**1. Introduction**

The *Cladosporium* fungi, one of the largest genera of dematiaceous hyphomycetes, have attracted considerable attention of natural products researchers in recent years [1,2]. Versatile bioactive metabolites, such as cladosporin [3], macrolide [4], sulfur-containing diketopiperazines [5], indole alkaloids [6], hybrid polyketides [7], and diterpenes with 5-8-5 ring system [8], have been isolated from the *Cladosporium* strains. As part of our research on discovering structurally novel and biologically active natural products, a series of interesting metabolites have been obtained from marine-derived fungal strains [9,10], including those from *Cladosporium* species [11]. Our current chemical investigation on *C. cladosporioides* MA-299, an endophytic fungus obtained from the fresh inner leaves of the marine mangrove plant *Bruguiera gymnorrhiza*, led to the discovery of five new polyketides, namely, 5*R*-hydroxyrecifeiolide (**1**), 5*S*-hydroxyrecifeiolide (**2**), *ent*-cladospolide F (**3**) [12], cladospolide G (**4**), and cladospolide H (**5**) (Figure 1), as well as two known analogues, including *iso*-cladospolide B (**6**) [13,14], and pandangolide 1 (**7**) [13,15] (Figure 1). Herein, we report the isolation, structure assignment, and biological evaluation of the isolated compounds.

**Figure 1.** Structures of the isolated compounds **1**–**7**.

#### **2. Results and Discussion**

#### *2.1. Structure Elucidation of the New Compounds*

5*R*-Hydroxyrecifeiolide (**1**) was isolated as a colorless oil and the molecular formula C12H20O3 was deduced from the (+)-HRESIMS data, indicating three degrees of unsaturation. The 1H and 13C NMR spectra of **1** showed the signals for one ester/lactone carbonyl, two olefinic and two oxygenated sp3 methines, six sp<sup>3</sup> methylenes, and one methyl group (Table 1). In addition, the 1H NMR data of **1** were quite similar to those of recifeiolide (11-hydroxy-*trans*-8-dodecenoic acid lactone) [16,17], except that one methylene (δ<sup>H</sup> 1.5–2.3 ppm) in recifeiolide was replaced by an oxygenated methine (δ<sup>H</sup> 3.51 ppm) in **1**. The key COSY correlations elucidated the connectivity from H-2 through H-12 (Figure 2). Key HMBC correlations from H-2 to C-1 and C-4, from H-3 to C-1 and C-5, and from H-11 to C-1, connected C-1 and C-2 and determined the 12-membered macrolide skeleton of **1** (Figure 2). The relative configuration at C-5 and C-11 for **1** was established by the NOESY experiment (Figure S8). The NOESY correlations (Figure 3) from H-2β to H-5 and H-11 revealed a β orientation of these protons [18]. The coupling constants between H-8 and H-9 (*J*H-8/H-9 = 15.3 Hz) suggested the *E*-configuration of the C-8/C-9 double bond. The absolute configuration of C-5 of **1** was assigned by application of the modified Mosher's method [19]. The Δδ values obtained for the (*S*)- and (*R*)-MTPA esters (**1a** and **1b**, respectively) of **1** (Figure 4) suggested that the absolute configuration of C-5 is *R*. Furthermore, the electronic circular dichroism (ECD) spectrum of **1** was recorded and then computed with the time-dependent density function theory (TD-DFT) method at the gas-phase B3LYP/6-31G (d) level [20,21]. The calculated ECD spectra were produced by SpecDis software [22]. The experimental ECD spectrum for **1** matched well with the calculated spectrum for 11*R* (Figure 5). Therefore, the 5*R*, 11*R* configuration of **1** was established, and the trivial name 5*R*-hydroxyrecifeiolide was assigned.



 DMSO-*d*6. DMSO-*d*6. CDCl3. DMSO-*d*6. DMSO-*d*6. Measured at 125 MHz in CDCl3.

> *a*

*f*

**Figure 2.** Key COSY (bold lines) and HMBC (red arrows) correlations for **1**–**6**.

**Figure 3.** Key NOESY correlations for **1** and **5**.

**Figure 4.** Δδ values (Δδ (in ppm) = δ*<sup>S</sup>* − δ*R*) obtained for the (*S*)-and (*R*)-MTPA esters (**1a** and **1b**, respectively) of **1**.

**Figure 5.** Comparison of experimental and calculated ECD spectra of **1** and **2**.

The molecular formula of **2** was determined as C12H20O3, which was the same as that of **1,** according to its (+)-HRESIMS data. The 1H and 13C NMR spectra (Table 1) of **2** were similar to those of **1**, except for the different 13C chemical shifts at C-5 (δ<sup>C</sup> 70.2 in **1**, and δ<sup>C</sup> 65.4 in **2**) and its adjacent positions (C-2–C-4 and C-6–C-8), which indicated that compound **2** was the 5-epimer of **1**. The chemical shifts at C-2 and C-8 (γ-position of C-5) exhibited obvious difference in **2** and **1** probably due to the space effect. As expected, the experimental ECD spectrum of **2** matched well with the calculated spectrum of 11*R* (Figure 5). The trivial name 5*S*-hydroxyrecifeiolide was assigned to **2**.

Compound **3** was initially obtained as pale yellow powder and possessed a molecular formula C12H22O4 by (+)-HRESIMS, implying two degrees of unsaturation. The 1H and 13C NMR data (Table 2) exhibited signals attributed to one ester carbonyl, three oxygenated sp<sup>3</sup> methines, seven sp<sup>3</sup> methylenes, and one methyl group. These data were very similar to those of cladospolide F [12], suggesting that they had the same planar structure, which was also confirmed by the COSY and HMBC correlations (Figure 2). However, the signs of the optical rotations of **3** (−29.41, MeOH) and cladospolide F (+15.7, MeOH) were opposite, indicating that the absolute configurations of their stereogenic carbons were different. The relative configuration at C-3, C-4, and C-11 could not be concluded by NOESY experiment. Nevertheless, suitable crystals were obtained for X-ray diffraction analysis using Cu Kα radiation which confirmed the absolute configuration of C-3, C-4 and C-11 as 3*R*, 4*S*, and 11*R* (Figure 6). The *ent*-cladospolide F was therefore assigned as a trivial name for **3**.

**Figure 6.** Ortep diagrams of *ent*-cladospolide F (**3**) and *iso*-cladospolide B (**6**).

Compound **4** was obtained as a pale yellow oil and its molecular formula was determined as C14H24O5 on the basis of (+)-HRESIMS, requiring three degrees of unsaturation. The 1H and 13C NMR data for **4** (Table 2) were quite similar to those of **3,** except for the presence of additional ester carbonyl (C-13) and methyl (C-14) groups, which indicated the replacement of 11-OH group in **3** by an OAc group in **4**, and thus caused the down-field shift of 11-H from δ<sup>H</sup> 3.55 in **3** to δ<sup>H</sup> 4.78 in **4**. Detailed interpretation of the COSY and HMBC spectra revealed that **4** was an analogue of **3**, with the hydroxyl group at C-11 in **3** being replaced by an acetoxyl group in **4**. The HMBC correlation from H-11 to C-13 established the presence of an acetoxyl group at C-11, and the planar structure of **4** was hence confirmed as shown (Figure 2). In a biogenetic perspective, it was tentatively assigned the same relative configuration as that of **3**. The similar optical rotations of **4** (−24.56, MeOH) and **3** (−29.41, MeOH) also supported that the absolute configurations of the stereogenic carbons in **4** were the same as those in **3**. Therefore, the absolute configurations of the stereogenic carbons in **4** were tentatively assigned as 3*R*, 4*S*, and 11*R*, and the trivial name cladospolide G was assigned. Acetylation of compounds **3** and **4** using acetyl chloride yielded the same diacetylated derivative, which further correlated the structure relationship of compounds **3** and **4**.



*a* *f*

Compound **6** was isolated as colorless crystals and gave ion peaks at *m*/*z* 229.1432 [M + H]<sup>+</sup> and 246.1699 [M + NH4] <sup>+</sup> in the (+)-HRESIMS, corresponding to a molecular formula C12H20O4, indicating three degrees of unsaturation. All the 1H and 13C NMR data of **6** were quite similar to those of the previously reported polyketide metabolite *iso*-cladospolide B [13]. The COSY and HMBC correlations (Figure 2) confirmed that the planar structure of **6** was the same as that of *iso*-cladospolide B. The high similarity of specific rotations of **6** ([α] 25 <sup>D</sup> = −90.91 (*c* 0.11, MeOH) ) and *iso*-cladospolide B ([α] 25 <sup>D</sup> = −90 (*c* 0.23, MeOH)) [13] suggested that they may have the same relative and absolute stereochemistry. However, neither the relative nor the absolute configuration was determined [13]. In 2001, Franck et al. carried out the first synthesis of *iso*-cladospolide B and proposed that it has the 4*S*, 5*S*, and 11*R* configuration [14]. Later, in 2005, the absolute configuration of *iso*-cladospolide B, isolated from *Cladosporium sp*. isolated from the Red Sea sponge *Niphates rowi*, was assigned to be 4*S*, 5*S*, and 11*S* ([α] 28 <sup>D</sup> = −61 (*c* 16.6, MeOH)) [15]. It was later stated that both diastereomers appear to be natural products and (4*S*, 5*S*, 11*S*)-isomer referred to as 11-*epi*-*iso*-cladospolide B [23,24]. The relative configuration of **6** could not be assigned by NOESY experiments but the coupling constant for C-4 (*J* = 1.5 Hz) confirmed the *threo* relative configuration [14]. Upon slow evaporation of the solvent (MeOH-H2O), compound **6** was crystallized and the X-ray analysis was carried out, which was first reported for *iso*-cladospolide B (Figure 4). The Cu Kα Flack parameter 0.5 (7) allowed preliminary confirmation of the relative configurations of **6** as 4*S\**, 5*S\**, 11*R\**.

Compound **5** was obtained as a pale yellow oil and possessed a molecular formula of C12H18O3 by (+)-HRESIMS, implying four degrees of unsaturation. The 1D NMR data (Table 2) and HSQC spectrum (Figure S38) suggested signals attributed to one ester and one olefinic quaternary carbons, one oxygenated and three olefinic methines, five sp3 methylenes, and one methyl group. These NMR data were similar to those of *iso*-cladospolide B (**6**) [13]. However, resonances for two oxygenated methines (C-4 and C-5) in **6** were not detected in the NMR spectra of **5**. Instead, two additional olefinic signals including one quaternary sp<sup>2</sup> (C-4, δ<sup>C</sup> 149.4) and one methine sp<sup>2</sup> (C-5, δ<sup>C</sup> 117.1/δ<sup>H</sup> 5.53) carbons were observed in the NMR spectra of **5** (Table 2). These data indicated that **5** was a reduced analogue of **6**, and this deduction was supported by the molecular formula. The COSY and HMBC spectra established the structure of **5** as shown in Figure 1. In the NOESY experiment, the correlation between H-3 and H-5 indicated the *Z*-conformation of the double bond between C-4 and C-5 (Figure 3). The absolute stereochemistry of **5** could not be determined by Mosher's method because of the limited amount of material available. From a biogenetic point of view, **5** was putatively produced by reduction of **6**. Therefore, it was tentatively assigned the absolute configuration of C-11 of **5** as 11*R*. From these data, the name cladospolide H was assigned for **5**.

Compound **7** was acquired as white powder and showed ion peaks at *m*/*z* 267.1197 [M + Na]<sup>+</sup> in the positive HRESIMS, corresponding to a molecular formula of C12H20O5. A literature search indicated that all the 1H and 13C NMR data of **7** were almost the same as those of previously reported compound pandangolide 1 [13,15]. The almost exactly the same specific rotations of **7** ([α] 25 <sup>D</sup> = −30.16 (c 1.22, MeOH)) and pandangolide 1 ([α] 25 <sup>D</sup> = −30 (c 2.3, MeOH)) [15] revealed that they may have the same relative and absolute configurations.

## *2.2. Biological Activities of the Isolated Compounds*

Compounds **1**–**7** were tested for antimicrobial activities against two human pathogens (*Escherichia coli*, *Staphylococcus aureus*), ten aquatic bacteria (*Aeromonas hydrophilia*, *Edwardsiella ictarda*, *E. tarda*, *Micrococcus luteus*, *Pseudomonas aeruginosa*, *Vibrio alginolyticus*, *V. anguillarum*, *V. harveyi*, *V. parahaemolyticus*, and *V. vulnificus*), and 15 plant pathogenic fungi (*Alternaria solani*, *Bipolaris sorokiniana*, *Ceratobasidium cornigerum*, *Colletotrichum glecosporioides*, *Coniothyrium diplodiella*, *Fusarium graminearum*, *F. oxysporum* f. sp. *cucumerinum*, *F. oxysporum* f. sp. *momodicae*, *F. oxysporum* f. sp. *radicis lycopersici*, *F. solani*, *Glomerella cingulate*, *Helminthosporium maydis*, *Penicillium digitatum*, *Physalospora piricola Nose*, and *Valsa mali*). As shown in Table 3, **3** exhibited moderate inhibitory activities against human pathogenic bacteria S. aureus with MIC value of 8.0 μg/mL. Compound **4** showed potent inhibitory activities

against plant-pathogenic fungi (*G. cingulate* and *F. oxysporum* f. sp. *cucumerinum*), each with an MIC value of 1.0 μg/mL, while **7** showed activity against aquatic bacterium (E. ictarda) and plant-pathogenic fungus (*G. cingulate*), with MIC values of 4.0 and 1.0 μg/mL, respectively.


**Table 3.** Antimicrobial Activities of **1**–**7** (MIC, μg/mL) a.

*<sup>a</sup>* (–) = MIC > 64 μg/mL, *<sup>b</sup>* Chloramphenicol as positive control, *<sup>c</sup>* Amphotericin B as positive control.

Compounds **1**–**7** were also evaluated for acetylcholinesterase inhibitory activity. Compound **3** exhibited potent activity against acetylcholinesterase with the IC50 value of 40.26 μM. The other compounds have a weak activity (IC50 > 50 μM).
