**2. Results**

#### *2.1. Structural Elucidation of New Compounds*

A series of column chromatography (CC) methods were used during the isolation of *H. jecorina* H8. As a result, 11 compounds were isolated, including five new compounds, trichodermolide C (**1**), trichodermolide D (**2**), 7,7-,9--hydroxy-trichodimerol (**3**), 1-(2,4-dihydroxy-3,5-dimethylphenyl)-3,4,5-trihydroxyhexan-1-one (**4**), and isobisvertinol A (**5**). At the same time, by comparing NMR spectral data with those published in literatures, the six known compounds were determined to be 2-,3--dihydrosorbicillin (**6**) [9], 6-demethylsorbicillin (**7**) [10], sorbicillin (**8**) [9], (2*E*,4*E*)-1-(2,4-dihydroxy-3,5-dimethylphenyl)-6-hydroxyhexa-2,4-dien-1-one (**9**), trichodimerol (**10**), and bisvertinol (**11**) [11] (Figure 1).

**Figure 1.** Structures of compounds **1**–**11** isolated from an extract of *Hypocrea jecorina* H8; the relative configuration of **5** is reported in this article.

Compound **1** was obtained as yellow amorphous powder. The molecular formula of **1** was established as C21 H26 O6 based on the HR-ESI-MS peak at *m/z* 375.1798 [M + H]+ (calcd for 375.1729 C21 H27 O6 +), requiring nine degrees of unsaturation. The IR spectrum of **1** indicated the presence conjugated lactone carbonyl signal of at 1691 cm<sup>−</sup><sup>1</sup> [6]. In high chemical shifts region of 1H NMR, four olefinic protons were observed at *δ*H 6.20 (d, 15.4 Hz, H-20), 6.35 (dt, 15.2, 4.8 Hz, H-21), 6.43 (dd, 15.4, 10.2 Hz, H-18), and 7.25 (dd, 15.2, 10.8 Hz, H-19) (Table 1).


Their corresponding olefinic carbon signals were found in the *sp*<sup>2</sup> region of the 13C NMR spectrum at *δ*C 128.8 (C-20), 143.2 (C-21), 127.9 (C-18), and 143.0 (C-19). The *sp*<sup>3</sup> low chemical shifts region of the 1H NMR spectrum displayed four notable methyl proton signals at *δ*H 1.48 (9-CH3), 1.77 (8-CH3), 2.27(14-CH3) linked to quaternary carbons, and *δ*H 0.95 (t, 7.2 Hz, 11-CH3) linked to a secondary carbon.

The 13C NMR data showed four carbonyl signals at *δ*C 204.0 (C-13), 196.5 (C-17), 191.3 (C-6), and 174.9 (C-2) attributing to a ketone carbonyl, two conjugated ketone carbonyl, and ester carbonyl, respectively. In addition, the 13C NMR spectrum and DEPT data of **1** showed the presence of 21 carbons, sorted into four methyls, four methylenes, five methines (four olefinic carbons), and eight quaternary carbons (four carbonyl carbons and two olefinic carbons).

The above spectroscopic data showed high similarities to those of trichodermolide B, a known compound isolated from *Trichoderma reesei* (HN-2016-018) [12], except for the presence of signal for a hydroxyl group, a signal for methylene group at C-22 (*δ*H 4.32, *δ*C 62.7) and the lack of signal for a methyl group (*δ*C 19.1). Thus, **1** was deduced to be a hydroxylated derivative of trichodermolide B at C-22, validated by the COSY correlations of *δ*H 6.35 (H-21) with *δ*H 4.32 (H2-22) (Figure 2a).

The two double bonds in sorbyl side chain for **1** were assigned both as *E* configuration based on their coupling constants (*J*H-18/H-19 = 15.2 Hz, *J*H-20/H-21 = 15.4 Hz) and the NOESY correlation between H-18/H-20. For the bridged bicycle lactone ring system, it was only possible if the CH3-9 and CH2-10 were oriented equatorially. In addition, the 1H NMR chemical shifts of H-15 (*δ*H 3.37 for trichodermolide B compared to *δ*H 3.55 for **1**), H-16 (*δ*H 2.54, 2.99 for trichodermolide B compared to *δ*H 2.43, 3.13 for **1**), CH3-9 (*δ*H 1.37 for trichodermolide B compared to *δ*H 1.48 for **1**), and CH2-10 (*δ*H 1.25, 2.03 for trichodermolide B compared to *δ*H 1.30, 2.08 for **1**) suggested the same relative configurations of C-3, C-15, and C-7 in 1 as that in trichodermolide B [12].

Therefore, the relative configuration of **1** was assumed as 3*S*\*,15 *R*\*,7 *R*\*. The ECD curve of **1** showed a negative Cotton effect around 220 nm and a positive Cotton effect around 270 nm, respectively. These were the same as trichodermolide B (Figure S41: Calculated and experimental ECD spectra of trichodermolide B). The absolute configuration of **1** was assigned as 3*S*,15*R*,7*R* (Figure 3a). As a result, the structure of **1** was determined and named as trichodermolide C (Figure 1).

**Figure 2.** Key 1H–1H COSY, HMBC, and NOESY correlations of compounds **1**–**5**. (**a**) Key COSY, HMBC, and NOESY correlations of compounds **1**; (**b**) Key COSY, HMBC, and NOESY correlations of compounds **2**; (**c**) Key COSY, HMBC, and NOESY correlations of compounds **3**; (**d**) Key COSY, HMBC correlations of compounds **4**; (**e**) Key COSY, HMBC, and NOESY correlations of compounds **5**.

Compound **2** was also obtained as yellow amorphous powder. The molecular formula was deduced to be C21H30O5 by interpretation of the HR-ESI-MS peak at *m/z* 363.2162 [M + H]+ (calcd for 363.2093 C21H31O5+), implying seven degrees of unsaturation. Compound **2** presented 1H and 13C NMR signals similar to those of compounds **1**, especially those on the bridged bicyclic ring moiety. The structural differences in the side chains could be revealed by the DEPT spectra, in which two carbonyl signals vanished and two oxygenated methine signals emerged, together with the signal different from the oxygenated methylene (C-22) to methyl.

The corresponding 1H NMR signals in **2** were found at *δ*H 4.17 (H-17), 4.10 (H-13), and 1.76 (CH3-22), respectively. Thus, compound **2** is considered to be a reduction product of compound **1**. On the aids of 1H-1H COSY spectrum, two continuous connected spin systems revealed the presence of two side chain as: CH3-CH=CH-CH=CH-CH(O-)-CH2- and -CH2-CH(O)-CH2-. The key HMBC correlations from *δ*H 2.60 (CH2-12) to *δ*C 133.8 (C-5), 152.1 (C-4), 51.3 (C-3) and from *δ*H 1.87 (CH3-8) to *δ*C 191.9 (C-6), C-4, C-5 allowed the elucidation of structure for compound **2**.

The *E* configurations among the two double bonds in the side chains of **2** could be confirmed by the large coupling constant (*J*H-18/H-19 = *J*H-20/H-21 = 15.2 Hz). The relative configurations of three stereocenters, C-3, C-15, and C-7 of **2** were confirmed by comparison of its 1H NMR data with compound **1**. The NOESY correlations between H-15 and CH3-11, H-17/CH3-11, and CH3-8/CH2-12 established the relative configuration of **2**. Therefore, the relative configuration of **2** was assumed as 3*S*\*,15*R*\*,7*R*\*. The calculated ECD curve of 3*S*,15*R*,7*R*-**2** was consistent with the experimental data (Figure 3b), and hence the absolute configuration of **2** was assigned as 3*S*,15*R*,7*R*.

Compound **3** was obtained as a yellow amorphous powder. The results from the HR-ESI-MS peak at *m/z* 537.2090 [M + Na]+ (calcd for 537.2100 C28H34O9Na+) suggested that the molecular formula of **3** was C28H34O9, thus, implying twelve degrees of unsaturation. The 1H and 13C NMR data of **3** showed 30 protons and 28 carbons signals, and these carbon signals were classified into twelve quaternary carbons (four ethylenic bonds), two carbonyls, six ethylenic bonds, one oxygenated methines, two methines without an oxygen link, one methylene, and six methyls.

Comparison of the NMR data of **3** with those of **10**, a known metabolite isolated from a strain of the same genus [6], indicated that **3** possessed an identical trichodimerol [13] skeleton to **10**. Although compound **10** afforded only 14 signals because of a symmetric structure, some 13C signals of compound 3 split, suggesting an asymmetric structure (Table 2). The major difference were found at the signals due to C-8-, C-10- moiety, suggesting a hydration at C-8-/C-9- double bond in compound **3**. In 13C NMR and DEPT, the double bond signals of C-8- and C-9- in **10** turn to a methylene (*δ*C 40.9, C-8-) and an oxygen-linked methine (*δ*C 80.7, C-9-) in **3**. The conclusion was confirmed by the COSY correlation from *δ*H 2.38 (H-8-) to *δ*H 4.31 (H-9-), and to *δ*H 5.67 (H-10-). In the HMBC spectrum of **3**, the correlations from H-8- to C-7-, C-9- were also found.

The NOESY correlations between *δ*H 1.41 (CH3-14) and *δ*H 3.31 (H-1-) indicated the same orientation of these signals. In addition, the protons *δ*H 1.41 (CH3-14-) and *δ*H 2.94 (H-1) were simultaneously correlated with *δ*H 1.35 (CH3-13), reflecting that these signals locate at the same orientation. (Figure 2c). The absolute configuration of **3** was established as 1*S*,2*R*,3*S*, 4*S*, 1-*S*,2-*R*, 3-*S*, 4-*S* by comparison on experimental and calculated ECD spectra (Figure 3c).

Compound **4** was isolated as a colorless amorphous powder. The molecular formula of C14H20O6, which gave five unsaturation degrees, was established by the positive HR-ESI-MS ion peak at *m/z* 285.1342 [M + H]+ (calcd for 285.1338 C14H21O6+). The UV maximum absorption bands at *λ*max (log *ε*): 216 (3.66) nm were assigned to a conjugated carbonyl, which was confirmed by the 13C NMR data at *δ*C 192.1 (C-1-). In the 1H NMR, three methyl peaks at *δ*H 1.19 (d, 6.0 Hz, CH3-12), 2.04 (s, CH3-13), and 2.12 (s, CH3-14) were assigned. One olefinic proton was also observed at 7.34 (H-6).


**Table 2.** 1H NMR (600 Hz) and 13C NMR (150 Hz) data of **3**, **5** (CDCl3) and 13C NMR data of Compounds **10** and **11**.

In the 13C NMR, except for the carbonyl, six *sp*<sup>2</sup> carbons at *δ*C 160.3 (C-4), 159.9 (C-2), 125.1 (C-6), 118.4 (C-5), 113.6 (C-1), and 111.7 (C-3) and three oxygenated methines at *δ*C 77.1 (C-3-), 76.8 (C-3-), and 65.9 (C-5-) were assigned, respectively. The benzene ring signals in compound **4** closely resembled those of the known compound 2-,3--dihydrosorbicillin (**6**) [14]; however, they were different regarding the side chain. The COSY correlation data suggested the side chain of **4** was -CH2-CH(OH)-CH(OH)-CH(OH)-CH3. In addition, the HMBC correlation from *δ*H 2.89 (CH2-8) to *δ*C 192.1 (C-7) indicated that the carbonyl was at C-7. Comprehensive HSQC, COSY, and HMBC established the structure of **4** (Figure 1).

Compound **5** was obtained as a white amorphous powder with positive HR-ESI-MS ion peaks at m/z 499.2312 [M + H]+ indicating 12 degrees of unsaturation. According to the HR-ESI-MS data, compound **5** and **11** shared the same molecular formula (Figure S42). The NMR data of **5** had similar features compared to those of **11**, which suggests that they are stereoisomers. The planar structure of **5** was determined by the COSY and HMBC data. The main differences of NMR signals were attributable to C-1, C-4, and C-4a, with the 13C NMR deference of [*δ*C 191.8 (C-1), 73.9 (C-4), and 110.2 (C-4a) in **5** vs. *δ*C 194.3 (C-1), 72.7 (C-4), and 108.4 (C-4a) in **11**].

Thus, compounds **5** and **11** should be epimers around either C-4 or C-4a. In the NOESY spectrum, key cross peaks were observed between *δ*H 3.63 (H-8a) and *δ*H 1.34 (CH3-1a) and *δ*H 1.48 (CH3-5a) (Figure 2e), indicating that the methyls CH3-1a and CH3-5a were at the same side with H-8a. Furthermore, NOESY were observed between CH3-1a and *δ*H 1.29 (CH3-4). These results indicated that the relative configuration of C-4 was same as that of **11**. Therefore, we concluded that **5** is a stereoisomer of **11** on C-4a. The relative configuration of **5** was defined as shown in Figure 1.

#### *2.2. Evaluation of Antifungal Activity*

Compounds **1** to **11** were evaluated for antifungal activities by a paper disc inhibition assay. Compounds **5**, **6**, **8**, **9**, and **10** possessed significant activities against the tea pathogenic fungus *P. theae* (Table 3). The ED50 values of **5**, **6**, **8**, **9**, and **10** were 9.13 ± 1.25, 2.04 ± 1.91, 18.22 ± 1.29, 1.83 ± 1.37, and 4.68 ± 1.44 μg/mL, respectively. Compared to positive control hexaconazole (24.25 ± 1.57 μg/mL), compounds **5**, **6**, **8**, **9**, and **10** exhibited more potent antifungal activity. Particularly, new compound **5** had nearly 3-fold more, and known compound **9** had a 13-fold stronger anti-tea pathogenic fungus effect.


**Table 3.** Antifungal activities of **5**, **6**, **8**, **9**, and **10** (ED50, μg/mL).

\* hexaconazole serves as positive control.

#### *2.3. Evaluation of Toxicity*

The zebrafish is a small teleost that is becoming increasingly popular in many biomedical and environmental studies [15]. This model has shown sensitivity to a broad variety of contaminants (such as endocrine disruptors and organic pollutants), indicating their suitability as a biological method for environmental monitoring in risk assessment.

We evaluated the toxicity of compounds **5**, **6**, **8**, **9**, and **10** in a zebrafish model (Figure 4). Figure 4a showed that these compounds, except compound **8**, killed zebrafish embryo less than 50% when treated with a concentration of 10 μM. When the treatment time was prolonged to 72 h (Figure 4b), the mortality rate of zebrafish embryo caused by compound **8** at 0.625 μM increased to nearly 60%, whereas the effects of compounds **5**, **6**, **9**, and **10** did not change greatly.

In addition, the impact on the malformation of zebrafish by these compounds was observed using a Leica stereomicroscope. Figure 4c showed graphically under the same treatment that compounds **5** and **8** had greater effects than other compounds on the mortality rate and malformation of zebrafish both at a concentration of 0.625 μM for 24 h and at a concentration of 10 μM for 72 h. In summary, our data demonstrated that compounds **6**, **9**, and **10** were of low toxicity and could be used against tea pathogenic fungi agents and deserve further optimization.

**Figure 4.** Embryotoxicity and developmental toxicity assay; 15 zebrafish embryos per condition were exposed to compounds at the concentrations of 10, 5, 2.5, 1.25, and 0.625 μM, and 0.1% DMSO was used as blank control. (**a**) The mortality rate of 24 embryo treated with compounds; (**b**) The mortality rate of 72 embryo treated with compounds; (**c**) The impact on the malformation of zebrafish treated with compounds) The statistics of 24 and 72 h mortality rate; © Morphology of 24 h embryo or 72 h zebrafish larvae treated with compounds or control.
