New Hydroxydecanoic Acid Derivatives Produced by an Endophytic Yeast Aureobasidium pullulans AJF1 from Flowers of Aconitum carmichaeli

Endophytes have been recognized as a source for structurally novel and biologically active secondary metabolites. Among the host plants for endophytes, some medicinal plants that produce pharmaceuticals have been reported to carry endophytes, which could also produce bioactive secondary metabolites. In this study, the medicinal plant Aconitum carmichaeli was selected as a potential source for endophytes. An endophytic microorganism, Aureobasidium pullulans AJF1, harbored in the flower of Aconitum carmichaeli, was cultured on a large scale and extracted with an organic solvent. Extensive chemical investigation of the extracts resulted in isolation of three lipid type compounds (1–3), which were identified to be (3R,5R)-3,5-dihydroxydecanoic acid (1), (3R,5R)-3-(((3R,5R)-3,5-dihydroxydecanoyl)oxy)-5-hydroxydecanoic acid (2), and (3R,5R)-3-(((3R,5R)-5-(((3R,5R)-3,5-dihydroxydecanoyl)oxy)-3-hydroxydecanoyl)oxy)-5-hydroxydecanoic acid (3) by chemical methods in combination with spectral analysis. Compounds 2 and 3 had new structures. Absolute configurations of the isolated compounds (1–3) were established using modified Mosher’s method together with analysis of NMR data for their acetonide derivatives. All the isolates (1–3) were evaluated for antibiotic activities against Escherichia coli, Staphylococcus aureus, Bacillus subtilis, Pseudomonas aeruginosa, and their cytotoxicities against MCF-7 cancer cells. Unfortunately, they showed low antibiotic activities and cytotoxic activities.


Introduction
Endophytes spend all or part of their lives colonizing inside of healthy tissues of a host plant. They cause no apparent disease symptoms in host plant, and rather give abilities including tolerance to biotic and abiotic stress, nutrient acquisition, and plant growth promotion to the host plant [1]. Various types of secondary metabolites with significant bioactivities have been produced by endophytes as a result of interaction with host plant [2]. Biosynthesis of many metabolites has been considered to be caused by a result of environmental signals such as insects and plant pathogen [3]. In a search for endophytes to produce secondary metabolites with intriguing structures and potent bioactivities, an endophytic yeast Aureobasidium pullulans AJF1 was isolated from flowers of Aconitum carmichaeli. Some studies have revealed that A. pullulans produce various types of metabolites such as polysaccharides, mannitol oils, poly malic acid, 2-propylacrylic acid, 8,9-dihydroxy-2-methyl-4H,5H-pyrano [3,2-c]chromon-4-one, 2-methylenesuccinic acid, hexane-1,2,3,5,6-hexol, etc. [4][5][6][7]. Roots of Aconitum carmichaeli Debeaux have been treated as one of the important crude drugs in oriental medicine due to its potent activities to treat shock caused by acute and chronic heart failure and low blood pressure as well as toxicities [8]. Even though a broad spectrum of research on the secondary metabolites produced by endophytes from medicinal plants has been carried out since discovery of paclitaxel production by the endophytic fungus, Taxomyces andreanae [9,10], only a few studies have been conducted on the endophytes isolated from A. carmichaeli [11][12][13]. Herein, three metabolites (1)(2)(3) were isolated from cultures of the endophytic yeast A. pullulans AJF1 and investigated for their activities (Figure 1). They were all hydroxy fatty acid derivatives and compound 1, 3,5-dihydroxydecanoic acid, was previously reported to be one of main components of extracellular lipids produced by Aureobasidium sp. [14]. However, its absolute as well as relative stereochemistry was elucidated for the first time in this study.
Compound 1 was isolated as colorless oil. The molecular formula was determined to be C 10 4.33 (quint, J = 3.7 Hz), a terminal methyl at δ H 0.85 (t, J = 6.9 Hz), together with several methylene signals. Its 13 C-NMR spectrum exhibited the presence of ten carbons including a carboxylic carbon at δ C = 171.2, two oxymethine carbons at δ C 76.1 and 62.5, six methylene carbons at δ C 38.5, 35.7, 35.4, 31.5, 24.5, and 22.5, and a terminal methyl carbon at δ C 13.9, suggesting that compound 1 was a dihydroxy decanoic acid. The positions of two hydroxyl groups were confirmed by 2D-NMR data including 1 H-1 H COSY, HSQC, and HMBC. A C-CH 2 -CH(O)-CH 2 -CH(O) spin system was obtained by analysis of its 1 H-1 H COSY spectrum ( Figure 2). HMBC correlations of the methylene protons at δ H 2.67 and 2.58 with the carboxylic carbon (δ C 171.2), oxymethine carbon (δ C 62.5), and a methylene carbon (δ C 35.7) indicated that the spin system was directly linked to carboxylic acid, that is, 3,5-dihydroxy decanoic acid ( Figure 1). This structure was further confirmed by HMBC correlations of the oxymethine proton at δ H 4.33 with the carboxylic carbon, two methylene carbons (δ C 38.5 and 35.7), and the oxymethine carbon (δ C 76.1).  13 C-NMR spectra of compound 2 were similar to those for compound 1, differing in that compound 2 has ten more carbons including two more oxymethine groups and one more carboxylic acid than 1 (Table 1). Two units of C-CH 2 -CH(O)-CH 2 -CH(O) spin system were observed in its 1 H-1 H COSY spectrum ( Figure 1). Thus, it was presumed that two units of dihydroxy fatty acid might be esterified to form compound 2, which was supported by the downfield shifted oxymethine proton signal (δ H 5.27). The connection of two units was verified by HMBC correlations. HMBC correlations of the oxymethine proton at δ H 5.27 with a carboxylic acid carbon (δ C 169.2), a carboxylic ester carbon (δ C 171.1), and two methylene carbons (δ C 35.3 and 32.7) supported that one dihydroxy fatty acid moiety was esterified to the other dihydroxy fatty acid moiety at the C-3 position (Figure 1). The possibility that two dihydroxy fatty acids might have different carbon numbers was not excluded. However, the presence of two decanoic acid moieties was confirmed by ten carbon atoms observed in the 13 C-NMR spectrum for hydrolysate of compound 2. Thus the planar structure of compound 2 was elucidated to be 3-((3,5-dihydroxydecanoyl)oxy)-5-hydroxydecanoic acid. Compound 3 was also isolated as colorless oil. Its molecular formula was established to be C 30 H 56 O 10 from the positive HRFABMS at m/z 581.3657 [M -H 2 O + Na] + (Calcd for C 30 H 54 O 9 Na, 581.3660). The 1 H and 13 C-NMR of compound 3 were also similar to those for compound 2 ( Table 1). As in the cases of compounds 1 and 2, ten more carbons including two additional oxymethines (δ C 69.8 and 72.4), one more carboxylic ester carbon (δ C 172.3), six methylene carbons (δ C 42.7, 42.4, 37.9, 25.0, 31.5, 22.5), and a terminal methyl group (δ C 14.0) appeared compared with compound 2, suggesting that three dihydroxy fatty acid moieties were linked together through ester linkage. The positions of the esterification were confirmed by HMBC correlations (Figure 2). HMBC correlations of the oxymethine H-3 at δ H 5.29 with a carboxylic acid carbon C-1 at δ C 171.5 and a carboxylic ester carbon C-1 at δ C 169.0) indicated that one dihydroxy fatty acid was esterified at the C-3 position of another dihydroxy fatty acid moiety. In addition, HMBC correlations of the oxymethine H-5 at δ H 5.04 with the last carboxylic ester carbon at δ C 172.3 and the oxymethine carcon C-3 at δ C 66.3 indicated that the last dihydroxy fatty acid moiety was esterified to C-5 position of one of the dihydroxy fatty acid moiety. As shown in compound 2, the 13 C-NMR spectrum of hydrolysate of compound 3 indicated the presence of three dihydroxy decanoic acids, suggesting the planar structure of compound 3. Relative configurations of compounds 1-3 were determined on the basis of semi-synthetic structure modification and their spectroscopic data analyses. The relative configuration of the acyclic 1,3-diols in compounds 1-3 were assigned by using the [ 13 C]acetonide method [15]. Compound 2 was hydrolyzed with 1 N HCl in MeOH to yield the 3,5-dihydroxydecanoic acid methyl ester (2a) and 3,5-dihydroxydecanoic acid (2b) as depicted in Figure 3. Analogous treatment of compound 3 also afforded the 3,5-dihydroxydecanoic acid methyl ester (3a) and 3,5-dihydroxydecanoic acid (3b). Treatments of 1, 2a, 2b, 3a, and 3b with 2,2-dimethoxypropane and pyridinium-p-toluensulfonate in methanol afforded the acetone ketal (1a, 2c, and 3c) at their 1,3-diol positions ( Figure 3). Two methyl resonances observed at δ C 30.1 and 19.7 for the acetone ketals suggested the relative configurations of the 1,3-diol to be in syn configurations. Modified Mosher's method was employed to determine the absolute configurations of hydroxyl groups in these compounds. Treatment of 1 with (R) and (S)-MTPA-Cl and a catalytic amount of DMAP (4-dimethylaminopyridine) in pyridine-d 5 afforded its (S) and (R)-MTPA ester, respectively. Low field shifted proton resonance at δ H 4.33 for H-3 implied that the (R) and (S)-MTPA-Cl reacted only with the hydroxyl group at the C-3 position not that at C-5. Analysis of 1 H-NMR chemical shift differences (∆δ S-R ) between (R) and (S) Mosher's ester of 1 revealed the absolute configuration of C-3 and C-5 to be R and R, respectively, as shown in Figure 2. Consequently, the structure of 1 was determined to be (3R,5R)-3,5-dihydroxydecanoic acid. Likewise, compound 2 was derivatized to its (R) and (S)-MTPA ester using (S) and (R)-MTPA-Cl, respectively, in which MTPA groups were confirmed to be attached to all the hydroxyl groups of compound 2 by chemical shifts of oxymethine protons. Analysis of 1 H-NMR chemical shift differences (∆δ S-R ) between (R) and (S) tri-Mosher's ester of 2 revealed that C-3, 5, 3 , and 5 were all in R configurations as shown. Thus, compound 2 was identified to be (3R,5R)-3-(((3R,5R)-3,5-dih ydroxydecanoyl)oxy)-5-hydroxydecanoic acid. Determination of the absolute configurations of 3 using Mosher's method was quite challenging since the proton resonances for methylene groups neighboring oxymethines were too difficult to be distinguished from others since many proton resonances were overlapped [16]. Nevertheless, Mosher's esterification of the hydrolysates of compound 3 allowed the determination of the absolute stereochemistry, since compound 3 was composed of three units of 3,5-dihydroxydecanoic acid moiety. To obtain each monomer of compound 3, the acetonide derivative 3C was treated with pyridinium-p-toluensulfonate at 40 • C in MeOH to afford 3d. Analysis of 1 H-NMR chemical shift differences (∆δ S-R ) of (S) and (R) esters of 3d elucidated the absolute configuration of C-3 and 5 to be R and R configurations, respectively ( Figure 4). Finally, chemical structure of 3 was determined as (3R,5R)-3-(((3R,5R)-5-(((3R,5R)-3,5-dihydroxydecanoyl)oxy)-3-hydroxydecanoyl)oxy)-5 -hydroxydecanoic acid.  Compound 1, (3R,5R)-3,5-dihydroxydecanoic acid was previously reported as a not only synthetic intermediate but also natural product [14,17]. 3,5-Dihydroxydecanoic acid was reported to be a main component of the lipophilic moieties in lipid produced by Auresobasidium sp. [14]. However, its relative and absolute stereochemistry was established in this study for the first time. The (3R,5R)-3,5-dihydroxydecanoic acid moiety is extremely rare in nature. As one example of natural compounds to bear this moiety, exophilin A was isolated from a marine microorganism Exophiala pisciphila [18]. Exophilin A is also a trimer of the (3R,5R)-3,5-dihydroxydecanoic acid with the ester linkages at 5-OH and 5 -OH while the esters in compound 3 at 3 and 5 . The relative and absolute stereochemistry of exophilin A were elucidated by comparison of the optical rotation value for its hydrolysate with that for (3R,5R)-3-hydroxy-5-decanolide [18]. In this study, the acetonide method, Mosher's method, and acid hydrolysis together with NMR analysis were employed to elucidated the stereochemistry of the isolated compounds. All the isolates (1-3) were evaluated for antibiotic activities against Escherichia coli, Staphylococcus aureus, Bacillus subtilis, Pseudomonas aeruginosa, and their cytotoxicities against MCF-7 cancer cells. Unfortunately, they showed low antibiotic activeities and cytotoxic activities. While exophilin A was reported to be active against Staphylococcus aureus with an MIC of 50 µg/mL, compound 3 did not show antibacterial activities even at 128 µg/mL.

General Experimental Procedures
The high-resolution fast atom bombardment mass spectrometry (HRFABMS) data were obtained on gas chromatography high resolution mass spectrometer (JMS-700, Jeol, Tokyo, Japan). The nuclear magnetic resonance (NMR) spectra were acquired with a 300 Ultra shield spectrometer ( 1 H, 300 MHz; 13

Fungal Materials
AJF1 was isolated from the flowers of Aconitum carmichaeli collected from Jangbaek Mountain, Gangwon-do, South Korea, in 2016. The collected flowers were sterilized with 70% ethanol for 1 min and 7% H 2 O 2 for 1 min, and then washed with sterilized distilled water. The sterilized flowers were cut into small pieces, and then diluted with sterilized water. The diluted solution was poured onto YME agar media (Yeast Ex, 4 g/L; Malt Ex, 10 g/L; Dextrose, 4 g/L; and sterilized water 1L), which was then incubated at 28 • C for 1 month. A pure strain of incarnadine colored colony (AJF1) was obtained and the strain AJF1 was identified to be Aureobasidium pullulans based on 16S/18S rDNA sequence analysis (97.72%, similarity to Aureobasidium pullulans strain YY20).

Conclusions
An endophytic yeast, Aureobasidium pullulans AJF1, was isolated from flowers of the Aconitum carmichaeli in this study and cultivated on a large scale for chemical investigation. Extensive chemical investigation of the extracts resulted in isolation of three lipid type compounds (1-3). They were (3R,5R)-3,5-dihydroxydecanoic acid unit or the esterified complexes of the unit. Relative and absolute stereochemistries of the isolated compounds (1-3) were established using modified Mosher's method together with analysis of NMR data for their acetonide derivatives. Even though compounds 2 and 3 were all unique lipid type new compounds, they did not show potent antibiotic activities against Escherichia coli, Staphylococcus aureus, Bacillus subtilis, Pseudomonas aeruginosa, and no cytotoxicities against MCF-7 cancer cells. Regardless of their low potencies of activities, it is a good finding that endophytes from medicinal plants could be good sources for new chemistries.