Eremophilane-Type Sesquiterpenes from a Marine-Derived Fungus Penicillium Copticola with Antitumor and Neuroprotective Activities

Chemical examination of a marine sponge-associated Penicillium copticola fungus resulted in the isolation of ten undescribed eremophilanes, namely copteremophilanes A–J (1–10), along with two new glycosides, 5-glycopenostatin F (11) and 5-glucopenostatin I (12). Their structures were determined by extensive spectroscopic data, in association with ECD data and chemical conversions for configurational assignments. Analogs 1, 2, and 10 represent a group of uncommon skeletons of eremophilanes with an aromatic ring and a methyl migration from C-5 to C-9, and analogs 11 and 12 are characteristic of a PKS scaffold bearing a glucose unit. The incorporation of a chlorinated phenylacetic unit in 3–9 is rarely found in nature. Analog 7 showed neuroprotective effect, whereas 8 exhibited selective inhibition against human non-small cell lung cancer cells (A549). This study enriched the chemical diversity of eremophilanes and extended their bioactivities to neuroprotection.


Introduction
Farnesyl diphosphate (FPP) biogenetically generates sesquiterpenes with diverse scaffolds via various ring rearrangements. Eremophilanes are one class of sesquiterpenes characterized by the presence of a bicyclic backbone with irregular rule to assemble FPP in association with a methyl migration from C-10 to C-5 [1,2]. Eremophilane analogues extensively distribute in nature. Apart from the eremophilanes derived by plants of which the genus Ligularia is dominated to produce eremophilanes [3], terrestrial-or marine-originated fungi emerge as the additional sources to produce relevant analogs. It is noteworthy that most fungal eremophilanes are enantiomeric to the corresponding entities from plants [1], and marine-associated fungi have the potential to generate structurally unique analogs [4][5][6]. Eremophilanes possess wide range of bioactivities, such as phytotoxins, antimicrobials, protein inhibitors, immunomodulators and cytotoxins. In marine-derived fungi, eremophilanes with chemical diversity are widely distributed in fungal genera of Acremonium [6], Penicillium [7][8][9], Cochliobolus [10], Phomopsis [11], and Cryptosphaeria [12]. These findings imply marine-derived fungi as a potential source to generate structurally unique and bioactive eremophilanes. With the aim to continue our discovery of bioactive natural products from marine-associated organisms, a marine sponge (Xestospongia testudinaria)-associated fungus strain Penicillium copticola WZXY-m122-9 was selected for chemical examination. The LC-MS/MS data of the EtOAc extract of the cultured fungus were processed into a molecular network using MZmine and the GNPS platform (http://gnps.ucsd.edu, accessed on 20 June 2022), which allowed the formation of the spectral nodes into clusters ( Figure S103). Annotation of the nodes in a cluster with m/z values of 200 to 300 by GNPS MS/MS spectral library matched eremophilanes,

Structure Elucidation of New Compounds
Copteremophilane A (1) has a molecular formula of C14H18O2 as determined by the HRESIMS and NMR data, requiring six degrees of unsaturation. The 13 C NMR and DEPT data afforded a total of 14 carbon resonances, including six aromatic carbons for a phenyl group, a ketone, as well as seven alkyl carbons, of which three were classified into methyl groups. A tetra-substituted aromatic ring was recognized by the presence of two meta-coupled aromatic protons, H-6 (δH 7.61, brs) and H-8 (δH 7.55, brs). The substitution of a methyl group at C-9 (δC 136.4) and an acetyl group at C-7 (δC 140.8) was confirmed by the HMBC correlations from H3-14 (δH 2.24, s) to C-8 (δC 126.8), C-9, and C-10 (δC 134.8), and both H-6 and H-8 to carbonyl carbon C-11 (δC 198.3), along with H3-12 (δH 2.53, s) to C-7 (δC 140.8) and C-11. The COSY correlations established a segment in ring A from C-1 to C-4, and this unit was fused to C-5 (δC 141.1) and C-10 positions of the aromatic ring to form a cyclohexene ring due to the HMBC correlations of C-5 and C-10 to both H2-1 (δH 2.62, 2.69) and dq,J = 2.6,7.2 Hz). In addition, a methyl group at C-4 and a hydroxy group at C-3 were confirmed by the COSY correlations from H-4 to H-3 (δH 3.66, dt, J =2. 6,4.9 Hz) and d,J = 7.2 Hz), along with the HMBC correlations from H3-15 to C-3 (δC 69.8), C-4 (δC 41.2), and C-5. The NOE correlation between H3-15 and H-3 was indicative of a trans-orientation of H3-15 toward OH-3. Based on the modified Mosher method, the (R)-MPA and (S)-MPA esters of 1 were synthesized. Calculation

Structure Elucidation of New Compounds
Copteremophilane A (1) has a molecular formula of C 14 H 18 O 2 as determined by the HRESIMS and NMR data, requiring six degrees of unsaturation. The 13 C NMR and DEPT data afforded a total of 14 carbon resonances, including six aromatic carbons for a phenyl group, a ketone, as well as seven alkyl carbons, of which three were classified into methyl groups. A tetra-substituted aromatic ring was recognized by the presence of two metacoupled aromatic protons, H-6 (δ H 7.61, brs) and H-8 (δ H 7.55, brs). The substitution of a methyl group at C-9 (δ C 136.4) and an acetyl group at C-7 (δ C 140.8) was confirmed by the HMBC correlations from H 3 -14 (δ H 2.24, s) to C-8 (δ C 126.8), C-9, and C-10 (δ C 134.8), and both H-6 and H-8 to carbonyl carbon C- 11 (δ C 198.3), along with H 3 -12 (δ H 2.53, s) to C-7 (δ C 140.8) and C-11. The COSY correlations established a segment in ring A from C-1 to C-4, and this unit was fused to C-5 (δ C 141.1) and C-10 positions of the aromatic ring to form a cyclohexene ring due to the HMBC correlations of C-5 and C-10 to both H dq,J = 2.6,7.2 Hz). In addition, a methyl group at C-4 and a hydroxy group at C-3 were confirmed by the COSY correlations from H-4 to H-3 (δ H 3.66, dt, J =2. 6,4.9 Hz) and H [3][4][5][6][7][8][9][10][11][12][13][14][15]d,J = 7.2 Hz), along with the HMBC correlations from H 3 -15 to C-3 (δ C 69.8), C-4 (δ C 41.2), and C-5. The NOE correlation between H 3 -15 and H-3 was indicative of a trans-orientation of H 3 -15 toward OH-3. Based on the modified Mosher method, the (R)-MPA and (S)-MPA esters of 1 were synthesized. Calculation of the ∆δ (δ R − δ S ) values resulted in 3R configuration ( Figure 2). Thus, a 4R configuration was suggested with the help of the NOE data. The similar data of the experimental ECD compared to that calculated for (3R, 4R)-1 ( Figure 3A) further supported the configurational assignment. of the Δδ (δR − δS) values resulted in 3R configuration ( Figure 2). Thus, a 4R configuration was suggested with the help of the NOE data. The similar data of the experimental ECD compared to that calculated for (3R, 4R)-1 ( Figure 3A) further supported the configurational assignment. Δδ RS (L 1 )<0 The molecular formula (C15H20O2) of copteremophilane B (2) was afforded by the HRESIMS and NMR data. The NMR data of both 1 and 2 (Table 1) were comparable, except for the distinction of the substituent at C-7 (δC 136.3). A hydroxyisopropene unit was identified by the NMR resonances for two olefinic carbons and a hydroxymethyl group, in association with the HMBC correlations from the hydroxymethyl protons H2-12 at δH 4.28 (s) to C-11 (δC 148.3) and C-13 (δC 110.1) and between H2-13 (δH 5.22, 5.34) and C-12 (δC 63.0). The location of the hydroxyisopropene unit at C-7 was confirmed by the HMBC correlations of H-6 (δH 7.06, brs) and H-8 (δH 7.03, brs) to C-11 and from H2-13 to C-7. The same relative configuration of both 1 and 2 was evident from the NOE correlations between H-3 and H3-15. The similar ECD data suggested the same absolute configuration for both 1 and 2. This was supported by the calculated ECD data ( Figure 3B), in which the experimental ECD data were consistent with those calculated for (3R, 4R)-2.  Δδ RS (L 1 )<0 The molecular formula (C15H20O2) of copteremophilane B (2) was afforded by the HRESIMS and NMR data. The NMR data of both 1 and 2 (Table 1) were comparable, except for the distinction of the substituent at C-7 (δC 136.3). A hydroxyisopropene unit was identified by the NMR resonances for two olefinic carbons and a hydroxymethyl group, in association with the HMBC correlations from the hydroxymethyl protons H2-12 at δH 4.28 (s) to C-11 (δC 148.3) and C-13 (δC 110.1) and between H2-13 (δH 5.22, 5.34) and C-12 (δC 63.0). The location of the hydroxyisopropene unit at C-7 was confirmed by the HMBC correlations of H-6 (δH 7.06, brs) and H-8 (δH 7.03, brs) to C-11 and from H2-13 to C-7. The same relative configuration of both 1 and 2 was evident from the NOE correlations between H-3 and H3-15. The similar ECD data suggested the same absolute configuration for both 1 and 2. This was supported by the calculated ECD data ( Figure 3B), in which the experimental ECD data were consistent with those calculated for (3R, 4R)-2.  The molecular formula (C 15 H 20 O 2 ) of copteremophilane B (2) was afforded by the HRESIMS and NMR data. The NMR data of both 1 and 2 (Table 1) were comparable, except for the distinction of the substituent at C-7 (δ C 136.3). A hydroxyisopropene unit was identified by the NMR resonances for two olefinic carbons and a hydroxymethyl group, in association with the HMBC correlations from the hydroxymethyl protons H 2 -12 at δ H 4.28 (s) to C-11 (δ C 148.3) and C-13 (δ C 110.1) and between H 2 -13 (δ H 5.22, 5.34) and C-12 (δ C 63.0). The location of the hydroxyisopropene unit at C-7 was confirmed by the HMBC correlations of H-6 (δ H 7.06, brs) and H-8 (δ H 7.03, brs) to C-11 and from H 2 -13 to C-7. The same relative configuration of both 1 and 2 was evident from the NOE correlations between H-3 and H 3 -15. The similar ECD data suggested the same absolute configuration for both 1 and 2. This was supported by the calculated ECD data ( Figure 3B), in which the experimental ECD data were consistent with those calculated for (3R, 4R)-2. Copteremophilane C (3) was determined to have a molecular formula of C 23 H 25 ClO 7 by the HRESIMS and NMR data, requiring 11 degrees of unsaturation. Diagnostic 2D NMR data revealed 3 to be assembled by two moieties. Based on the 2D NMR data, one of them was identified as a phomenone unit, a co-isolated eremophilane-type sesquiterpene which was structurally characterized by the presence of an α,β-unsaturated ketone with an epoxy group at ring B [13]. The second moiety contained six aromatic carbons for a phenyl unit, a methylene and a carbonyl carbon, along with two meta-coupling aromatic protons H-2 (δ H 6.70, d, J = 2.1 Hz) and H-6 (δ H 6.67, d, J = 2.1 Hz) and the methylene protons H 2 -7 (δ H 3.50, brs). The HMBC correlations from H 2 -7 to a carbonyl carbon at C-8 (δ C 171.2), C-2 (δ C 121.0) and C-6 (δ C 115.5) and from H-2 and H-6 to the aromatic carbons established a tri-substituted phenylacetic segment. Additional HMBC correlations from two phenol protons OH-4 (δ H 9.01, brs) and brs) to the aromatic carbons, respectively, in association with the molecular composition, allowed the assignment of a 3-chloro-4,5-dihydroxyphenylacetic moiety. The linkage of the acyl unit to C-3 for an ester formation was evident from the HMBC correlation between H-3 (δ H 4.82, dt, J = 4.4, 11.2 Hz) and C-8 . The same relative configuration of 3 as phomenone was identified by the NOE relationships of both H 3 -14 (δ H 1.22, s) and H [3][4][5][6][7][8][9][10][11][12][13][14][15]d,J = 6.8 Hz) with H-3 and H-6 (δ H 3.40, s) and between H-6 and H 2 -13 (δ H 5.08, 5.20) (Figure 4). Comparison of the experimental and calculated ECD data ( Figure 3C) suggested 3 possessing R configuration for C-3, C-4, C-5, C-6, and C-7.
The structure of copteremophilane F (6) was determined to be a 5′-methoxylated analog of 4 due to the comparable NMR data of both analogs, with the exception of an additional methoxy group in 6. The methoxy protons at δH 3.78 (s) showed HMBC correlation to C-5′ (δC 148.4) and the NOE relationship with H-6′ (δH 6.77, d, J = 1.9 Hz) confirmed a 3-chloro-4-hydroxy-5-methoxyphenylacetic unit, which was linked to C-12 (δC 64.5) on the basis of the HMBC correlation between H2-12 (δH 4.70, 4.76) and the carbonyl carbon C-8′ (δC 170.6). The similar NOE data of 6 as those of 4 and 5 in association with the comparable ECD data of 6 to those calculated for (3R, 4R, 5R, 6R, 7R)-6 identified the same configuration of 6 as 4.
Copteremophilane G (7) has a molecular formula of C23H27ClO6 as determined by the HRESIMS data. The NMR data demonstrated 7 possessing two moieties. Analyses of 2D NMR data ( Figure S104) revealed the sesquiterpene unit being identical to JBIR-27 [9], The molecular composition of copteremophilane D (4) showed the same as that of 3 according to the HRESIMS data. The NMR data of both compounds (Tables 2 and 3) were very similar, and the 2D NMR data established the gross structure of 4 also containing two segments which were identical to those of 3. The distinction was attributed to the 3-chloro-4,5-dihydroxyphenylacetic moiety location. The chemical shifts of C-3 (δ C 69.0 vs. δ C 73.2 of 3) and C-12 (δ C 64.9 vs. δ C 62.1 of 3) in addition to the HMBC correlation between H 2 -12 (δ H 4.70, 4.71) and a carbonyl carbon C-8 (δ C 171.1) allowed the location of the acyl group at C-12. The similar NOE data of both 3 and 4, along with the comparable ECD data of 4 to those calculated for (3R, 4R, 5R, 6R, 7R)-4 ( Figure 3D), assigned the same configuration of both 3 and 4.
The molecular formula of copteremophilane E (5) (Tables 2 and 3) were comparable to those of 4, and the NMR resonances for phomenone moiety of both compounds were identical. The acyl moiety was identified as 3-chloro-4-hydroxyphenylacetic group based on the presence of an ABX spin system for the aromatic protons at H-3 (δ H 6.88, d, J = 8.2 Hz), H-2 (δ H 6.97, dd, J = 1. 5,8.2 Hz), and H-6 (δ H 7.18, d, J = 1.5 Hz), and the HMBC correlations from the methylene protons H 2 -7 (δ H 3.54, s) to C-1 (δ C 126.1), C-2 (δ C 131.0), C-6 (δ C 129.4) and the carbonyl carbon C-8 (δ C 171.2), along with the correlations of H-2 to the nonprotonated carbons C-3 (δ C 119.8) and C-4 (δ C 152.5). The similar NOE and ECD data ( Figure 3E) suggested the same configuration for both 4 and 5. The structure of copteremophilane F (6) was determined to be a 5 -methoxylated analog of 4 due to the comparable NMR data of both analogs, with the exception of an additional methoxy group in 6. The methoxy protons at δ H 3.78 (s) showed HMBC correlation to C-5 (δ C 148.4) and the NOE relationship with H-6 (δ H 6.77, d, J = 1.9 Hz) confirmed a 3-chloro-4-hydroxy-5-methoxyphenylacetic unit, which was linked to C-12 (δ C 64.5) on the basis of the HMBC correlation between H 2 -12 (δ H 4.70, 4.76) and the carbonyl carbon C-8 (δ C 170.6). The similar NOE data of 6 as those of 4 and 5 in association with the comparable ECD data of 6 to those calculated for (3R, 4R, 5R, 6R, 7R)-6 identified the same configuration of 6 as 4.
Copteremophilane J (10) was determined to have a molecular formula of C 23 H 27 ClO 6 by the HRESIMS data, containing ten degrees of unsaturation. The NMR data featured a homolog of 3, and the 2D NMR data assigned a 3-chloro-4,5-dihydroxyphenylacetic group at C-3 as the case of 3. In regard to the sesquiterpene moiety, a phenyl ring was assigned to ring B due to the presence of six aromatic carbons as well as the HMBC correlations of the aromatic protons H-6 (δ H 7.12, brs) and H-8 (δ H 7.06, brs) to the aromatic carbons. The HMBC correlations of methyl protons H 3 -14 to C-8, C-9, and C-10 in association with the NOE correlation between H 2 -1 and H 3 -14 substituted a methyl group at C-9. The remaining NMR resonances were attributed to a 1,2-dihydroxypropane unit, which was deduced by the HMBC correlations from H 2 -12 (δ H 3.37, s) to C-11 (δ C 73.9) and C-13 (δ C 22.6). The location of this unit at C-7 (δ C 145.0) was confirmed by additional HMBC correlations from H 3 -13 (δ H 1.36, s) and H [2][3][4][5][6][7][8][9][10][11][12]ddd,J = 2.3,4.4,6.3 Hz) and H 3 -15 (δ H 1.21, d, J = 7.2 Hz) suggested a cis-orientation of H-3 toward H 3 -15. The absolute configuration of C-12 was determined as S on the basis of the Mo(AcO) 4 induced ECD data (ICD) (Figure 3J), in which a positive Cotton effect at 330 nm was observed.
Compound 11 has a molecular formula of C 28 H 42 O 8 as established by the HRESIMS data. The DEPT 13 C NMR spectrum exhibited a total of 28 carbon resonances, involving 6 olefinic carbons, a ketone carbon, 7 sp 3 oxygenated carbons, and 14 other alkyl carbons. Analyses of the 2D NMR data established a nucleus structure which was identical to penostatin F [13]. This assignment was evident from the COSY correlations along with the HMBC correlations to establish a 4-methylcycloocta-3,5-dien-1-one ( Figure 5). Additional COSY relationship between H-7 and H-8 in association with the HMBC correlations from H-7 to C-2 and C-3 and from H-2 to C-1 and C-9 fused a cyclohexene across C-1 and C-8. A spin system from H 2 -4 to H-7 via H-5 and H 2 -6 along with the HMBC correlation between H-5 and C-3 clarified a cyclopentane ring fused to C-3 and C-7, and C-5 (δ C 78.7) was oxygenated. The location of an n-heptane unit at C-14 was based on additional COSY and HMBC data. The remaining six oxygenated sp3 carbons were attributed to a sugar unit, which was identified as glucose according to the comparison of the NMR data with those of authentic sample. It was linked to C-5 based on the HMBC correlation between H-5 (δ H 4.38, t, J = 5.0 Hz) and C-1 (δ C 102.1). The NOE correlations between H-5 and H-8 (δ H 2.86, t, J = 5.6 Hz) and between H-7 (δ H 3.11, m) and H-14 ( Figure 6) suggested the relative configuration of the nucleus moiety to be identical to penostatin F. The J H-1 /H-2 value (7.9 Hz) suggested a β-form of Glc unit. Acidic hydrolysis of 11 derived an aglycone and a sugar, the former was identical to penostatin F based on the comparable NMR data and specific rotation and the latter was identical to authentic glucose due to the TLC and HPLC chromatographic comparison.
Compound 11 has a molecular formula of C28H42O8 as established by the HRESIMS data. The DEPT 13 C NMR spectrum exhibited a total of 28 carbon resonances, involving 6 olefinic carbons, a ketone carbon, 7 sp 3 oxygenated carbons, and 14 other alkyl carbons. Analyses of the 2D NMR data established a nucleus structure which was identical to penostatin F [13]. This assignment was evident from the COSY correlations along with the HMBC correlations to establish a 4-methylcycloocta-3,5-dien-1-one ( Figure 5). Additional COSY relationship between H-7 and H-8 in association with the HMBC correlations from H-7 to C-2 and C-3 and from H-2 to C-1 and C-9 fused a cyclohexene across C-1 and C-8. A spin system from H2-4 to H-7 via H-5 and H2-6 along with the HMBC correlation between H-5 and C-3 clarified a cyclopentane ring fused to C-3 and C-7, and C-5 (δC 78.7) was oxygenated. The location of an n-heptane unit at C-14 was based on additional COSY and HMBC data. The remaining six oxygenated sp3 carbons were attributed to a sugar unit, which was identified as glucose according to the comparison of the NMR data with those of authentic sample. It was linked to C-5 based on the HMBC correlation between H-5 (δH 4.38, t, J = 5.0 Hz) and C-1′ (δC 102.1). The NOE correlations between H-5 and H-8 (δH 2.86, t, J = 5.6 Hz) and between H-7 (δH 3.11, m) and H-14 (Figure 6) suggested the relative configuration of the nucleus moiety to be identical to penostatin F. The JH-1′/H-2′ value (7.9 Hz) suggested a β-form of Glc unit. Acidic hydrolysis of 11 derived an aglycone and a sugar, the former was identical to penostatin F based on the comparable NMR data and specific rotation and the latter was identical to authentic glucose due to the TLC and HPLC chromatographic comparison. Compound 12 has the same molecular formula as 11 according to the HRESIMS data. The NMR data of 12 (Table 4) were almost superimposed to those of 11, suggesting the structural similarity of both compounds. Analysis of 2D NMR data established the same planar structure of 11 and 12. The similar NOE data with the exception of the correlation between H-5 and H-7 instead of between H-7 and H-1′ of 11 suggested the stereoisomers of both compounds ( Figure 6). However, the opposite rotation ([α] 20 D −10 for 11 and [α] 20 D +12 for 12) and the opposite Cotton effects (Figure 7) suggested an enanti- Compound 12 has the same molecular formula as 11 according to the HRESIMS data. The NMR data of 12 (Table 4) were almost superimposed to those of 11, suggesting the structural similarity of both compounds. Analysis of 2D NMR data established the same planar structure of 11 and 12. The similar NOE data with the exception of the correlation between H-5 and H-7 instead of between H-7 and H-1 of 11 suggested the stereoisomers of both compounds ( Figure 6). However, the opposite rotation ([α] 20 D −10 for 11 and [α] 20 D +12 for 12) and the opposite Cotton effects (Figure 7) suggested an enantiomeric form with the exception of the chiral center at C-5, implying the nucleus part of 12 to be consistent with penostatin I [13].  In addition, the phenylacetic acids, including sporogen AO-1 [14], phomenone [15], JBIR28 [9], JBIR27 [9], 3-chloro-4,5-dihydroxyphenylacetic acid, 3-chloro-4-hydroxyphenylacetic acid, and 3-chloro-4-hydroxy-5-methoxyphenylacetic acid [16], were isolated and identified on the basis of spectroscopic data.

Bioassays
Fungus-derived eremophilanes gain potential activities against cancer cell lines for specific aspects in cancer therapies [17,18]. To evaluate the relevant effects of the isolated compounds, an MTT method was performed to detect the cytotoxic activities of compounds toward human non-small cell lung cancer cells (A549), human colon cancer cells  and human breast cancer cells (MCF-7). As shown in Table 5, eremophilane analogs possessing specific selection to inhibit tumor cell lines were observed. Copteremophilane H (8) with an acyl unit at C-14 selectively inhibited A549 cells with IC 50 value of 3.23 µM. However, analog 7 with an olefinic rearrangement from C-7/C-11 of 8 to C-11/C-13 totally attenuated the inhibitory effect. Analogs 4 and 5 with a phenylacetic unit at C-12 of phomenone showed selective inhibition against HCT-8 cell line. It is noteworthy that 4 with a 3-chloro-4,5-dihydroxy-phenylacetic unit exhibited higher activity than 5 which possesses a 3-chloro-4-hydroxyphenylacetic unit, whereas 6 with a 5-methoxy-4-hydroxy-3-chlorophenylacetic unit was inactive toward the three tumor cell lines. These findings suggested that the substituents at the phenylacetic unit are sensitive to affect the activities. In contrast with the bioactive analogs, the eremophilane nuclei and the phenylacetic units alone had weak to no antitumor effects, implying the necessity for the combination of eremophilane moiety with phenylacetic unit to improve antitumor effects. In addition, the irregular eremophilanes 1, 2 and 10 with an aromatic ring B attenuated the inhibitory activities. Table 5. Inhibitory effects of 1-10 against tumor cell lines.

General Experimental Procedures
Optical rotations were recorded on an Autopol-III automatic polarimeter (Rudolph Research Co., Ltd., Hackettstown, NJ, USA). UV spectra were recorded on a Cary 300 spectrometer. IR spectra were measured on a Thermo Nicolet Nexus 470 FT-IR spectrometer. ECD spectra were measured on a JASCO J-815 spectropolarimeter. 1 H and 13 C NMR spectra together with 2D NMR spectra were measured on Bruker Advance NMR spectrometers (400 MHz or 600 MHz for 1 H and 100 MHz or 125 MHz for 13 C, respec-

General Experimental Procedures
Optical rotations were recorded on an Autopol-III automatic polarimeter (Rudolph Research Co., Ltd., Hackettstown, NJ, USA). UV spectra were recorded on a Cary 300 spectrometer. IR spectra were measured on a Thermo Nicolet Nexus 470 FT-IR spectrometer. ECD spectra were measured on a JASCO J-815 spectropolarimeter. 1 H and 13 C NMR spectra together with 2D NMR spectra were measured on Bruker Advance NMR spectrometers (400 MHz or 600 MHz for 1 H and 100 MHz or 125 MHz for 13 C, respec-Scheme 1. Biogenetic relationships of the isolated eremophilanes.

General Experimental Procedures
Optical rotations were recorded on an Autopol-III automatic polarimeter (Rudolph Research Co., Ltd., Hackettstown, NJ, USA). UV spectra were recorded on a Cary 300 spectrometer. IR spectra were measured on a Thermo Nicolet Nexus 470 FT-IR spectrometer. ECD spectra were measured on a JASCO J-815 spectropolarimeter. 1 H and 13 C NMR spectra together with 2D NMR spectra were measured on Bruker Advance NMR spectrometers (400 MHz or 600 MHz for 1 H and 100 MHz or 125 MHz for 13 C, respectively). Chemical shifts are expressed in δ referenced to the solvent peaks of DMSO-d 6 ( 1 H at δ H 2.50 and for 13 C at δ C 39.5). HRESIMS spectra were obtained on a Bruker APEX IV 70 eV FT-MS spectrometer and on a Waters Xevo G2 Q-TOF spectrometer fitted with an ESI source (acquisition range:100-1000, acquisition: start time 0 and end time 4, voltages: ESI + 2 kV and ESI − 1.5-2 kV, external standard: HCOONa). DAD HPLC was performed on Waters e2695 separations module with Waters 2998 photodiode array detector and Thermo BDS column (250 × 4.6 mm, 5 µm). Column chromatography (CC) was performed using ODS (50 µm, Daiso), silica gel Qingdao Marine Chemistry Co.,Ltd.,Qingdao,China), and Sephadex LH-20 (Amersham Pharmacia Biotech AB, Staffanstorp, Sweden). Precoated silica gel plates (Qingdao Marine Chemistry Co., Ltd.) were used for TLC analyses. Solvents used for isolation are all analytical grade. Semipreparative HPLC was performed on an Alltech 426 pump using a Uvis-201 detector, and the Prevail C 18 column (semipreparative, 5 µm) was used for separation.

Fungal Strain and Identification
Fungal strain WZXY-m122-9 was isolated from the sponge of Xestospongia testudinaria, which was collected from Weizhou island in May 2016. The fungal identification was based on the ITS gene sequence which showed 100% similarity to a Penicillium copticola clone using Blast. The ITS gene sequence was deposited in GenBank. The strain is preserved at the State Key Laboratory of Natural and Biomimetic Drugs, Peking University.
Copteremophilane A (1) was added into each well and incubated for 2 h at 37 • C. Then, the medium was changed with the same volume of DMSO. After incubation, absorbance is read at 570 nm for MTT by a spectrophotometer and the quantity of formazan product is directly proportional to the number of living cells in culture.

Cell Viability Assay
Rat pheochromocytoma (PC12) cells were cultured in Dulbecco's modified Eagle medium (DMEM), and supplemented with 10% FBS, 100 U/mL penicillin, and 100 µg/mL streptomycin in a humidified atmosphere at 37 • C with 5% CO 2 . The cells were passaged every 3 days. PC12 cells (1 × 10 4 ) were cultured for 24 h, and then the medium was replaced with serum-free DMEM medium. The PC12 cells were divided into three groups. In the model group, PC12 cells were treated using Aβ [25][26][27][28][29][30][31][32][33][34][35] with different concentrations. Then, the cells were incubated for 12 h, 24 h and 48 h until there was a cell viability of up to 50%. PC12 cells without pretreatment were set up as control. The third group of PC12 was treated with different concentrations of compounds to detect the cell viability by CCK8 method. The model groups (Aβ 25-35 induced PC12 cells) were treated with compounds in different concentrations, and the viability of PC12 cells was measured using cell counting kit-8 (Dojindo, Japan), according to the manufacturer's instructions. Then, the medium was aspirated and cells in each group were incubated with 10 µL of CCK8 at 37 • C for 2 h. Afterwards, the absorbance was determined at a wavelength of 450 nm with a microplate reader (Thermo, MuLTiSKAN MK3, Waltham, MA USA).

Measurement of Malondialdehyde (MDA)
Cultured PC12 cells were initially seeded in 6-well plates for 24 h. The cells were then pre-incubated with or without compound, followed by incubation with Aβ 25-35 (40 µM) for 24 h. The cultures were washed with ice-cold PBS and homogenized. The homogenate was centrifuged at 4 • C. The protein concentration in each sample was determined by the MDA Protein Assay Kit as a reference standard. The levels of MDA were determined according to the manufacturer's instructions. Concentrations were normalized to the protein concentration expressed as a percentage of control samples.

Lactate Dehydrogenase (LDH) Release Assay
Cell injury was assessed through measuring the LDH activity in the supernatant of PC12 cells using an LDH kit according to the manufacturer's protocol. In brief, doubledistilled H 2 O, 0.2 µM pyruvic acid, matrix buffer and coenzyme I buffer were added in sequence at 48 h after the treatment by compound. The supernatant was collected after incubation for 30 min at room temperature. The absorbance at 450 nm was then measured with a microplate reader.

Statistical Analysis
Data are expressed as mean ± standard deviation (SD). Biostatistical analyses were conducted with SPSS 16.0 software. Statistical differences among groups were assessed by one-way analysis of variance (ANOVA). Differences were considered to be statistically significant at p < 0.05.

Conclusions
In summary, this work reports a chemical examination of the marine sponge-derived P. copticola fungus to afford 12 undescribed natural products, of which 10 are identified as eremophilanes and 2 are determined as glucosides. Eremophilanes 1, 2 and 10 are characteristic of the eremophilanes with a phenyl moiety in ring B accompanying a methyl migration to C-9, which are uncommonly found in nature. Eremophilanes 3-9 feature the esterification of a chlorinated phenylacetic unit in backbone, and this is the second isolation of relevant analogs from fungi. Analogs 11 and 12 are characteristic of an unprecedented PKS scaffold bearing a glucose unit. The bioassay results revealed the inhibitory effects of the isolated eremophilanes against tumor cell lines related to the structure variation. Analog 8 only inhibited human non-small cell lung cancer cells (A549), and analogs 4 and 5 selectively inhibited HCT-8. The eremophilane nuclei or acyl moieties showed inactivity towards tumor cell lines. These findings suggest the incorporation of chlorinated phenylacetic units significantly enhances the antitumor effects of eremophilanes, and the substituted positions and the acyl compositions are sensitive for the selective inhibition. In addition, the noncytotoxic analogs, such as 7, showed a neuroprotective effect. This study implies that eremophilanes have potential for development as antitumor or neuroprotective agents after structure modification.

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Conflicts of Interest:
The authors declare no conflict of interest.