Fungi are prolific producers of bioactive secondary metabolites and have contributed to improvements in human and animal health in spectacular and indispensable ways (Bills and Gloer, 2016; El-Elimat et al., 2021). They are ubiquitous in nature and often provide nutrients or protection for the host (Wani et al., 2015). Many endophytic fungi produce compounds with a novel structure and specific bioactivity for their long period of living in host tissues, which have been potential resources for novel compounds and new drugs (Jia et al., 2016; Kaul et al., 2012). Leptospermum brachyandrum is a famous ornamental and medicinal plant that belongs to the Myrtaceae family. In our previous research, three endophytes were isolated from this plant, such as Diaporthe foeniculina, Eutypella scoparia, and Rhytidhysteron sp., and a number of novel natural products with antibacterial or cytotoxic activity have been discovered from these endophytes (Zhang et al., 2021a; Zhang et al., 2021b; Zhang et al., 2020; Zhang et al., 2021). Simultaneously, five new 2-pyrones were isolated from D. foeniculina (Yu et al., 2021). With the aim of seeking new bioactive natural products from medicinal plant endophytes, chemical investigation of strain D. foeniculina BZM-15 was further researched and afforded to find three new metabolites, including one γ-butyrolactone derivative, diaportone A (1), one cyclopentenone derivative, diaportone B (3), and one monoterpene, diaportone C (5), along with six known compounds, colletolides A (2) (Li et al., 2019), phomotenone (4) (Ahmed et al., 2011), altiloxin B (6) (Hemberger et al., 2013), dankasterone A (7) (Jiao et al., 2015), 14α-hydroxyergosta-4,7,22-triene-3,6-dione (8) (Liu et al., 2018), and fortisterol (9) (Kitchawalitet al., 2014) (Figure 1). Herein, this report describes the isolation, structural elucidation, and biological activity of compounds 1–9.

Compound 1 was isolated as colorless oil and assigned the molecular formula C₁₅H₂₂O₅ as inferred from its HRESIMS ion peak at m/z 305.1361 [M + Na] ⁺ (calcd 305.1359 for C₁₅H₂₂NaO₅). The ¹H NMR spectral data (Table 1), in combination with the HSQC spectrum, displayed two ester carbonyl groups [δ _C 178.5 (C-2) and 172.2 (C-1′)], one conjugated diene [δ _H 5.92 (H-6), 6.48 (H-7), and 6.10 (H-8); δ _C 133.3 (C-6), 126.4 (C-7), 126.1 (C-8), and 139.2 (C-9)], two oxygenated methines [δ _H 5.25 (H-10) and 3.80 (H-4); δ _C 76.4 (C-10) and 82.5 (C-4)], one methine [δ _H 2.48 (H-3); δ _C 42.6 (C-3)], one oxygen quaternary carbon [δ _C 86.6 (C-5)], and five methyl groups [δ _H 1.22 (H-11), 1.51 (H-12), 1.76 (H-13), 1.31 (H-14), and δ _H 2.03 (H-2′); δ _C 12.4 (C-11), 25.2 (C-12), 12.7 (C-13), 19.4 (C-14), and 21.1 (C-2′)].

Compared to the NMR spectra of compound 2 (Table 1), they shared a typical γ-butyrolactone ring bearing two methyl groups at C-3 and C-5 and a hydroxyl group at C-4. However, the spectra of compound 1 exhibited additional signals for the acetyl group, which was connected with 10-OH to form an ester. This conclusion can be proved by the correlations of H-10 with C-1′ in the HMBC spectrum. Thus, the planar structure of compound 1 was elucidated, as shown in Figure 1.

The relative configuration was determined by NOESY spectrum and coupling constants. The NOE correlations of H-11/H-4/H-12 indicated that 11-H₃, 4-H, and 12-H₃ were on the same side of the lactone ring. A conjugated diene moiety was determined by the large coupling constant J = 15.5 Hz between H-6 and H-7 that was trans-oriented and J = 10.5 Hz between H-7 and H-8 that was cis-oriented. In addition, the E confirmation of 8,9-diene was deduced from the NOE cross peaks of H-6/H-8/H-10 and H₃-13/H-7 (Figure 2). The chiral HPLC analysis of compound 1 revealed that it should be optically pure. The calculated ECD spectrum was consistent with its experimental ECD spectrum, suggesting the absolute configuration of compound 1 as 3R,4R,5R,10S (Figure 3). Thus, the structure of compound 1 was determined as a new γ-butyrolactone derivative with 6,8-hexadien-1-ol,1-acetate side chain and was named diaportone A. Notably, diaportone A possess conjugated double bond, which might be light sensitive. Moreover, it might be an artificial compound generated from the known compound 2 through acylation, although much more evidence was needed.

Compound 3 was isolated as yellow oil. It was determined to have a molecular formula of C₁₁H₁₈O₂ by a combination of its HRESIMS ion peak at m/z 183.1388 [M + H] ⁺ (calcd 182.1307 for C₁₁H₁₉O₂). The ¹H and ¹³C NMR spectroscopic data (Table 2) of compound 3 exhibited characteristic signals assignable to a conjugated ketone carbonyl, a tetrasubstituted olefin, two methines (one oxygenated), three methylene groups, and three methyl groups. A detailed analysis of the NMR data exhibited that compound 3 was similar to the known compound 4 (Ahmed et al., 2011) (Table 2) with the only difference in compound 3 being the substitution position of CH₃-11, which was deduced based on the HMBC cross peaks of H₃-11 (δ _H 1.08)/C-1 (δ _C 212.2), C-5 (δ _C 39.9), and C-4 (δ _C 34.8) (Figure 2). The absolute configuration of compound 3 was confirmed by the similarity between the calculated ECD curve of 5S,6S and its experimental ECD spectrum (Figure 4). Therefore, the structure of compound 3, diaportone B, was defined as shown.

Compound 5 was isolated as a white solid. The molecular formula C₁₀H₁₆O₂ was determined by the HRESIMS ion peak at m/z 169.1224 [M + H] ⁺ (calcd 169.1229 for C₁₀H₁₇O₂). The 1D NMR data (Table 2) of compound 5 exhibited three methyl groups, two methylene groups, two methines, two sp³ nonprotonated carbons (one oxygenated), and one ketone carbonyl. Comparison of NMR data with those of dihydroxysabinane (Lin et al., 2009) revealed a high degree of similarity skeleton, where the only obvious difference is in the presence of a carbonyl group at C-1 in compound 5 instead of a hydroxyl group in dihydroxysabinane. This deduction was confirmed by the HMBC correlations of C-1 with H-2 (δ _H 1.88 and 2.36), H-6 (δ _H 1.88), and H-10 (δ _H 1.36 and 1.10) and obvious low-field chemical shift of C-1 (δ _C 210.9) (Figure 2).

The relative configuration of compound 5 was determined by the analysis of ROESY spectral data (Figure 2). The methylene (CH₂-10) and the hydroxyl groups at the chiral carbons (C-3) were assigned as β-oriented and the methyl groups (CH₃-9) and the methine (CH-6) should be α-oriented due to the presence of the ROESY correlation of H-4 with H₃-9. The above data supported the presence of two possible enantiomers (3S,4R,5R)-5 and (3R,4S,5S)-5. To determine the absolute configuration of compound 5, the ECD calculation was performed. Its experimental ECD spectrum of compound 5 was in good agreement with the calculated ECD spectrum for 3S,4R,5R (Figure 5). Therefore, the structure of compound 5 was determined and given the trivial name diaportone C.

Finally, the cytotoxic activities of all compounds inhibited three human cancer cell lines (SF-268, MCF-7, and HepG2), and antibacterial activities against MRSA and E. coli were evaluated. As show in Table 3, compounds 7 and 8 showed different potency of cytotoxicity against the three cell lines, with IC₅₀ values ranging from 3.6 to 15.8 μM, whereas they were not active on the normal cell line LX-2. Unfortunately, none of the compounds showed any antibacterial activities against MRSA and E. coli at a concentration of 100 µg/ml.

A phytochemical investigation into D. foeniculina BZM-15 resulted in the isolation and structural elucidation of three undescribed and six known compounds. Their structures including absolute configurations were determined by extensive physicochemical and spectroscopic analysis, as well as by ECD calculation. Cytotoxicity assays found that compounds 7 and 8 showed good cell inhibition against three human cancer cell lines (SF-268, MCF-7, and HepG2). This result enriched the study on the chemical constituents of D. foeniculina and validated that endophytic fungi remained a rich source of structurally/biologically new compounds.