Structure and Antimicrobial Activity of Rare Lactone Lipids from the Sooty Mold (Scorias spongiosa)

Abstract

Two new lactone lipids, scoriosin (1) and its methyl ester (2), with a rare furylidene ring joined to a tetrahydrofurandione ring, were isolated from Scorias spongiosa, commonly referred to as sooty mold. The planar structure of these compounds was assigned by 1D and 2D NMR. The conformational analysis of these molecules was undertaken to evaluate the relative and absolute configuration through GIAO NMR chemical shift analysis and ECD calculation. In addition to the potent antimicrobial activities, compound 2 strongly potentiated the activity of amphotericin B against Cryptococcus neoformans, suggesting the potential utility of this compound in combination therapies for treating cryptococcal infections.

Graphical Abstract

Microorganisms, such as fungi, are a proven source of structurally diverse metabolites possessing a wide array of biological activity.^1-3 As part of an effort focused on screening for biologically active natural products, the secondary metabolites produced by Scorias spongiosa have been explored. This is a type of sooty mold in the phylum Ascomycota. Species from this phylum have produced a diverse array of biologically active secondary metabolites.^4-8 We now report the isolation and structure elucidation of scoriosin (1), a new biologically active lactone lipid, as well as the methyl ester (2), along with a known compound previously isolated from Aspergillus sp.

The methanol extract of S. spongiosa displayed antimicrobial and antifungal activity in disc-diffusion assays. Bioassay-guided fractionation of the extract using normal phase flash chromatography, followed by reversed phase HPLC, resulted in the isolation of several new compounds. Compound 1 was assigned the molecular formula C₂₃H₃₄O₆ (7 double-bond equivalents) on the basis of the HRESIMS data in the Experimental Section. The ESI-MS spectrum of 1 also displayed a strong [M − 1] signal at m/z 405. These data coupled with IR resonances at 3079 (broad) and 1699 (strong) cm⁻¹ suggested the presence of a carboxylic acid functional group.

The IR spectra also displayed strong signals at 1757 and 1699 cm⁻¹, supporting the presence of multiple carbonyl groups. Resonances in the UV spectra at 215 and 325 nm supported the presence of a conjugated carbonyl system. The ¹³C NMR data and DEPT (see Table 1) analysis indicated the presence of 23 carbon signals, including two methyl groups, 12 methylenes, three methines, and six fully substituted carbons. The carbon resonance at δ 173.9 was attributed to the carboxylic function in agreement with the mass spectrometry and IR data. Signals were also assigned to an ester (δ 167.9) and an α,β-unsaturated ketone (δ 196.4), which was also consistent with the IR spectral data. Other low-field signals included a 1,2-disubstituted olefin unit (δ 123.0 and 160.8) and three fully substituted carbons (δ 180.3, 102.1, and 93.3). These assignments accounted for four of the seven degrees of unsaturation in compound 1.

Table 1.

¹H and ¹³C NMR Spectra Data of 1 and 2

	1	2	1	2
carbon	δC (ppm)	δC (ppm)	δH (ppm) (mult., J in Hz)	δH (ppm) (mult., J in Hz)
1	167.9, C	169.7, C
2	93.3, C	93.7, C
3	196.4, C	196.4, C
4	78.6, CH	78.7, CH	4.87 dd (7, 3)	4.83 dd (8,4)
5	36.1, CH2	36.0, CH2	2.98 dd (17, 3)	2.95 dd (17, 4)
			2.73 dd (17, 7)	2.77 dd (17, 7)
6	173.9, C	170.2, C
7		52.1, CH3		3.66(s)
1′	180.3, C	180.0, C
2′	123.0, CH	123.2, CH	7.45 d (5)	7.49 d (6)
3′	160.8, CH	160.33, CH	7.48 d (5)	7.42 d (6)
4′	102.1, C	102.4, C
5′	37.9, CH2	38.0, CH2	1.91 m	1.92 m
6′	24.0, CH2	24.0, CH2	1.18 m	1.19 m
7′–13′	29.5, CH, CH2	29.5, CH, CH2	1.18 m	1.19 m
14′	31.9, CH2	31.9, CH2	1.18 m	1.19 m
15′	22.6, CH2	22.7, CH2	1.18 m	1.19 m
16′	14.1, CH2	14.1, CH2	0.81 t (6)	0.82 t (6)
17′	23.2, CH3	23.3, CH3	1.56 s	1.57 s

The ¹H NMR data of 1 included signals assignable to two olefin doublets at δ 7.45 (J = 5 Hz) and 7.48 (J = 5 Hz); a heterosubstituted methine at δ 4.87 (dd, J = 7, 3 Hz), a diasterotopic methylene at δ 2.98 (dd, J = 17, 3 Hz) and 2.73 (dd, J = 17, 7 Hz), a methyl singlet at δ 1.56, and a methyl triplet at δ 0.81. The broad signal at δ 1.18, which integrated for 20 protons, was characteristic of a straight chain alkyl unit.

Analysis of the COSY data allowed the construction of three partial structures corresponding to C4–C5, C2′–C3′, and C5′–C16′. Analysis of the HMBC data generated two partial structures (Chart 1). Two- and three-bond HMBC correlations from H(4) to C(1), C(3), C(5), and C(6) and from H(5) to C(3), C(6), and C(4) afforded one partial structure. The alkane chain was linked with the diene subunit by analysis of three- and two-bond HMBC correlations from Me-17′ to C(3′) and C(5′) and from H(2′) and H(3′) to C(4′). However, these two subunits could not be connected due to a lack of sufficient HMBC correlations to unambiguously place C(1′) and C(2). The connectivity of these subunits and placement of C(1′) and C(2) were accomplished by comparison of 1 with lowdenic acid (4). The structure of 4, which had the same absence of correlations, was elucidated using X-ray crystallography.⁹

Chart 1.

Key HMBC Correlations of 1

Compound 2 was structurally similar to 1 except for an additional singlet at 3.66 ppm in the proton spectra and the addition of one carbon at 52.1 ppm. This was an indication of the presence of a methyl ester in place of the carboxylic acid functionality.

The X-ray diffraction data for 4 only allowed for the relative configuration to be assigned.¹⁰ For 2, GIAO shielding chemical shift analysis¹⁰ was performed using Macromodel (version 9.9, Schrodinger LLC) to perform the conformational searches and the Gaussian 09 package to calculate the chemical shift values. The data for the two diastereomers were then submitted to DP4 probability analysis¹¹ to determine the correct relative configuration of C-4 and C-4′. The results showed that there was a 97.7% probability when both the proton and the carbon data were taken into account that the relative configuration was opposite for the two stereogenic centers (Figure 1).

Figure 1.

DP4 analysis to determine the relative configuration of C-4 and C-4′.

To determine the absolute configuration of the molecules, the ECD spectra were measured and based on comparison with calculated ECD data, position 4′ possessed the R absolute configuration. This together with the DP4 calculations indicated that the absolute configuration of the molecule is 4S,4′R (Figure 2). This absolute configuration agrees with the absolute configuration of pestalotic acid, which was assigned in 2012 using a similar ECD approach and crystallography techniques that established the relative configuration.

Figure 2.

Experimental vs computational ECD data for 2.

The structure of 1 is highly conjugated with the presence of multiple carbonyl and double-bond functionalities. Such a compound has the potential to undergo intramolecular tautomerization to give isomers that vary in selected configuration. In the present study, it was observed that 1 did undergo tautomerization, which was reflected by changes in the NMR spectrum. The NMR spectrum of 1 indicated the presence of a set of minor signals when recorded after a period of time. Repurification of the sample on silica gel gave pure 1 with a distinctive NMR spectrum initially. However, a similar set of minor NMR signals emerged when the sample was analyzed again. The possible tautomerization pathway of 1 is shown in Figure 3. This paper deals with the structure determination of only one isomer, compound 1. Chemical shifts for 1 were assigned as the Z configuration based on comparison to lowdenic acid.⁹

Figure 3.

Tautomerization pathway of scoriosin (1).

Scoriosin and its methyl ester displayed significant antifungal activity with IC₅₀ value of 0.37 μg/mL and MIC value of 0.63 μg/mL against Cryptococcus neoformans, respectively (Table 2). These values are readily comparable to the control drug amphotericin B, which has an IC₅₀ value of 0.35 μg/mL. It also displayed antibacterial activity using a disc-diffusion assay (SI).

Table 2.

Antifungal and Antibacterial Activity of Compounds 1–3

		C. albicans	C. glabrata	C. krusei	C. neo formans	S. aureus	MRSA	E. coli
scoriosin (1)	IC50a (μg/mL)	15.29	2.85	6.31	0.37	0.44	0.39	NA
	MICb (μg/mL)	NT	10.00	10.00	0.63	1.25	0.63
	MFC/MBCc (μg/mL)	NT	10.00	10.00	0.63	20.00	10.00
scoriosin methyl ester (2)	IC50a (μg/mL)	NA	NA	20.00	0.63	1.08	0.71	NA
	MICb (μg/mL)			NT	1.25	2.50	1.25
	MFC/MBCc (μg/mL)			NT	1.25	10.00	5.00
methyl asterrate (3)		NA	NA	NA	NA	NA	NA	NA
amphotericin B	IC50a (μg/mL)	0.14	0.21	0.74	0.35
	MICb (μg/mL)	0.63	0.63	1.25	0.63
	MFC/MBCc (μg/mL)	1.25	0.63	1.25	0.63
ciprofloxacin	IC50a (μg/mL)					0.13	0.12	0.01
	MICb (μg/mL)					0.50	0.50	0.03
	MFC/MBCc (μg/mL)					0.50	0.50	0.13

^aThe concentration that affords 50% inhibition of growth.

^bMinimum inhibitory concentration is the lowest test concentration that allows no detectable growth.

^cMinimum fungicidal/bactericidal concentration is the lowest test concentration that kills the organism. NT = not tested; NA = not active.

Cryptococcal infections are treated using a combination of amphotericin B (AMB) and 5-flucytosine (5-FC); both drugs are toxic, and 5-FC is not widely available in developing countries.¹¹ Thus, new compounds are needed to potentiate AMB activity so that AMB toxicity can be reduced and 5-FC can be replaced in the AMB + 5-FC combination. To determine if compound 2 potentiated AMB activity, a checkerboard assay was performed. As shown in Figure 4A, compound 2 was strongly synergistic with AMB (FIC = 0.3), improving AMB’s MIC by 16-fold. In addition, compound 2 improved AMB’s fungicidal activity by 8-fold (Figure 4B). Thus, there is a potential utility for compound 2 in combination therapy with AMB to treat cryptococcal infections.

Figure 4.

Synergistic and fungicidal effect of compound 2 on AMB activity. (A) A checkerboard assay was performed to evaluate the combined effects of compound 2 and AMB in C. neoformans. Cells were grown in the presence of 2-fold serially diluted concentrations of compound 2 and AMB, and the OD was measured on a microplate reader. OD values were normalized relative to the solvent control. Data were quantitatively displayed using Java Treeview (see color bar). (B) To evaluate fungicidal effects, 2 μL aliquots from each well were spotted on a YPD agar plate.

Scoriosin is a member of a rare class of antifungal agents, which contain a furylidene ring joined to a tetrahydrofurandione by a carbon–carbon double bond. There are only three other natural products that contain this functionality. Italicic acid, isolated from Penicillium italicum in 1989, has no reported biological activity.¹² Nodulisporacid A,¹³ isolated from a marine-derived fungus Nodulisporium sp., shows cytotoxicity and moderate antiplasmodial activity.¹⁴ Pestalotic acids were isolated from the plant endophyte Pestalotiopsis yunnanensis and showed antimicrobial activity against Staphylococcus aureus and Streptococcus pneumonia.¹⁵ The closest analogue to scoriosin and its derivatives is lowdenic acid,⁹ which was isolated from a terrestrial fungus Verticillium sp. and possesses antifungal and Gram-positive antibacterial activity.

The biological activity of scoriosin is postulated to be due to the capability of the ring system to form six-membered-ring chelates¹⁶ and the long lipophilic side chain in the molecule. This structural feature is present in lowdenic acid, which displays antifungal activity, and pestalotic acids, which showed antimicrobial activity. As mentioned above, italicic acid,¹² nodulisporacid,¹³ pestalotic acids,¹⁵ and carolic acid¹⁷ are the only other compounds containing this or the unsaturated form of this ring system as in the case of carolic acid. None of these compounds contain a long lipophilic side chain; italicic acid¹² has a four-carbon side chain, nodulisporacid has a five-carbon chain, and carolic acid¹⁷ lacks a side chain. However, no activity has been reported for these compounds.

Methyl assterate (3) was isolated from the same sooty mold extract as the scoriosin molecules. This molecule has been previously isolated in many fungal samples including endophytic- and leaf litter-derived fungi.^18,19 It has shown a wide array of weak activity including antibacterial, vascular endothelial growth factor (VEGF)-induced tube formation,²⁰ and antinematodal.¹⁸

The elucidation of scoriosin from sooty mold gives us a unique look into the ecological relationship of this symbiotic fungus. The potent antifungal properties of this compound have a large significance for the development of novel antifungals for the treatment of C. neoformans infections with far-reaching implications for the immunocompromised community.

EXPERIMENTAL SECTION

General Experimental Procedures.

IR spectra were obtained using a PerkinElmer 1600 FT-IR spectrometer. The UV spectrum was obtained using a Hewlett-Packard 8452A diode array spectrometer. ¹H, ¹³C NMR, HSQC, and HMBC were recorded on a Bruker Avance DRX 400 spectrometer at 400 and 100 MHz, respectively. ESIMS and HRESIMS were obtained on a Bruker MicroTof mass spectrometer. The experimental ECD data were collected on an OLIS DSM-20 circular dichroism spectrometer using ACN with varying concentrations as stated. The calculated ECD data were obtained using Gaussian 09 with a calculation method in ground-state geometric optimization of DFT at 6-31G**, and the ECD simulation was completed using TDDFT at 6-31G**. Optical rotation was determined using an Autopol V polarimeter (Rudolph, Flanders, NJ, USA) with a 0.5 mL sample tube of path length 25 mm.

Collection of Sample.

Samples of sooty mold were collected in East Hampton during October 2003 and identified by the Cornell Plant Disease Diagnostic Clinic, Cornell University, Ithaca, NY, as Scorias spongiosa. The sooty mold was found growing on Fagus sylvaticus, commonly known as beech trees.

Extraction and Purification.

A 456 g sample of S. spongiosa was covered in methanol for 24 h, and the resulting extract filtered to remove insoluble debris. Bioassay-guided fractionation using normal phase column chromatography (SiO₂; methylene chloride–methanol) resulted in the isolation of 25 g of a mixture of scoriosin-like compounds. A small amount of this mixture was subjected to RP HPLC (Alltech C₁₈ 5 μm; 10 mm × 250 mm; flow rate, 2 mL/min; 100% CH₃CN over 20 min), affording compounds 1, 2, and 3.

Antimicrobial Assay.

All organisms were obtained from the American Type Culture Collection (Manassas, VA, USA) and include the fungi Candida albicans ATCC 90028, C. glabrata ATCC 90030, C. krusei ATCC 6258, and Cryptococcus neoformans ATCC 90113 and the bacteria Staphylococcus aureus ATCC 29213, methicillin-resistant S. aureus ATCC 33591 (MRS), and Escherichia coli ATCC 35218. All organisms were tested using modified versions of the CLSI (formerly NCCLS) methods. For all organisms, optical density was used to monitor growth. Samples (dissolved in DMSO) were serially diluted in 20% DMSO/saline and transferred (10 μL) in duplicate to 96-well flat-bottom microplates. Inocula were prepared by correcting the OD₆₃₀ of microbe suspensions in incubation broth [RPMI 1640/0.2% dextrose/0.03% glutamine/MOPS at pH 6.0 (Cellgro) for Candida spp., Sabouraud dextrose for C. neoformans, cation-adjusted Mueller-Hinton (Difco) at pH 7.3 for Staphylococcus spp. and E. coli] to afford an assay volume of 200 μL and final target inocula of Candida spp. and C. neoformans of 1.5 × 10³ and Staphylococcus spp. and E. coli: of 5.0 × 10⁵ CFU/mL. The final sample test concentrations were 1/100th the DMSO stock concentration. Drug controls [ciprofloxacin (ICN Biomedicals, OH) for bacteria and amphotericin B (ICN Biomedicals) for fungi] were included in each assay. All organisms are read at 530 nm using the Biotek Powerwave XS plate reader (Bio-Tek Instruments, VT) prior to and after incubation: Candida spp. at 35 °C for 46–50 h, Staphylococcus spp. and E. coli at 35 °C for 16–20 h, and C. neoformans at 35 °C for 70–74 h. IC₅₀ values (concentrations that afford 50% inhibition relative to controls) were calculated using XLfit 4.2 software (IDBS, Alameda, CA, USA) using fit model 201. The MIC was defined as the lowest test concentration that allows no detectable growth. Minimum fungicidal or bactericidal concentrations were determined by removing 5 μL from each clear well, transferring to fresh media, and incubating as previously mentioned. The MFC/MBC was defined as the lowest test concentration that kills the organism.

Checkerboard Assay.

In order to determine if compound 2 altered the sensitivity of C. neoformans to amphotericin B, a checkerboard assay was performed according to CLSI guidelines. An overnight culture of C. neoformans strain H99 was diluted in RPMI 1640 medium (buffered with MOPS at pH 7.0) to obtain a final inoculum of 1.0 × 10⁴ CFU/mL. After dilution, 180 μL of the inoculum was added to a microplate containing 10 μL of compound 2 and 10 (μL of AMB at varying concentrations. AMB dilutions were placed in consecutive rows of the microplate, and compound 2 dilutions were placed in consecutive columns. Appropriate controls included solvent control, media control, a row of only compound 2 dilutions, and a column of only AMB dilutions. The microplate was read at 490 nm prior to and after incubation at 30 °C for 72 h using a BioTek Powerwave XS microplate reader. The fractional inhibitory concentration (FIC) was calculated using the following formula: FIC = [A*]/[A] + [B*]/[B], where [A*] is the MIC of compound A in the presence of compound B, [A] is the MIC of compound A alone, [B*] is the MIC of compound B in the presence of compound A, and [B] is the MIC of compound B alone. FIC ≤ 0.5 indicates synergy, FIC between >0.5 and ≤4.0 indicates indifference, and FIC > 4 indicates antagonism. To monitor fungicidal effects of the drug treatments, the microplate was carefully shaken on a Microplate Genie mixer (Scientific Industries, Bohemia, NY, USA), and a representative aliquot of 2 (μL from each well was spotted on a yeast peptone dextrose (YPD) agar plate, which was incubated at 30 °C for 48 h.

Computational Calculations.

The calculation procedures were referenced in previous studies. All conformational searches were conducted employing the Micromodel (version 9.9, Schrodinger LLC.) program with “Mixed torsional/Low Mode sampling” in the MMFF force field. The searches were performed in the gas phase with a 50 kJ/mol energy window limit and 10 000 maximum number of steps to fully explore all low-energy conformers. All minimization processes were carried out utilizing the Polak–Ribiere conjugate gradient (PRCG) method, 10 000 maximum iterations, and a 0.001 convergence threshold. The RS configuration was produced as the mirror image of the SR configuration results.

The GIAO shielding constants of all conformers within 10 kJ/mol of the global minimum were calculated utilizing the Gaussian 09 package (Gaussian Inc.) at the B3LYP/6-31G(d,p) level in the gas phase. The CS values were calculated according to the equations described in previous studies. The calculation of DP4 probability was facilitated using an applet available at http://www-jmg.ch.cam.ac.uk/tools/nmr/DP4/. The predicted CS values of diastereotopic protons were compared with the closet experimental values of those protons unless sufficient NMR information was available for the assignment of each proton.

Compound 1: white amorphous solid; [α]²⁶_D −20 (c 0.02, CHCl₃); IR (dry film) 3450, 3079, 1757, 1699, and 1589 cm⁻¹; ¹H and ¹³C NMR data, see Table 1; ESIMS m/z 405 [M − H]⁻ and 437 [M − H]⁺ MeOH; HRESIMS m/z 405.2297 (calcd for [C₂₃H₃₄O₆ − H]⁻, 405.2277); UV (95% ethanol), 215 and 325 nm.

Compound 2: yellowish-white amorphous solid; [α]²⁶_D −32 (c 0.02, CHCl₃); ¹H and ¹³C NMR data, see Table 1; ESIMS m/z 421 [M + H]⁺ and 863 [2M + Na]⁺; HRESIMS m/z 421.2580 (calcd for [C₂₄H₃₆O₆ + H]⁺, 421.2590); UV (95% ethanol), 263 and 325 nm.

Compound 3: white amorphous solid; ¹H and ¹³C NMR and MS data are in agreement with published literature values.