Alkaloids with Cardiovascular Effects from the Marine-Derived Fungus Penicillium expansum Y32

Three new alkaloids (1, 4 and 8), together with nine known analogues (2, 3, 5–7, and 9–12), were isolated from the marine-derived fungus Penicillium expansum Y32. Their structures including the absolute configurations were elucidated by spectroscopic and Mosher’s and Marfey’s methods, along with quantum electronic circular dichroism (ECD) calculations. Each of the compounds was evaluated for cardiovascular effects in a live zebrafish model. All of the compounds showed a significant mitigative effect on bradycardia caused by astemizole (ASM) in the heart rate experiments. Compounds 4–6 and 8–12 exhibited potent vasculogenetic activity in vasculogenesis experiments. This is the first study to report that these types of compounds show cardiovascular effects in zebrafish. The results suggest that these compounds could be promising candidates for cardiovascular disease lead compounds.


Introduction
According to the World Health Organization, cardiovascular diseases (CVDs) are the number one cause of death globally: more people die annually from CVDs than from any other cause. In 2012, 17.5 million people, representing 31% of all global deaths, died from this disease. CVDs are disorders of the heart and blood vessels and include coronary heart disease, cerebrovascular disease, peripheral arterial disease, rheumatic heart disease, congenital heart disease and other conditions [1]. Unfortunately, the number of people suffering from CVDs is on the rise, but only few and expensive drugs are available to treat the diseases. It is of great significance to search new and effective drugs to combat CVDs. In order to discover related lead compounds from marine-derived fungi, a fungus Penicillium expansum Y32 has been isolated from a seawater sample collected from the Indian Ocean. A chemical investigation of the ethyl acetate extract of fermentation broth of Y32 led to the identification of three new alkaloids, named communesin I (1), fumiquinazoline Q (4) and protuboxepin E (8), along with nine known analogues, communesin A and B (2 and 3) [2][3][4], cottoquinazoline A (5) [5], prelapatin B (6), glyantrypine (7) [6], protuboxepin A and B (9 and 10) [7], chaetoglobosin C (11) [8] and penochalasin E (12) [9] (Figure 1).

Figure 1. Structures of Compounds 1-12.
Compounds 1-12 were screened for cardiovascular effects in a model of live zebrafish, which provides a system for drug screening that combines the biological complexity of in vivo models with the ability for much higher throughput screening than other available animal models [10]. In the heart rate experiments, all of the compounds showed significant mitigative effects on bradycardia caused by astemizole (ASM) at different concentrations. In the vasculogenesis experiments, Compounds 4-6 and 8-12 (at concentrations of 20 μg/mL, 50 μg/mL and 100 μg/mL) exhibited potent vasculogenetic activity regarding vessel numbers, and four hits (4, 6, 8 and 10) displayed remarkable promoting functions regarding vessel length. Additionally, 12 compounds were also evaluated in antiangiogenic experiments, and no obvious activity was observed (the results are not listed). The results indicated that these compounds might be used for screening for new natural cardiovascular effect candidates.
The relative configuration was determined by nuclear Overhauser effect spectroscopy (NOESY) experiments and Murata's J-based method [3,11]. The NOESY correlations of H-1/H-9 (δH 5.05)/H-11 (δH 4.11) /H-12 (δH 6.05), H3-1′ (δH 2.84)/H-6 (δH 4.70)/H2-19 (δ 2.37/2.28), and Ha-19 (δH 2.37)/Ha-18 (δH 2.71) confirmed the two -CH2-CH2-bridges, H-6 and H3-1″, to be on the same side, which displayed a similar correlative pattern to those of 2 ( Figure 3). Application of Murata's JH-H-based method [3] enabled to determine the relative stereochemistry at C-21. The large coupling constant between H-21 and H-11 (JH-H = 9 Hz) indicated an approximate 180° (anti arrangement) of the dihedral angles of H11-C11-C21-H21 ( Figure 4). Additionally, the NOE correlations of H3-23/H3-24/H-12 and the absent correlations of the geminal methyl groups to H-9 or H2-19/20 supported the theory that the epoxide oxygen was oriented syn to N-10 as shown in conformation A in Figure 4.     [4], indicating the same 6R, 7R, 8R, 9S, 11S and 21R configurations. This was further confirmed by quantum chemical equivalent circulating density (ECD) calculation of 1 and ent-1 using Gaussian 09 ( Figure 5) [12]. The preliminary conformational distribution search was performed using the HyperChem 7.5 software. The corresponding minimum geometries were further fully optimized using the density functional theory (DFT) at the B3LYP/6-31G(d) level as implemented in the Gaussian 09 program package [12]. The stable conformers obtained were submitted to ECD calculation using the time-dependent DFT (TDDFT) method (B3LYP/6-31G(d)). The overall predicted ECD spectrum of 1 was subsequently compared with the measured one. Finally, the measured CD curve matched well with the calculated curve for 1 and was opposite to that of ent-1 ( Figure 5). The absolute configuration at C-3″ was determined using the modified Mosher's method [13].    [14]. In the 1 H NMR spectrum, the signals for two ortho-disubstituted benzene ring system were present at   Figure 2) correlations from H2-3 (δH 4.70/4.49) to C-4 (δC 148.0) and C-1 (δC 169.9). Thus, the planar structure of 4 was established as shown in Figure 2. The relative configuration of 4 could be deduced from the NOESY data ( Figure 3). The NOESY correlations of Ha-15 (δ 2.37)/H-17 (δ 5.25)/H-19 (δ 4.25) suggested that H-17/H-19 was in cis-configuration and H-17/OH-16 was in trans-configuration. The absolute configuration at C-19 was confirmed by acidic hydrolysis of 4, which afforded L-Ala that was determined by Marfey's method [16]. The absolute configuration at C-14 was also determined by ECD calculations of 4 and ent-4 [12]. The result showed that CD curve of 4 was consistent with the calculated ECD curve of 4 but opposite to that of ent-4 ( Figure 7). Therefore, the absolute configuration of 4 was unambiguously assigned as (14R, 16R, 17S, 19S).  , which was further confirmed by COSY (Figure 2). The 13 C NMR data (Table 1) revealed the presence of 22 carbon signals, sorted by DEPT experiments into two methyls, two methylenes, eleven methines including eight aromatic methines, and seven quaternary carbons including two conjugated carbonyl groups (δC 167.4, C-2; 160.0, C-13). The NMR data of 8 implied that it was an analogue of protuboxepin A [7], except for the obvious upfielded shift for C-7 and downfielded shifts for C-8 and C-12 in 8. Meanwhile, C-9 in protuboxepin A had been transformed from a methine into a quaternary carbon in 8, which implied a replacement of the 7-membered ring A in protuboxepin A by a hexatomic benzene ring in 8 with the ether bond in protuboxepin A changed to a hydroxyl in 8. Marfey's analysis of the acid hydrolysis of 8 afforded a D-phenylalanine residue. Therefore, the absolute configurations of C-1 and C-4 were determined as R and S, respectively. Murata's J-based method and Marfey's analysis were applied to assign the absolute stereochemistry at C-22, but it was unfortunately unsuccessful because of the uncertainty of the single bond C4-C22, which could rotate freely. Thus, the configuration of C-22 has not been determined.
The partial absolute stereostructure of the known Compound 5 was first determined by the Marfey's method [5]. According to the L-Ala residue in acidic hydrolysis of 5, the partial absolute stereostructure of 5 was unambiguously assigned as (16R, 17S, 19S).
Zebrafish embryos of the AB wild-type strain and TG (VEGFR2: GFP) type strain with fluorescent blood vessels were used to screen Compounds 1-12 for cardiovascular effects [18][19][20]. In the heart rate experiments, all of the compounds showed a significant mitigative effect on bradycardia caused by astemizole (ASM) at different concentrations ( Figure 8). Figure 9 gives representative examples of the vasculogenetic effect observed in zebrafish by the test compounds. Regarding number of vessels, Compounds 4-6 and 8-12 (at concentrations of 20 μg/mL, 50 μg/mL and 100 μg/mL) exhibited potent vasculogenetic activity and Compounds 1, 2 and 7 had moderate effects, while Compound 3 was ineffective. Regarding vessels length, four hits (4, 6, 8 and 10) displayed remarkable promoting function; mild effects were identified for Compounds 1, 2, 5, 7 and 11. Compounds 3, 9 and 12 did not show any relevant activity. As additional investigations aiming at screening for cardiovascular effects in the model of live zebrafish, 12 compounds were also evaluated in an antiangiogenic experiment, and no obvious activity was observed. This is the first report showing cardiovascular effects of these compounds in zebrafish.  shown as red arrows). Significant difference between compound treatment and PTK787 (* p < 0.05; ** p < 0.01). The 1‰ DMSO-treated group is represented as "Control".

Fungal Material and Fermentation
The fungus Penicillium expansum Y32 was originally obtained from a seawater sample collected from a depth of about 30 m in the Indian Ocean (88°59′51″ E, 2°59′54″ S) in 2013. The sample was cultured in a Malt Extract Agar (MEA: Malt extract, 17 g; Peptone, 3 g; Agar, 20 g; sea water, 1 L) plate using chloramphenicol (100 μg/mL) as a bacterial inhibitor. A single colony was transferred onto another MEA plate and was identified according to its morphological characteristics and 18S rRNA gene sequences (GenBank access No. KP872504, supplementary materials). The fungus Penicillium expansum Y32 was cultured in 1000 mL conical flasks containing 300 mL fermentation media consisting of 1.7% malt extract and 0.3% peptone at 28 °C for 30 days.

Extraction and Purification
The whole fermentation broth (15 L) of cultivated medium was extracted exhaustively with EtOAc, while the mycelia were extracted with 80% volume aqueous acetone. After removing the acetone by evaporation under vacuum, the obtained aqueous acetone solution was extracted three times with EtOAc. The combined EtOAc extracts were dried under vacuum to produce 10.1 g of extract.

Marfey's Analysis
A solution of Compound 4 (0.5 mg)/Compound 5 (0.5 mg) in 6 M HCl (0.5 mL) was heated at 110 °C for 19 h. Then the solution was evaporated to dryness. The residue, L-Ala and D-Ala were dissolved in H2O (250 μL each), respectively. One hundred microliters of each solution was treated with 100 μL of 1% solution of L-FDAA in acetone followed by 1.0 M NaHCO3 (40 μL). The reaction mixture was incubated for 1 h at 45 °C before being quenched by 1.0 M HCl (40 μL). The derivatives of the hydrolysates and standard amino acids were analyzed by HPLC column (YMC-pack ODS-A, 10 × 250 mm, 5 μm, 1 mL/min) at 30 °C using the following gradient program: solvent A, water + 0.2% TFA; solvent B, MeCN; linear gradient: 0 min 25% B, 40 min 60% B, 45 min 100% B; UV detection at 340 nm. HPLC analysis showed that the retention times for L-FDAA derivatives of hydrolysates of 4/5, standard L-Ala and standard D-Ala were 11.7, 11.7 and 14.3 min, respectively ( Figures S26 and S27).
Similarly, Compound 8, standard L-phenylalanine and D-phenylalanine were also derivatized with L-FDAA and analyzed by HPLC column. HPLC analysis revealed that acid hydrolysates of 8 displayed the same retention time (24.1 min) as the D-phenylalanine but was different from L-phenylalanine (21.8 min) ( Figure S28).

Zebrafish Embryos
Zebrafish (Danio rerio) of the AB wild-type strain and TG (VEGFR2:GFP) type strain were maintained under a 14 h light/10 h dark cycle in an automatic circulating tank system and fed with artificial granular bait and fresh Artemia nauplii [21]. Adult mating pairs were placed in a breeding tank in the evening, and fertilized eggs were collected in the next morning (9 h). After disinfected, fertilized eggs raised in culture solution (5.0 mM NaCl, 0.17 mM KCl, 0.4 mM CaCl2, and 0.16 mM MgSO4) in a light-operated incubator at 28.0 °C ± 0.5 °C [22].

Heart Rate Experiments
Zebrafish embryos cultured for 48 h were arrayed into 24-well microtiter plates (six to eight embryos and 1000 μL culture solution per well). Each test compound was diluted with DMSO, and 1 μL of the working stock was added to each well, resulting in a final screening concentration of 1, 10 and 100 μg/mL, respectively. Then 2 μM ASM was added to all model groups. Two micromols of ASM was used as positive control and 1 ‰ DMSO was tested as negative control. After 24 h incubation with test compounds, the heart rate of each zebrafish was recorded by the inverted microscope (XSJ-D) (Figure 8) [10].

Vasculogenesis Experiments
Egg membranes were removed from embryos by pronase E solution (1.0 mg/mL) (Shanghai, China) at 24 h post fertilization. Then zebrafish embryos were added to 24-well microtiter plates (six to eight embryos per well) treated with 20, 50 and 100 μg/mL of each test compound and 4 μg/mL of vatalanib (PTK787, Basel, Switzerland). The positive control was 4 μg/mL PTK787 and the negative control was 1‰ DMSO (Shanghai, China). Before arraying the plate, unhealthy or developmentally delayed embryos were removed by examination under a stereoscopic microscope. After 24 h incubation in a light-operated incubator at 28.0 °C ± 0.5 °C, the number and length of intersegmental vessels were captured using a fluorescent microscope (SZX16 Tokyo, Japan) or image acquisition systems (DP2-BSW, Tokyo, Japan) ( Figure 9) [10].

Antiangiogenic Vessel Growth Experiments
Zebrafish embryos cultured for 24 h were placed in 24-well microtiter plates, with six to eight embryos and 1000 μL culture solution per well. Embryos were exposed to the compound solutions at 1, 10 and 100 μg/mL. At 48 hpf (hours post fertilization), the influence on antiangiogenic vessel growth of each zebrafish was observed using inverted microscope (XSJ-D, Chongqing, China) and an automated custom algorithm was used to quantify the number of angiogenic vessels in the zebrafish.

Conclusions
Three new alkaloids (1, 4 and 8), together with nine known analogues (2, 3, 5-7, and 9-12), were isolated from the marine-derived fungus Penicillium expansum Y32. Their structures including the absolute configurations were elucidated by spectroscopic and Mosher's and Marfey's methods, along with quantum ECD calculations. Each of the compounds was evaluated for cardiovascular effects with live a zebrafish model. All of the compounds showed significant mitigative effects on bradycardia caused by astemizole (ASM) in the heart rate experiments. Compounds 4-6 and 8-12 exhibited potent vasculogenetic activity in the vasculogenesis experiments. The results suggested that these compounds could be promising candidates for cardiovascular disease lead compounds.