**1. Introduction**

Andrastins are meroterpenoids characterized by a 6,6,6,5-tetra-carbocyclic skeleton. They are biogenetically derived from 3,5-dimethylorsellinic acid (DMOA) and farnesyl diphosphate (FPP), synthesized via a mixed polyketide-terpenoid pathway, and usually possess a keto-enol tautomerism at the cyclopentane ring [1–4]. To date, over 40 andrastins have been reported with multiple potential biological activities, including cytotoxic [5], anti-inflammatory [6], antiproliferative [7] and antimicrobial activity [4]. The complex structures and potential biological activities of andrastins have attracted much attention in recent years [8–10].

Marine fungus is known to be a natural source of structurally diverse and biologically active metabolites for drug discovery [11–16]. Recently, a series of novel bioactive natural products from marine fungi were reported by our group [17–22]. In our ongoing search for new bioactive secondary metabolites from marine fungi, the fungus *Penicillium* sp. N-5, isolated from the rhizosphere soil of mangrove plant *Avicennia marina*, led to the isolation of four new compounds, hemiacetalmeroterpenoids A–C (**1**–**3**) and astellolide Q (**15**). Especially, hemiacetalmeroterpenoid A (**1**) was a new andrastin-type meroterpenoid containing a unique 6,6,6,6,5,5-hexa-cyclic skeleton. Meanwhile, eleven known compounds, including 3-deacetyl-citreohybridonol (**4**) [23] citreohybridone A (**5**) [24], 3,5-dimethylorsellinic acid-based meroterpenoid 2 (**6**) [5], andrastins A–C (**7**, **10**, **13**) [25], andrastone C (**8**) [26], penimeroterpenoid A (**9**) [4], 23-deoxocitreohybridonol (**11**) [1], 6α-hydroxyandrastin B (**12**) [1], and compound V (**14**) [27] were also obtained from the fungus N-5 (Figure 1). All the isolated compounds were investigated for their antimicrobial activity against two phytopathogenic fungi and four bacterial strains. Herein, we report the isolation, structural characterization and antibacterial activity of these compounds.

**Citation:** Chen, T.; Yang, W.; Li, T.; Yin, Y.; Liu, Y.; Wang, B.; She, Z. Hemiacetalmeroterpenoids A–C and Astellolide Q with Antimicrobial Activity from the Marine-Derived Fungus *Penicillium* sp. N-5. *Mar. Drugs* **2022**, *20*, 514. https://doi.org/ 10.3390/md20080514

Academic Editor: Dehai Li

Received: 1 August 2022 Accepted: 12 August 2022 Published: 13 August 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** Structure of compounds **1**–**15**.

#### **2. Results**

#### *2.1. Structure Identification*

Hemiacetalmeroterpenoid A (**1**) was obtained as a white powder. It molecular formula was assigned as C26H34O7 according to HRESIMS analysis at *m*/*z* 459.23709 [M+H]+ (calcd. 459.23773), indicating ten degrees of unsaturation. In the 1H NMR spectrum, the signal for one olefinic proton (*δ*<sup>H</sup> 5.63), one methoxyl (*δ*<sup>H</sup> 3.60), one methine (*δ*<sup>H</sup> 1.47), four methylenes (*δ*<sup>H</sup> 1.42, 1.62, 1.85, 1.91, 1.94, 2.11, 2.33 and 2.61) and six methyls (*δ*<sup>H</sup> 1.03, 1.07, 1.19, 1.23, 1.33 and 1.49). The 13C NMR data exhibited 26 carbon resonances, including two olefinic carbons for one double bond (*δ*<sup>C</sup> 127.2, 150.1), three carbonyl carbons for two ketone (*δ*<sup>C</sup> 203.7 and 203.8), and one ester carbonyl (*δ*<sup>C</sup> 169.3), one methine (*δ*<sup>C</sup> 49.1), five methylenes (one oxygenated), seven methyls (one oxygenated), eight quaternary carbons (one highly oxygenated: *δ*<sup>C</sup> 99.8) (Table 1). The NMR data established a nucleus of meroterpenoid characterized by an andrastin scaffold, structurally similar to the citreohybriddione C (Figures S2 and S3) [28]. Analysis of the 1H-1H COSY data led to the identification of two isolated spin-systems of C-1/C-2 and C-5/C-6/C-7. The HMBC from H2-1 to C-3, C-10, from H1-5 to C-10, and from H3-22 to C-3, C-4, C-5, C-23, ring A was formed. The HMBC correlations from H3-24 to C-7, C-8, C-9, from H1-11 to C-8, C-10 and from H2-21 to C-5, C-10 completed ring B. Then, the HMBC cross-peaks from H2-1, H2-6 to C-10 and from H3-23, H2-7 to C-5 indicated ring A and ring B were fused at C-5 and C-10. Then, HMBC correlations from H3-24 to C-14, C-15, from H3-20 to C-11, C-12, C-13, C-17, from H3-19 to C-12, C-13, C-14, C-17 and from H3-18 to C-15, C-16, C-17, ring C and ring D were constituted, and they were blended at C-13 and C-14. The HMBC correlations from H2-6, H1-11 to C-8, from H1-11 to C-10, suggested ring B and ring C were tightly connected. In addition, the HMBC correlation from H3-26 to C-25 implied the presence of a methyl carboxylate. A weak HMBC correlation from H3-26 to C-14 located the methyl carboxylate at C-14. Except one double bond, three carbonyls and four rings, ten degrees of unsaturation indicated that two new rings were required. According to the HMBC correlation from H2-21 to C-3 (*δ*<sup>C</sup> 99.8), a 6-membered ring was confirmed between C-1, C-2, C-3, C-10, and C-21. Finally, another new 5-membered ring was formed by intramolecular dehydration of hydroxyl groups at C-12 and C-16. Thus, the planar structure of **1** was established as shown in Figure 2.


**Table 1.** 1H NMR (600 MHz) and 13C NMR (150 MHz) of **1**-**3** in CD3OD.

**Figure 2.** Key HMBC and COSY correlations of **1**–**3** and **15**.

The relative configuration of compound **1** was defined by the NOESY correlations. The correlations of H2-21 with H3-23, H3-20 with H3-19, H3-24, and H3-18 with H3-19, H3-26 were observed in the NOESY spectrum, which means H3-18, H3-19, H3-20, H2-21, H3-23, H3-24 and H3-26 were on the same side. The NOESY correlations of H1-5 with H3-22 suggested that H1-5 and H3-22 were in the opposite face (Figure 3). The absolute configuration of **1** was determined by comparing the calculated ECD spectra generated by the time-dependent density functional theory (TDDFT) for two enantiomers 3*R*, 5*S*, 8*S*, 10*S*, 12*R*, 13*S*, 14*R*, 16*R*-**1a** and 3*S*, 5*R*, 8*R*, 10*R*, 12*S*, 13*R*, 14*S*, 16*S*-**1b** with the experimental one. Finally, the experimental ECD spectrum of **1** was nearly identical to the calculated ECD spectrum for **1a** (Figure 4), clearly suggesting the 3*R*, 5*S*, 8*S*, 10*S*, 12*R*, 13*S*, 14*R*, 16*R* absolute configuration for **1**.

**Figure 3.** Key NOE correlations of **1**–**3** and **14**–**15**.

**Figure 4.** ECD spectra of compounds **1** (**A**), **2** and **3** (**B**), **14** and **15** (**C**) in CH3OH.

Hemiacetalmeroterpenoid B (**2**) was isolated as a white powder and had a molecular formula of C26H36O6, determined by HRESIMS data *m*/*z* 445.25772 [M+H]+ (calcd. 445.25847) with nine degrees of unsaturation. The 1H NMR spectrum of **2** displayed the signal for one olefinic proton (*δ*<sup>H</sup> 5.42), one methoxyl (*δ*<sup>H</sup> 3.56), two methines (*δ*<sup>H</sup> 1.33 and 1.89), four methylenes (*δ*<sup>H</sup> 1.12, 1.33, 1.55, 1.75, 2.07, 2.12, 2.19 and 2.76) and six methyls (*δ*<sup>H</sup> 1.01, 1.04, 1.18, 1.19, 1.57 and 1.82). The 13C NMR data revealed 26 carbon resonances, involving four olefinic carbons for two double bonds (*δ*<sup>C</sup> 113.5, 124.2, 137.7, 190.7), two carbonyl carbons for one ketone (*δ*<sup>C</sup> 201.4), one ester carbonyl (*δ*<sup>C</sup> 172.6) (Table 1). According to 1D NMR and 2D NMR data, the planar structure of **2** was similar to the co-isolated andrastin B (**13**). The obvious difference is that the acetyl group at the C-3 position of compound **2** disappears. Meanwhile, the HMBC from H2-21 to C-3 (*δ*<sup>C</sup> 99.5) also indicated that a new 6-membered ring was formed between C-1, C-2, C-3, C-10 and C-21 (Figure 2).

The NOESY spectrum indicated that H1-5, H1-9 and H3-22 were on the same side based on the correlations of H1-5 with H1-9 and H3-22. On the contrary, it was suggested that H3-19, H3-21, H3-23, H3-24, and H3-26 were on the other side based on the NOESY correlations of H2-21 with H3-23 and H3-24, along with H3-19 with H3-24 and H3-26 (Figure 3). Thus, the relative configuration of **2** was determined to be 3*R*, 5*S*, 8*S*, 9*R*, 10*S*, 13*R* and 14*R*. The absolute configuration of the stereogenic centers in **2** was assigned as 3*R*, 5*S*, 8*S*, 9*R*, 10*S*, 13*R* and 14*R* by comparing its experimental ECD spectrum with that of the calculated model molecule (Figure 4).

Hemiacetalmeroterpenoid C (**3**) was also purified as a white powder. The molecular formula was specified as C28H38O7 (ten degrees of unsaturation) by HRESIMS (*m*/*z* 509.25015 [M+Na]+), which is 42 mass units higher than that of **2** (Figure S17). Analysis of its NMR data (Table 1) revealed the presence of the same partial structure as that found in compound **2**. The only difference was **3** has an additional acetyl fragment. Finally, a weak HMBC correlation from Ac-CH3 to C-15 suggested that the acetyl fragment was attached to C-15 (Figure 2).

Because compound **3** has the same chiral center as **2**, the NOESY correlation and experimental ECD spectrum of compound **3** were in agreement with those of **2** (Figures 3 and 4). Thus, the absolute configuration of **3** was identified as 3*R*, 5*S*, 8*S*, 9*R*, 10*S*, 13*R* and 14*R*.

Compound **14** was obtained as a yellow powder. Analysis of its 1H NMR and 13C NMR data showed that the planar structure of **14** was the same as compound V, which was the product of the alkaline hydrolysis of parasiticolide A [27]. However, the absolute configuration of compound V was ambiguous.

The relative configuration of **14** was also defined by the NOESY correlation. The correlations of H3-14 with H1-5 and H1-6, and H2-13 with H2-15 were found in the NOESY spectrum, which means H1-5, H1-6, and H3-4 were on the same side, and H2-13 and H2-15 were on the opposite face (Figure 3). Thus, the absolute configuration of the stereogenic centers in **14** was assigned as 4*R*, 5*R*, 6*S*, 10*S* by comparing its experimental ECD spectrum with that of the calculated model molecule (Figure 4). Finally, compound **14** was named as astellolide J.

Astellolide Q (**15**) was also acquired as a yellow powder. It molecular formula was determined as C17H24O6 according to HRESIMS analysis at *m*/*z* 347.14578 [M+Na]<sup>+</sup> (calcd. 347.14651), indicating six degrees of unsaturation. The 1H NMR of **15** showed two methyls (*δ*<sup>H</sup> 1.15 and 2.08), four methylenes (*δ*<sup>H</sup> 1.16, 1.45, 1.54, 1.78, 1.91, 2.05, 2.34 and 2.50), one methines (*δ*<sup>H</sup> 1.74) one hydroxymethine (*δ*<sup>H</sup> 4.55) and three hydroxy-methylenes (*δ*<sup>H</sup> 3.34, 3.92, 4.09, 4.33, 4.84 and 5.04). In addition, according to the HSQC data, the 13C NMR data showed the presence of 17 carbon signals, including two ester carbonyl carbons (*δ*<sup>C</sup> 173.1, 177.0) and two olefinic carbons (*δ*<sup>C</sup> 124.0, 169.0), one methyl, seven methylenes (three oxygenated), two methines (one oxygenated), two aliphatic quaternary carbons (Table 2). Analysis of its 1H NMR and 13C NMR data in association with the 2D NMR data established a nucleus of drimane sesquiterpenoid characterized by an astellolide scaffold, structurally similar to the co-isolated compound **14** (Figures S26 and S27). It can be clearly observed

that compound **15** has an additional acetyl fragment. Furthermore, the HMBC from H2-13 to Ac-OCO indicated that the acetyl fragment was linked to C-13 (Figure 3).


**Table 2.** 1H NMR (400 MHz) and 13C NMR (100 MHz) of **15** in CD3OD.

Finally, the NOESY correlation and experimental ECD spectrum of compound **15** were identical to those of **14** (Figures 3 and 4). Thus, the absolute configuration of **15** was also assigned as 4*R*, 5*R*, 6*S*, 10*S*.

#### *2.2. Antimicrobial Assay*

Compounds **1**–**15** were investigated for their antimicrobial activities against two phytopathogenic fungi and four bacterial strains. As shown in Table 3, andrastin-type meroterpenoids have better antimicrobial activities against phytopathogenic fungus than against bacteria. Most of all the tested compounds (9 compounds out of total 15 compounds) displayed potent antimicrobial activities (MIC < 50 μg/mL). Among them, compounds 1, 5 and 10 exhibited remarkable antimicrobial activities against *Penicillium italicum* and *Colletrichum gloeosporioides* with MIC values of 6.25, 1.56, 6.25 and 6.25, 3.13, 6.25 μg/mL. Moreover, compound 1 showed inhibitory activities against *Bacillus subtilis* under concentration of 6.25 μg/mL. Compound 10 also displayed significant antimicrobial activity against *Salmonella typhimurium* with an MIC value of 3.13 μg/mL. Notably, compound 5 revealed potential antimicrobial activity against all the strains, the MIC values were lower than 25 μg/mL.

**Table 3.** Antimicrobial activity of compounds **1**–**15**.



a: The deviation value of three parallel experiments; -: No test.

As for the study of the structure–activity relationship (SAR), it was found that the degree of oxidation at C-21 had different effects on the activities of the compounds. The compound with methyl (**10**) at C-21 has significantly antimicrobial activity, followed by the aldehyde group (**7**), and hydroxymethyl (**13**) was the weakest. In-depth analysis showed that apart from the degree of oxidation at C-21, keto-enol tautomerism at the cyclopentane ring also had obvious influences on the antimicrobial activities of compounds. Compared to compounds **7** and **13** (enol form), compounds **8** and **9** (keto form) showed no activities against all strains (Table 3).
