Marine-derived fungi are considered as rich sources of various classes of bioactive secondary metabolites, of which meroterpenoids are shown to have unique structures with a wide range of chemical diversity. Meroterpenoids that formed via a mixed terpenoid–polyketide biosynthetic pathway were found in many fungal species and marine organisms [1]. In the course of our ongoing searching for the bioactive secondary metabolites from marine-derived fungal isolates [2], we recently reported the isolation of two new meroterpenoids named furanoaustinol and 7-acetoxydehydroaustinol from the extracts obtained from cultures of a marine-derived isolate of Penicillium sp. SF-5497 [3]. In our continuing efforts to explore the chemistry of this isolate, five meroterpenoids including two new preaustinoid-related meroterpenoids were encountered, and this study describes the isolation, structural elucidation, and PTP1B inhibitory effects of these metabolites from the EtOAc extract of Penicillium sp. SF-5497.

PTP1B, a member of the protein tyrosine phosphatase (PTP) superfamily, is considered as a critical regulator of multiple signaling networks involved in human disorders, such as diabetes, obesity, and cancer [4]. Biochemical, genetic, and pharmacological studies have indicated that PTP1B plays as a negative regulator in both insulin and leptin signal pathways. PTP1B down-regulates insulin signaling cascades via the tyrosine dephosphorylation of the insulin receptor or insulin receptor substrates 1 and 2, preventing their interactions with downstream signaling molecules. PTP1B also negatively regulates the leptin signaling pathway by dephosphorylating phosphorylated tyrosine kinase JAK2 [5]. In vivo studies have demonstrated an increase in insulin sensitivity, glycemic control, and resistance to a high fat diet in PTP1B-deficient mice [6]. Reduction of PTP1B level decreases the adipose tissue mass, plasma insulin, and blood glucose levels in diabetic animal models [7]. These biochemical and pharmacological evidences suggested that the inhibition of PTP1B could be an effective strategy in the treatment of obesity, diabetes, and cancer, and that the development of new and effective PTP1B inhibitors is important and necessary. In line with this, PTP1B inhibitory effects of the isolated compounds were evaluated herein.

The fungal strain SF-5497 was cultured on 45 Fernbach flasks of potato dextrose agar (PDA) media containing 3% NaCl at 25 °C for 14 days. The culture plates were extracted with EtOAc, and the resultant extract was concentrated in vacuo to provide a residue. The residue was subjected to multiple chromatographic steps, including reversed-phase C18 column chromatography, and semi-preparative HPLC, to yield compounds 1–5 (Fig. 1a).

Fig. 1
figure 1

a Chemical structures of compounds 1–5. b Key HMBC or HMQC (JCH = 8 Hz) and COSY correlations of compounds 1 and 2. c Key NOESY correlations of compounds 1 and 2

Full size image

Compound 1 was isolated as a colorless gum. Its molecular formula was established as C26H38O8 based on the observation of a sodium adduct ion peak at m/z 501.2438 [M + Na]+ in the HRESITOFMS, along with analysis of 1H and 13C NMR data, implying eight degrees of unsaturation. The 1H NMR spectrum (Table 1) displayed signals for two olefinic protons [δH 5.37 and 4.84 (each s, H-1′a and H-1′b)], six tertiary methyl groups [δH 1.23 (s, H3-12), δH 0.94 (s, H3-13), δH 1.17 (s, H3-14), δH 1.24 (s, H3-15), δH 1.45 (s, H3-9′), and δH 1.36 (H3-10′)], and one methoxy group at δH 3.70 (s). The 13C NMR and DEPT spectra (Table 1) indicated the presence of 26 carbon signals, including seven methyls, six methylenes, two methines, and 11 non-protonated carbons (including four carbonyl carbons). In addition, analysis of the 13C NMR spectrum suggested the presence of signals corresponding to two ketone carbonyl carbons [δC 208.1 (C-4′) and δC 203.8 (C-6′)], one carboxylic methyl ester [δC 168.7 (C-8′) and δC 52.5 (8′-OCH3)], an exocyclic double bond [δC 112.4 (C-1′) and δC 145.8 (C-2′)], two quaternary carbons [δC 50.9 (C-3′) and δC 72.7 (C-7′)], a quaternary carbinolic carbon (δC 80.1, C-5′), and two methyl groups [δC 22.1 (C-9′) and δC 15.0 (C-10′)], suggesting the presence of a 2-hydroxy-2-methyl-1,3-dioxo partial structure (D ring), bearing methyl ester functionality [8, 9]. The assignment of this partial structure was confirmed by the HMBC correlations from H2-1′ to C-2′, C-3′, and C-7′; from H3-9′ to C-2′, C-3′, and C-4′; from H3-10′ to C-4′, C-5′, and C-6′; and from δH 3.70 to C-8′ (Fig. 1b). Further detailed analysis of the 1H and 13C NMR spectroscopic data of 1 and comparison with those of the reported meroterpenoid, preaustinoid A1 revealed that the structures of these two compounds are very similar, but the difference was clearly presented at the A ring of preaustinoid A1 [9]. The up-field shifted 13C chemical shifts of two methylene groups [δC 33.6 (C-1) and δC 27.8 (C-2)] and a non-protonated carbon (δC 75.8, C-4), as well as the down-field shifted chemical shift of the carbonyl carbon (δC 178.4, C-3) suggested the presence of a 4-hydroxy-3,4-seco-3-oic acid structural moiety in 1. The partial structure corresponding to a 4-hydroxy-3,4-seco-3-oic acid moiety in 1 was assigned with the aid of HMBC correlations from H3-14 and H3-15 to C-4 and C-5; from H2-1 to C-2 and C-3; and from H2-2 to C-3, along with the COSY correlation between H2-1 and H2-2. This was further confirmed by a conversion of 1 to methyl ester 1a. Upon methylation reaction in the presence of MeOH and trimethylsilyldiazomethane (TMSCHN2), and followed by preparative TLC, compound 1 was converted to a methylated product 1a, which was recently isolated from Penicillim sp. [10] and Eupenicillium sp. [11]. The relative configuration of 1 was determined by NOESY spectrum as well as the computer-generated lower energy conformation using MM2 force field calculations (Fig. 1c). In the NOESY spectrum, H-5 showed NOE correlation with H-9, indicating that H-5 and H-9 are both β-oriented. NOE correlations between H3-13/H-11a and H3-12; between H3-12/H-11a and H3-9′; and between H3-9′/H-11a revealed that H3-12, H3-13, H3-9′, and H-11a are α-oriented. Furthermore, NOE correlations observed between H-5/H-11b and H3-10′ indicated the β-orientation of H3-10′. Thus, the structure of 1 was established as shown in Fig. 1a, and was given the trivial name preaustinoid A6. The numbering system shown in 1 was chosen by analogy to that used in the literature for preaustinoids [8, 9].

Table 1 1H and 13C NMR data for preaustinoid A6 (1) and preaustinoid A7 (2)
Full size table

Compound 2 was isolated as a colorless gum. Its molecular formula was determined to be C26H36O7 (nine degrees of unsaturation) based on the observation of an ion peak at m/z 483.2356 [M + Na]+ in the HRESITOFMS data. The 1H NMR data (Table 1) showed signals for four olefinic protons [δH 5.28, δH 4.76 (each s, H-1′a, H-1′b); δH 4.84 and δH 4.64 (each s, H-14a and H-14b)], five tertiary methyl groups [δH 1.30 (s, H3-12), δH 0.81 (s, H3-13), δH 1.70 (s, H3-15), δH 1.44 (s, H3-9′), and δH 1.34 (s, H3-10′)], and one methoxy group at δH 3.63 (s). Analysis of 13C NMR, DEPT, and HMQC (JCH = 8 Hz) correlations (Table 1) displayed the presence of 26 carbon signals, including six methyls, seven methylens, two methines, and 11 non-protonated carbons (including four carbonyl carbons). These data for 2 were similar to those of 1 except for the presence of signals corresponding to the terminal olefinic group [δC 148.0 (C-4) and δC 114.6 (C-14)], and the absence of a methyl group and one oxygenated carbon. The position of the terminal olefinic group was confirmed by 1H–13C long-range correlations of H-14a and H-14b with C-4, C-5, and C-15; H-15 with C-4 and C-14 (Fig. 1b). All the other key 1H–13C long-range correlations for 2 (Fig. 1b) enabled to assign the complete structure of this compound as shown in Fig. 1a, and named preaustinoid A7. The relative configuration of 2 was deduced by NOESY spectroscopic analysis. As with the case of 1, the relative configuration of 2 was determined on the basis of the NOESY data. NOE correlations between H-5/H-9 and H-9/H3-10′ suggested that these groups are on the β-oriented. On the other hand, NOE correlations of H3-12 with H3-13, H3-9′, and H-14; of H3-13 with H-14 and H3-9′; and of H-1′a with H3-9′ suggested that these groups are on the α-oriented. Thus, the relative configuration of 2 was suggested as shown in Fig. 1a.

The structures of three known compounds were identified as berkeleyone C (3) [12], preaustinoid A2 (4) [9], and preaustinoid A3 (5) [13] by analyses of NMR and MS, along with comparisons of these data with those reported in the literature (Fig. 1a). The absolute configurations of 1a, 3, and 5 have been determined by ECD spectrum [11]. Since we observed negative specific rotations for 1a and 3, and positive-specific rotations for 5, consistent with those in the literature, it was suggested that compounds 1a, 3, and 5 have the same absolute configurations as those defined in the literature [11]. In addition, the absolute configuration of compounds 1 and 2 were proposed to be analogous to those for 1a, 3, and 5 because they were produced by the same fungal strain.

As evidenced by many biochemical, genetic, and pharmacological studies, inhibition of PTP1B could be helpful in the treatment of human disorders, such as diabetes, obesity, and cancer [14]. To date, more than 300 PTP1B inhibitors have been discovered and developed from natural resources, of which two groups of phenolics and terpenoids have emerged as potential PTP1B inhibitors [15]. Although many natural PTP1B inhibitors exhibited promising clinical potential, there are no clinically used PTP1B inhibitors, which is most likely due to relatively low activities or lack of selectivity. Thus, searching for more potent and selective PTP1B inhibitors is still necessary. In the present study, the isolated compounds were evaluated for their inhibitory effects on PTP1B activity, and the result showed that compounds 1 and 3 inhibited PTP1B activity in a dose dependent manner, with IC50 values of 17.6 and 58.4 µM, respectively, while compounds 1a, 2, 4, and 5 exhibited no inhibitory activity up to 100 μM. It is interesting to note that the conversion of acid functionality to methyl ester or minor structural modification at the A part of the compound 1 led to the significant reduction in PTP1B inhibitory activity.

To identify the PTP1B inhibitory characteristic of the active compound 1, kinetic analysis was conducted using different concentrations of the compounds and p-NPP. The initial rate was determined on the basis of the rate of increase in absorbance at 405 nm. The Michaelis–Menten constant (Km) and maximal velocity (Vmax) of PTP1B were determined by Lineweaver–Burk plot analysis for competitive inhibition and the intercept on the vertical axis for noncompetitive inhibition (GraphPad Prism 5.01). As shown in Fig. 2, compound 1 lowered the apparent value of Vmax and increased the Km value, indicating that 1 inhibited PTP1B in a noncompetitive manner, with the Ki value of 17.0 µM. In addition, the plot in Fig. 2 displayed the lines converged to the left of the 1/[V] axis and above the 1/[S] axis (α > 1). This observation suggested that compound 1 preferentially bound to the free enzyme than to the enzyme–substrate complex [16]. It is noted that the discovery of noncompetitive PTP1B inhibitors targeting the active site or allosteric of the enzyme could be potentially developed into effective and bioavailable PTP1B inhibitors, which are most challenging features of PTP1B inhibition-based drugs [17, 18]. Previously, berkeleyone C (3) was reported to possess inhibitory effect against the production of interleukin-1β from induced inflammasomes in THP-1 cell line [12]. Although the PTP1B inhibitory effects of several meroterpenoids have been reported [19], to the best of our knowledge, this is the first case reporting the PTP1B inhibitory effects of the preaustinoid-related meroterpene derivatives. Further studies to assess the selectivity, bioavailability, and efficacy of these compounds are necessary to evaluate their potential as lead compounds in the treatment of diabetes and obesity based on the inhibition of PTP1B activity.

Fig. 2
figure 2

Lineweaver–Burk plots for inhibition of PTP1B-catalyzed hydrolysis of p-NPP by compound 1. Data are expressed as mean initial velocity for triplicate experiments (n = 3) at each substrate concentration

Full size image