Veraguamides A–G, Cyclic Hexadepsipeptides from a Dolastatin 16-Producing Cyanobacterium Symploca cf. hydnoides from Guam

Abstract

Cytotoxicity-directed purification of a Symploca cf. hydnoides sample from Cetti Bay, Guam, afforded seven new cyclic depsipeptides, veraguamides A–G (1–7), together with the known compound dolastatin 16. The planar structures of 1–7 were elucidated using NMR and MS experiments, while enantioselective HPLC and Mosher's analysis of acid and base hydrolysates, respectively, were utilized to assign the absolute configurations of the stereocenters. Veraguamides A–G (1–7) are characterized by the presence of an invariant proline residue, multiple N-methylated amino acids, an α-hydroxy acid, and a C₈-polyketide derived β-hydroxy acid moiety with a characteristic terminus as either an alkynyl bromide, alkyne, or vinyl group. These compounds and a semisynthetic analog (8) showed moderate to weak cytotoxic activity against HT29 colorectal adenocarcinoma and HeLa cervical carcinoma cell lines. Preliminary structure-activity relationship analysis identified several sensitive positions in the veraguamide scaffold that affect the cytotoxic activity of this compound class. Dolastatin 16 showed only weak cytotoxic activity on both cell lines tested. The complete stereostructure of dolastatin 16 was proposed for the first time through degradation followed by a combination of advanced Marfey's analysis and modified Mosher's analysis using phenylglycine methyl ester as a chiral anisotropic reagent.

Marine cyanobacteria have provided both structurally diverse and potent cytotoxic compounds with varying mechanisms of action as lead structures for drug discovery. Most of these natural products are peptide–polyketide hybrids and are derived from the genus Lyngbya, marked by N-methylation, modified amino acids, α- and β-hydroxy acids, and fatty acid-type moieties.^1,2 We have had particular success with our drug discovery efforts based on the genus Symploca. From Micronesian collections of Symploca spp. we previously isolated the linear peptides dolastatin 10 and its analog symplostatin 1.^3,4 These compounds are microtubule disrupting agents that interfere with mitosis and have pico- to nanomolar IC₅₀ values against cancer cells.⁵ From a Floridian Symploca sp. collection we recently discovered the HDAC inhibitor, largazole, which showed strong cytotoxic activity against cancer cells with superior selectivity over nontransformed cells.^6,7 In this light, we screened several Symploca sp. collections from Guam and Florida to find new cytotoxic agents. In this paper, we present the cytotoxicity-directed fractionation of a S. cf. hydnoides collection from Cetti Bay, Guam, which afforded the known compound dolastatin 16,⁸ together with seven new cyclic depsipeptides, given the trivial names veraguamides A–G (1–7).⁹

Results and Discussion

The freeze-dried cyanobacterium was extracted with EtOAc–MeOH (1:1) and subsequently solvent-partitioned into hexanes-, n-BuOH, and H₂O-soluble fractions. The n-BuOH soluble fraction showed biological activity at 10 μg/mL in an initial cytotoxicity screen using the HT29 colorectal adenocarcinoma cell line. This fraction was further purified by silica column chromatography, with the cytotoxic activity concentrated in the fraction eluting with 20% i-PrOH in CH₂Cl₂. Reversed-phase HPLC purification of this silica fraction yielded dolastatin 16 and veraguamides A–G (1–7).

HRESIMS of the major compound in this series, veraguamide A (1), showed the distinctive 1:1 isotopic cluster for a Br-containing compound for the [M + H]⁺ peak at m/z 767.3675/769.3660, suggesting a molecular formula of C₃₇H₅₉BrN₄O₈. The ¹H NMR spectrum of 1 displayed characteristic peptide resonances for a secondary amide proton (δ_H 6.25), two tertiary amide N-CH₃s (δ_H 3.00, δ_H 2.94) and several α-protons (δ_H 3.85 – 4.95). 2D NMR analysis (Table 1) in CDCl₃ using HSQC, COSY, TOCSY, and HMBC established the presence of four amino acids (Pro, Val, 2 × N-Me-Val) and an α-hydroxy acid [(2-hydroxy-3-methylpentanoic acid (Hmpa)]. The last spin system consisted of a CH₃ doublet (δ_H 1.25) that showed a COSY correlation to a methine (δ_H 3.11) and HMBC correlations to a carbonyl (δ_C 170.8) and an oxymethine (δ_C 76.4). Further extension of this unit using HMBC and COSY established the presence of a 8-bromo-3-hydroxy-2-methyl-7-octynoic acid (Br-Hmoya) moiety in 1. This was supported by HMBC correlations of the methylene (δ_C-6 19.2/δ_H2–6 2.23) with two quaternary carbons at δ_C 38.4 and δ_C 79.3 and by the large difference in chemical shifts between these quaternary carbons characteristic for an alkynyl bromide.¹⁰ The linear sequence of N-Me-Val-1–Pro–Hmpa–N-Me-Val-2–Val–Br-Hmoya was established based on HMBC correlations between α-protons and carbonyl groups (Table 1) and was verified by MS/MS fragmentation (Figure 1). The deshielded C-3 methine of the Br-Hmoya unit suggested acylation with the carbonyl of N-Me-Val-1 to form a cyclic hexadepsipeptide, corroborated by HMBC and consistent with the molecular formula requirements based on HRESIMS.

Table 1.

NMR Data for Veraguamide A (1) in CDCl₃

unit	C/H no	δ Ca	δH (J in Hz)b	COSYb	HMBCb
Br-Hmoya	1	170.8, C
	2	42.4, CH	3.11, br q (7.4)	H-3, H3-9	1, 3, 4, 9
	3	76.4, CH	4.85, dt (10.2, 2.5)	H-2, H-4a, H-4b	1, 1 (N-Me-Val-1)
	4a	27.4, CH2	2.06, m	H-3, H-4b, H-5a, H-5b	5, 6
	4b		1.59, m	H-3, H-4a, H-5a, H-5b
	5a	25.0, CH2	1.60, m	H-4a, H-4b, H-5b, H2-6	7, 8
	5b		1.41, m	H-4a, H-4b, H-5a, H2-6	6, 8
	6	19.2, CH2	2.23, m	H-5a, H-5b	7, 8
	7	38.4, C
	8	79.3, C
	9	14.1, CH3	1.25, d (7.4)	H-2	1, 2, 3
N-Me-Val-1	1	170.6, C
	2	65.0, CH	3.93, d (10.3)	H-3	1, 3, 4, N-Me, 1 (Pro)
	3	28.3, CH	2.30, m	H-2, H3-4, H3-5	2, 4, 5
	4	19.56, CH3	0.98, d (6.6)	H-3	2
	5	19.51, CH3	0.91, d (6.6)	H-3	2
	N-Me	28.6, CH3	3.00, s		2, 1 (Pro)
Pro	1	172.1, C
	2	57.3, CH	4.94 dd (8.9, 4.5)	H-3a, H-3b	1, 3, 4, 1 (Hmpa)
	3a	29.5, CH2	2.30, m	H-2, H-3b, H-4a	1, 2, 5
	3b		1.79, m	H-2, H-3a, H-4a	1, 2, 5
	4a	24.9, CH2	2.03, m	H-3a, H-3b, H-4b, H-5a, H-5b	2, 3, 5
	4b		1.98, m	H-4a, H-5a, H-5b	2, 3, 5
	5a	47.3, CH2	3.84, dt (−17.0, 7.1)	H-4a, H-4b, H-5b	2, 3, 4
	5b		3.61, dt (−17.0, 7.1)	H-4a, H-4b, H-5a	3, 4
Hmpa	1	165.9, C
	2	76.1, CH	4.90, d (9.1)	H-3	1, 3, 4, 5, 1 (N-Me-Val-2)
	3	35.7, CH	1.97, m	H-2, H3-6
	4a	24.81, CH2	1.54, m	H-4b, H3-5
	4b		1.13, m	H-4a, H3-5
	5	10.5, CH3	0.87, t (7.3)	H-4a, H-4b	3, 4
	6	13.8, CH3	1.01, d (6.8)	H-3	2, 3, 4
N-Me-Val-2	1	169.6, C
	2	66.0, CH	4.15, d (9.5)	H-3	1, 3, 5, N-Me, 1(Val)
	3	28.5, CH	2.27, m	H-2, H3-4, H3-5	1
	4	20.4, CH3	1.00, d (7.0)	H-3	2
	5	20.2, CH3	1.11, d (7.0)	H-3	2, 3
	N-Me	30.0, CH3	2.94, s		2, 1 (Val)
Val	1	173.4, C
	2	52.8, CH	4.70, dd (8.7, 6.5)	H-3, NH	1, 3, 4, 5, 1 (Br-Hmoya)
	3	32.1, CH	1.96, m	H-2, H3-4, H3-5	5
	4	20.3, CH3	0.95, d (6.7)	H-3	2, 3, 5
	5	17.5, CH3	0.88, d (6.7)	H-3	2, 3
	NH		6.25, d (8.7)	H-2	1, 1 (Br-Hmoya)

^a100 MHz

^b600 MHz

Figure 1.

MS/MS fragmentation of veraguamide A (1), veraguamide D (4), and veraguamide E (5).

Veraguamide B (2) showed a 1:1 isotopic pattern for the pseudomolecular ion [M + H]⁺ at m/z 753.3517/755.3508, suggesting the presence of a Br as in 1 with a negative difference of 14 amu corresponding to one less CH₂ unit and thus a molecular formula of C₃₆H₅₇BrN₄O₈. Comparison of the ¹H NMR spectrum of 1 and 2 showed differences in the splitting pattern in the CH₃ region atδ_H 0.93 ppm and the chemical shift of the α-proton (δ_H 4.85) of the α-hydroxy acid (Table 2). The vicinal methine (δ_H 2.17) of the α-hydroxy acid showed COSY correlations to two methyl groups (δ_H 0.93, δ_H 1.02) instead of COSY correlations to methylene and methyl protons in 1. Therefore, 2 possesses a 2-hydroxyisovaleric acid (Hiva) instead of the Hmpa unit as in 1.

Table 2.

NMR Data for Veraguamides B (2) and C (3) in CDCl₃

			2		3
unit	C/H no.	δ Ca	δH (J in Hz)b	δ Ca	δH (J in Hz)b
Br-Hmoyac/Hmoyad	1	170.8, C		170.8, C
	2	42.3, CH	3.13, br q (7.4)	42.4, CH	3.10, br q (7.2)
	3	76.4, CH	4.85, d (8.7)	76.4, CH	4.86, dt (10.4, 2.5)
	4a	27.5, CH2	2.07, m	27.4, CH2	2.07, m
	4b		1.60, m		1.62, m
	5a	24.93, CH2	1.61, m	25.2, CH2	1.62, m
	5b		1.42, m		1.44, m
	6	19.2, CH2	2.21, m	18.0, CH2	2.19, m
	7	38.4, C		83.6, C
	8	79.4, C		68.8, CH	1.93 t (2.5)
	9	14.6, CH3	1.25, d (7.4)	14.5, CH3	1.25, d (7.2)
N-Me-Val-1	1	170.7, C		170.7, C
	2	65.0, CH	3.94, d (10.4)	65.0, CH	3.93, d (11.0)
	3	28.26, CH	2.29, m	28.3, CH	2.28, m
	4	19.57, CH3	0.98, d (6.5)	19.58, CH3	0.98, d (6.8)
	5	19.55, CH3	0.92, d (6.5)	19.56, CH3	0.91, d (6.8)
	N-Me	28.7, CH3	3.00, s	28.7, CH3	3.00, s
Pro	1	172.1, C		172.2, C
	2	57.2, CH	4.95 dd (8.6, 4.8)	57.3, CH	4.94 dd (8.4, 5.0)
	3a	29.4, CH2	2.28, m	29.5, CH2	2.28, m
	3b		1.79, m		1.79, m
	4a	24.99, CH2	2.04, m	24.89, CH2	2.03, m
	4b		1.98, m		1.99, m
	5a	47.3, CH2	3.80, dt (−16.4, 6.9)	47.3, CH2	3.84, dt (−16.8, 7.1)
	5b		3.60, dt (−16.4, 6.9)		3.60, dt (−16.8, 7.1)
Hivac/Hmpad	1	165.8, C		165.9, C
	2	77.2, CH	4.85, d (8.7)	76.0, CH	4.89, d (9.4)
	3	29.6, CH	2.17, m	35.7, CH	1.98, m
	4	18.1, CH3	1.02, t (6.6)	24.81, CH2	1.54, m
					1.13, m
	5	18.5, CH3	0.93, d (6.6)	10.5, CH3	0.86, t (7.3)
	6			13.8, CH3	1.01, d (6.7)
N-Me-Val-2	1	169.6, C		169.6, C
	2	66.1, CH	4.15, d (10.2)	66.0, CH	4.13, d (10.0)
	3	28.34, CH	2.28, m	28.5, CH	2.28, m
	4	20.3, CH3	0.99, d (6.8)	20.4, CH3	0.99, d (6.4)
	5	20.1, CH3	1.11, d (6.8)	20.2, CH3	1.10, d (6.4)
	N-Me	30.0, CH3	2.94, s	30.0, CH3	2.93, s
Val	1	173.5, C		173.4, C
	2	52.8, CH	4.71, dd (8.6, 6.4)	52.8, CH	4.70, dd (8.6, 6.2)
	3	32.2, CH	1.98, m	32.1, CH	1.96, m
	4	20.3, CH3	0.94, d (6.4)	20.3, CH3	0.94, d (6.7)
	5	17.5, CH3	0.88, d (6.4)	17.6, CH3	0.87, d (6.7)
	NH		6.26, d (8.6)		6.26, d (8.6)

^a100 MHz

^b600 MHz

^cRefers to Compound 2

^dRefers to Compound 3

The HRESIMS spectrum of veraguamide C (3) showed a negative deviation of 79 amu compared with 1 which indicated the lack of Br and a molecular formula of C₃₇H₆₀N₄O₈. This was supported by the absence of the 1:1 isotopic pattern for the [M + H]⁺ peak when compared to 1. The ¹H NMR spectrum of 3 showed an additional triplet at δ_H 1.93 with J_H,H = 2.5 Hz (Table 2); otherwise it was virtually identical to that of 1. This proton correlated to a methine at δ_C 68.8 and a quaternary C (δ_C 83.6) in the HSQC and HMBC spectra, respectively. These signals are indicative of a terminal alkyne; hence 3 had to bear a 3-hydroxy-2-methyl-7-octynoic acid (Hmoya) moiety in lieu of the Br-Hmoya present in 1 and 2.

Veraguamide D (4) appeared closely related to 3 as its ¹H NMR spectrum showed the acetylenic proton at δ_H 1.93 (Table 3). In comparison to 3, the HRESIMS spectrum of 4 showed a positive difference of 14 amu, corresponding to an additional CH₂ unit and in agreement with a molecular formula of C₃₈H₆₂N₄O₈. The ¹H and ¹³C NMR spectra of 4 showed a high-field CH₃ at δ_C 11.4/δ_H 0.93, characteristic of an Ile or Ile-derived moiety. COSY (δ_H/δ_H 0.93/1.46, 1.46/2.02, 2.02/4.01) and (δ_C/δ_H 15.8/4.01, 28.7/4.01) correlations established that the N-Me-Val-1 residue is replaced by an N-Me-Ile in veraguamide D (4).

Table 3.

NMR Data for Veraguamides D (4) and E (5) in CDCl₃

			4		5
unit	C/H no.	δ Ca	δH (J in Hz)b	δ a	δ (J in Hz)b
Hmoya	1	170.8, C		170.71, C
	2	42.4, CH	3.13, br q (7.2)	42.4, CH	3.08, br q (7.4)
	3	76.4, CH	4.86, dt (10.8, 2.6)	76.5, CH	4.85, d (9.0)
	4a	27.5, CH2	2.07, m	27.5, CH2	2.06, m
	4b		1.63, m		1.62, m
	5a	25.2, CH2	1.63, m	25.2, CH2	1.61, m
	5b		1.47, m		1.43, m
	6	17.5, CH2	2.18, m	18.0, CH2	2.18, m
	7	83.6, C		83.6, C
	8	68.8, CH	1.93, t (2.5)	68.8, CH	1.93, t (2.3)
	9	14.4, CH3	1.24, d (7.2)	14.5, CH3	1.23, d (7.4)
N-Me-Ilec/N-Me-Vald	1	170.7, C		170.69, C
	2	64.0, CH	4.01, d (10.6)	64.9, CH	3.93, d (10.0)
	3	34.6, CH	2.02, m	28.3, CH	2.29, m
	4	25.7, CH2	1.46, m	19.59, CH3	0.91, d (6.4)
	5	11.4, CH3	0.93, t (6.5)	19.56, CH3	0.98, d (6.4)
	6	15.8, CH3	0.94, d (6.8)
	N-Me	28.7, CH3	2.99, s	28.6, CH3	3.00, s
Pro	1	172.7, C		172.2, C
	2	57.2, CH	4.94 dd (8.9, 4.8)	57.3, CH	4.94 dd (9.0, 5.3)
	3a	28.8, CH2	2.26, m	29.5, CH2	2.29, m
	3b		1.78, m		1.78, m
	4a	24.9, CH2	2.03, m	24.89, CH2	2.01, m
	4b		1.97, m		1.99, m
	5a	47.2, CH2	3.82, dt (−17.0, 7.3)	47.3, CH2	3.86, dt (−17.0, 7.0)
	5b		3.60, dt (−17.0, 7.3)		3.60, dt (−17.0, 7.0)
Hmpa	1	165.9, C		166.0, C
	2	76.0, CH	4.90, d (9.2)	76.1, CH	4.85, d (9.0)
	3	35.7, CH	1.98, m	35.1, CH	1.98, m
	4	24.8, CH2	1.53, m	24.86, CH2	1.54, m
			1.12, m		1.13, m
	5	10.5, CH3	0.86, t (7.6)	10.5, CH3	0.86, t (7.0)
	6	13.9, CH3	0.99, d (6.9)	13.8, CH3	1.01, d (6.8)
N-Me-Valc/N-Me-Iled	1	169.6, C		169.7, C
	2	66.0, CH	4.15, d (9.4)	65.2, CH	4.22, d (9.6)
	3	28.4, CH	2.28, m	35.7, CH	1.98, m
	4	20.3, CH3	1.10, d (6.8)	26.6, CH2	1.54, m
					1.06, m
	5	20.2, CH3	0.99, d (6.8)	11.7, CH3	0.96, t (7.2)
	6			16.5, CH3	1.04, d (6.9)
	N-Me	30.0, CH3	2.92, s	30.1, CH3	2.93, s
Valc/Iled	1	173.4, C		173.5, C
	2	52.8, CH	4.70, dd (8.6, 6.6)	52.4, CH	4.70, dd (8.4, 6.7)
	3	32.1, CH	1.98, m	38.6, CH	1.69, m
	4	19.0, CH3	0.94, d (6.6)	23.9, CH2	1.47, m
	5	17.6, CH3	0.87, d (6.6)		1.05, m
				11.5, CH3	0.85, d (6.6)
	6			16.3, CH3	0.91, d (6.6)
	NH		6.24, d (8.6)		6.26, d (8.4)

^a125 MHz

^b600 MHz

^cRefers to Compound 4

^dRefers to Compound 5

Compound 5 (C₃₉H₆₄N₄O₈) exhibited a close relationship to both 3 and 4, showing a positive deviation of 28 amu and 14 amu, respectively, and also having the Hmoya moiety. The NMR data of 5 (Table 3) indicated the presence of two high-field methyl (δ_C 11.7/δ_H 0.96, δ_C 11.5/δ_H 0.85) and additional methylene (δ_C 26.6/δ_H 1.54, 1.06; δ_C 23.9/δ_H 1.47, 1.05) groups in comparison with 3, which suggested that two isopropyl groups in the latter are replaced with sec-butyl groups in the former. This is further corroborated by HMBC and COSY correlations which established the replacement of Val and N-Me-Val-2 moieties with Ile and N-Me-Ile, respectively, in veraguamide E (5). The N-Me-Val residue that was replaced by N-Me-Ile was located at different positions in 4 and 5; with N-Me-Val-1 replaced in the former and N-Me-Val-2 in the latter. This NMR result was verified by MS/MS fragmentation of both 4 and 5 (Figure 1).

The ¹H NMR spectrum of veraguamide F (6) (Table 4) showed additional resonances for aromatic protons at δ_H 7.2 – 7.4 ppm, upfield-shifted N-Me protons to δ_H 2.60 ppm presumably due to shielding by the aromatic ring, and a low-field α-proton of the hydroxy acid (δ_H 5.47), with the acetylenic proton still present (δ_H 1.93). COSY correlations of δ_H 5.47 to diastereotopic CH₂ protons at δ_H 3.17/δ_H 2.91, together with HMBC correlations of the latter to aromatic carbons at δ_C 136.2, δ_C 129.3 (Table 4) established the presence of phenyllactic acid (Pla) as the α-hydroxy acid in 6. These NMR-derived conclusions fulfilled the molecular formula requirements for C₄₀H₅₈N₄O₈ based on HRESIMS of 6.

Table 4.

NMR Data for Veraguamide F (6) in CDCl₃

unit	C/H no.	δCa	δ (J in Hz)b
Hmoya	1	170.9, C
	2	42.2, CH	3.24 br q (7.3)
	3	76.6, CH	4.89, dt (10.6, 2.4)
	4a	27.5, CH2	2.08, m
	4b		1.64, m
	5a	25.2, CH2	1.65, m
	5b		1.45, m
	6	18.0, CH2	2.20, m
	7	83.5, C
	8	68.9, CH	1.93, t (2.5)
	9	14.6, CH3	1.29, d (7.3)
N-Me-Val-1	1	170.7, C
	2	65.3, CH	3.95, d (10.4)
	3	28.3, CH	2.32, m
	4	19.6, CH3	1.00, d (6.5)
	5	19.7, CH3	0.93, d (6.5)
	N-Me	28.7, CH3	3.06, s
Pro	1	172.3, C
	2	57.1, CH	4.97, dd (9.0, 4.5)
	3a	29.1, CH2	2.27, m
	3b		1.83, m
	4a	25.0, CH2	2.07, m
	4b		1.97, m
	5a	47.0, CH2	3.62, m
	5b		3.56, m
Pla	1	165.4, C
	2	72.8, CH	5.47, dd (9.8, 4.0)
	3	36.7, CH2	3.17, m
			2.91, m
	4	136.2, C
	5//9	129.3, CH	7.18, d (8.0)
	6/8	128.5, CH	7.28, m
	7	126.7, CH	7.20, m
N-Me-Val-2	1	168.9, C
	2	65.8, CH	4.05, d (10.5)
	3	27.4, CH	2.08, m
	4	19.8, CH3	0.89, d (6.8)
	5	20.0, CH3	0.91, d (6.8)
	N-Me	29.0, CH3	2.60, s
Val	1	173.4, C
	2	52.7, CH	4.76, dd (8.5, 6.1)
	3	32.3, CH	1.98, m
	4	20.3, CH3	0.91, d (6.6)
	5	17.7, CH3	0.88, d (6.6)
	NH		6.28, d (8.5)

^a100 MHz

^b600 MHz

Veraguamide G (7) lacked the acetylenic signal (δ_C 68.8/δ_H 1.93) observed for 3–6 and instead showed downfield resonances of a terminal methylene (δ_C 114.9/δ_H 4.97) and a methine (δ_C 138.2/δ_H 5.74). These signals indicated that the terminal alkyne group of the C₈-polyketide derived moiety is replaced by a terminal vinyl group. This conclusion was further supported by the positive deviation of 2 amu compared to 3 and a molecular formula of C₃₇H₆₂N₄O₈. Hence,the Hmoya unit present in 3–6 was replaced by 3-hydroxy-2-methyl-7-octenoic acid (Hmoea) in 7.

Enantioselective HPLC analysis coupled with mass spectrometry or UV detection of the acid hydrolysates of 1–7 allowed us to assign the absolute configuration of all the amino acids and α-hydroxy acid components as L and S, respectively. To determine the absolute configuration at C-2 and C-3 of the Br-Hmoya unit, veraguamide A (1) was subjected to methanolysis to yield the linear fragment 9 (Figure 2). The observed coupling constant of 3.2 Hz was characteristic for a syn configuration, whereas a coupling constant near 6.3 Hz would have been expected for the anti configuration.¹¹ The absolute configuration at C-3 and consequently for C-2 of the Br-Hmoya unit of 9 was determined using Mosher's analysis. The derived Δδ values (Figure 2) predicted an R configuration at C-3 and hence from the relative configuration, C-2 should have an S configuration. Of note, comparison of the ³J_H,H values of H-2 and H-3 with a model system to assign the relative configuration could only be applied when C-3 bears a free hydroxy group. This moiety is involved in intramolecular H-bonding with the adjacent carbonyl group, thus hindering free bond rotation across C-2 and C-3.^12,13 Accordingly, the corresponding MTPA-esters (10, 11) did not show the same ³J_H,H values for H-2 and H-3 as that of 9. The same absolute configuration at C-2 and C-3 for Hmoya, Hmoea and 3-hydroxy-2-methyl-octanoic acid (Hmoaa) is expected based on virtually identical ¹³C NMR shifts and specific rotations observed for 1–8.

Figure 2.

Assignment of absolute configuration of veraguamide A (1) using methanolysis and subsequent Mosher's analysis. Δδ = δ(S-MTPA ester) − δ(R-MTPA ester).

To gain insight into structure–activity relationships, veraguamide A (1) was partially (Lindlar catalyst, H₂) and fully (Pd/C, H₂) hydrogenated to yield the semisynthetic veraguamide G (7) and tetrahydroveraguamide A (8), respectively. The cytotoxic activities of 1–8, semisynthetic veraguamide G (7), and dolastatin 16 were evaluated for effects on viability of HT29 colorectal and HeLa cervical adenocarcinoma cells (Table 6). The IC₅₀ values of the natural and semisynthetic veraguamide G (7) were comparable suggesting the activities of these compounds were not likely due to traces of highly biologically active impurities. The most active in this series of compounds are veraguamides D (4) and E (5), with IC₅₀ values more than 5-fold lower than those for their related congener veraguamide C (3). This suggested that increased hydrophobicity of specific units (II, IV, V, VI) increased the cytotoxicity of this compound class, with the position having minimal effect on the bioactivity as 4 and 5 showed comparable IC₅₀s. However, modification with bulkier groups is detrimental to the activity, as exemplified by a close to 10-fold decrease in the cytotoxic activity of 6 (Table 6) compared to 3, where a phenyllactic acid (Pla) moiety is introduced at position IV of the former instead of the Hmpa unit. The C₈-polyketide derived moiety also plays a role in the cytotoxicity of these compounds. Comparing the biological activities of related compounds 1, 3, 7, and 8, weaker cytotoxicity was observed for compounds with a Br-Hmoya or Hmoaa unit, while compounds with Hmoya or Hmoea were about equally potent. This then suggests the importance of a π-system combined with the presence of acetylenic or vinylic protons in this moiety for cytotoxic activity.

Table 6.

Cytotoxic Activity (IC₅₀, μM) of Natural and Semisynthetic Veraguamides

Compound	HT29	HeLa
Veraguamide A (1)	26	21
Veraguamide B (2)	30	17
Veraguamide C (3)	5.8	6.1
Veraguamide D (4)	0.84	0.54
Veraguamide E (5)	1.5	0.83
Veraguamide F (6)	49	49
Veraguamide G (7)	2.7	2.3
Tetrahydroveraguamide A (8)	33	48
Dolastatin 16	78	58

The veraguamides are reminiscent of other cyanobacterial compounds such as hantupeptins,^13,14 antanapeptins,¹⁵ and trungapeptins.¹⁶ These compounds are also cyclic hexadepsipeptides with the characteristic C₈-polyketide derived units as Hmoya, Hmoea, or Hmoaa. Veraguamide F (6) is a constitutional isomer of antanapeptin D,¹⁵ where an N-Me-Phe and Hiva are present in the latter instead of Pla and N-Me-Val as in 6. It is interesting that subtle changes in structure of these compounds have a profound effect on the cytotoxicity. Antanapeptins A–D (brine shrimp) as well as trungapeptin A (KB and LoVo cells) did not display cytotoxicity at the reported concentrations (10 μg/mL),^15,16 while hantupeptins A–C were cytotoxic against MOLT-4 leukemia and MCF7 breast cancer cells, with hantupeptin A being the most active in this series with IC₅₀s of 32 nM and 4.0 μM, respectively.¹³ Trungapeptins B and C were not tested for cytotoxicity.¹⁶

Although dolastatin 16 has been isolated previously from sea hares⁸ and cyanobacteria,^15,17 its stereostructure has been only partially assigned. Therefore we analyzed the remaining dolaphenvaline (Dpv) and dolamethylleucine (Dml) units. The absolute configuration of the Dpv unit in dolastatin 16 was determined as 2S,3R using enantioselective HPLC-MS analysis of the FDLA-derivatized acid hydrolysate in comparison with that for pitiprolamide.¹⁸ The configuration of the Dpv unit in dolastatin 16 is the same as in pitiprolamide and kulokekahilide-1 determined from X-ray crystallography and chemical synthesis, respectively.^18,19 To determine the configuration of the stereocenters in the Dml unit, we carried out acid and base hydrolysis of dolastatin 16 to yield the Dml unit and the linear fragment 12, respectively. The Dml unit is related to the 3-amino-2-methylpentanoic acid (Map) unit, present in dolastatin 12^20,21 and majusculamide C,²² as well as the 3-amino-2-methylhexanoic acid (Amha) unit in kulokekahilide-1,¹⁹ lyngbyastatin 1,²¹ and the ulongamides.²³ For all these α-substituted β-amino acids, the corresponding Marfey's adducts of the two 3R isomers (2S,3R and 2R,3R) consistently eluted later by reversed-phase C₁₈ HPLC than those of the two 3S isomers (2S,3S and 2R,3S), while the C-2 stereoisomers are poorly resolved and variable.^19–23 Extrapolation to the elution order we obtained for the Dml suggested the presence of 3R in the Dml unit of dolastatin 16.²⁴ To assign the configuration at C-2 of Dml, modified Mosher's analysis using phenylglycine methyl ester (PGME)²⁵ derivatization of 12 (Figure 3) suggested a 2R configuration. The applicability of this chiral derivatization technique has been demonstrated for α- and β- substituted carboxylic acids,²⁵ but may not have been widely used for α,β-disubstituted carboxylic acids. From the Δδ values that we obtained (Figure 3), it is predicted that there is some deviation from the presumed conformation of the PGME amide, but nonetheless a systematic arrangement of positive and negative Δδ values was obtained. Therefore, we propose a 2R,3R configuration for the Dml unit of dolastatin 16. The ¹³C NMR chemical shifts for the Dpv and Dml unit in both dolastatin 16 and homodolastatin 16 were similar, suggesting the latter would also have a (2S,3R)-Dpv and (2R,3R)-Dml units.^8,26 Dolastatin 16 showed weak cytotoxic activity with IC₅₀ of 78 μM and 58 μM for HT29 and HeLa cell lines, respectively. These values are significantly higher than those reported by Pettit et al. against a different panel of cell lines.⁸

Figure 3.

Assignment of absolute configuration dolastatin 16 by base hydrolysis and modified Mosher's analysis using PGME. Δδ = δ(S-PGME amide) − δ(R-PGME amide).

Experimental Section

General Experimental Procedures

Optical rotations were measured on a Perkin-Elmer 341 polarimeter. UV spectra were recorded on SpectraMax M5 (Molecular Devices). ¹H and 2D NMR spectra were recorded in CDCl₃ on a Bruker Avance II 600 MHz spectrometer equipped with a 5-mm TXI cryogenic probe using residual solvent signals (δ_H 7.26; δ_C 77.0) as internal standards. HSQC and HMBC experiments were optimized for ¹J_CH = 145 and ⁿJ_CH = 7 Hz, respectively. TOCSY experiments were done using a mixing time of 100 ms. ¹³C NMR spectra were recorded on Varian 400 MHz or Bruker 500 MHz NMR spectrometers. HRESIMS data was obtained using an Agilent LC-TOF mass spectrometer equipped with an APCI/ESI multimode ion source detector. LRESIMS measurements and MS/MS fragmentation were done on an ABI 3200Q TRAP.

Biological Material

The Symploca cf. hydnoides cyanobacterium was collected by hand while snorkeling in the shallow waters of the southern fore-reef (1–3 m) of Cetti Bay, Guam on April 17, 2009. A voucher specimen, which is preserved in 100% EtOH, is deposited in the University of Guam Herbarium (accession no. GUAM-GH11446). A voucher specimen is also retained at the Smithsonian Marine Station, Fort Pierce, FL.

Extraction and Isolation

The freeze-dried cyanobacterium (142.0 g) was extracted with EtOAc–MeOH (1:1) to yield the nonpolar extract (11.6 g). This was partitioned between hexanes and 20% aqueous MeOH, the latter concentrated under reduced pressure and further partitioned between n-BuOH and H₂O. The n-BuOH fraction was concentrated to dryness (2.84 g) and chromatographed on Si gel eluting first with CH₂Cl₂, followed by increasing concentrations of i-PrOH, while after 100% i-PrOH, increasing gradients of MeOH were used. The 20% i-PrOH fraction was subjected to a C₁₈ SPE eluting with 25%, 50%, 75%, and 100% MeOH in H₂O. The 100% MeOH fraction was purified by semipreparative reversed-phase HPLC (Phenomenex Synergi-Hydro RP, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of MeOH–H₂O (70%–100% MeOH in 60 min and then 100% MeOH for 10 min) to yield dolastatin 16 (t_R 32.6 min, 21.6 mg), semipure veraguamide C (t_R 36.4 min, 15.1 mg), semipure veraguamide F (t_R 37.4 min, 10.0 mg), a mixture of veraguamides B and D (t_R 40.0 min, 25.0 mg), veraguamide A (1) (t_R 42.6 min, 25.9 mg), and a mixture of veraguamides E and G (t_R 43.7 min, 9.0 mg).

The final purification of the semipure veraguamide C (3) was achieved using semipreparative HPLC (Phenomenex Phenyl-hexyl, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of MeOH–H₂O (85%–100% MeOH in 40 min and then 100% MeOH for 5 min) to yield veraguamide C (3) (t_R 19.9 min, 10.7 mg). Using the same chromatographic conditions, purification of the semipure veraguamide F yielded veraguamide F (6) (t_R 21.7 min, 6.8 mg). The mixture of veraguamides B and D was resolved using the same chromatographic condition with a different linear gradient (70%–100% MeOH in 45 min and then 100% MeOH for 10 min) to yield veraguamide D (4) (t_R 36.6 min, 4.0 mg) and veraguamide B (2) (t_R 37.3 min, 11.5 mg). The mixture of veraguamides E and G was further purified using the same chromatographic conditions to yield veraguamide G (7) (t_R 39.9 min, 4.4 mg) and veraguamide E (5) (t_R 40.5 min, 3.6 mg).

Hydrogenation of 1

A catalytic amount of 10% Pd/C was added to a methanolic solution of 1 (1.8 mg/mL). The reaction was left to stir for 6 h under a hydrogen balloon. The catalyst was filtered through a Celite pad, and the filtrate, upon concentration, was purified by semipreparative HPLC (Phenomenex Phenyl-hexyl, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of MeOH–H₂O (70%–100% MeOH in 45 min and then 100% MeOH for 10 min) to yield 8 (t_R 40.9 min, 1.2 mg).

Partial hydrogenation of 1 was carried out with Lindlar catalyst, using the same reaction and chromatographic conditions stated above. This afforded the semisynthetic veraguamide G (t_R 39.9 min, 1.7 mg). The LRESIMS and ¹H NMR spectra of the semisynthetic veraguamide G were in good agreement with the spectra for the natural product (7).

Veraguamide A (1)

colorless, amorphous solid; [α]²⁰_D −44 (c 0.44, MeOH); UV (MeOH); λ_max (log ε) 202 (6.29); IR (film) ν_max 3428, 2966, 2877, 2361, 2334, 1736, 1647, 1190 cm⁻¹; ¹H NMR ,¹³C NMR, COSY, and HMBC data, see Table 1; HRESIMS m/z 767.3675 [M + H]⁺ (calcd for C₃₇H₆₀⁷⁹BrN₄O₈, 767.3594), m/z [M + H]⁺ 769.3660 (calcd for C₃₇H₆₀⁸¹BrN₄O₈, 769.3574) (100:100 [M + H]⁺ ion cluster).

Veraguamide B (2)

colorless, amorphous solid; [α]²⁰_D −40c 0.16, MeOH); UV (MeOH); λ_max (log ε) 202 (4.30); IR (film) ν_max 3378, 2966, 2886, 1738, 1657, 1190 cm⁻¹; ¹H NMR and ¹³C NMR data, see Table 2; HRESIMS m/z 753.3517 [M + H]⁺ (calcd for C₃₆H₅₈⁷⁹BrN₄O₈, 753.3438), m/z [M + H]⁺ 755.3508 (calcd for C₃₆H₅₈⁸¹BrN₄O₈, 755.3418) (100:100 [M + H]⁺ ion cluster).

Veraguamide C (3)

colorless, amorphous solid; [α]²⁰_D −44 (c 0.31, MeOH); UV (MeOH); λ_max (log ε) 202 (4.17); IR (film) ν 3429, 2965, 2871, 1732, 1650, 1190 cm⁻¹;¹H NMR and ¹³C NMR data, see Table 2; HRESIMS m/z 689.4486 [M + H]⁺ (calcd for C₃₇H₆₁N₄O₈, 689.4490).

Veraguamide D (4)

colorless, amorphous solid; [α]²⁰_D −57 (c 0.11, MeOH); UV (MeOH); λ_max (log ε) 202 (4.30); IR (film) ν_max 3378, 2958, 2878, 1732, 1647 cm⁻¹; ¹H NMR and ¹³C NMR data, see Table 3; HRESIMS m/z 703.4639 [M + H]⁺ (calcd for C₃₈H₆₃N₄O₈, 703.4646).

Veraguamide E (5)

colorless, amorphous solid; [α]²⁰_D −56 (c 0.22, MeOH); UV (MeOH); λ_max (log ε) 202 (4.30); IR (film) ν_max 3320, 2958, 2877, 1733, 1647 cm⁻¹; ¹H NMR and ¹³C NMR data, see Table 3; HRESIMS m/z 717.4799 [M + H]⁺ (calcd for C₃₉H₆₅N₄O₈, 717.4802).

Veraguamide F (6)

colorless, amorphous solid; [α]²⁰_D −41 (c 0.13, MeOH); UV (MeOH); λ_max (log ε) 206 (4.33); IR (film) ν_max 3450, 3306, 2965, 2871, 1737, 1661, 1647, 1490 cm⁻¹; ¹H NMR and ¹³C NMR data, see Table 4; HRESIMS m/z 723.4411 [M + H]⁺ (calcd for C₄₀H₅₉N₄O₈, 723.4333).

Veraguamide G (7)

colorless, amorphous solid; [α]²⁰_D −41 (c 0.17, MeOH); UV (MeOH); λ_max (log ε) 202 (4.26); IR (film) ν_max 3436, 2965, 2879, 1740, 1661, 1646 cm⁻¹; ¹H NMR and ¹³C NMR data, see Table 5; HRESIMS m/z 691.4649 [M + H]⁺ (calcd for C₄₀H₅₉N₄O₈, 691.4646).

Table 5.

NMR Data for Veraguamide G (7) and Tetrahydroveraguamide A (8) in CDCl₃

			7		8
unit	C/H no.	δ Ca	δH (J in Hz)b	δCb,c	δH (J in Hz)b
Hmoead/Hmoaae	1	170.9, C		170.7, C
	2	42.4, CH	3.10, br q (7.4)	42.1, CH	3.08, br q (7.4)
	3	76.8, CH	4.85, dt (10.6, 2.4)	76.8, CH	4.86, dt (10.1, 2.1)
	4a	27.9, CH2	1.98, m	31.0, CH2	1.21, m
	4b		1.45, m		1.26, m
	5a	25.5, CH2	1.48, m	28.2, CH2	1.39, m
	5b		1.30, m
	6a	33.2, CH2	2.05, m	25.8, CH2	1.39, m
	6b				1.20, m
	7	138.2, CH	5.74, m	22.2, CH2	1.26, m
	8	114.9, CH2	4.97, m	13.6, CH3	0.85, t (6.9)
	9	14.4, CH3	1.22, d (7.4)	14.0, CH3	1.23, d (7.4)
N-Me-Val-1	1	170.7, C		170.7, C
	2	65.0, CH	3.93, d (9.8)	64.9, CH	3.93, d (10.7)
	3	28.3, CH	2.28, m	28.3, CH	2.28, m
	4	19.59, CH3	0.98, d (6.4)	19.2, CH3	0.98, d (6.5)
	5	19.54, CH3	0.92, d (6.4)	19.3, CH3	0.91, d (6.5)
	N-Me	28.6, CH3	3.01, s	28.4, CH3	3.00, s
Pro	1	172.2, C		172.0, C
	2	57.3, CH	4.95, dd (8.7, 5.0)	57.1, CH	4.94, dd (9.0, 5.0)
	3a	29.4, CH2	2.29, m	29.1, CH2	2.28, m
	3b		1.79, m		1.79, m
	4a	24.9, CH2	2.03, m	24.6, CH2	2.03, m
	4b		1.98, m		1.99, m
	5a	47.3, CH2	3.84, dt (−16.7, 7.1)	47.0, CH2	3.84, dd (−17.0, 7.3)
	5b		3.61, dt (−16.7, 7.1)		3.60, dd (−17.0, 7.3)
Hmpa	1	165.9, C		165.7, C
	2	76.0, CH	4.90, d (8.7)	76.6, CH	4.90, d (8.8)
	3	35.7, CH	1.98, m	35.4, CH	1.98, m
	4	24.8, CH2	1.54, m	24.5, CH2	1.54, m
			1.13, m		1.13, m
	5	10.5, CH3	0.86, t (7.3)	10.2, CH3	0.86, t (7.1)
	6	13.8, CH3	1.00, d (6.0)	13.5, CH3	1.00, d (6.4)
N-Me-Val-2	1	169.6, C		169.5, C
	2	66.0, CH	4.15, d (10.2)	65.8 CH	4.14, d (9.6)
	3	28.6, CH	2.28, m	28.1, CH	2.28, m
	4	20.4, CH3	1.00, d (6.1)	20.0, CH3	0.99, d (6.6)
	5	20.2, CH3	1.10, d (6.1)	19.9, CH3	1.10, d (6.6)
	N-Me	30.0, CH3	2.93, s	29.7, CH3	2.93, s
Val	1	173.4, C		173.3, C
	2	52.7, CH	4.70, dd (8.6, 6.7)	52.5, CH	4.70, dd (8.6, 6.2)
	3	32.1, CH	1.98, m	31.7, CH	1.96, m
	4	20.3, CH3	0.93, d (6.8)	19.3, CH3	0.93, d (6.3)
	5	17.6, CH3	0.86, d (6.8)	17.2, CH3	0.87, d (6.3)
	NH		6.23, d (8.6)		6.26, d (8.6)

^a125 MHz

^b600 MHz

^cBased on HSQC and HMBC

^dRefers to Compound 7

^eRefers to Compound 8

Tetrahydroveraguamide A (8)

colorless, amorphous solid; [α]²⁰_D −43 (c 0.05, MeOH); UV (MeOH); λ_max (log ε) 202 (4.33); ¹H NMR and ¹³C NMR data, see Table 5; HRESIMS m/z 693.4791 [M + H]⁺ (calcd for C₃₇H₆₅N₄O₈, 693.4802).

Dolastatin 16

colorless, amorphous solid; [α]²⁰_D +22 (c 0.18, MeOH) {lit.⁸ [α]²⁰_D +15.5 (c 0.20, MeOH)}; UV (MeOH); λ_max (log ε) 202 (4.51); IR (film) ν_max 3407, 2965, 2871, 1733, 1647 cm⁻¹; ¹H NMR and ¹³C NMR data are identical with literature values⁸ (Supporting Information); LRESIMS m/z 879.6 [M + H]⁺, 901.3 [M + Na]⁺.

Acid Hydrolysis of Veraguamides and Enantioselective Analysis

Portions of 1–7 (100 μg) were acid-hydrolyzed (200 μL 6 N HCl, 110 °C, 20 h), the product mixtures dried, reconstituted in 100 μL H₂O and analyzed by enantioselective HPLC-UV and enantioselective HPLC-MS. The absolute configurations of the amino acids N-Me-Ile, Ile, N-Me-Val, Val, and Pro were determined by enantioselective HPLC-MS [column, Chirobiotic TAG (250 × 4.6 mm), Supelco; solvent, MeOH–10 mM NH₄OAc (40:60, pH 5.30); flow rate, 0.5 mL/min; detection by ESIMS in positive ion mode (MRM scan)]. The acid hydrolysates of 1–4, 6, and 7 showed retention times at 7.8, 11.6, and 13.6 min corresponding to L-Val, N-Me-L-Val, and L-Pro, respectively. The acid hydrolysate of 4 in addition showed a retention time at 12.4 min corresponding to N-Me-L-Ile. The acid hydrolysate of 5 showed retention times at 8.4, 11.6, 12.4, and 13.6 min, corresponding to L-Ile, N-Me-L-Val, N-Me-L-Ile, and L-Pro, respectively. The retention times (t_R, min; MRM ion pair) of the authentic amino acids were as follows: NMe-L-Val (11.6; 132→86), N-Me-D-Val (34.3), L-Val (7.8; 118→72), D-Val (13.7), N-Me-L-Ile (12.4; 146→100), N-Me-L-allo-Ile (15.0), N-Me-D-Ile (49.0), N-Me-D-allo-Ile (51.0), L-Ile (8.4; 132→86), L-allo-Ile (8.6), D-allo-Ile (17.6), D-Ile (20.2), L-Pro (13.6; 116→70), D-Pro (36.0). Compound-dependent parameters used were as follows: N-Me-Val: DP 29.4, EP 4.2, CE 17.4, CXP 2.7, CEP 10.6; Val: DP 5.7, EP 9.0, CE 40.0, CXP 8.0, CEP 10.0; N-Me-Ile: DP 35.0, EP 7.0, CE 17.0, CXP 2.0, CEP 10.0; Ile: DP 40.0, EP 9.0, CE 15.0, CXP 3.0, CEP 8.0; Pro: DP 35.0, EP 7.7, CE 22.7, CXP 5.0, CEP 10.3. Source gas parameters used were as follows: CUR 40, CAD Medium, IS 4500, TEM 750, GS1 65, GS2 65.

The absolute configurations of the α-hydroxy acids [2-hydroxyisovaleric acid (Hiva), 2-hydroxy-3-methylpentanoic acid (Hmpa), and phenyllactic acid (Pla)] were determined using enantioselective HPLC [column, CHIRALPAK MA (+) (50 × 4.6 mm); solvent, CH₃CN–2 mM CuSO₄ (10:90); flow rate, 1.0 mL/min; detection by UV (254 nm)]. The acid hydrolysates of 1, 3–5, and 7 each showed peaks at 33.0 min corresponding to (2S,3S)-Hmpa. The acid hydrolysate of 2 contained a component that had a retention time at 10.0 min corresponding to (2S)-Hiva, while 6 gave a peak at 51.0 min corresponding to (2S)-Pla. The retention times of the authentic standards were as follows: (2R)-Hiva (6.0), (2S)-Hiva (10.0), (2R,3S)-Hmpa (16.0), (2R,3R)-Hmpa (19.0), (2S,3R)-Hmpa (26.0), (2S,3S)-Hmpa (33.0), (2R)-Pla (33.5), (2S)-Pla (51.0). All other amino acid units eluted within less than 5.0 min using this chromatographic condition.

Methanolysis of 1

Compound 1 (5.0 mg) was dissolved in 2.0 mL of 5% (w/w) methanolic KOH solution and stirred for 24 h at room temperature. The solvent was evaporated and the residue was partitioned between CH₂Cl₂ and H₂O. The organic layer was collected, dried over anhydrous MgSO₄ and concentrated to dryness under nitrogen. The crude methanolysis product was further purified by semipreparative reversed-phase HPLC (Phenomenex Synergi-Hydro RP, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of MeOH–H₂O (70%–100% MeOH in 60 min and then 100% MeOH for 10 min) to yield 9 (t_R 19.3 min, 1.4 mg).

9

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 6.30 (d, J = 8.1 Hz, 1H), 4.92 (d, J = 10.5 Hz, 1H), 4.76 (dd, J = 9.1, 6.6 Hz, 1H), 3.78 (ddd, J = 8.6, 4.4, 3.2 Hz, 1H), 3.49 (s, 3H), 3.08 (s, 3H), 2.42 (qd, J = 6.8, 3.2 Hz , 1H), 2.23 (m, 2H), 2.03 (m, 1H), 1.74 (m, 2H), 1.56 (m, 1H), 1.46 (m, 2H), 1.01 (d, J = 6.8 Hz, 3H), 0.98 (d, J = 6.8 Hz, 3H), 0.93 (d, J = 6.8 Hz, 3H), 0.85 (d, J = 6.8 Hz, 3H); HRESIMS m/z 497.1640 [M + Na]⁺ (calcd for C₂₁H₃₅⁷⁹BrN₂O₅Na, 497.1627), m/z [M + Na]⁺ 499.1616 (calcd for C₂₁H₃₅⁸¹Br N₂O₅Na, 499.1607) (100:100 [M + Na]⁺ ion cluster).

Preparation of MTPA Esters

The methanolysis product 9 was dissolved in 50 μL CDCl₃, and was divided into two equal portions and to each was added 0.75 mL triethylamine. To one portion was added 10 μL of (R)-MTPA-Cl and to another was added 10 μL of (S)-MTPA-Cl to give the (S)-MTPA ester (10) and (R)-MTPA ester (11), respectively. Each reaction was allowed to stir for 24 h and 10 μL of N,N-dimethylaminopropylamine was added to quench the reactions. The reaction products were dried under N2 and applied onto silica SPE eluting with EtOAc – hexanes (1:1). The semipure product was further purified by semipreparative HPLC (Phenomenex Phenyl-hexyl, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of MeOH–H₂O (70%–100% MeOH in 45 min and then 100% MeOH for 10 min) to yield 10 (t_R 38.0 min, 0.1 mg) or 11 (t_R 37.8 min, 0.1 mg).

10

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 7.57 (dd, J = 6.4, 2.7 Hz, 2H), 7.41 (m, 3H), 6.14 (d, J = 9.2 Hz, 1H), 5.28 (q, J = 6.9 Hz, 1H), 4.93 (d, J = 10.6 Hz, 1H), 4.68 (dd, J = 9.1, 7.6 Hz, 1H), 3.69 (s, 3H), 3.58 (s, 3H), 3.06 (s, 3H), 2.46 (quintet, J = 7.1 Hz , 1H), 2.21 (m, 1H), 2.15 (t, J = 6.8 Hz, 2H), 1.95 (m, 1H), 1.68 (m, 1H), 1.47 (m, 2H), 1.09 (d, J = 7.0 Hz, 3H), 1.01 (d, J = 6.6 Hz, 3H), 0.92 (d, J = 6.9 Hz, 3H), 0.86 (d, J = 6.9 Hz, 3H), 0.81 (d, J = 6.9 Hz, 3H); HRESIMS m/z 729.1746 [M + K]⁺ (calcd for C₃₁H₄₂⁷⁹BrF₃N₂O₇K, 729.1759), m/z [M + K]⁺ 731.1747 (calcd for C₃₁H₄₂⁸¹BrF₃N₂O₇K, 731.1755) (100:100 [M + K]⁺ ion cluster); LRESIMS m/z 691/693 (100:100 [M + H]⁺ ion cluster), 713/715 (100:100 [M + Na]⁺ ion cluster).

11

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 7.56 (dd, J = 4.3, 3.6 Hz, 2H), 7.41 (m, 3H), 6.27 (d, J = 8.3 Hz, 1H), 5.27 (q, J = 6.0 Hz, 1H), 4.94 (d, J = 10.4 Hz, 1H), 4.73 (dd, J = 9.1, 7.5 Hz, 1H), 3.69 (s, 3H), 3.57 (s, 3H), 3.07 (s, 3H), 2.54 (quintet, J = 6.7 Hz, 1H), 2.22 (m, 1H), 2.08 (td, J = 7.2, 2.9 Hz, 2H), 2.00 (m, 1H), 1.63 (m, 1H), 1.31 (m, 2H), 1.19 (d, J = 6.7 Hz, 3H), 1.01 (d, J = 6.3 Hz, 3H), 0.95 (d, J = 6.8 Hz, 3H), 0.90 (d, J = 7.0 Hz, 3H), 0.83 (d, J = 6.5 Hz, 3H); HRESI/APCIMS m/z 691.2202 [M + H]⁺ (calcd for C₃₁H₄₃⁷⁹BrF₃N₂O₇, 691.2206), m/z [M + H]⁺ 693.2186 (calcd for C₃₁H₄₃⁸¹BrF₃N₂O₇, 693.2186) (100:100 [M + H]⁺ ion cluster).

Advanced Marfey's Analysis of Dpv Unit in Dolastatin 16

To liberate the authentic (2S,3R)-Dpv standard, pitiprolamide (500 μg) was hydrolyzed (400 μL of 6 N HCl, 110 °C, 20 h). The mixture was dried and reconstituted in 400 μL H₂O. Half of this sample was used to epimerize the α-amino acids in this mixture (40 μL Et₃N and 40 μL Ac₂O, heating at 60 °C for 1 h)²⁷ to yield a mixture of (2S,3R)-Dpv and (2R,3R)-Dpv standards. These standards were derivatized with 1% (w/v) Marfey's reagent (L-FDLA and DL-FDLA).²⁸ Dolastatin 16 was hydrolyzed and derivatized with 1% (w/v) L-FDLA according to the procedure stated above. The absolute configuration of the Dpv unit was determined by reversed-phase HPLC-MS [column, Phenomenex Synergi Hydro-RP (150 × 4.6 mm) using a linear gradient of 0.1% HCOOH in MeOH– 0.1% aqueous HCOOH (40%–100% for 45 min and then 100% MeOH for 10 min); flow rate, 0.5 mL/min; detection by ESIMS in negative ion mode (MRM scan)] of the FDLA adducts. The acid hydrolysate of dolastatin 16 showed a peak at 36.1 min corresponding to (2S,3R)-Dpv. Coinjection of the acid hydrolysate with the DL-FDLA derivatized epimerization product showed enrichment of the peak corresponding to (2S,3R)-Dpv. The retention times (t_R, min; MRM ion pair) of the authentic amino acid L-FDLA adducts were as follows: (2S,3R)-Dpv (36.1; 486→191), (2S,3S)-Dpv [(36.6); (= D-FDLA derivatized (2R,3R)-Dpv)], (2R,3S)-Dpv [(43.5); (= D-FDLA derivatized (2S,3R)-Dpv)], (2R,3R)-Dpv (44.3). Compound-dependent parameters used were as follows: DP −95.0, EP −6.0, CE −31.0, CXP −2.0, CEP −26.0. Source gas parameters used were as follows: CUR 40, CAD High, IS 4500, TEM 450, GS1 40, GS2 40.

Advanced Marfey's Analysis of β-Amino Acid Unit in Dolastatin 12 and Dolastatin 16

Samples of dolastatin 12 and dolastatin 16 (0.2 mg each) were hydrolyzed (200 μL of 6 N HCl, 110 °C, 20 h) to liberate the β-amino acids Map and Dml, respectively. Another portion of dolastatin 12 and dolastatin 16 (0.2 mg each) was first heated with 3 N NaOH (80 °C, 3 h), the resulting product mixture acidified with 5% (v/v) HCl, dried under N₂, and hydrolyzed (200 μL of 6 N HCl, 110 °C, 20 h) to liberate the partially C-2 epimerized β-amino acid mixture. The reaction products were derivatized using L-FDLA and DL-FDLA.²⁸ The order of elution of the β-amino acids Dml and Map liberated from dolastatin 16 and dolastatin 12, respectively, was determined by reversed-phase HPLC-MS [column, Phenomenex Synergi Hydro-RP (150 × 4.6 mm) using a linear gradient of 0.1% HCOOH in MeOH− 0.1% aqueous HCOOH (25%–80% for 50 min and then 100% MeOH for 5 min); flow rate, 0.5 mL/min; detection by ESIMS in negative ion mode (MRM scan)] of the FDLA adducts. The enantiomers bearing the 3R configuration consistently eluted later than those with the 3S configuration. The retention times (t_R, min; MRM ion pair) of the Dml-L-FDLA adducts were as follows: (2S,3S)-Dml [(29.3; 438.2→310.3); (= D-FDLA derivatized (2R,3R)-Dml)], (2R,3S)-Dml [(30.8); (= D-FDLA derivatized (2S,3R)-Dml)], (2R,3R)-Dml (33.7), (2S,3R)-Dml (35.6). The retention times (t_R, min; MRM ion pair) of the Map-L-FDLA adducts were as follows: (2R,3S)-Map [(27.5; 424.5→310.4); (= D-FDLA derivatized (2S,3R)-Map)], (2S,3S)-Map [(29.6); (= D-FDLA derivatized (2R,3R)-Dml)], (2R,3R)-Dml (31.0), (2S,3R)-Dml (32.5). Compound-dependent parameters used were as follows: DP −95.0, EP −6.0, CE −31.0, CXP −2.0, CEP −26.0. Source gas parameters used were as follows: CUR 40, CAD Medium, IS 4500, TEM 450, GS1 40, GS2 40.

Base Hydrolysis of Dolastatin 16

Dolastatin 16 (4.0 mg) was hydrolyzed using 2 N KOH–MeOH (1:1) for 3 h at 80 °C. The base hydrolysate was neutralized with 2 N HCl, partitioned between EtOAc:H₂O, the organic layer collected and dried under N₂. The crude base hydrolysate product was purified by semipreparative HPLC (Phenomenex Synergi Hydro-RP, 250 × 10 mm, 4 μm; flow rate, 2.0 mL/min) using a linear gradient of ACN–H₂O (40%–100% MeOH in 45 min) to yield 12 (t_R 22.7 min, 1.0 mg).

12

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 7.13 – 7.26 (m), 6.81 (d, J = 9.0 Hz, 1H), 6.65 (d, J = 10.4 Hz, 1H), 4.83 (d, J = 10.6 Hz, 1H), 4.73 (dd, J = 8.7, 3.8 Hz, 1H), 4.59 (m, 2H), 4.48 (m, 1H), 4.28 (m, 1H), 3.79 (m, 1H), 3.56 – 3.69 (m, 2H), 3.50 (m, 1H), 3.36 (m, 1H), 3.16 (s, 3H), 2.70 (m, 2H), 2.24 – 2.48 (m, 6H), 1.94 – 2.19 (m, 2H), 1.77 – 1.94 (m), 1.14 (d, J = 6.8 Hz, 3H), 1.03 (d, J = 7.3 Hz, 3H), 0.96 (m, 6H), 0.90 (m, 6H), 0.80 (m, 6H); HRESI/APCIMS m/z 750.4407 [M + Na]⁺ (calcd for C₃₉H₆₁N₅O₈Na, 750.4418).

Modified Mosher's Analysis Using PGME

The base hydrolysate product was dissolved in 600 μL DMF-d₇ and divided into two portions. To one portion was added 0.7 mg (S)-PGME, and the solution cooled and stirred at 0 °C before the successive addition of PyBop (1.8 mg), HOBt (0.5 mg), and N-Me morpholine (5.0 μL). The reaction was left to stir for another 3 h at room temperature and quenched with the addition of EtOAc. The resulting solution was successively washed with 5% HCl, saturated NaHCO₃, and brine. The organic layer was collected and dried over anhydrous MgSO₄ before drying under N₂ to yield 13 (0.8 mg). The same procedure was used to prepare the (R)-PGME derivatized product 14 (0.6 mg). ¹H NMR chemical shifts of the Dml spin system were assigned by COSY and TOCSY analyses.

13

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 7.27 – 7.38 (m, 12H, incl. NH Dml at 7.27), 5.57 (dd, J = 14.0, 7.4 Hz, 1H), 5.02 (d, J = 10.2 Hz, 1H), 4.42 (m, 2H), 4.24 (d, J = 2.6 Hz, 1H), 3.74 (s, 3H), 3.56 – 3.65 (m, 3H, incl. H-3 Dml at 3.62), 3.45 (m, 2H), 3.29 (m, 1H), 2.99 (s, 3H), 2.75 (m, 2H, incl. H-2 Dml), 2.30 – 2.40 (m, 5H), 2.02 – 2.09 (m, 4H), 1.97 – 2.00 (m, 4H), 1.45 (m, 1H, H-5 Dml), 1.19 (d, J = 7.2 Hz, 3H, H₃-7 Dml), 1.04 (d, J = 6.6 Hz, 3H), 0.95 (d, J = 6.8 Hz, 3H), 0.94 (d, J = 6.7 Hz, 3H, H₃-6 Dml), 0.85 (d, J = 6.8 Hz, 3H), 0.82 (d, J = 6.7 Hz, 3H, H₃-5 Dml), 0.80 (d, J = 6.6 Hz, 6H); HRESI/APCIMS m/z 897.5099 [M + Na]⁺ (calcd for C₄₈H₇₀F₃N₆O₉Na, 897.5102).

14

colorless, amorphous solid; ¹H NMR (CDCl₃) δ 7.30 – 7.40 (m, 10H), 7.14 – 7.24 (m, 4H, incl. NH Dml at 7.21), 5.46 (d, J = 7.3 Hz, 1H), 4.99 (d, J = 10.2 Hz, 1H), 4.52 (dd, J = 6.4, 2.3 Hz, 1H), 4.42 (t, J = 6.4 Hz, 1H), 4.27 (d, J = 2.0 Hz, 1H), 4.01, (m, 1H), 3.72 (s, 3H), 3.56 – 3.67 (m, 3H, incl. H-3 Dml at 3.64), 3.02 (s, 3H), 2.65 – 2.68 (m, 2H, incl. H-2 Dml at 2.68), 2.30 – 2.40 (m, 5H), 1.89 – 2.02 (m, 8H), 1.62 (m, 1H, H-5 Dml), 1.13 (d, J = 6.8 Hz, 3H, H₃-7 Dml), 1.03 (d, J = 6.7 Hz, 3H), 0.98 (d, J = 6.8 Hz, 3H, H₃-6 Dml), 0.96 (d, J = 6.7 Hz, 3H), 0.89 (d, J = 6.7 Hz, 3H, H₃-5 Dml), 0.86 (d, J = 6.8 Hz, 3H), 0.81 (d, J = 6.7 Hz, 6H); HRESI/APCIMS m/z 897.5101 [M + Na]⁺ (calcd for C₄₈H₇₀F₃N₆O₉Na, 897.5102).

Cell Viability Assay

HT29 colorectal adenocarcinoma and HeLa cervical carcinoma cells were cultured in Dulbecco's modified Eagle medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum (FBS, Hyclone) under a humidified environment with 5% CO₂ at 37 °C. HeLa (3,000) and HT29 (12,500) cells were seeded in 96-well plates and treated with varying concentrations of 1–8, and dolastatin 16, positive control (paclitaxel), and solvent control (DMSO) after 24 h of seeding. The cells were incubated for an additional 48 h before the addition of the MTT reagent. Cell viability was measured according to the manufacturer's instructions (Promega). IC₅₀ values for positive control were 3.0 and 2.6 nM for HT29 and HeLa cells, respectively. IC₅₀ calculations were done by GraphPad Prism 5.03 based on duplicate experiments.