.81, 2.79 (s, 3H, rotamers), 1.47, 1.33 (s, 9H, rotamers). ESIMS m/z calcd for C24H29NO6: 427.49; found: 450.02 [M + Na]+.

Boc Deprotected Didepsipeptide (4)

Compound 3 (1.73 g, 4.04 mmol) was dissolved in dry DCM (6 mL) and allowed to cool to 0 °C while stirring. TFA (2 mL, 26.04 mmol) was added dropwise, and the reaction was slowly allowed to warm to room temperature and stirred for 4 h. The product was concentrated under a high vacuum pressure to ensure complete removal of TFA. The obtained brownish crude mass was used directly for the next reaction without purification.

Synthesis of Tridesipetide (5)

N-Boc-allo-Ile-OH (0.86 g, 4.4 mmol) and HATU (1.84 g, 4.8 mmol) were added to a solution of 4 (1.1 g, 3.4 mmol) in dry DCM (30 mL). The mixture was then cooled to 0 °C, and DIPEA (6.0 mL, 34.0 mmol) was added dropwise. The reaction was slowly allowed to warm to room temperature and was stirred for 12 h. It was then quenched by adding saturated ammonium chloride (2 mL) dropwise, which was followed by the removal of the solvent under reduced pressure. The residue was then dissolved in 120 mL of EtOAc, and the organic phase was washed with saturated NaHCO3 (2 × 100 mL) and brine (100 mL), dried with MgSO4, and filtered, and the solvent was evaporated under reduced pressure. The obtained brownish-yellow oily mass was then purified via column chromatography using EtOAc/heptane (3:2) as a mobile phase and silica gel as a stationary phase. Yield (88%), R =0.48 in EtOAc/heptane (1:1), ESIMS m/z calcd for C30H40N2O7: 540.65; found: 563.08 [M + Na]+.

Synthesis of Benzyl Deprotected Tridepsipeptide (6)

Compound 5 (0.89 g, 1.65 mmol) was dissolved in EtOAc (80 mL) and kept in a hydrogenator reaction vessel. Pd (0.44 g, 10%) catalyst was added to the solution to form a suspension. The reaction mixture was then subjected to 5 atm hydrogen gas and stirred overnight. After the end of the reaction, the catalyst was removed by filtration and the filtrate was concentrated in vacuum to give 6. Yield (92%), ESIMS m/z calcd for C23H34N2O7: 450.53; found: 449.23 [M – H]−.

Synthesis of Boc Deprotected Tridepsipetide (7)

Compound 5 (0.7 g, 1.3 mmol) was dissolved in dry DCM (10 mL) and allowed to cool to 0 °C while stirring. TFA (2.0 mL, 26.14 mmol) was added dropwise, and the reaction was slowly allowed to warm to room temperature and was stirred for 4 h. The product was concentrated under a high vacuum pressure to ensure complete removal of TFA. The obtained brownish crude mass was used directly for the next reaction without purification.

Synthesis of Fully Protected Hexadepsipeptide (8)

Compound 6 (0.57 g, 1.3 mmol) and BOP-Cl (0.36 g, 1.43 mmol) were added to a solution of 7 (0.56 g, 1.3 mmol) in dry DCM (30 mL). The mixture was then allowed to cool to 0 °C, and DIPEA (2.1 mL, 13.0 mmol) was added dropwise. The reaction was slowly allowed to warm to room temperature and stirred for 18 h. It was then quenched by adding saturated ammonium chloride (2.5 mL) dropwise. The solvent was evaporated under reduced pressure, and the residue was dissolved in 50 mL of EtOAc under vigorous stirring. Saturated NaHCO3 (50 mL) was added to the flask and continuously stirred until all of the contents dissolved. The content was then transferred to a 250 mL separatory flask. The EtOAc layer was again washed with 50 mL of saturated NaHCO3 and then with brine (50 mL), dried with MgSO4, and filtered, and the solvent was evaporated under reduced pressure. The obtained brownish-yellow oily mass was then purified by column chromatography using heptane/EtOAc (3:2) as a mobile phase and silica gel as a stationary phase. Yield (45%), R = 0.13 in heptane/EtOAc (3:2), ESIMS m/z calcd for C48H64N4O11: 873.04; found: 896.15 [M + Na]+.

Removal of the Benzyl Protecting Groups from 8

Compound 8 (0.39 g, 0.44 mmol) was dissolved in EtOAc (60 mL) and kept in a hydrogenator reaction vessel. A 10% Pd (0.2 g) catalyst was added to the solution to form a suspension. The reaction mixture was subjected to 5 atm hydrogen gas and stirred overnight. Afterward, the catalyst was removed by filtration and the filtrate was concentrated in vacuum; yield: 92%. It was then used directly for the next reaction.

Removal of the Boc Protecting Groups from 8

The compound obtained in the previuos step (0.32 g, 0.4 mmol) was dissolved in dry DCM (8 mL) and allowed to cool to 0 °C while stirring. TFA (2.0 mL, 26.14 mmol) was added dropwise, and the reaction was slowly allowed to warm to room temperature and stirred for 5 h. The product was concentrated under strong vacuum pressure to ensure complete removal of TFA. The obtained brownish crude mass was used directly for the next reaction without purification.

Synthesis of Cyclic Hexadepsipeptide (9)

HATU (0.46 g, 1.2 mmol) and HOBt (0.16 g, 1.2 mmol) were added to a solution of the obtained crude mass in the above step (0.27 g, 0.4 mmol) in DMF (195 mL). The mixture was then allowed to cool to 0 °C, and DIPEA (0.7 mL, 4.0 mmol) was added dropwise. The reaction was slowly allowed to warm to room temperature and stirred for 3 days. Following that, the solvent was concentrated under reduced pressure, the residue was dissolved in 50 mL of EtOAc, washed by saturated NaHCO3 (2 × 100 mL) and brine (100 mL), dried with MgSO4, and filtered, and the solvent was removed under reduced pressure. The obtained brownish-yellowish solid mass was then purified by column chromatography using heptane/EtOAc (1:1) as a mobile phase and silica gel as a stationary phase. Yield (50%). RT 1.29 min, mp 254–257 °C, UPLC–MS (UV) purity: 98%, 1H NMR (400 MHz, CDCl3-d) δ 7.53 (d, J = 9.7 Hz, 2H), 7.31–7.08 (m, 10H), 5.60 (dd, J = 11.7, 3.0 Hz, 2H), 4.87 (t, J = 10.1 Hz, 2H), 4.40 (d, J = 17.0 Hz, 2H), 3.65 (dd, J = 14.2, 3.0 Hz, 2H), 3.26 (s, 6H), 3.17 (d, J = 17.1 Hz, 2H), 2.73 (dd, J = 14.2, 11.8 Hz, 2H), 2.28–2.15 (m, 2H), 1.42–1.32 (m, 2H), 1.10–1.01 (m, 2H), 0.94–0.86 (m, 12H). 13C NMR (101 MHz, cdcl3) δ 174.18, 168.82, 166.76, 136.13, 129.07, 128.56, 127.10, 74.10, 52.95, 51.88, 38.72, 37.89, 36.39, 25.96, 14.32, 11.05. HRMS (ESI) m/z: [M + H]+ calculated for C36H49N4O8: 665.3550; found: 665.3535.

Synthesis of Compounds 10–13

Syntheses of the depsipeptides 10–13 were achieved following a similar methodology used for the synthesis of compound 9. Physical properties data, purity assay, and spectral characteristics of these target compounds are included in the Supporting Information (Section S2).

Synthesis of the Target Cyclic Hexapeptides (14–20)

The target cyclic hexapeptides were synthesized in two steps. First, the linear hexapeptide precursors were assembled on 2-chlorotrityl chloride resin support using the solid-phase peptide synthesis (SPPS) approach, according to a protocol described by Chatterjee et al.[17]

Following this, macrocyclization of these precursors was done in the solution phase under a highly diluted environment. This was also done according to a protocol previously reported by Asfaw et al.[7] Details regarding the synthesis protocols are also described in the Supporting Information (Section S1). The structures of the final compounds were confirmed by a combination of ESIMS and NMR spectral analyses, and their purities were asserted using UPLC–MS.

Biological Assay

Mtb Inhibition Assay

Bacterial Cells and Culture Media

M. tuberculosis H37Rv (ATCC 118138) was routinely grown in 7H9 broth (Difco Middlebrook) supplemented with 10% (v/v) ADC (5% bovine albumin fraction, 2% dextrose, 0.004% catalase, and 0.8% sodium chloride solution) and 0.025% (v/v) Tween 80 at 37 °C in standing cultures.

MIC Determination

The MIC determination of the target compounds was performed in 96-well flat-bottom polystyrene microtiter plates. First, the test compounds were prepared in 10 twofold dilutions in dimethyl sulfoxide (DMSO) using V-bottom microtiter wells. Similarly, the positive control isoniazide was prepared in eight twofold dilutions, starting at a concentration of 160 μg/mL. Following this, 5 μL of the drug and test solutions were transferred to 95 μL of Middlebrook 7H9 medium (Difco catalog ref 271310). Similarly, 5 μL of the vehicle (DMSO) was used as the growth and blank controls. The final inoculum was prepared by 1 to 100 dilution of prestandardized inoculum (approximately 1 × 107 CFU/mL) in Middlebrook 7H9 broth (10% ADC; Becton Dickinson BBL catalogue ref 211887) and 0.025% Tween 80. This inoculum (100 μL) was then transferred to the entire plate, except for the blank controls. The plates were then incubated for 6 days at 37 °C after being sealed in a box to prevent drying out of the peripheral wells. At the end of the incubation period, a resazurin tablet (Resazurin Tablets for Milk Testing; ref 330884Y, VWR International Ltd.) was dissolved in sterile phosphate-buffered saline, and 25 μL of this solution was added to each well and the plates were then further incubated for 48 h. Endpoint readout was done by measuring fluorescence using Spectramax M5, Molecular Devices (excitation 530 nm, emission 590 nm). For each compound, the average value of the duplicate samples was calculated.

Antiplasmodial Assay

P. falciparum Strain Culture

The chloroquine-sensitive 3D7 strain of P. falciparum was originally obtained from Parasitology Unit, Center for Infectious Diseases, University of Heidelberg, Germany, and were cultivated using the previously reported standard protocols.[32]

Determination of IC50 Values

The drug susceptibility of P. falciparum was studied using a modified method[33] of the protocol described previously for the 3H-hypoxanthine incorporation-based assay.[22] All assays included CQ diphosphate as standard and control wells with untreated infected and uninfected erythrocytes. The IC50 values were derived by sigmoidal regression analysis (Microsoft xlfit). All data from in vitro tests were recorded three times in triplicate and were given with 95% confidence limits.

Hep2G Cytotoxicity Assay

Hep2G cells obtained from Biocat 118138 were maintained in Eagle’s minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS), 1% nonessential amino acids (NEAA), and 1% penicillin/streptomycin in a 5% CO2 incubator at 37 °C. The cells were removed from a T-175 TC flask using the same medium and adjusted to a final density of 1.2 × 105 cells/mL. Using a Multidrop instrument, 25 μL of this cell suspension (3000 cells per well) was dispensed into the wells of 384-well clear-bottom plates. Prior to addition of the cell suspension, the screening compounds (250 nL) were dispensed into the plates with an Echo 555 instrument. The plates were then incubated at 37 °C for 48 h under 5% CO2 and relative humidity of 80%. At the end of incubation, the plates were equilibrated for 30 min before the addition of 25 μL (in each well) of a luminescent signal developer (CellTiter-Glo, Promega). The plates were then left for 10 min at room temperature for stabilization and were subsequently read using a ViewLux instrument (PerkinElmer). For each compound, the average value of the duplicate samples was calculated.

The human biological samples were sourced ethically and their research use was in accordance with the terms of the informed consents. The use of human biological samples was approved by human biological samples management (HBSM) core team of GSK and conforms to HBSM GSK policies.

ADME Pharmacokinetic Profiling

Physicochemical properties including aqueous kinetic solubility, passive artificial membrane permeability, human serum albumin (HSA) binding, and microsomal and plasma stabilities of the synthesized compounds were determined according to the methods we previously reported.[7]