Bis-Indolyl Benzenoids, Hydroxypyrrolidine Derivatives and Other Constituents from Cultures of the Marine Sponge-Associated Fungus Aspergillus candidus KUFA0062

A previously unreported bis-indolyl benzenoid, candidusin D (2e) and a new hydroxypyrrolidine alkaloid, preussin C (5b) were isolated together with fourteen previously described compounds: palmitic acid, clionasterol, ergosterol 5,8-endoperoxides, chrysophanic acid (1a), emodin (1b), six bis-indolyl benzenoids including asterriquinol D dimethyl ether (2a), petromurin C (2b), kumbicin B (2c), kumbicin A (2d), 2″-oxoasterriquinol D methyl ether (3), kumbicin D (4), the hydroxypyrrolidine alkaloid preussin (5a), (3S, 6S)-3,6-dibenzylpiperazine-2,5-dione (6) and 4-(acetylamino) benzoic acid (7), from the cultures of the marine sponge-associated fungus Aspergillus candidus KUFA 0062. Compounds 1a, 2a–e, 3, 4, 5a–b, and 6 were tested for their antibacterial activity against Gram-positive and Gram-negative reference and multidrug-resistant strains isolated from the environment. Only 5a exhibited an inhibitory effect against S. aureus ATCC 29213 and E. faecalis ATCC29212 as well as both methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci (VRE) strains. Both 1a and 5a also reduced significant biofilm formation in E. coli ATCC 25922. Moreover, 2b and 5a revealed a synergistic effect with oxacillin against MRSA S. aureus 66/1 while 5a exhibited a strong synergistic effect with the antibiotic colistin against E. coli 1410/1. Compound 1a, 2a–e, 3, 4, 5a–b, and 6 were also tested, together with the crude extract, for cytotoxic effect against eight cancer cell lines: HepG2, HT29, HCT116, A549, A 375, MCF-7, U-251, and T98G. Except for 1a, 2a, 2d, 4, and 6, all the compounds showed cytotoxicity against all the cancer cell lines tested.


Introduction
Aspergillus candidus (Family Aspergillaceae) is a member of Aspergillus section Candidi [1]. This species frequently contaminates stored food and feeding stuff [2], and is one of the most frequently encountered mold in cereal grains and flour [3]. It also occurs in soil, usually on seeds or in the rhizosphere, and also in milk [4]. Strains of A. candidus produce a variety of secondary metabolites including chlorflavonin, a chlorine containing flavone antifungal antibiotic [5], p-terphenyl derivatives such as terphenyllin [6], deoxyterphenyllin [7], 3-hydroxyterphenyllin [8], candidusins A and B [9], immunosuppressant terprenins [10]. On the other hand, there are only a few reports on the chemical investigation of marine-derived A. candidus. Wei et al. [11] reported the isolation of cytotoxic prenylterphyllin, 4"-deoxyprenylterphyllin, 4"-deoxyisoterprenin and 4"-deoxyterprenin from A. candidus IF10, isolated from marine sediment. Recently, Wang et al. [12] described the isolation of spiculisporic acid derivatives including the new compounds, spiculisporic acids F and G.
In our pursuit for antibiotic and anticancer compounds from marine-derived fungi from the tropical sea, we have investigated secondary metabolites from cultures of A. candidus KUFA 0062, which were isolated from the marine sponge Epipolasis sp., collected from the coral reef at the Similan Island National Park in Phang-Nga province, Southern Thailand.
Compounds 1a, 2a-e, 3, 4, 5a-b, and 6 were tested for their antibacterial activity against four reference bacterial strains consisting of two Gram-positive (Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212) and two Gram-negative bacteria (Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853), three multidrug-resistant isolates from the environment, MRSA S. aureus 66/1, VRE E. faecalis B3/101, a colistin-resistant E. coli 1418/1, and a clinical isolate ESBL E. coli SA/2. The isolated compounds were also investigated for their capacity to inhibit biofilm formation in the four reference strains as well as for their potential synergism with the clinically used antibiotics against multidrug-resistant isolates from the environment. Moreover, these compounds were also evaluated for their cytotoxic effect against eight cancer cell lines, i.e., Hep G2 (human hepatocellular carcinoma), HT29 (human colorectal adenocarcinoma), HCT116 (human colorectal carcinoma), A549 (human lung carcinoma), A375 (human malignant melanoma), MCF7 (human mammary gland adenocarcinoma), U251 (human glioblastoma multiforme), and T98G (human glioblastoma astrocytoma), by MTT assay.
Compound 2e was isolated as a white solid (m.p. 299-300 • C), and its molecular formula C 28  ), indicating sixteen degrees of unsaturation. The IR spectrum showed absorption bands for amine (3346 cm −1 ), aromatic (1625, 1579 cm −1 ), and ether (1291 cm −1 ) groups. The general features of the 1 H and 13 C NMR spectra of 2e (Table 1, Supplementary Materials, Figures S18 and S19) resembled those of kumbicin A (2d), petromurin D (2b), and two bis-indolyl benzenoids isolated from the soil fungus Aspergillus kumbius [15]. The 13 C NMR spectrum (Table 1) exhibited twelve carbon signals which, in combination with DEPTs and HSQC spectra, can be categorized in six quaternary sp 2 (δ C 153.1, 147.6, 131.0, 127.4, 122.0, 106.7), four methine sp 2 (δ C 126.0, 111.9, 110.9, 102.1), and two methoxyl (δ C 60.3, 55.3) carbon signals. As the number of the carbon signals is less than the number of carbon atoms found in HRMS, the molecule must be symmetrical. The presence of the 3,5-disubstituted indole moiety with the methoxyl group on C-5 was supported by the presence of an amine proton at δ H 11.15 (d, J = 2.2 Hz), a doublet at δ H 7.43 (J = 2.5 Hz, H-2; δ C 126.0), and the proton signals of the 1,2,3 substituted benzene ring at δ H 6.78, dd (J = 8.8, 2.5 Hz, H-6; δ C 110.9), 6.88, d (J = 2.5 Hz, H-4; δ C 102.1), 7.33, dd (J = 8.8, H-7; δ C 111.9) as well as a singlet of OMe at δ H 3.72 (δ C 55.3). That another portion of the molecule was 2,3,5,6-tetramethoxy-1,6-disubstituted benzene ring was supported by the presence of the carbon signals at δ H 122.0 (C-1 and C-6) and δ C 147.6 (C-2, C-3, C-5, C-6) as well as four chemically and magnetically equivalent methoxyl groups at δ H 3.46 (δ C 60.3). Putting together the HRMS and NMR data, the structure of 2e was established as 5 -methoxykumbicin B. However, instead of having one methoxyl group and one hydroxyl group on C-5 and C-5" of the indole ring system like petromurin D, the substituents on both C-5 and C-5" are methoxyl groups in 2e. To the best of our knowledge, 2e is a new analogue of bis-indolyl benzenoids, which was named candidusin D.  Figures S20 and S21), revealed that this compound has the same planar structure as that 2"-oxoasterriquinol D methyl ether, an oxidized bis-indolyl benzenoid previously isolated from extracts of the sclerotia of Aspergillus sclerotiorum (NRRL 5167) [17]. Although Whyte et al. [17] have found that 2"-oxoasterriquinol D methyl ether was optically active, with [α] 25 D − 13.9 (c 0.001 g/mL, MeOH), they did not determine the absolute configuration of the stereogenic carbon (C-3") of the oxidized indole ring system of this compound. In order to fully elucidate the structure of 3, we have attempted to obtain suitable crystal for X-ray analysis. The results obtained, as seen in the ORTEP (Oak Ridge Thermal-Ellipsoid Plot Program) view in Figure 2, revealed that although 3 was levorotatory, it was not optically pure since both enantiomers were present in the crystal. The (+)-HRESIMS, 1 H, 13 C NMR (Table 2), COSY, HSQC, and HMBC (Supplementary Materials, Figures S24 and S25) data of 5a revealed that its planar structure is equal to that of the antifungal compound L-657, 398, first isolated from the fermentation of Aspergillus ochraceus ATCC 22947 [18]. However, the authors failed to establish the relative stereochemistry of this compound by 1 H NMR data. The same compound, called preussin, was later isolated as a yellow oil ([α] 25 D + 22.0, c 1.0, CHCl 3 ), from the mycelial cake obtained from fermentation of a fungus Preussia sp. [19]. By comparison of the chemical shift values of H 2 -1 and H 2 -6 of the (S)-and (R)-O-methylmandelate esters with those of the acetate derivative of preussin, Johnson et al. [19] determined the absolute configurations of C-2, C-3, and C-5 of preussin as 2S, 3S, and 5R.
Since we were able to obtain 5a as a suitable crystal, X-ray analysis was carried out to determine if the absolute configurations of its stereogenic carbons were the same as those of preussin. The ORTEP view of a protonated 5a, shown in Figure 3, not only determined the number of carbon atoms in the alkyl side chain, but also confirmed the absolute configurations of C-2, C-3, and C-5 as 2S, 3S, and 5R. To the best of our knowledge, this is the first X-ray crystal structure of preussin. Compound 5b was isolated as a white crystal (m.p. 219-220 • C), and its molecular formula was established as C 20 H 33 NO on the basis of its (+)-HRESIMS m/z 304.2647 [M + H] + , (calculated 304.2640 for C 20 H 34 NO), which indicated five degrees of unsaturation. The IR spectrum showed absorption bands for hydroxyl and amine (3441 cm −1 ) and aromatic (1634, 1581, 1539 cm −1 ). The general features of the 1 H and 13 C NMR spectra of 5b resembled those of 5a (Table 2, Supplementary Materials, Figures S26, S27 and S29), except for the absence of the N-methyl singlet at δ H 2.81 (δ C 36.7), which was confirmed by its molecular formula. Moreover, H 2 -1, H-3, H-4, and H 2 -6 appeared at lower frequencies while H-2 and H-5 resonated at higher frequencies when compared to the corresponding protons in 5a. Additionally, the chemical shift values of C-2 and C-5 in 5b are around 10 ppm lower than those of the corresponding carbons in 5a, while C-6 in 5b is around 3 ppm higher than that of C-6 in 5a. Therefore, 5b is a N-demethyl analog of 5a. This hypothesis was confirmed by COSY correlations from H-2 (δ H 3.46, m; δ C 65.2) to H-1a (δ H 3.11, dd, J = 13.  (Table 2 and Supplementary Materials, Figures S28 and 29). This was supported by HMBC correlations from H-1a and H-1b to C-2 and C-3; H-2 to C-1; H-3 to C-2, C-5; OH-3 to C-2, C-3, C-4; H-4 to C-3, C-5 and C-6, and from H 2 -6 to C-4, C-5, C-7 (Table 2 and Supplementary Materials, Figure S30). The relative stereochemistry of the pyrrolidine ring of 5b was established, based on NOESY experiments, to be the same as that of 5a. The NOESY spectrum (Table 2, Supplementary Materials, Figure S31) exhibited correlations from OH-3 to H-1a, H-3 and H-4α; H-3 to H-2, OH-3 and H-4β; H-4β to H-3, H-4α and H-5, and from H-6 at δ H 1.73, m to H-4α. Since 5a can be hypothesized to derive from 5b through N-methylation by SAM, it is legitimate to presume that the absolute configurations of C-2, C-3, and C-5 in 5b should be the same as those of 5a, i.e., 2S, 3S, 5R.
Since 5b could not be obtained as a suitable crystal for X-ray analysis, we have performed a calculated Electronic Circular Dichroism (ECD) spectrum to compare with its experimental ECD spectrum to determine unequivocally the absolute configurations of its C-2, C-3, and C-5. Experimental NOESY data of 5b, as well as the data obtained from X-ray crystallographic and Trost's O-methylmandelate method [19] of 5a, suggest that all the pyrrolidine substituents of 5b are directed to the same side of the ring. This formally narrows the investigation of the stereochemistry of 5b to the enantiomer pair (2S, 3S, 5R) and (2R, 3R, 5S), combined with the two possible configurations of the secondary amine. However, comparison of ECD simulated spectra for the above-mentioned configurations of 5b, in its neutral forms, with the experimental spectrum in methanol was inconsistent with the crystallographic configuration previously determined for its N-CH 3 analogue (5a). Since this analogue, most probably a metabolic derivative of 5b and therefore sharing the same (2S, 3S, 5R) configuration, precipitated and crystalized as an ammonium salt, the same was assumed for 5b and ECD simulations were therefore redirected to its protonated (positively charged) form.
This was expected to be a major difference since the calculated SCF HOMO and LUMO orbitals show that the electronic transitions happen between the two rings. The new charge in the five-membered ring alters significantly the electron density distribution and the molecular orbitals energies. The electronic transitions associated with the hydrocarbon side-chain, however, do not involve the five-membered ring and should remain unaffected. Additionally, these transitions are predicted to be caused by wavelengths below 200 nm, hence outside the observational window. Models were constructed based only on the protonated (2S, 3S, 5R) configuration because deprotonation upon dissolution of the secondary ammonium group is expected to be negligible. The typically large pKa values (≈10) associated with these very weak acid groups in water is maintained in methanol [24]. To have a good representation of 5b's conformational space, the three staggered orientations of C-1 (relative to C-2) and of 3-OH (relative to C-3) were combined with two five-membered ring conformations to give 18 models. One, the lowest MM2 energy, staggered conformation of C-6 was set in all models, as well as a staggered linear conformation for the carbon chain. The orientation of the phenyl group was always determined as a result of energy minimization. To have a notion of its energetic ranking, the conformational energy of each of the 18 models was first minimized using the semi-empirical PM6 method. Since the six lowest PM6-energy conformations accounted for 98% of the conformer population, as determined by a Boltzmann distribution of computed PM6 conformational energies, these were further minimized with the Amplitude Probability Density Function (APFD) Density Functional Theory (DFT) method. The four lowest-energy APFD conformers were selected for ECD simulation, also by the Boltzmann distribution significance. Figure 4 shows the lowest energy model. The ECD transitions of the four models were calculated to a minimum wavelength of 170 nm with the TDDFT method at the same level of theory than the final minimizations (APFD, with the same basis set and solvent model). The four sets of transitions were Boltzmann weighed using the calculated populations, Gaussian broadened by 0.17 eV and added to give the averaged simulated spectrum in Figure 5. As is apparent, the fit with the experimental spectrum is good, representing well its three main negative intensities (roughly at 215, 225 and 255 nm). This evidence supports the claim that the isolated 5b is the (2S, 3S, 5R) enantiomer ( Figure 4) in the ammonium form.  Chrysophanic acid (1a), asterriquinol D dimethyl ether (2a), petromurin C (2b), kumbicin B (2c), kumbicin A (2d), 2"-oxoesterriquinol D methyl ether (3), kumbicin D (4), preussin (5a), preussin C (5b), and (3S, 6S)-3,6-dibenzylpiperazine-2,5-dione (6) were tested for their antibacterial activity and their minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) for four reference strains consisting of three multidrug-resistant isolates from the environment and one clinical isolate. In the range of concentrations tested, none of the compounds were active against Gram-negative bacteria. However, 5a displayed an inhibitory effect against Gram-positive bacteria, with MIC values of 32 µg/mL for S. aureus ATCC 29213 and E. faecalis ATCC 29212. Moreover, 5a consistently showed a MIC value of 32 µg/mL for both methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci (VRE) strains. Compound 5a also displayed a MBC value of 64 µg/mL for the VRE E. faecalis B3/101 strain, however, it was not possible to determine the MBC for the other Gram-positive strains. This result suggests that 5a might not only have a bacteriostatic effect but also a bactericidal action.
The effect of 1a, 2a-e, 3, 4, 5a, b, and 6 on biofilm formation in four reference bacterial strains was also evaluated. Four concentrations of 5a, ranging from 2 × MIC to 1 4 MIC were tested against S. aureus ATCC 2913 and E. faecalis ATCC 2912. Since it was not possible to determine a MIC of the rest of the other compounds, the previously tested highest concentration that did not inhibit bacterial growth was used. Although none of the compounds tested showed an inhibitory effect on biomass production in P. aeruginosa ATCC 27853, the anthraquinone chrysophanic acid (1a) induced significant reduction in biofilm formation (67.7 ± 8.3% of control; One-sample t test: * p < 0.05, significantly different from 100%) in E. coli ATCC 25922. Interestingly, 5a also caused an almost 50% reduction of biofilm mass production (42.8 ± 32.7% of control) at a sub-inhibitory concentration in E. coli ATCC 25922. Although this inhibitory effect was not statistically significant, this result shows an interesting aspect in that 5a is active against both Gram-positive and Gram-negative bacteria. In fact, among the compounds tested, only 5a was capable of interfering with biofilm formation in S. aureus ATCC 29213 and E. faecalis ATCC 29212 at concentrations equal to or above the MIC (Figure 6), however, at lower concentrations, 5a had no effect on the biomass production of both strains. One-sample t test: ** p < 0.01 and *** p < 0.001, significantly different from 100%. MIC, minimum inhibitory concentration.
The tested compounds were also screened for their potential synergies with the clinically relevant antibiotics and some of them revealed small to moderate synergistic associations with antibiotic as determined by the disk diffusion method. Except 5a, which presented an inhibition zone of 8 mm for VRE E. faecalis B3/101 and MRSA S. aureus 66/1, the rest of the tested compounds showed no inhibition zone (0 mm), when tested alone. The combination of 2a-d, 3, 4, 5a, and 6 with cefotaxime in impregnated disks resulted in a minor synergistic effect, as can be observed by a small increase in the zone of inhibition when compared with the halo of inhibition produced by cefotaxime alone in an ESBL E. coli strain (SA/2). In the case of VRE E. faecalis B3/101, all the tested compounds induced a small increase in the halo of a partial inhibition of vancomycin when compared with vancomycin alone, however, this effect was more pronounced with 5a. On the other hand, only 2b and 5a revealed a synergistic effect with oxacillin against MRSA S. aureus 66/1. Interestingly, although 2b alone did not exhibit the antibacterial effect, it showed a marked synergistic effect with oxacillin, increasing the halo of inhibition from 0 mm, when oxacillin was used alone, to 7 mm when tested in combination. These results were also confirmed by the checkerboard method since the combination of 5a with vancomycin and oxacillin showed the values of ΣFIC ≤ 0.5 against VRE and MRSA isolates, respectively (Table 3). Interestingly, 5a also demonstrated a strong synergistic effect with colistin, reducing MIC of colistin from 8 µg/mL to less than 0.008 µg/mL, i.e., at least 100-fold, when tested against E. coli 1410/1 (Table 4). Following this result, lower concentrations of 5a were tested and it was found that a concentration of 8 µg/mL of 5a was enough to alter E. coli 1410/1 resistance to colistin, thus decreasing colistin MIC to 1 µg/mL (susceptibility breakpoint of ≤0.2 µg/mL) [25]. With the exception of 2b, the modest synergistic effects observed by the disk diffusion method were not replicated when the MIC of the antibiotics was determined in the presence of the tested compounds. The MIC of oxacillin, in a combination with 2b, was four folds less than that of oxacillin alone (from 64 µg/mL to 16 µg/mL) when tested against MRSA S. aureus 66/1. Although no effect on MIC of the antibiotics was observed with any other compounds, it is likely that since the method used for determination of the MIC requires two-fold serial dilutions, the synergistic effect between the antibiotics and the test compounds might not be enough to decrease the MIC of the antibiotics by a factor of two, in case the absolute difference between the two consecutive concentrations is large. This reasoning can justify the difference in the results obtained using two different methods.
The cytotoxic effect of the ethyl acetate crude extract of A. candidus KUFA 0062 (200 µg/mL) and 1a, 2a-e, 3, 4, 5a, b, and 6, at only one concentration (100 µM), were tested against eight cancer cell lines, i.e., Hep G2, HT29, HCT116, A549, A375, MCF7, U251, and T98G by MTT assay. At 200 µg/mL, the crude extract significantly decreased cell viability in all cell lines tested, with higher cytotoxic effect in A375 > HT29 > HCT116 > HepG2 > A549 > T98G > MCF7 > U251 cells. At 100 µM, 5a was found to decrease the cell viability in all cell lines tested, reaching the lowest % of cell viability in colon cancer cell lines (6.4% and 8.6% in HCT116 and HT29, respectively). Compound 5b also decreased cell viability in all of the cell lines tested, however, it was less effective than 5a. Except for human malignant glioma U251 and T98G cells, 2b significantly decreased the number of viable cells in all the cell lines tested (with cell viability less than 50). Similarly, 2e and 3 also exhibited a significant decrease of cell viability in all cell lines tested except for T98G and HepG2 for 2e and HepG2 for 3. Compound 2c significantly decreased cell viability in the HepG2, HT29, HCT116, A549, and A375 cells. Compounds  1a, 2a, 2d, 4 and 6, at 100 µM, did not change cell viability in any of the cancer cell lines tested (Table 5). In view of these results, dose-response curves were constructed to determine IC 50 values only for compounds and cell lines whose significant cytotoxic effects were observed. Doxorubicin (Dox) was used as a positive control. The IC 50 values and respective 95% confidence intervals are summarized in Table 6. Compound 5a was the most effective in all the cell lines tested with IC 50 ranging from 12.3 µM in HT29 cells to 74.1 µM in U251 cells while 2b also induced a significant decrease of cell viability with IC 50 values ranging from 34.8 µM in H29 cells to 94.8 µM in MCF7 cells. On the other hand, 3 was more cytotoxic in HT29 cells with IC 50 value of 43.2 µM but less cytotoxic in U251 cells with IC 50 value of 120.2 µM. Compound 5b showed IC 50 values ranging from 57.2 µM (in HT29 cells) to 215.7 µM (in A549 cells). Compound 2c and 2e significantly decreased cell viability but with higher IC 50 values ranging from 72.9 µM (2c in HCT116 cells) to 186.6 µM (2e in MCF7 cells). Over all, it can be observed that while 2b, 3, 5a, and 5b were more effective in HT29 cells, 2c and 2e were more active in HCT116 cells. This suggests that these compounds may act by different mechanisms of action since each cell line has different genetic characteristics. As to the positive control, Dox, we have observed a decrease in cell viability in all cell lines with IC 50 values ranging from 0.05 µM (in A375 cells) to 15.4 µM (in T98G cells). Our findings are in agreement with the literature showing that HT29 was more resistant to the cytotoxic effect of Dox than the HCT116 cells [26]. Table 5. Percentage of cell viability relative to control of 1a, 2a-e, 3, 4, 5a, b, 6 (100 µM) and crude extract of A. candidus KUFA 0062 (200 µg/mL) in eight human cancer cell lines after 48 h of incubation as assessed by MTT assay.
The results are the mean (SD) of at least four independent experiments, each in duplicate. Significant differences (** p < 0.01; *** p < 0.001 and **** p <0.0001) when compared with control cells were evaluated by one-way ANOVA, followed by the post-hoc Dunnett's test. # Indicates significant differences when 5b is compared with 5a, as evaluated by a t-test. Compounds/extract are marked in light gray when cell viability is equal to or greater than 50% and in dark gray when cell viability is lower than 50% relative to control. n.d.-not determined. The results are the mean (SD) of at least four independent experiments, each in duplicate. Significant differences (** p < 0.01; *** p < 0.001 and **** p < 0.0001) when compared with control cells were evaluated by one-way ANOVA, followed by the post-hoc Dunnett's test. # Indicates significant differences when 5b is compared with 5a, as evaluated by a t-test. Compounds/extract are marked in light gray when cell viability is equal to or greater than 50% and in dark gray when cell viability is lower than 50% relative to control. n.d.-not determined.

General Experimental Procedures
Melting points were determined on a Bock monoscope and are uncorrected. Optical rotations were measured on an ADP410 Polarimeter (Bellingham + Stanley Ltd., Tunbridge Wells, Kent, UK). Infrared spectra were recorded in a KBr microplate in a FTIR spectrometer Nicolet iS10 from Thermo Scientific (Waltham, MA, USA) with Smart OMNI-Transmission accessory (Software 188 OMNIC 8.3). 1 H and 13 C NMR spectra were recorded at ambient temperature on a Bruker AMC instrument (Bruker Biosciences Corporation, Billerica, MA, USA) operating at 300 or 500 and 75 or 125 MHz, respectively. High resolution mass spectra were measured with a Waters Xevo QToF mass spectrometer (Waters Corporations, Milford, MA, USA) coupled to a Waters Aquity UPLC system. A Merck (Darmstadt, Germany) silica gel GF 254 was used for preparative TLC, and a Merck Si gel 60 (0.2-0.5 mm) was used for column chromatography.

Fungal Material
The strain KUFA 0062 was isolated from the marine sponge Epipolasis sp., which was collected, by scuba diving at a depth of 15-20 m from the coral reef at Similan Island National Park (8 • 39 09" N, 97 • 38 27" E), Phang-Nga province, Southern Thailand, in April 2010. The sponge was washed with 0.06% sodium hypochlorite solution for 1 min, followed by sterilized seawater three times, and then dried on sterile filter paper under a sterile aseptic condition. The sponge was cut into small pieces (5 × 5 mm), and placed on Petri dish plates containing 15 mL malt extract agar (MEA) medium with 70% seawater and 300 mg/L of streptomycin sulfate and incubated at 28 • C for 7 days. The hyphal tips emerging from the sponge pieces were individually transferred onto a MEA slant and maintained as pure culture for further identification.
The fungus was identified as Aspergillus candidus, based on morphological characters as described by Varga et al. [27].This identification was supported by molecular techniques using ITS primers. DNA was extracted from young mycelia following a modified Murray and Thompson method [28]. Primer pairs ITS1 and ITS4 [29] were used for ITS gene amplification. PCR reactions were conducted on Thermal Cycler and the amplification process consisted of initial denaturation at 95 • C for 5 min, 34 cycles at 95 • C for 1 min (denaturation), at 55 • C for 1 min (annealing) and at 72 • C for 1.5 min (extension), followed by a final extension at 72 • C for 10 min. PCR products were examined by Agarose gel electrophoresis (1% agarose with 1 × TBE buffer) and visualized under UV light after staining with ethidium bromide. DNA sequencing analyses were sequenced using the dideoxyribonucleotide chain termination method [30] by Macrogen Inc. (Seoul, Korea). The DNA sequences were edited using FinchTV software and submitted into the BLAST program for alignment and compared with that of fungal species in the NCBI database (http://www.ncbi.nlm.nih.gov/). Its gene sequences were deposited in GenBank with accession numbers KX 431210. The pure cultures were deposited as KUFA 0062 at Kasetsart University Fungal Collection, Department of Plant Pathology, Faculty of Agriculture, Kasetsart University, Bangkok, Thailand.

Electronic Circular Dichroism
Experimental ECD spectra were obtained with a Jasco J-815 CD spectropolarimeter in a 0.1 mm cuvette. The model construction, dihedral driver search, and MM2 minimizations were done in Chem3D Ultra (Perkin-Elmer Inc., Waltham, MA, USA). All other conformational energy minimization and spectral calculations (TDDFT method) were performed with Gaussian 16W (Gaussian Inc., Wallingford, CT, USA) at the PM6 or APFD/6-311+G(2d,p) level [31] with the IEFPCM methanol solvation model. The simulated spectral lines were obtained by summation of Gaussian curves, as recommended in [32]. ECD spectra were added using Boltzmann weights derived from its minimal APFD energies [33].

X-ray Crystal Structure of 3 and 5a
Single crystals were mounted on cryoloops using paratone. X-ray diffraction data were collected at room temperature with a Gemini PX Ultra equipped with CuK α radiation (λ = 1.54184 Å). The structures were solved by direct methods using SHELXS-97 and refined with SHELXL-97 [34]. Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were either placed at their idealized positions using appropriate HFIX instructions in SHELXL and included in subsequent refinement cycles or were directly found from difference Fourier maps and were refined freely with isotropic displacement parameters.
Full details of the data collection and refinement and tables of atomic coordinates, bond lengths and angles, and torsion angles have been deposited with the Cambridge Crystallographic Data Centre. The absolute structure was established with confidence (Flack parameter = 0.018 (15)). The refinement converged to R (all data) = 3.42% and wR2 (all data) = 8.07%. (CCDC 1823438).

Bacterial Strains and Growth Conditions
Four reference and three multidrug-resistant bacterial strains were used in this study. The Gram-positive bacteria comprised Staphylococcus aureus ATCC 29213, Enterococcus faecalis ATCC 29212, MRSA S. aureus 66/1 isolated from public buses [35], and VRE E. faecalis B3/101 isolated from river water [36]. The Gram-negative bacteria used were Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, a colistin-resistant E. coli 1418/1 strain isolated from rabbit feces, and a clinical isolate ESBL E. coli SA/2. Frozen stocks of all strains were grown in Mueller-Hinton agar (MH-BioKar diagnostics, Allone, France) at 37 • C. All bacterial strains were sub-cultured in MH agar and incubated overnight at 37 • C before each assay.

Antimicrobial Susceptibility Testing
The minimum inhibitory concentration (MIC) was used for determining the antibacterial activity of each compound, in accordance with the recommendations of the Clinical and Laboratory Standards Institute (CLSI) [37]. With the exception of 6, 10 mg/mL stock solutions of each compound were prepared in dimethylsulfoxide (DMSO-Applichem GmbH, Darmstadt, Germany). For 6, which was less soluble in DMSO than the other compounds, a stock solution of 1 mg/mL was prepared. Two-fold serial dilutions of the compounds were prepared in Mueller-Hinton broth 2 (MHB2-Sigma-Aldrich, St. Louis, MO, USA) within the concentration range of 0.062-64 µg/mL, except for 6 for which the highest concentration tested was 32 µg/mL. The highest concentration tested was chosen in order to maintain DMSO in-test concentration below 1% as recommended by the CLSI [37]. At this concentration DMSO did not affected bacterial growth. Cefotaxime (CTX) ranging between 0.031 and 16 µg/mL was used as a control. A purity check and colony counts of the inoculum suspensions were also evaluated in order to ensure that the final inoculum density closely approximates the intended number (5 × 10 5 CFU/mL). The MIC was determined as the lowest concentration at which no visible growth was observed. The minimum bactericidal concentration (MBC) was assessed by spreading 10 µL of culture collected from wells showing no visible growth on MH agar plates. The MBC was determined as the lowest concentration at which no colonies grew after 16-18 h incubation at 37 • C. These assays were performed in duplicate.

Biofilm Formation Inhibition Assay
The effect of all compounds on biofilm formation was assessed using crystal violet staining as previously described [38]. Briefly, the highest concentration tested in the MIC assay was added to bacterial suspensions of 1 × 10 6 CFU/mL prepared in Tryptic Soy broth (TSB-BioKar diagnostics, Allonne, France). When it was possible to determine a MIC, four concentrations were tested: 2 × MIC, MIC, 1 2 MIC, and 1 4 MIC. A control without any compound as well as a negative control (TSB alone) were included. The stabilized biofilm mass was quantified after 24 h incubation at 37 • C. The absorbance was measured at 595 nm on an iMark™ microplate spectrophotometer (Bio-Rad Laboratories, Hercules, CA, USA). The background absorbance (TSB without inoculum) was subtracted from the absorbance of each sample and data are shown as a percentage of control. Three independent experiments were performed in triplicate for each experimental condition.

Antibiotic Synergy Testing
In order to evaluate the combined effect of the compounds and clinically relevant antimicrobial drugs, a screening was conducted using the disk diffusion method, as previously described [38,39]. A set of antibiotic disks (Oxoid, Basingstoke, UK) to which the isolates were resistant was selected: cefotaxime (CTX, 30 µg) for E. coli SA/2, oxacillin (OX, 1 µg) for S. aureus 66/1, and vancomycin (VA, 30 µg) for E. faecalis B3/101. Antibiotic disks alone (controls) and antibiotic disks impregnated with 15 µg of each compound were placed on MH agar plates seeded with the respective bacteria. Sterile 6 mm blank paper disks (Oxoid, Basingstoke, UK) impregnated with 15 µg of each compound alone were also tested. A blank disk with DMSO was used as a negative control. MH inoculated plates were incubated for 18-20 h at 37 • C. Potential synergism was recorded when the halo of an antibiotic disk impregnated with a compound was greater than the halo of the antibiotic or compound-impregnated blank disk alone.
The potential synergy between the most promising compounds and clinically relevant antibiotics (oxacillin and vancomycin-Sigma-Aldrich, St. Louis, MO, USA) was also evaluated by the checkerboard method as previously described [38]. For this purpose multidrug-resistant strains were selected based on their resistance towards those antibiotics. The fractional concentration (FIC) was calculated as follows: FIC of drug A (FIC A) = MIC of drug A in combination/MIC of drug A alone, and FIC of drug B (FIC B) = MIC of drug B in combination/MIC of drug B alone. The FIC index was calculated as the sum of each FIC and interpreted as follows: ΣFIC ≤ 0.5, synergy; 0.5 < ΣFIC ≤ 4, no interaction; 4 < ΣFIC, antagonism [40]. When it was not possible to determine a MIC value for the test compound, the potential synergy between the compounds and clinically relevant antibiotics was evaluated by determining the antibiotic MIC in the presence of each compound. Briefly, the MIC of CTX, OX, VA, and colistin (Sigma-Aldrich, St. Louis, MO, USA) for the respective multidrug-resistant strains was determined in the presence of the highest concentration of each compound tested, unless otherwise stated, that did not affect bacterial growth when the compound was used alone. The antibiotic tested was serially diluted whereas the concentration of each compound was kept fixed. In the case of 6 the concentration used was 32 µg/mL. For all other compounds the concentration used was 64 µg/mL. Antibiotic MICs where determined as described above.

Cell Lines
Eight cancer cell lines were used to assess the anticancer activity of the extract and isolated compounds of the marine sponge-associated fungus Aspergillus candidus KUFA 0062. HT29, HCT116 cells were kindly provided by Prof. Carmen Jerónimo, from CI-IPO, Porto and HepG2 cells were provided by Prof. Rosário Martins, from ESTSP and CIIMAR, Porto. A375, A549, MCF7, U-251 and T98G cell lines were obtained from the European Collection of Cell Cultures. Cells were maintained as monolayer cultures in DMEM supplemented with 10% FBS, 1% antibiotic solution (100 U/mL penicillin and 100 µg/mL streptomycin), 10 mM N- [2-hydroxyethyl] piperazine-N -[2-ethane-sulfonic acid] and 0.1 mM sodium pyruvate in a humidified incubator with 5% CO 2 at 37 • C. Cells were trypsinized near the confluence. Stock solutions of the isolated compounds, extract and Dox were prepared in dimethyl sulphoxide (DMSO) and aliquots and kept at −20 • C. For experiments, the final concentration of DMSO in the medium was < 0.1% (v/v) and the controls received DMSO only.

MTT Reduction Assay
To assess the effects of the isolated compounds and the extract on cell viability, cells were plated in 96-multiwell culture plates at a density of 0.1 × 10 6 cells/mL. After cell adhesion, cells were incubated for 48 h with fresh medium containing the isolated compounds, at 100 µM, and extract, at 200 µg/mL, and cell viability was measured by MTT reduction assay as described previously [41]. Briefly, after 48 h of treatment MTT was added at a final concentration of 0.5 mg/mL and incubated for 2 h at 37 • C. The formazan crystals were dissolved in a DMSO: EtOH solution (1:1) (v/v) and the absorbance was measured at 570 nm in a microplate reader (Multiskan Ex, Labsystems, Milford, MA, USA).
The determination of IC 50 (half-maximal inhibitory concentration) was performed only for the isolated compounds that significantly decrease cell viability at 100 µM. For this, cancer cell lines were incubated with different concentrations of isolated compounds (0-200 µM) or Dox (0.001-10 µM), as a positive control, and the MTT assay was performed as described above. The IC 50 values were determined using GraphPad Prism v6.0 software (GraphPad Software, La Jolla, CA, USA). This software was also used to conduct the statistical analysis, using a one-way ANOVA, followed by the post-hoc Dunnett's test. The normality and homogeneity of variance were confirmed by the Shapiro-Wilk test and Levine test, respectively. The significance level was set at the conventional 95%.

Conclusions
Most of the investigations of the secondary metabolites of the fungus Aspergillus candidus were carried out with its terrestrial strains. Except for a chlorine containing flavone antifungal antibiotic chlorflavonin, the secondary metabolites produced by this fungal species were p-terphenyl and bis-indolyl benzenoids derivatives, some of which immunosuppressive activity, cytotoxicity against sea urchin embryos and various tumor cell lines. Interestingly, to the best of our knowledge, only two marine-derived A. candidus were investigated chemically. While A. candidus, isolated from the marine sediment furnished p-terphenyl derivatives, the strain isolated from the gut of sea urchin produced a different group of metabolites, i.e., spiculisporic derivatives. This fact has prompted us to investigate the secondary metabolites from a marine-sponge associated strain of A. candidus to verify if the marine host can influence the type of secondary metabolites produced by this fungus. Our results were very interesting since we isolated, besides the commonly found fungal metabolites such as palmitic acid, ergosterol-5,8-endoperoxide, the anthraquinones chrysophanic acid and emodin, a diketopiperazine derivative, and bis-indolyl derivatives that are the chemical signature of this fungus, one previously reported hydroxypyrrolidine alkaloid preussin (5a) and a new congener, preussin C (5b). Although preussin (5a) was isolated from A. ochraceus and other fungi (Preussia sp. and Simplicillium lanosoniveum), this is the first report of the isolation of hydroxypyrrolidine alkaloids from A. candidus. From the biological perspective, it is interesting to note that although most of the tested secondary metabolites exhibited a cytotoxic effect against eight human cancer cell lines, their activity varied depending on the cell lines. The most notable aspect was that the presence of the N-methyl group on the pyrrolidine ring in preussin (5a) was always crucial in both antibacterial and cytotoxic activities. This was demonstrated by the effectiveness of preussin (5a) to decrease cell viability when compared to the N-demethyl analogue, preussin C (5b). For the antibacterial activity assay, it is important to point out that, among the compounds tested, only preussin (5a) stood out since it displayed a relevant inhibitory effect against Gram-positive bacteria as well as methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci (VRE) strains. Moreover, preussin (5a) was not only able to interfere with biofilm formation in S. aureus ATCC 29213 and E. faecalis ATCC 29212 at concentrations equal to or above the MIC but also showed synergistic effect with vancomycin and oxacillin against VRE and MRSA isolates and a strong synergistic effect with colistin when tested against E. coli 1410/1. Therefore, our study reveals that preussin (5a) can represent a new model for the development of a new class of antibiotic and anticancer agents.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-3397/16/4/119/s1, Figure S1: Structures of palmitic acid, clionasterol and ergosterol-5,8-endoperoxide, isolated from the marine sponge-associated fungus Aspergillus candidus KUFA0062, Figures S2-S35: 1D and 2D NMR spectra of isolated compounds, Costa and Ângela Inácio performed and interpreted the results of antibacterial assays; Alice A. Ramos and Eduardo Rocha performed and interpreted the results of in vitro anticancer assays; Nazim Sekeroglu assisted in the elaboration of the manuscript; Michael Lee provided HRMS; Artur M. S. Silva provided the NMR spectra.

Conflicts of Interest:
The authors declare no conflict of interest.