*2.1. Structure Elucidation of Compounds* **1** *and* **2**

Euphonoid H (**1**) was obtained as a colorless oil and was shown to possess a molecular formula of C23H34O7 based on its HRESIMS ion at 445.2213 [M + Na]+ (calcd 445.2197). The 1H spectrum (Supplementary Material) showed signals for an olefinic hydrogen (δ<sup>H</sup> 6.71, H-14), one methoxyl (δ<sup>H</sup> 3.68, 16-OCH3) and four methyl groups (δ<sup>H</sup> 2.02, 0.91, 0.79 and 0.69). The 13C and HSQC NMR showed resonances assignable to one ketone (δ<sup>C</sup> 196.0), two ester (δ<sup>C</sup> 170.8 and 171.1), two olefinic carbons (δ<sup>C</sup> 133.7 and 154.2), four methyl (δ<sup>C</sup> 34.0, 22.1, 17.8 and 21.0), one methoxyl group (δ<sup>C</sup> 34.0), three oxygenated carbons (δ<sup>C</sup> 72.0, 69.5 and 62.5) and ten additional sp<sup>3</sup> carbons (δ<sup>C</sup> between 18.5 and 60.4). Comprehensive analysis of the 1D and 2D-NMR data (Table 1) revealed that compound **1** possessed, except for an acetoxyl group and a methoxy group, an abietane diterpene skeleton similar to that of methyl-8β,11β-dihydroxy-12-oxo-ent-abietadi-13,15(17)-ene-16-oate previously isolated from this plant [13]. However, the 13C NMR data for the Δ15(17) (δ<sup>C</sup> 137.3 for C-15 and 128.8 for C-17) of the latter were replaced by signals for a methine (δ<sup>C</sup> 45.1, C-15) and an oxymethylene (δ<sup>C</sup> 62.5, C-17). These observations implied that compound **1** was a hydrogenated derivative of the known compound.


**Table 1.** 1H (400 MHz) and 13C (100 MHz) NMR data for **1** and **2** in CDCl3.


**Table 1.** *Cont.*

Detailed 2D-NMR (1H-1H COSY, HSQC, HMBC and NOESY) data analysis further confirmed the above deduction and fulfilled the structural assignment. The 1H-1H COSY revealed four spin systems, CH2-1/CH2-2/CH2-3, H-5/CH2-6/CH2-7, H-9/H-11 and H-15/CH2-17 (Figure 2). HMBC correlations from H3-20 (to C-1 and C-10), H2-1 (to C-9), H3-18 (to C-3, C-4 and C-5), H2-6 (to C-4 and C-10), H-11 (to C-8 and C-13) and H-14 (to C-7, C-9 and C-12) to their corresponding carbons not only connected the former three fragments, but suggested that compound **2** shared the same ABC rings with methyl-8β,11βdihydroxy-12-oxo-ent-abietadi-13,15(17)-ene-16-oate. In addition, HMBC correlations from H-15 to C-12, C-13 and C-14 and from H2-17 to C-13 located the Δ15(17) double bond at C-13. HMBC correlations from the methoxyl group to C-16 suggested the presence of a methoxyformyl group, while the HMBC correlation from H2-17 to C-16 revealed its position at C-15. The acetoxyl group was connected to the abietane skeleton at C-17 by the key HMBC cross-peaks from H2-17 and H3-2 to C-1- . Thus, the gross structure of **1** was established as depicted.

**Figure 2.** Key 1H-1H COSY ( ) and HMBC ( ) correlations of **1** and **2**.

The NOESY correlations (Figure 3) of H-5/H-9 indicated that these protons were cofacial and were arbitrarily assigned to be *β*-oriented, while the NOESY correlation of H3-20/H-11 indicated that these protons were *α*-oriented. However, the NOESY spectrum did not give useful signals to determine the relative configuration of C-8 and C-15. To establish the relative configuration, the chemical shifts of four conformers were predicted at the B3LYP/6-311+G (d, p) level in chloroform (Figure 4). The results showed that the calculated chemical shifts of conformer **1b** was in the best agreement with the experimental values among those predicted for **1a**, **1b**, **1c** and **1d**. Further DP4+ analyses verified that conformer **1b** was assigned with a 99.99% probability among all the conformers (Figure 4). These results suggested that compound **1** had the structure of conformer **1b** with the relative stereochemistry of 5*R*\*, 8*R*\*, 9*R*\*, 10*R*\*, 11*R*\*, 15*S*\*.

**Figure 3.** Key NOE correlations ( ) of compounds **1** and **2**.

**Figure 4.** 13C NMR calculation results of compound **1** at the mPW1PW91/6-311+G(d,p) level. (**a**) Structures of conformer **1a**–**1d**. (**b**) Key parameters of the calculated chemical shifts of conformers **1a**–**1d**.

The absolute configuration of **1** was established by comparing its experimental ECD spectrum with those calculated at the CAM-B3LYP/6-31+G(d) level in acetonitrile. As shown in Figure 5, the experimental ECD curve of **1** showed first negative and second positive Cotton effects around 250 and 213 nm, respectively, which matched well the calculated ECD spectrum of 5*R*, 8*R*, 9*R*, 10*R*, 11*R*, 15*S-***1** (Figure 5a). Thus, compound **1,** as an *ent*-abietane diterpenoid, was established as depicted and named euphonoid H.

The molecular formula of **2** was determined to be C20H24O5 by the HRESIMS ion peak at *<sup>m</sup>*/*<sup>z</sup>* 343.1550 ([M − H]−, calcd 343.1551). The 1H NMR data (Table 1) of **<sup>2</sup>** indicated the presence of three methyls [*δ*<sup>H</sup> 0.98 (3H, Me-18), 0.85 (3H, Me-19), 0.80 (3H, Me-20)] and one aldehyde [*δ*<sup>H</sup> 9.97 (s, H-17)]. The 13C NMR (Table 1) and HSQC data of **2** revealed the presence of three methyls, five methylenes, five methines (including two oxygenated ones at *δ*<sup>C</sup> 64.8 and 55.3 and seven quaternary carbons (including two olefinic ones at *δ*<sup>C</sup> 166.2 and 127.5). The above NMR characteristic features of **2** resembled those of jolkinolide B [14], the major differences being the replacement of the 17-CH3 group in jolkinolide B by an aldehyde group (*δ*<sup>C</sup> 185.0) in **2**. HMBC correlations from H-17 (*δ*<sup>H</sup> 9.97, s) to C-13 (*δ*<sup>C</sup> 166.2) and C-15 (*δ*<sup>C</sup> 127.5) further confirmed the above deduction. The NOESY correlation H-5/H-9 suggested that H-5 and H-9 were *β*-oriented, whereas the NOESY correlations H3-20/H-11 and H3-20/H-14 indicated that H-11, H-14 and CH3-20 were *α*-oriented. Subsequently, quantum chemical calculation of NMR chemical shifts was run on the proposed structure of **2**. As indicated by R2 ( 13C: 0.9979), CMAD (13C: 1.88 ppm) and CLAD (13C: 4.79 ppm) good consistency was observed between the theoretically predicted and experimental chemical shifts, which validated the proposed structure for **2** (Figure 6). Subsequently, ECD calculation (Figure 5b) of the two enantiomers of **2** enabled the establishment of the absolute configuration of **2** to be 5*S*, 8*S*, 9*R*, 10*R*, 11*R*, 12*R*, 14*R*. The structure of **2** was therefore established as depicted and named euphonoid I.

**Figure 5.** (**a**) Experimental and calculated ECD spectra of **1**; (**b**) Experimental and calculated ECD spectra of **2**.

**Figure 6.** Linear correlation plots of predicted versus experimental 13C NMR chemical shifts.

The other two known diterpenoids (**3**–**4**) were identified to be *ent*-abietane diterpenoids raserranes A (**3**) and B (**4**) by comparison of their NMR data with those reported in the literature, these four diterpenoids were discovered for the first time from this species [15].

#### *2.2. Biological Activity of Isolated Compounds*

The anticancer effects of the isolates **1**–**4** were evaluated against human breast cancer cells MDA-MB-231, human colon cancer cells HCT-15 and RKO and human prostate cancer cells C4-2B and C4-2B/ENZR (enzalutamide-resistant C4-2B cells). The IC50 values (Table 2) indicated that the two new compounds exhibited varying degrees of growth inhibition against the five cancer cell lines. Compound **1** showed significant inhibitory activities against C4-2B and C4-2B/ENZR cell lines with IC50 values of 5.52 ± 0.65 μM and 4.16 ± 0.42 μM, respectively. Compound **2** exhibited marked inhibitory activity towards the five human cancer cell lines (IC50 values ranging from 4.49 ± 0.78 to 12.45 ± 3.24 μM) and was particularly active against C4-2B and C4-2B/ENZR cell lines (IC50 values: 4.49 ± 0.78 and 5.74 ± 0.45, respectively).


**Table 2.** IC50 data of compounds **1**–**4** for the indicated cell lines.

<sup>1</sup> Each IC50 value was the mean ± standard deviation from three experiments; <sup>2</sup> Positive control: doxorubicin.

Macrocyclic and polycyclic diterpenes were usually encountered in the genus of *Euphorbia* and macrocyclic diterpenes were characteristic components of *Euphorbia* plants, while polycyclic diterpenes were nonspecific in this genus. Although polycyclic diterpenes were not the characteristic components of *Euphorbia* plants, some polycyclic diterpenes showed great potential in the development of anticancer drugs [16–18]. Jolkinolide B, a typical *ent*-abietane diterpene first isolated from *Euphorbia jolkini*, induced apoptosis and sensitized bladder cancer to mTOR inhibitors [19,20]; 17-hydroxy-jolkinolide B, a potent inhibitor of JAK/STAT3 signaling, is a promising anticancer drug candidate [21]. In this study, compounds **1**–**2** sharing the same abietane diterpene skeleton (6/6/6 carbon ring system) were shown to be promising anti-prostate cancer candidates. Among the four compounds isolated, compound **2** that possessed an α,β-unsaturated γ-lactone ring at C-12 and C-13, was very active against almost the test cancer cells. This observation was consistent with our previous discovery that such an α,β-unsaturated γ-lactone ring was beneficial for the anticancer activity of this type of diterpenoids [12]. Despite the fact that several antitumor abietane diterpenoids were reported in recent years, the pharmacophores and structure-activity relationship of abietane diterpenoids as anticancer agents were rarely investigated. Thus, synthesis of these diterpenoids and study of their structure–activity relationship and potential molecular mechanisms were of great significance for the design and development of anticancer agents.
