**1. Introduction**

Marine fungi have formed different metabolic pathways and adaptation mechanisms within the peculiar marine environment. Hence, marine fungi can produce natural secondary metabolites that are characterized by unique chemical structures and high biological activities [1,2]. Alkaloids derived from marine-derived compounds have received extensive attention in recent years. Indole alkaloids, as an important class of secondary metabolites produced by marine-derived fungi [3], showed excellent biological activities, including cytotoxic [4], antibacterial [5], quorum sensing inhibitory [6], anti-Zika virus [7], and protein tyrosine phosphatase inhibitory activities [8,9].

Marine-derived fungus *Penicillium* sp. KFD28 was isolated and identified from bivalve shellfish, *Meretrix lusoria*, collected from Haikou Bay, China. Our previous study on the secondary metabolites of this fungus discovered a series of indole alkaloids with novel structures and intriguing bioactivities, e.g., protein tyrosine phosphatase inhibitory activity [8,9]. The OSMAC (one strain, many compounds) approach is highly efficient for inducing structural diversity by the variation of cultivation conditions [10]. Most of the indole alkaloids precursors are related to L-tryptophan [11]. To find more new alkaloids from marine fungi, adding amino acids to the culture media is becoming a viable

**Citation:** Dai, L.-T.; Yang, L.; Kong, F.-D.; Ma, Q.-Y.; Xie, Q.-Y.; Dai, H.-F.; Yu, Z.-F.; Zhao, Y.-X. Cytotoxic Indole-Diterpenoids from the Marine-Derived Fungus *Penicillium* sp. KFD28. *Mar. Drugs* **2021**, *19*, 613. https://doi.org/10.3390/md19110613

Academic Editor: Khaled A. Shaaban

Received: 28 September 2021 Accepted: 24 October 2021 Published: 28 October 2021

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

**Copyright:** © 2021 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/).

strategy [12,13]. In order to explore the metabolic potential of the fungus *Penicillium* sp. KFD28, we recultured this fungus by adding L-tryptophan to the solid rice culture and found that the HPLC profiles of the extract are different from that of the first time obtained from the liquid medium. Later chemical investigation of the fermentation broth led to the isolation of four new indole-diterpenoids, penpenes K-N (**1**–**4**), along with 12 known ones, including paxilline (**5**) [14], dehydroxypaxilline (**6**) [15], 7-hydroxyl-13-dehydroxypaxilline (**7**) [16], 3-deoxo-4b-deoxypaxilline (**8**) [17], epipaxilline (**9**) [18], 7-hydroxypaxilline-13-ene (**10**) [19], paspaline (**11**) [20], 4a-demethylpaspaline-4a-carboxylic acid (**12**) [17], paspalinine (**13**) [21], PC-M6 (**14**) [22], emindole SB (**15**) [23], emeniveol (**16**) [24] (Figure 1). Herein, the isolation, structure elucidation, and bioactivities of these compounds are reported.

**Figure 1.** The chemical structures of compounds **1**–**16**.

#### **2. Results and Discussions**

## *2.1. Structure Elucidation*

Compound **1** was obtained as a yellow oil. Its formula was determined as C28H39NO3 on the basis of HRESIMS data (*m*/*z* 476.2565 for [M+K]+), indicating ten degrees of unsaturation. The 1H and 13C NMR data of **1**, with the aid of its HSQC spectrum (Supplementary Materials, Figures S1–S4), showed a total of 28 carbon signals comprising eight olefinic

or aromatic carbons (four protonated) for a 2,3-disubstituted indole moiety, four methyls, eight sp<sup>3</sup> methylenes with one oxygenated, four sp<sup>3</sup> methines with two oxygenated, and four non-protonated sp<sup>3</sup> carbons with one oxygenated. These NMR data allowed the construction of the carbon skeleton of indole-diterpenoid. Careful contrast of the similar NMR spectral data between **1** and paspaline (**11**) [20] revealed that they had the same planar structure except that the methyl group at C-12 (*δ*C 36.6) in paspaline was oxidized to oxymethylene in **1**. Consistent with the introduction of the hydroxyl at C-30 (*δ*C 58.5), it showed a distinctive deshielding of the C-12 signal (*δ*C 48.8) in **1** compared to those of paspaline (*δ*C 36.6). The linkage of a hydroxyl at C-30 was further supported by the COSY correlation of 30-OH/H2-30 (Figure 2; Supplementary Material, Figure S8) and ROESY correlation of H3-26/H2-30 (Supplementary Material, Figure S9). Detailed analysis of 2D NMR data further confirmed that **1** and paspaline share the same indole-diterpenoid skeleton. The relative configuration of 6/6/6 tricyclic rings in indole-diterpenoid were determined by the ROESY spectrum (Figure 3), in which the sequential correlations of H-16/H3-26/H2- 30/H-10*β*/H3-28 suggested the same face of H-16, CH3-26, CH2-30, and C-27 in the 6/6/6 tricyclic ring system, while the correlations of H3-25/H-13/H-7/H-9 suggested CH3-25, H-13, H-7, and H-9 were on the opposite face of this system. It has been reported that the strong Cotton effect (CE) around 220 nm was related to the absolute configurations of the chiral carbons around the indole chromophore in the paxilline-type indole-diterpene [17]. Thus, the strong negative CE at 223 nm in the experimental ECD spectrum of **1** (Figure 4) suggested its (3*S*,4*S*,7*S*,9*S*,12*S*,13 *R*,16*S*)-**1** absolute configuration [17]. As for the absolute configuration of **1**, the ECD spectrum of (3*S*,4*S*,7*S*,9*S*,12*S*,13 *R*,16*S*)-**1** was calculated; this deduction was further supported, establishing the absolute configuration of **1** as presented in Figure 1.

Compound **2** was isolated as a yellow oil. Its formula was determined as C27 H35NO4 on the basis of HRESIMS data, indicating 11 degrees of unsaturation. Its 13C NMR data showed a total of 27 carbon signals comprising eight aromatic or olefinic carbons for an indole moiety, four methyls, six sp<sup>3</sup> methylenes, four sp<sup>3</sup> methines with two oxygenated, and five non-protonated sp<sup>3</sup> carbons with three oxygenated. Analysis of the NMR spectra (Tables 1 and 2, Supplementary Materials, Figures S12–S18) of **2** suggested that its structure was related to that of 4a-demethylpaspaline-3,4,4a-triol [17], and the main difference being C-9 (*δ*C 109.8), C-11 (*δ*C 43.7), and C-12 (*δ*C 84.6) in **2** instead of C-9 (*δ*C 82.6), C-11 (*δ*C 73.5), and C-12 (*δ*C 76.1) in 4a-demethylpaspaline-3,4,4a-triol, indicating the replacement of two oxygenated carbons by a methylene (C-11) and an acetal carbon (C-9) in **2**. The HMBC correlations (Supplementary Material, Figure S16) from H-11, H-7, H-29, and H-28 to C-9 confirmed this deduction. The presence of an indole moiety, together with the HRESIMS data (Supplementary Material, Figure S19), indicated that **2** has a heptacyclic ring system. At this point, one more ring was required to fulfill the 11 double-bond equivalents, and an oxygen bridge was proposed according to the molecular formula. The oxygen bridge was assigned to connect C-9 and C-12 as deduced from distinctive deshielding of the C-12 (*δ*C 84.6) and C-9 (*δ*C 109.8) in **2** compared to corresponding C-12 (*δ*C 76.1) and C-9(*δ*C 96.4-96.5) [25,26] with a free hydroxyl group. In the ROESY spectrum (Supplementary Material, Figure S18), correlations of H-16/H3-26/H-5*β*, and H-5*α*/H3-25/H-13/H-7/H-10/H3-28 determined the relative configuration of **2**, as shown in Figure 3. In addition, the *β* orientation of the oxygen bridge was also proved by the ROESY correlation of H-7/H2-11 with the aid of the 3D ball-and-stick molecular model. Thus, the planar structure of compound **2** was established and named penerpene L. The ECD curve (Figure 4) of compound **2** is similar to **1**, indicating that the absolute configurations for the chiral carbons C-3, C-4, C-16, and C-13 in **1** were the same as those of **2**. The ECD calculation experiment also confirmed the above deduction (Figure 4), establishing the (3*S*,4*S*,7*S*,9*S*,10 *R*,12 *R*,13 *R*,16*S*)-**2** absolute configuration.

Compound **3** was assigned the molecular formula of C28 H41NO2 by HRESIMS, indicating nine degrees of unsaturation. The 1H NMR spectrum (Supplementary Material, Figure S21) displayed the typical pattern of a 3-substituted indole moiety with five aromatic protons

at *δ*H 7.04 (s, H-2), 7.30 (d, *J* = 8.0, H-7), 6.93 (t, *J* = 7.6, H-5), 7.02 (t, *J* = 7.6, H-6) and 7.55 (d, *J* = 8.0, H-4), as well as one clear olefinic proton at *δ*H 5.05 (t, H-21), one oxygenated proton *δ*H 3.27 (1H, H-17), and five methyls at *δ*H 1.62 (s, 24-Me), 1.55 (s, 23-Me), 0.62 (s, 27-Me), 1.11 (s, 26-Me), 0.94 (s, 25-Me). The 13C NMR data of **3** (Supplementary Materials, Figures S22 and S23) displayed 28 carbon signals, including ten aromatic or olefinic carbons (six protonated), five methyls, seven sp<sup>3</sup> methylenes, three sp<sup>3</sup> methines with one oxygenated, three non-protonated sp<sup>3</sup> carbons with one oxygenated. These NMR spectra data indicated that **3** was very similar to emeniveol (**16**) [24] except that a methyl in C-26 (*δ*C/H 16.3/1.11) and an oxygenated non-protonated sp<sup>3</sup> carbon C-14 (*δ*C 76.4) in **3** replaced two olefinic carbons in emeniveol, suggesting that the exocyclic double bond in emeniveol changed to a methyl and an oxygenated non-protonated sp<sup>3</sup> carbon. This obvious difference was supported by the HMBC correlations (Supplementary Material, Figure S25) from the H3-26 to C-14, C-13 (*δ*C 42.0) and C-9 (*δ*C 42.9). The relative configuration of **3** was determined by the ROESY spectrum (Figure 3), and the correlations of 14-OH/H3-25/H3-27/17-OH and H-9/H3-25 suggested the same face of these protons, while the correlation of H-12/H-17 indicated that these protons were on the opposite face to 17-OH. Thus, the planar structure of **3** was assigned, as shown in Figure 1. The absolute configuration of **3** was established as (9*S*,12*S*,13*S*,14*S*,17*S*,18*S*)-**3** by comparison of its experimental ECD spectrum with the calculated ECD curves (Figure 4).

The molecular formula of compound **4** was established as C27H33NO4 on the basis of HRESIMS data, indicating 12 degrees of unsaturation. Analysis of the 1H and 13C NMR spectra (Tables 1 and 2, Supplementary Materials, Figures S30 and S31) of **4** suggested that its structure was closely related to that of penerpene G [9], a previously reported indole diterpene with an unusual 6/5/5/6/6/7 hexacyclic ring system bearing a 1,3-dioxepane ring. The main difference between them was the replacement of an oxygenated nonprotonated sp<sup>3</sup> carbon C-14 (*δ*C 78.3) in penerpene G by that of a sp<sup>3</sup> methine (*δ*C 42.8) in **4**. This assignment was confirmed by HMBC (Figure 2) correlations from H-12 (*δ*H 5.54) and H-27 (*δ*H 0.82) to C-14 and the COSY correction of H-14/H-15. The ROESY corrections (Figure 3) of H3-27/H-17/H-16*β* and H-16*α/*H3-26/H-14/H-7/H-9 assigned the same relative configuration of **4** as that of penerpene G. The ECD curves of compound **4** (Supplementary Material, Figure S43) are similar to that reported for penerpene G, containing strong positive CEs at a shorter wavelength (219 nm in **4** and 220 nm in penerpene G) and strong negative CEs at a longer wavelength (238 nm in **4** and 234 nm in penerpene G), leading to the determination of the absolute configuration of **4** as (3*S*,4*S*,7*S*,9*S*,14*R*,17*S*)-**4**.

**Figure 2.** Key COSY and HMBC correlations of compounds **1**–**4**.

**Figure 3.** Key ROESY correlations of compounds **1**–**4**.

**Figure 4.** Experimental and calculated ECD curves for compound **1** (**A**); compound **2** (**B**); compound **3** (**C**).




**Table 1.** *Cont.*

Compound **1** was measured with 500 MHz 1H NMR. Compounds **2**, **3**, and **4** were measured with 600 MHz 1H NMR.

**Table 2.** 13C NMR (125 and 150 MHz) data of **1**–**4** in DMSO-*d*6.



Compound **1** was measured with 125 MHz 13C NMR. Compounds **2**, **3***,* and **4** were measured with 150 MHz 13C NMR.

## *2.2. Biological Assay*

The cytotoxic activities of compounds **1**–**16** were conducted by the MTT assay method [27] using cisplatin as the positive control. All compounds were tested against the human cervical cancer cell line HeLa, human gastric cancer cell line SGC-7901, human lung carcinoma cell line A549, and human liver cancer cell line BeL-7402. The results (Table 3) indicated that compound **9** exhibited the most pronounced activity against BeL-7402 with an IC50 value of 5.3 μM and was comparable to that of positive control cisplatin (IC50 4.1 μM). Compound **9** also displayed moderate cytotoxic activity against A549. While compounds **4** showed low cytotoxicity against HeLa. Compound **15** displayed mild inhibitory activity against HeLa, A549, and BeL-7402 (IC50 = 24.4–40.6 μM). The remaining compounds were found to be inactive against the three cell lines. All the tested compounds **1**–**16** were inactive against the cell line SGC-7901. The loss of a hydroxyl at C-14 suggested being a determinant of cytotoxicity shown by compound **4** against HeLa (**4** vs. penerpene G [9]). It is worth mentioning that it was the first-time report of the cytotoxicity of epipaxilline (**9**) [18]. Arintari et al. [19] and Sallam et al. [28] demonstrated the cytotoxicities of emindole SB (**15**) against human breast cell line MCF-7, murine lymphoma cell line L5178Y, and the human embryonic kidney cell line HEK-293, which provides emindole SB (**15)** a meaningful pharmacophore for further biological studies.

**Table 3.** Cytotoxicity of compounds **4**, **9**, and **15**.


a Positive control.

Compounds **1**–**16** were also tested for their antibacterial activity against *Escherichia coli* ATCC 25922, *Staphylococcus aureus* ATCC 6538, *Listeria monocytogenes* ATCC 1911, and *Bacillus subtilis* ATCC 6633 using the 96-well microtiter plates method [29] reported previously and using ampicillin as a positive control. The results (Table 4) revealed that six compounds **5**, **7**, **10**, **12**, **14**, and **15** showed moderate inhibitory activity against *S. aureus* ATCC 6538. Emindole SB (**15**) displayed reported selectivity toward *S. aureus* ATCC 33591 (MIC = 6.25 μg/mL) [21]. Compound **7** showed reasonable antibacterial activity against *B. subtilis* ATCC 6633 (MIC = 16 μg/mL), but compounds **5**, **10**, and **12** exhibited lower inhibitory. The results of the rest ten compounds (**1**–**4**, **6**, **8**–**9**, **11**, **13**, and **16**) did not show remarkable antibacterial activities (MIC > 128 μg/mL) against *S. aureus* and *B. subtilis*. In this assay, none of these compounds showed inhibitory activity against *E. coli* ATCC 25922 and *L. monocytogenes* ATCC 1911 (MIC > 128 μg/mL).


**Table 4.** Antibacterial activities of compounds **5**, **7**, **10**, **12**, **14**, and **15**.

a Positive control.
