Tryptoquivalines W and X, two new compounds from a Hawaiian fungal strain and their biological activities

Abstract

Two new compounds tryptoquivalines W (1) and X (2) were isolated from a Hawaiian soil fungal strain Aspergillus terreus FS107. The soil sample was collected on the top of Mauna Kea, the tallest mountain in Hawaii. The structures of compounds 1 and 2 were determined on the basis of MS spectroscopic and NMR analysis, and NMR calculation. The absolute configuration (AC) was determined by ECD calculations. Compounds 4 and 5 showed inhibition against NF-κB with IC₅₀ values of 3.45 and 6.76 μM, respectively.

Fungi, including unicellular yeast, multicellular mold, mildews, rusts, smuts, and mushrooms, are widely distributed all over the world. It was suggested that there may be as many as 120,000 species of microfungi within the United States of America and 1.5 million worldwide. Fungi in the genus Aspergillus are well-known as a source of biologically active secondary metabolites [1]. The Aspergillus Secondary Metabolites Database (A2MDB) provides a phylogenetic representation of over 2500 strains, catalogs 807 secondary metabolites from 675 Aspergillus species, and presents a detailed chemical information of secondary metabolites [2]. Among the widely distributing Aspergillus strains are A. flavus, A. fumigatus, A. niger, A. tubingensis, A. oryzae, A. versicolor, and A. terreus etc [2]. The common producers of secondary metabolites in the genus Aspergillus include A. niger, A. terreus, A. versicolor, A. flavus, A. tubingensis, and A. fumigatus, etc [2]. Aspergillus species produce not just mycotoxins, but also biosynthesize diverse types of secondary metabolites. A. terreus is a saprotrophic fungal species found in soil sources throughout the world. They are prevalent in tropical and subtropical regions, but they are also found in very harsh environmental conditions. In our continuous search of new and biologically active compounds from Hawaiian fungi [3–16], we studied a fungal strain, Aspergillus terreus FS107 that was isolated from a soil sample collected on the top of the Mauna Kea mountain, the highest mountain in the State of Hawaii, and isolated six secondary metabolites, including two new compounds (1 and 2) and four known molecules (3–6) (Fig. 1). Herein, we present the isolation, structural elucidation, plausible biosynthesis and biological evaluation of these metabolites.

Fig. 1.

Structures of compounds 1–6.

Compound 1 [17] was isolated as a colorless powder. Its molecular formula was determined to be C₂₇H₃₀N₄O₇ by HR-ESIMS (m/z 523.2194, calcd for [M + H]⁺ 523.2193), with fifteen degrees of unsaturation. The ¹H NMR and HSQC spectra (Table 1) of 1 exhibited the presence of eight aromatic protons in the low field, four methyl groups in the high field, and six protons between them including four methines and one methylene. In the ¹H–¹H COSY spectrum of 1 (Fig. 2), four spin systems were identified including two AA’BB’ spin systems, one 1-oxygenated 2-methylpropyl group [(CH₃)₂CH-CH(O-], and a –(COO)CH(N)–CH₂-spin system that accounted for the α- and β-protons of an amino acid. In the HMBC spectrum, the two methyl singlets showed correlations to a carbonyl carbon at δ_C 173.1 and a nirogenated tertiary carbon at δ_C 70.8 indicating a 2-methyl alanine (MeAla) or 2-amino-isobutanoic acid (AIBA) unit in the molecule. Further HMBC analysis indicated the presence of two 1,2-disubstituted aromatic rings, one of which should be derived from tryptophan with the indole being reduced to form an indoline moiety. Based on the four fragments identified (Fig. 2), a search on the “Dictionary of Natural Products” with the molecular weight defined from 500 to 550 Dalton revealed that 1 was very similar to tryptoquivaline A (3) [18], N-dehydroxy tryptoquivaline A (deoxytryptoquivaline) (4) [19], and especially O-deacetyl-tryptoquivaline A (5) [18], which were also isolated in this study. In the HMBC of 3, H-7 and H-23 showed correlations to C-9 and C-13, respectively, which further confirmed the presence of anthranilic acid and tryptophan moieties in these analogs. Since the molecular weight of 1 was 522 Dalton, which is 18 units more than that of 5, it was readily concluded that 1 was an acid with a carboxyl and a hydroxyl group at 13-position instead of a γ-lactone, which was supported by an IR band in the region of 1600–1700 cm⁻¹. From the biosynthetic point of view, we argued that compound 1 should have the same configuration [11(S), 13(R), 14 (R), and 25(R] as that of the known compound, tryptoquivaline A (3). To verify our hypothesis, we carried out NMR calculations of 1 [11(S), 13(R), 14(R), and 25(R)] and its 25-epimer [1-epimer: 11(S), 13(R), 14(R), and 25(S)] [20]. Each configuration was submitted for the configurational search by using the OPLS_2005 force field in MacroModel with an energy window of 5.02 kcal/mol. Geometry optimization and frequency calculation were performed at the B3LYP/6–31 + G(d,p) level in gas phase. NMR shielding tensors were computed with the GIAO method at the B3LYP/6–311 + G (2d,p) level in methanol. For the ECD calculation, both geometry optimization and TDDFT calculation were performed at the APFD/6–311 + G(2d,p) level with methanol as solvent. Results showed that the calculated NMR data of 1 matched the experimental NMR data of 1 better than 1-epimer (Table 2). To confirm our hypothesis, we carried out ECD calculations for 1, 1-epimer and 1e (enantiomer of 1) [21,22]. The experimental ECD spectrum of compound 1 showed a strong negative Cotton effect at 220250 nm. The calculated weighted ECD spectrum of the stereoisomer 1 coincides well with the recorded ECD spectrum of 1 rather than 1-epimer and 1e (Fig. 3). Hence, the structure of 1 was determined as shown.

Table 1.

¹H and ¹³C NMR data of 1 and 2.

	1		2
no.	δH, J (Hz)a	δCb	δH, J (Hz)a	δCb
1		168.9		168.9
3
4	7.83 d (7.8)	127.9	7.83 d (7.8)	127.9
5	7.49 overlapped	131.4	7.54 overlapped	131.6
6	7.21 t (7.8)	125.1	7.22 overlapped	125
7	8.45 d (7.8)	120.8	8.50 d (7.8)	120.8
8
9		175.4		174.4
10
11	5.04 dd (3.1, 8.7)	50.6	5.04 dd (3.1, 8.70)	50.1
12	2.51 dd (3.1, 15.0)	39.8	2.50 dd (3.1, 15.0)	38.5
	2.77 dd (8.7, 15.0)		2.80 dd (8.7, 15.0)
13		76.6		76.5
14	5.26 s	85.1	5.13 s	84.2
15
16		70.8		70.5
17		173.1		173
18
19		137.3		137.3
20	7.48 d (7.8)	115.1	7.50 overlapped	115.1
21	7.36 t (7.8)	129.4	7.39 t (7.8)	129.5
22	7.20 t (7.8)	123.3	7.21 overlapped	123.1
23	7.53 d (7.8)	124.5	7.52 overlapped	124.3
24		137.3		137.3
25	3.97 d (3.4)	76.2	3.98 d (3.4)	76.3
26	2.18 m	31.5	2.14 m	32
27	0.85 d (7.0)	14.8	0.89 d (7.0)	15.22
28	1.05 d (7.0)	18.2	1.04 d (7.0)	18.4
29	1.38 s	21.6	1.34 s	21.8
30	1.45 s	15.9	1.40 s	15.9
31		172.6		172.6
32			3.77 s	51.3

^aSpectra recorded at 400 MHz.

^bSpectra recorded at 100 MHz. Data based on ¹H, ¹³C, HSQC, and HMBC experiments.

Fig. 2.

A. Key COSY (Bold) and HMBC (Single headed) correlations of 1; B. Four fragments identified in 1.

Table 2.

Calculated ¹H and ¹³C NMR data of 1 and 1-epimer.

	1		1-epimer
no.	δH, J (Hz)	δδC	δH, J (Hz)	δC
1		160.91		157.69
3		147.30		147.93
4	7.67	127.05	7.72	127.91
5	7.80	135.99	7.86	136.05
6	7.53	127.59	7.55	127.86
7	8.27	127.52	8.29	128.07
8		121.37		120.75
9		163.19		165.03
10
11	5.34	58.51	6.63	56.39
12	2.63/3.50	37.39	3.00/3.64	38.30
13		82.86		80.08
14	5.39	84.25	5.34	88.87
15
16		80.55		76.04
17		173.83		173.10
18
19		136.94		138.31
20	7.51	113.87	7.41	115.87
21	7.32	130.64	7.34	130.31
22	7.08	125.10	7.14	125.52
23	7.23	124.68	7.43	124.65
24		141.21		140.31
25	5.40	75.94	4.50	77.01
26	2.20	39.32	2.54	37.77
27	1.08	15.52	1.23	17.23
28	1.12	19.36	1.09	17.23
29	1.49	21.28	1.51	18.04
30	1.44	19.44	1.36	22.37
31		172.60		173.90
32			3.75	51.0

Fig. 3.

ECD spectra of experimental and calculated 1, 1e, and 1-epimer.

Compound 2 [17] was isolated as a colorless solid. Its molecular formula was determined to be C₂₈H₃₂N₄O₇ by HR-ESIMS (m/z 537.1501, calcd for [M + H]⁺ 537.2349), which was 14 units more than that of 1. Analysis of NMR data (Table 1) indicated the presence of an extra methoxy group when compared with that of compound 1. Hence, the structure of compound 2 was readily determined as the methyl ester of compound 1, and it was named as tryptoquivalines X.

Besides the new compounds 1 and 2, four known compounds tryptoquivaline A (3) [18], deoxytryptoquivaline (4) [19], O-deacetyl-tryptoquivaline A (5) [18], and epifiscalin E (6) [23] were also purified. The structures of these known compounds were determined based on comparisons of NMR and HRESIMS data with previously reported data.

Biogenetically, all the compounds isolated in this study could be derived from a cyclic tripeptide-like precursor (I, valine-tryptophan-anthranilic acid) [24]. Nucleophilic attack from the nitrogen at the α-position of tryptophan on the carbonyl carbon of valine would yield intermediate II. Hydrolysis of the amide bond in the lactam ring with an isopropyl group could produce intermediate III. Oxidation followed by coupling between IVb and 2-methyl alanine would generate V. Hydroxylation at 15-position of V followed by oxidative deamination would produce the new compound 1, which could be esterified to compound 2. Cyclization between 13-hydroxy group and the carboxyl group at 11-position would generate compound 5, which would yield compound 3 via acetoxylation of the 25-hydroxy group. Compound 4 could be produced from V after oxidative deamination, acetoxylation of the 25-hydroxy group and cyclization between 13-hydroxyl and the carboxyl group at 11-position. Coupling of the oxidative product of II with MaAla followed by hydroxylation 25-position would generate compound 6 (Fig. 4).

Fig. 4.

Proposed biosynthetic pathway for compounds 1–6.

Tryptoquivalines are widely distributed in nature. So far, twenty two tryptoquivalines (A-V) [18,25–34] have been isolated. These tryptoquivalines were isolated from two fungal genera. Tryptoquivalines A-O were obtained from Aspergillus clavatus (or Aspergillus fumigatus), while tryptoquivalines P-V from Neosartorya species (N. laciniosa, N. takakii, and N. pseudofischeri). It was reported that tryptoquivalines A and B showed tremorgenic property, and tryptoquivaline O demonstrated antifungal activity. Besides the two subclasses (1–5, and 6) as identified in this study, some tri- or tetra-peptide precursors formed different or even more complicated molecules with diverse ring systems, for examples, fumiquinazoline K (7) [27], fumiquinazoline D (8) [29,35], quinadoline B (9), [36] fumiquinazoline C (10) [35], fumiquinazoline J (11) [35,37], and isochaetominine C (12) [34] (Fig. 5).

Fig. 5.

Some analogs (7–12) of compounds 1–6.

Compounds 1–6, which were pure enough for NMR analysis and bioassays, were tested for their anti-proliferative activity against A2780 human ovarian cancer cells [38] and antibacterial activity against S. aureus and E. coli [39], but none was active at 40 μM, and 100 μM, respectively. When evaluated in a mammalian cell-based assay designed to monitor TNF-α-induced NF-κB activity [40], compounds 4 and 5 showed NF-κB inhibition with IC₅₀ values of 3.45 and 6.76 μM, respectively. Compounds 1–6 were also evaluated for their cytotoxicity against the human embryonic kidney cells 293 (HEK 293) using the same conditions as the NF-κB assay, and no cytotoxicity was observed at a concentration of 50 μM. Hence, in the absence of cytotoxicity, inhibition of NF-κB activity suggests the potential of mediating a cancer chemopreventive response.