Biological and chemical guided isolation of 3,4-secograyanane diterpenoids from the roots of Pieris formosa

Abstract

Seventeen new 3,4-secograyanoids (1–17), together with seven known compounds (18–24), were isolated from the roots of Pieris formosa. Their structures with absolute configurations were characterized by a series of spectroscopic methods and X-ray diffraction. Compounds 1, 2, 4–8, 10–13, and 16–24 exhibited significant analgesic activity at 5.0 mg kg⁻¹ (ip) compared to vehicle-injected mice (p < 0.05). In particular, compounds 16 and 17 showed highly potent activities with inhibition rates of 63.5% and 69.9%, respectively.

---

Introduction

Grayanoids, a special type of diterpenoid, exist exclusively in the Ericaceae family with diverse structures and wide bioactivities.¹ To date, almost two hundred grayanoids with eleven carbon skeletons, including grayanane,^2,3 1,5-secograyanane,⁴ 3,4-secograyanane,⁵ 9,10-secograyanane,⁶ 1,10:2,3-disecograyanane,⁷ kalmane,⁸ 1,5-secokalmane,⁹ leucothane,¹⁰ micranthane,¹¹ mollane,¹² and rhodomollane,¹³ have been reported. Notably, 3,4-secograyanane is a rare type of grayanoid, and only seventeen 3,4-secograyanoids have been isolated and characterized. Several grayanoids display analgesic and insect antifeedant activities.

Pieris formosa (Wall) D. Don, a traditional Chinese medicine, has been used to treat tinea and scabies.¹⁴ In previous phytochemical investigations of P. formosa, fifteen grayananes and three diterpenes with an unusual carbon skeleton were isolated and characterized, and most compounds were found to exhibited an analgesic effect.^15,16 In our on-going efforts, biological and chemical guided fractionation led to the isolation of seventeen new 3,4-secograyananes (1–17), together with seven known 3,4-secograyanoids (18–24) (see Fig. 1). Herein, we report their isolation, structural elucidation, and bioactivity. The main text of the article should appear here with headings as appropriate.

Fig. 1 Chemical structures of the isolated compounds 1–24.

Results and discussion

Biological and chemical guided fractionation and isolation of the roots of P. formosa

The fractionation and isolation procedures were guided by acetic acid-induced writhing test and online HPLC/UV/ESIMS² analyses. The roots of P. formosa were extracted with 95% EtOH and the evaporation of the solvent under reduced pressure subsequently yielded a black residue. The residue was mixed with diatomite (1 : 2) and Soxhlet was successively extracted thereafter using a gradient of petroleum ether, CH₂Cl₂, EtOAc and CH₃OH. Acetic acid-induced writhing test showed analgesic activity for the CH₂Cl₂ portion (D) and the EtOAc portion (E), with 20% and 25% writhe inhibitions (100 mg kg⁻¹, op), respectively. Subsequently, the CH₂Cl₂ portion (D) and the EtOAc portion (E) were applied to a macroporous resin D101 column and eluted with a gradient of 30% EtOH–H₂O, 50% EtOH–H₂O and 70% EtOH–H₂O, to yield three fractions, respectively (D1–D3, and E1–E3). Fractions D1, D2, E1, and E2 showed analgesic activity with 20%, 33%, 73%, and 41% writhe inhibition (100 mg kg⁻¹, op), respectively. Fractions D1 and D2 were separated through a polyamide column and eluted with H₂O–EtOH (100 : 0, 30 : 70, 50 : 50, 70 : 30, v/v), to yield eight fractions (D1a–D1d and D2a–D2d). Fractions E1 and E2 were separated through a polyamide column and eluted with a gradient of H₂O–EtOH (100 : 0, 30 : 70, 60 : 40, v/v), to yield six fractions (E1a–E1c and E2a–E2c). Online HPLC/UV/ESIMS² analyses showed that there were many constituents that exhibited a maximum absorption band at λ 210–220 nm in the UV spectra and [M − 60]⁺, [M − 2 × 60]⁺, and [M − 3 × 60]⁺ fragment ions in the positive MS² mode in fractions D1a, D2a, and D2b (see Fig. S1–S3†), characteristic of the highly acylated 3,4-secograyanane diterpenoids.⁵ Similarly, there were many constituents that exhibited a maximum absorption band at λ 210–220 nm in the UV spectra and [M − 18]⁺, [M − 2 × 18]⁺, and [M − 3 × 18]⁺ fragment ions in the positive MS² mode in fractions E1a, E1b, and E2a (see Fig. S4–S6†), characteristic of the highly hydroxylated grayanane diterpenoids. Finally, seventeen new 3,4-secograyanoids (1–17), together with seven known 3,4-secograyanoids (18–24), were isolated from the fractions D1a, D2a, D2b, E1a, E1b, and E2a by a combination of different chromatographic methods. Acetic acid-induced writhing test showed that compounds 1, 2, 4–8, 10–13, and 16–24 exhibited significant analgesic activity at 5.0 mg kg⁻¹ (ip).

Characterization of compounds 1–17

Pieristoxinlactone A (1) was obtained as a colorless crystal, and its molecular formula was determined to be C₃₃H₄₄O₁₅ based on the HRESIMS [M + Na]⁺ ion peak m/z 703.2568 (calcd 703.2572), indicating twelve indices of hydrogen deficiency. Its ¹H NMR spectrum (see Table 1) showed one triplet methyl (δ_H 1.35, t, J = 7.0 Hz) and eight singlet methyls (δ_H 2.25, 2.24, 2.19, 2.18, 2.12, 2.00, 1.73, 1.70). In the low-field region, a pair of AB doublets (δ_H 6.41, 5.64, d, J = 10.0 Hz) and an exchangeable hydroxy proton (δ_H 7.26) were also observed. The ¹³C NMR (see Table 2) and DEPT spectra displayed 33 carbon resonances, ascribed to nine methyls, four methylenes (one olefinic at δ_C 113.9), eight methines (five oxygenated at δ_C 92.2, 75.6, 74.0, 68.6, 67.5), and twelve quaternary carbons (seven ester carbonyls at δ_C 174.8, 174.1, 171.9, 170.7, 170.3, 169.1, 168.9; one olefinic at δ_C 147.9; three oxygenated at δ_C 88.3, 88.2, 76.3). The ¹H–¹H COSY (see Fig. 3) and HSQC spectra revealed the presence of four spin systems: C(1)H–C(2)H₂, C(6)H–C(7)H, C(9)H–C(11)H–C(12)H₂–C(13)H–C(14)H, and a propionyl unit. These structural features suggested that 1 was a highly acylated 3,4-secograyanane. Further HSQC and HMBC experiments (see Fig. 3) enabled the full assignments of the ¹H and ¹³C NMR spectra of 1. Compound 1 showed the same planar structure as secorhodomollolide D⁵. NOESY correlations (see Fig. 3) from H-14 to H-1 and H-12a and from H-6 to H-1, H-14, and H₃-19 revealed that these protons were α-oriented. In contrast, the NOESY correlations from H-7 to H-9 and H₃-20, from H-15 to H-9 and H-12b, from H₃-17 to H-12b, suggested that these protons were β-oriented. Furthermore, the small coupling constant between H-9 and H-11 (³J_H9,H11 = 5.0 Hz) in the ¹H NMR spectrum, with the aid of NOESY correlations from H-11 to H-9 and H-12b, indicated that H-11 was β-orientated. To further corroborate the structural assignment and establish the absolute configuration, a single X-ray diffraction [Cu Kα radiation; Flack parameter: 0.00(8)] of 1 (see Fig. 2; CCDC: 1548366†) was successfully carried out, allowing an explicit assignment of the absolute configuration of 1 as 1S, 5S, 6R, 7R, 8S, 9R, 10R, 11R, 13R, 14R, 15R, 16S. Thus, the structure of 1 was defined, and it was named pieristoxinlactone A.

Table 1 ¹H NMR data of compounds 1–5 in pyridine-d₅

No.	1^a	2^a	3^b	4^a	5^b
a Recorded at 500 MHz.b Recorded at 600 MHz.
1	3.74, dd (13.5, 8.0)	4.45, dd (13.5, 7.0)	3.65, dd (10.2, 6.5)	4.49, dd (13.5, 7.0)	3.52, dd (13.2, 7.2)
2	a 3.12, dd (17.5, 13.5), b 2.51, dd (17.5, 8.0)	a 3.13, dd (17.5, 13.5), b 2.36, dd (17.5, 7.0)	a 3.12, dd (18.6, 10.2), b 2.50, overlap	a 3.16, dd (17.0, 13.5), b 2.41, dd (17.0, 7.0)	a 3.12, dd (17.4, 13.2), b 2.47, dd (17.4, 7.2)
6	5.64, d (10.0)	6.03, d (9.0)	5.74, overlap	4.88, d (9.5)	4.78, d (9.0)
7	6.41, d (10.0)	6.84, d (9.0)	6.63, d (10.8)	5.52, d (9.5)	5.45, d (9.0)
9	3.58, d (5.0)	2.48, overlap	3.21, brs	1.98, d (2.5)	2.40, brs
11	5.70, t (5.0)	4.71, dd (7.0, 5.0)	5.76, overlap	4.70, d (5.5, 2.5)	5.71, dd (4.8, 3.6)
12	a 2.46, brd (11.0), b 2.15, m	a 2.40, overlap, b 2.11, dd (16.0, 5.5)	a 2.45, overlap, b 2.18, overlap	a 2.15, dd (15.5, 9.5), b 2.03, dd (15.5, 5.5)	a 2.22, dd (16.8, 9.6), b 2.11, dd (16.8, 5.4)
13	3.62, d (10.0)	3.45, d (9.5)	3.48, overlap	2.59, d (9.5)	2.52, d (9.6)
14	7.04, brs	7.40, brs	6.92, brs	6.25, brs	5.57, brs
15	5.36, s	5.66, s	5.45, s	4.40, s	4.39, s
17	1.73, s	1.94, s	1.57, s	1.56, s	1.53, s
18	a 5.68, brs, b 5.12, brs	a 5.64, brs, b 5.05, brs	a 5.68, brs, b 5.14, brs	a 5.56, brs, b 5.04, brs	a 5.58, brs, b 5.12, brs
19	2.00, s	1.83, s	1.96, s	2.07, s	2.08, s
20	2.12, s	2.17, s	2.10, s	1.92, s	1.88, s
6-OAc	1.70, s	1.75, s	1.75, s
7-OAc	2.25, s	2.15, s	2.20, s
11-OAc	2.19, s		2.15, s		2.04, s
14-OPr	a 2.78, m, b 2.41, overlap, 1.35, t (7.0)	a 2.60, m, b 2.47, overlap, 1.30, t (7.5)	a 2.67, m, b 2.57, m, 1.27, t (7.0)
15-OAc	2.24, s	2.23, s	2.24, s
16-OAc	2.18, s	2.06, s
5-OH	7.26, s	7.18, s			6.30, s
11-OH		7.61, d (3.5)

---

Table 2 ¹³C NMR data of compounds 1–15 in pyridine-d₅

No.	1^a	2^a	3^b	4^a	5^b	6^b	7^a	8^b	9^b	10^b	11^a	12^b	13^b	14^b	15^b
a Recorded at 125 MHz.b Recorded at 150 MHz.
1	42.6	42.7	43.3	42.4	43.7	42.9	41.9	54.7	54.7	51.5	52.6	52.5	52.7	49.6	49.6
2	30.5	31.0	30.7	31.5	31.2	32.3	31.3	32.3	32.3	33.6	32.7	32.7	32.7	33.7	33.8
3	174.1	174.8	174.0	175.4	174.5	174.9	174.5	174.6	174.7	175.0	174.4	174.3	174.3	175.4	175.4
4	147.9	148.5	147.6	151.4	150.3	150.2	60.2	147.9	147.8	89.7	62.0	62.0	62.0	76.9	76.9
5	76.3	76.6	76.4	77.0	76.7	76.0	73.0	90.8	90.8	92.4	87.8	87.7	87.6	92.3	92.3
6	74.0	74.1	73.9	73.4	73.8	75.2	74.0	70.0	70.0	68.9	69.8	69.8	69.9	70.1	69.9
7	67.5	66.7	67.4	72.5	72.4	74.4	67.9	68.5	68.8	68.5	68.4	68.6	69.1	68.7	68.7
8	53.3	53.1	53.4	54.4	53.9	53.8	52.9	56.1	56.1	56.0	55.7	55.7	56.1	55.9	55.9
9	60.8	57.6	58.8	57.0	57.0	53.6	56.9	47.6	47.5	47.4	47.3	47.2	46.8	48.1	48.1
10	88.3	89.2	88.1	89.8	88.6	90.2	89.3	76.9	76.9	76.8	76.4	76.4	76.5	76.4	76.5
11	68.6	62.8	68.4	63.2	68.1	19.0	18.8	20.5	20.5	20.6	20.6	20.6	21.0	20.3	20.3
12	30.6	35.1	31.4	35.8	31.6	23.7	22.6	25.4	25.4	25.3	25.5	25.4	25.5	25.1	25.1
13	41.5	44.9	47.8	50.9	49.7	50.4	42.5	50.4	50.6	50.4	45.6	45.7	51.3	50.2	49.6
14	75.6	78.1	77.8	79.6	79.3	80.4	76.9	80.1	80.1	80.3	79.1	79.1	80.1	80.0	80.2
15	92.2	88.9	90.6	91.2	92.0	90.6	91.1	86.9	86.7	86.7	86.5	86.3	86.2	87.3	87.3
16	88.2	89.9	81.1	82.3	81.8	81.5	88.4	80.0	80.2	80.0	88.2	88.2	80.2	80.1	80.0
17	18.9	22.7	23.8	26.3	25.5	24.1	19.1	21.9	21.7	21.8	19.0	19.0	21.9	21.6	21.7
18	113.9	113.0	113.9	111.7	112.3	112.3	49.5	112.1	112.0	22.8	52.7	52.7	52.7	23.3	23.3
19	19.2	18.8	19.1	19.7	20.0	19.9	18.3	19.2	19.1	20.8	17.6	17.5	17.5	68.4	68.4
20	30.2	27.4	28.8	26.9	27.8	27.5	29.6	33.3	33.4	34.0	33.1	33.0	33.1	33.7	33.7
4-OAc										169.0, 22.5
6-OAc/OPr	169.1, 19.9	169.1, 20.0	169.2, 19.9				169.6, 20.4	169.2, 20.6	169.2, 20.5	169.2, 21.1	169.2, 20.6	169.2, 20.5	169.3, 20.9	168.8, 21.3	172.2, 27.9, 9.0
7-OAc/OPr	170.7, 21.5	170.1, 21.6	170.2, 21.5				170.5, 21.6	172.6, 28.1, 9.1	169.5, 21.4	169.4, 21.6	173.0, 28.4, 9.4	169.8, 21.5	169.6, 21.5	169.5, 21.6	169.4, 21.6
11-OAc	168.9, 21.3		169.0, 21.3		169.2, 21.6
14-OAc/OPr	174.8, 28.3, 9.2	173.6, 28.5, 9.5	173.2, 28.4, 9.1				174.8, 28.4, 9.3	170.0, 21.7	173.1, 28.5, 9.1	173.1, 28.6, 9.1	170.7, 22.0	170.6, 21.9	173.1, 28.4, 9.1	173.2, 28.6, 9.1	170.0, 22.1
15-OAc	171.9, 20.8	171.0, 21.0	171.5, 20.9				171.7, 20.8	171.9, 21.0	171.8, 20.9	171.7, 20.9	171.9, 21.0	171.8, 20.9	171.4, 20.5	171.9, 20.9	171.9, 20.9
16-OAc	170.3, 22.9	169.8, 22.5					170.3, 22.9				170.1, 22.8	170.0, 22.7
19-OAc														170.9, 20.8	170.9, 20.8

---

Fig. 2 X-ray crystal structure of 1.

Pieristoxinlactone B (2) was assigned the molecular formula C₃₁H₄₂O₁₄ based on the HRESIMS ion at m/z 661.2477 [M + Na]⁺, indicative of eleven degrees of unsaturation. The ¹H and ¹³C NMR data (see Tables 1 and 2) of 2 closely resembled those of 1, with the only difference between the two compounds was a hydroxy group replacing the acetoxyl group at C-11 of 2. The observation of upfield shift of H-11 (Δδ_H 1.69) and C-11 (Δδ_C 5.8) of 2 compared to those of 1 and the key HMBC correlation from 11-OH to C-11 confirmed the above deduction. The NOESY correlations from H-14 to H-1, H-6, and H-12a, and from H-1 to H₃-19 and 11-OH, suggested these protons are α-oriented. The NOESY correlations from H-9 to H-7, H-11, H-15, and H₃-17, and from H₃-20 to H-7 and 5-OH indicated that these protons are β-oriented. Thus, the structure of compound 2 was determined and it was given the trivial name pieristoxinlactone B.

Pieristoxinlactone C (3) exhibited the molecular formula of C₃₁H₄₂O₁₄, similar to 2, as deduced from the HRESIMS ion (m/z 661.2467 [M + Na]⁺). Analysis of the NMR data (see Tables 1 and 2) clearly indicated that 3 is closely related to 1. The major differences are that the signals due to an acetyl group was absent, and the resonances for C-13, C-15, C-16, and C-17 were shifted by Δδ_C + 6.3, −1.6, −7.1, and +4.9 ppm, respectively, compared to those of 1. This finding indicated that the acyloxy group at C-16 in 1 was replaced by a hydroxy group in 3. The NOESY correlations from H-14 to H-1, H-6, from H-6 to H₃-19 and 11-OH, suggested that these protons are α-orientated. The NOESY correlations from H-9 to H-15 and H₃-17, and from H₃-20 to H-7 indicated that these protons were β-orientated. Furthermore, NOESY correlations from H-9 to H-11, together with the small coupling constant of H-9 (brs), indicated that H-11 was also β-orientated. Thus, the structure of 3 was established as assigned, and it was named pieristoxinlactone C.

Pieristoxinlactone D (4) had the molecular formula C₂₀H₃₀O₉ based on its HRESIMS ion at m/z 415.1984 [M + H]⁺, indicative of six degrees of unsaturation. The ¹H NMR (see Table 1), together with the HSQC showed three tertiary methyls (δ_H 2.07, 1.92, 1.56), a terminal double bond (δ_H 5.56, 5.04, brs) and a pair of AB doublets (δ_H 5.52, 4.88, d, J = 9.5 Hz) in the low-field region. The ¹³C NMR (see Table 2) and DEPT spectra displayed 20 carbon resonances (one olefinic quaternary carbon resonance should appear at δ_C 151.4) ascribed to three methyls, three methylenes (one olefinic at δ_C 111.7), eight methines (five oxygenated at δ_C 91.2, 79.6, 73.4, 72.5, 63.2), and six quaternary carbons (one ester carbonyl at δ_C 175.4; one olefinic at δ_C 151.4; three oxygenated at δ_C 89.8, 82.3, 77.0). The ¹H–¹H COSY (see Fig. 3) and HSQC spectra established the presence of three spin systems: C(1)H–C(2)H₂, C(6)H–C(7)H and C(9)H–C(11)H–C(12)H₂–C(13)H–C(14)H. Analysis of the ¹H and ¹³C NMR data of 4 showed that its structure was closely related to 1 except for the absent of six acyl groups. A more detailed analysis of the 2D NMR data (see Fig. 3) revealed that 4 is based on the same carbon skeleton as 1. The NOESY correlations (see Fig. 3) from H-14 to H-1, H-6, and H-12a and from H-6 to H-19 established these protons were spatially and were α-orientated. On the other hand, the NOESY correlations from H-7 to H-9 and H₃-20, from H-9 to H-15 and H₃-17, and from H-12b to H-11 and H₃-17 suggested the β-orientation of H-7, H-9, H-11, H-12b, H-15, and H₃-17. The small coupling constant between H-11 and H-12b (³J_H11,H12b = 5.5 Hz) in the ¹H NMR spectrum also proved that H-11 was β-orientated. Therefore, the structure of 4 was assigned as shown, and it was named pieristoxinlactone D.

Fig. 3 Key ¹H–¹H COSY, HMBC and NOESY correlations of compounds 1, 7, 8, 11, and 14.

Pieristoxinlactone E (5) possessed a molecular formula of C₂₂H₃₂O₁₀ as assigned by the HRESIMS data (m/z 457.2075 [M + H]⁺). Comparison of the NMR data of 5 (see Tables 1 and 2) with those of 4 revealed that they are structural analogues. The only difference between the two compounds was the presence of an acetoxyl group at C-11 of 5 instead of a hydroxy group in 4, which was further confirmed by the HMBC correlation from H-11 to the acetate carbonyl carbon (δ_C 169.2). Therefore, the structure of 5 was assigned as shown, and it was named pieristoxinlactone E.

The molecular formula of 6 was assigned as C₂₀H₃₀O₈, by the HRESIMS ion at m/z 421.1831 [M + Na]⁺ and ¹³C NMR data. Analysis of the NMR data (see Tables 2 and 3) indicated that the structure of 6 closely resembled that of 4 with the only difference was the absence of the hydroxy group at C-11 in 6, which was confirmed by the spin system C(9)H–C(11)H₂–C(12)H₂–C(13)H–C(14)H verified by ¹H–¹H COSY and HSQC spectra. Further 2D NMR experiments enabled the full assignments of the ¹H and ¹³C NMR data of 6. Therefore, 6 was assigned as shown, and was named pieristoxinlactone F.

Table 3 ¹H NMR data of compounds 6–10 in pyridine-d₅

No.	6^b	7^a	8^b	9^b	10^b
a Recorded at 500 MHz.b Recorded at 600 MHz.
1	3.20, dd (13.2, 6.6)	3.43, dd (13.0, 7.0)	3.09, t (10.2)	3.09, t (10.2)	3.62, dd (12.0, 7.8)
2	a 3.13, dd (16.2, 13.2), b 2.55, dd (16.2, 6.6)	a 3.55, overlap, b 3.11, dd (17.5, 7.0)	a 2.94, overlap, b 2.91, overlap	a 2.94, overlap, b 2.91, overlap	a 3.11, dd (19.2, 12.0), b 2.94, dd (19.2, 7.8)
6	4.65, d (8.4)	5.60, d (10.0)	6.59, d (9.0)	6.60, d (9.0)	6.78, d (10.2)
7	5.27, d (8.4)	5.35, d (10.0)	5.78, d (9.0)	5.77, d (9.0)	5.77, d (10.2)
9	2.37, t (6.0)	3.23, dd (11.5, 7.0)	2.71, brd (7.2)	2.66, brd (7.8)	2.66, brd (7.8)
11	a 1.63, overlap, b 1.61, overlap	a 1.78, m, b 1.71, overlap	a 2.12, m, b 1.78, m	a 2.12, m, b 1.77, m	a 2.13, overlap, b 1.81, overlap
12	a 1.93, dd (15.0, 7.8), b 1.57, overlap	a 2.06, overlap, b 1.57, m	a 2.15, m, b 1.67, m	a 2.15, m, b 1.67, m	a 2.16, overlap, b 1.65, overlap
13	2.41, brd (7.8)	3.53, brd (7.0)	2.49, brs	2.47, brs	2.49, overlap
14	4.93, brs	6.15, brs	6.94, brs	6.99, brs	6.86, brs
15	4.26, s	5.35, s	5.44, s	5.46, s	5.48, s
17	1.53, s	1.73, s	1.45, s	1.46, s	1.46, s
18	a 5.54, s, b 5.08, s	a 3.29, d (5.5), b 2.48, d (5.5)	a 5.42, s, b 4.89, s	a 5.41, s, b 4.88, s	1.91, s
19	1.99, s	1.61, s	1.99, s	1.99, s	2.01, s
20	1.84, s	2.05, s	1.47, s	1.47, s	1.57, s
4-OAc					1.90, s
6-OAc		2.01, s	2.02, s	2.03, s	2.15, s
7-OAc/OPr		2.29, s	a 2.58, m, b 2.47, m, 1.23, t (6.0)	2.22, s	2.27, s
14-OAc/OPr		a 2.79, m, b 2.47, overlap, 1.36, t (7.5)	2.23, s	a 2.58, m, b 2.51, m, 1.26, t (6.0)	a 2.55, m, b 2.50, overlap, 1.27, t (7.8)
15-OAc		2.24, s	2.10, s	2.13, s	2.14, s
16-OAc		2.17, s
5-OH		5.65, s
10-OH			7.29, s	7.30, s	7.32, s
16-OH			3.64, s	3.94, s

---

The HRESIMS ion at m/z 661.2478 [M + Na]⁺ and the ¹³C NMR data of 7 suggested a molecular formula of C₃₁H₄₂O₁₄. The ¹H and ¹³C NMR data of 7 (see Tables 2 and 3) closely resembled those of pierisoid A¹⁷ except for the location of the propionyloxyl. Key HMBC correlations (see Fig. 3) from H-7 to one acetyl carbonyl (δ_C 170.5) and from H-14 to the propionyl carbonyl (δ_C 174.8) suggested that one acetyloxyl and the propionyloxyl were connected to C-7 and C-14, respectively. Further HSQC and HMBC experiments (see Fig. 3) enabled the full assignments of the ¹H and ¹³C NMR spectra of 7 to be carried out. The NOESY correlations (see Fig. 3) from H-14 to H-1, H-6, and H-12a established that these protons were located on the same face of the ring, while the NOESY correlations (see Fig. 3) from H-7 to H-9 and H₃-20, from H-9 to H-15 and H₃-17, and from 5-OH to H-9 and H₃-20 suggested these protons were on the other face of the ring. However, the chiral quaternary C-4 was in the side chain and its relative configuration couldn't be assigned. Subsequently, an X-ray single-crystal structure of 7 (see Fig. 4; CCDC: 1548369†), obtained using Cu Kα radiation [Flack parameter: 0.04(7)], not only permitted the unambiguous assignment of the propionyloxyl but also defined its absolute configuration as 1S, 4R, 5S, 6R, 7R, 8S, 9R, 10R, 13R, 14R, 15R, 16S. Thus, the structure of 7, named pieristoxinlactone G, was finally determined.

Fig. 4 X-ray crystal structure of 7.

The molecular formula C₂₉H₄₀O₁₂ of pieristoxinlactone H (8) was assigned according to the HRESIMS ion at m/z 603.2398 [M + Na]⁺. The ¹H and ¹³C NMR data of 8 (see Tables 2 and 3) revealed a close resemblance to those of secorhodomollolide A,⁵ and the only difference between the two compounds involved a hydroxy group at C-16 of 8 replacing an acetoxy group of secorhodomollolide A. Compared to secorhodomollolide A, the ¹H NMR resonances for H-13 and H₃-17 of 8 were shifted by Δδ_H −1.15 and −0.25, while the ¹³C NMR resonances for C-13, C-16, and C-17 of 8 were shifted by Δδ_C +5.5, −8.1, +3.0 ppm, respectively, suggesting the presence of a hydroxy group at C-16 in 8. The NOESY correlations (see Fig. 3) from H-9 to H-7 and H-15, from H-15 to H-11b and H₃-17, from H₃-17 to H-12b suggested the β-orientation of these protons. The large coupling constant between H-6 and H-7 (³J_H6,H7 = 9.0 Hz) in the ¹H NMR spectrum indicated that H-6 and H-7 were in the trans configuration. Therefore, H-6 was α-orientated. Moreover, the NOESY correlations from H-14 to H-12a, from H₃-19 to H-1 and H-6, and from H-1 to 10-OH established these protons to be α-orientated. Thus, the structure of pieristoxinlactone H (8) was assigned as drawn.

Compound 9 had the same molecular formula C₂₉H₄₀O₁₂ as 8, as assigned by the HRESIMS (m/z 603.2417 [M + Na]⁺) and ¹³C NMR data. Its NMR data were highly similar to those of 8 except for the slight variations in the chemical shifts of H-14 (Δδ_H +0.05), C-7 (Δδ_C +0.3), and the propionyl carbonyl (Δδ_C +0.5), suggesting that the propionyloxyl group may connect to C-14 in 9. The key HMBC correlations from H-7 to one acetyl carbonyl (δ_C 169.5) and from H-14 to the propionyl carbonyl (δ_C 173.1) confirmed the above assignment. The relative configuration of 9 was assigned as the same as 8 by their similar NMR and NOESY data. Therefore, the structure of 9 was assigned, and it was named pieristoxinlactone I.

The molecular formula C₃₁H₄₄O₁₄ of 10 was identical to 24, as determined by ¹³C NMR data and the HRESIMS ion at m/z 663.2618 [M + Na]⁺. The general features of the NMR data of 10 (see Tables 2 and 3) resembled those of 24, except for the substituents at C-4 and C-16. The observation of the upfield shift of H-13 (Δδ_H 1.11) and H₃-17 (Δδ_H −1.11), and downfield shift of H₃-18 (Δδ_H 0.24) and H₃-19 (Δδ_H 0.36) in the ¹H NMR spectrum of 10, as well as the downfield shift of C-4 (Δδ_C 13.8), C-13 (Δδ_C 5.3), and C-17 (Δδ_C 2.9) and an upfield shift of C-4 (Δδ_C 8.2) in the ¹³C NMR spectrum of 10 compared with those of 24 suggested that an acetyloxyl group and a hydroxy group were connected to C-4 and C-16, respectively. The HMBC cross-peaks from H-1 and H₃-19 to C-4 (δ_C 89.7) and from H-14, H-13 and H₃-17 to C-16 (δ_C 80.0) reconfirmed the above assignment. Compound 10, pieristoxinlactone J, was thereby structurally characterized.

The molecular formula C₃₁H₄₂O₁₄ was assigned to 11 with eleven indices of hydrogen deficiency by the HRESIMS ion at m/z 661.2469 [M + Na]⁺ and ¹³C NMR data. The NMR data of 11 (see Tables 2 and 4) showed many similarities to those of 23, revealing that these two compounds are homologues, except for the presence of a 4,18-oxirane group in 11. Compared to 23, the C-4 carbon resonance was shielded (Δδ_C 13.9) and the C-18 carbon resonance was deshielded (Δδ_C 26.9), suggesting that an oxirane group was present between C-4 and C-18. Moreover, the key HMBC correlations (see Fig. 3) from H₃-19 to C-4 (δ_C 62.0) and C-18 (δ_C 52.7) further confirmed this deduction. The NOESY correlations (see Fig. 3) from H-14 to H-6 and H-12a and from H₃-19 to H-1 and H-6 suggested that these protons are spatially close and α-orientated. In contrast, the NOESY correlations from H-9 to H-7, H-15, H₃-17, and H₃-20 demonstrated the β-orientation of H-7, H-9, H-15, H₃-17, and H₃-20. However, the relative configuration of C-4 still remains uncertain because C-4 was on the side chain. Finally, an X-ray diffraction study using the anomalous dispersion radiation of Cu Kα [Flack parameter: −0.07(15)] of 11 (see Fig. 5; CCDC: 1548367†) defined its absolute configuration as 1S, 4R, 5S, 6R, 7R, 8S, 9R, 10R, 13R, 14R, 15R, 16S. Thus, the structure of 11 was determined, and it was named pieristoxinlactone K.

Table 4 ¹H NMR data of compounds 11–15 in pyridine-d₅

No.	11^a	12^b	13^b	14^b	15^b
a Recorded at 500 MHz.b Recorded at 600 MHz.
1	3.42, brd (11.5)	3.41, dd (11.4, 7.8)	3.41, dd (11.4, 8.4)	4.09, t (10.2)	4.09, t (10.2)
2	a 3.00, dd (19.0, 11.5), b 2.81, dd (19.0, 8.0)	a 2.99, dd (18.6, 11.4), b 2.79, dd (18.6, 7.8)	a 3.01, dd (18.6, 11.4), b 2.80, dd (18.6, 8.4)	a 3.03, m, b 2.98, m	a 3.03, m, b 2.98, m
6	6.55, d (9.5)	6.55, d (9.6)	6.60, d (9.0)	6.62, d (8.4)	6.68, d (9.0)
7	5.80, d (9.5)	5.76, d (9.6)	5.84, d (9.0)	5.73, d (8.4)	5.73, d (9.0)
9	2.64, brd (8.0)	2.61, d (7.8)	2.52, overlap	2.74, brd (8.4)	2.74, brd (7.8)
11	a 2.09, overlap, b 1.78, m	a 2.11, m, b 1.75, m	a 2.08, m, b 1.79, m	a 2.12, overlap, b 1.79, m	a 2.11, overlap, b 1.79, m
12	a 2.18, overlap, b 1.66, overlap	a 2.17, m, b 1.65, m	a 2.14, m, b 1.65, m	a 2.18, overlap, b 1.67, m	a 2.17, overlap, b 1.67, m
13	3.61, t (4.0)	3.58, t (3.6)	2.46, t (3.0)	2.44, m	2.48, overlap
14	7.04, brs	7.03, brs	7.14, brs	6.78, brs	6.75, brs
15	5.53, s	5.52, s	5.51, s	5.45, s	5.45, s
17	1.71, s	1.71, s	1.47, s	1.46, s	1.47, s
18	a 3.35, d (5.5), b 2.59, d (5.5)	a 3.34, d (5.4), b 2.59, d (5.4)	a 3.35, d (4.8), b 2.60, d (4.8)	1.74, s	1.75, s
19	1.62, s	1.61, s	1.61, s	a 4.76, d (11.4), b 4.70, d (11.4)	a 4.78, d (11.4), b 4.69, d (11.4)
20	1.49, s	1.49, s	1.50, s	1.47, s	1.47, s
6-OAc	2.14, s	2.14, s	2.13, s	2.17, s	a 2.52, m, b 2.45, overlap, 1.16, t (7.2)
7-OAc/-OPr	a 2.68, m, b 2.56, overlap, 1.34, t (7.5)	2.29, s	2.25, s	2.26, s	2.24, s
14-OAc/-OPr	2.24, s	2.22, s	a 2.65, m, b 2.53, overlap, 1.28, t (7.2)	a 2.60, m, b 2.50, m, 1.26, t (7.2)	2.23, s
15-OAc	2.13, s	2.13, s	2.12, s	2.14, s	2.14, s
16-OAc	2.15, s	2.15, s
19-OAc				2.06, s	2.05, s
10-OH	7.55, s	7.51, s	7.50, s	7.33, s	7.32, s

---

Fig. 5 X-ray crystal structures of 11.

The molecular formula of 12, C₃₀H₄₀O₁₄, was determined by the HRESIMS ion at m/z 647.2307 [M + Na]⁺. A careful analysis of the NMR data of 12 revealed 12 to be based on a 3,4-secograyanane skeleton in a similar manner to 11, except for an acetoxy group substituent at C-7 in 12 instead of a propionyloxy group in 11. The key HMBC cross-peak from H-7 to an acetyl carbonyl carbon (δ_C 169.8) confirmed the assignment. The structure of compound 12, pieristoxinlactone L, was thus defined as depicted.

Compound 13 had a molecular formula of C₂₉H₄₀O₁₃, as assigned by HRESIMS (m/z 619.2355 [M + Na]⁺). The ¹H and ¹³C NMR spectra of compound 13 (see Tables 2 and 4) were similar to those of 11, suggesting that these two compounds were homologues. Compound 13 differs from 11 by substitutions at C-7, C-14, and C-16. The HMBC correlations from H-7 to one acetyl carbonyl (δ_C 169.6) and from H-14 to the propionyl carbonyl (δ_C 173.1) suggested that C-7 and C-14 were connected to an acetoxyl and propionyloxyl, respectively. In addition, compared to 11, the resonances of H-13 and H₃-17 were shifted by Δδ_H −1.15 and −0.24, and the resonances of C-13, C-16, and C-17 were shifted by Δδ_C +5.7, −8.0, and +2.9 ppm, respectively, suggesting a hydroxy group substituent at C-16 in 13. Thus, the structure of 13 was eventually determined, and it was named pieristoxinlactone M.

Compound 14 exhibited a molecular formula of C₃₁H₄₄O₁₅ based on the HRESIMS ion at m/z 679.2572 [M + Na]⁺ and with the aid of ¹³C NMR data. The ¹H and ¹³C NMR spectra of 14 (see Tables 2 and 4) were similar to those of 10, differing only in the presence of a hydroxy group and an oxygenated methylene group (δ_C 68.4) and the absence of a methyl group in 14. Compared to 10, the C-4 carbon resonance of 14 was shielded (Δδ_C 12.8), revealing that the hydroxy group substituent was at C-4. Moreover, the HMBC correlations (see Fig. 3) from H₂-19 to one acetyl carbonyl (δ_C 170.9) and from H₃-18 to C-4 (δ_C 76.9) and C-19 (δ_C 68.4) indicated that the hydroxy group and one acetyl group were placed at C-4 and C-19, respectively. The NOESY correlations (see Fig. 3) from H-14 to H-12a and H-6 and from H₃-18 to H-1 and H-6 established the α-orientations of H-1, H-6, H-12a, H-14, and H₃-18, while correlations from H-9 to H-7, H-15, H₃-17, and H₃-20 confirmed the β-orientations of these protons. The relative configuration of C-4 remains uncertain because C-4 is located on the side chain. Finally, the X-ray diffraction [Cu Kα; Flack parameter: 0.04(16)] of 14 (see Fig. 6; CCDC: 1548368†) was successfully performed, which not only confirmed the assigned structure but also established the absolute configuration as 1S, 4S, 5S, 6R, 7R, 8S, 9R, 10R, 13R, 14R, 15R, 16S. Thus, the structure of pieristoxinlactone N (14) was characterized as shown.

Fig. 6 X-ray crystal structures of 14.

The molecular formula of 15 was deduced to be C₃₁H₄₄O₁₅, identical to that of 14, as indicated by the HRESIMS ion at m/z 679.2572 [M + Na]⁺. NMR data analysis showed that its structure was closely related to that of 14, differing only in the location of the propionyloxyl group (at C-6 in 15; at C-14 in 14). The key HMBC correlations from H-6 to the propionyl carbonyl (δ_C 172.2) and from H-14 to one acetyl carbonyl (δ_C 170.0) confirmed that the propionyloxyl group was connected to C-6. Thus, the structure of compound 15 was determined as drawn, and it was named pieristoxinlactone O.

The HRESIMS ion peak at m/z 417.2123 [M + H]⁺ and the ¹³C NMR data suggested a molecular formula of C₂₀H₃₂O₉ for compound 16. Apart from the signals for the five acyl groups, the ¹H and ¹³C NMR data of 16 (see Table 5) and 10 were similar, suggesting that these two compounds were homologues. Further ¹H–¹H COSY, HSQC, and HMBC experiments allowed for the full assignments of the ¹H and ¹³C NMR spectra of 16. Compound 16 shared the same carbon skeleton as 10. The NOESY correlations from H-14 to H-12a, H₃-19 and from H-1 to H-2a, H-6, and H₃-19 demonstrated that the α-orientation of these protons, while the NOESY correlations from H-9 to H-7 and H-15, H₃-20 to H-2b, and from H-15 to H₃-17 confirmed the β-orientation of these protons. Compound 16, named pieristoxinlactone P, was thereby structurally characterized.

Table 5 ¹H and ¹³C NMR data (600 Hz) of compounds 16 and 17 in pyridine-d₅

No.	16		17
	δH (J in Hz)	δC	δH (J in Hz)	δC
1	3.80, dd (13.2, 5.4)	51.5	3.71, dd (11.4, 10.8)	51.1
2	a 3.10, dd (18.0, 13.2), b 2.72, dd (18.0, 5.4)	35.0	a 3.04, dd (18.0, 10.8), b 2.98, dd (18.0, 11.4)	34.7
3		176.6		176.2
4		76.7		75.6
5		95.2		95.1
6	5.60, brs	70.4	5.61, d (7.2)	71.6
7	4.65, brs	73.6	5.08, d (7.2)	70.9
8		57.6		57.6
9	2.02, overlap	47.2	2.41, d (4.2)	54.9
10		77.6		78.2
11	a 1.96, overlap, b 1.56, m	21.0	4.54, t (4.8)	68.1
12	a 1.91, overlap, b 1.60, m	26.0	a 2.39, overlap, b 2.02, dd (15.6, 4.2)	34.7
13	2.42, t (3.0)	51.2	2.58, brd (10.2)	46.8
14	5.73, brs	81.1	5.89, brs	77.8
15	4.20, s	88.2	4.04, s	90.7
16		81.2		82.2
17	1.50, s	22.5	1.61, s	23.3
18	2.05, s	28.1	1.86, s	27.9
19	1.82, s	25.9	1.76, s	26.8
20	1.47, s	34.7	1.67, s	33.6
11-OH	6.53, s

---

Compound 17 possessed the molecular formula of C₂₀H₃₂O₁₀, as assigned by the HRESIMS ion at m/z 433.2065 [M + H]⁺ and ¹³C NMR data. The NMR data of 17 resembled those of 16, except for the only difference of the presence of an oxygenated methine carbon and the absence of a methylene carbon in 17. The spin system C(9)H–C(11)H–C(12)H₂–C(13)H–C(14)H established by ¹H–¹H COSY and HSQC spectra, suggested that a hydroxy group was connected to C-11 (δ_C 68.1). The α-orientation of 11-OH was confirmed by the NOESY correlations from H-11 to H-9 and H₃-20. The structure of pieristoxinlactone Q (17) was thus defined as shown.

The seven known diterpenoids, namely, secorhodomollolide C (18),⁵ pierisoid B (19),¹⁷ secorhodomollolide B (20),⁵ secorhodomollolide A (21),⁵ pierisformotoxin G (22),¹⁸ pierisformotoxin B (23),¹⁹ and pierisformotoxin A (24),¹⁹ were identified by the comparison of the experimental and previously reported spectroscopic data.

The antinociceptive activities of compounds 1, 2, 4–8, 10–13, and 16–24 were evaluated using an acetic acid-induced writhing test. As shown in Fig. 7a, all evaluated compounds exhibited significant analgesic activity at 5.0 mg kg⁻¹ (ip) compared to vehicle-injected mice (p < 0.05). In particular, compounds 16 and 17 showed highly potent activities, with writhe inhibition rates of 63.5% and 69.9%, respectively. To further understand the action of the antinociceptive activities of compounds 16 and 17, three doses (5.0, 1.0, and 0.1 mg kg⁻¹) were employed in an acetic acid-induced writhing test. The results revealed that 16 and 17 were able to reduce acetic acid-induced writhes in a dose-dependent manner (Fig. 7b), with ID₅₀ values of 1.24 and 0.72 mg kg⁻¹, respectively.

Fig. 7 (a) Antinociceptive activities of the purified compounds from the roots of P. formosa in an acetic acid-induced writhing test. (b) Antinociceptive activities of various concentrations of 16 and 17 in an acetic acid-induced writhing test. Ibuprofen (Ib) and morphine (morph) were used as control drugs. Each bar and vertical line represents the mean ± SEM of the values obtained from 8 mice. *p < 0.05, **p < 0.01, ***p < 0.001, significant differences between the test compound or control drug groups and vehicle group (veh) (one-way ANOVA followed by the Bonferroni test).

Experimental

General experimental procedures

Optical rotations were measured on an AUTOPOL V polarimeter. IR spectra were recorded using a Nicolet 5700 FT-IR microscope (FT-IR microscope transmission). 1D and 2D NMR spectra were obtained using an INOVA-500 or a Bruker-600 spectrometer. Online HPLC/UV/ESIMS² analyses were recorded on an Agilent 1100 Series liquid chromatography system chromatograph. HRESIMS data were recorded on an Agilent 6520 Accurate-Mass Q-TOF LC/MS spectrometer. Preparative HPLC was performed using a Shimadzu LC-6AD instrument equipped with RID-20A and SPD-20A detectors, using a YMC-Pack RP-C18 column (250 × 10 mm, 5 μm) or a COSMOSIL packed column (250 × 10 mm, 5 μm). Column chromatography was performed using polyamide (30–60 mesh, Changzhou Changfeng Chemical Factory, China), macroporous resin D101 (Qingdao Marine Chemical Factory, China), MCI (Mitsubishi Chemical Corporation), and ODS (50 μm, Merck, Germany). TLC was carried out with glass precoated Si gel GF254 plates (Qingdao Marine Chemical Factory, China). Spots were visualized by spraying with 10% H₂SO₄ in EtOH–H₂O (95 : 5, v/v) followed by heating.

Plant material

Roots of Pieris formosa (120 kg) were collected in Chuxiong Yi Autonomous Prefecture in Yunnan Province, People's Republic of China, in June 2014 and were identified by Prof. Lin Ma (Department of Natural Products Chemistry, Institute of Materia Medica, Chinese Academy of Medical Sciences). A voucher specimen (ID-s-2747) was deposited in the herbarium at the Department of Medicinal Plants, Institute of Materia Medica, Chinese Academy of Medical Sciences.

Extraction and isolation

The isolation procedures were guided by acetic acid-induced writhing test and online HPLC/UV/ESIMS² analyses as described above. The roots of Pieris formosa (120 kg) were extracted with 95% EtOH (150 l × 3 h × 3). The evaporation of the solvent under reduced pressure yielded a black residue (5.0 kg). The residue was mixed with diatomite (1 : 2) and then successively Soxhlet extracted using a gradient of petroleum ether, CH₂Cl₂, EtOAc and CH₃OH. The CH₂Cl₂ portion (D) and the EtOAc portion (E) were then applied to a macroporous resin D101 column (10 kg) and eluted with a gradient of 30% EtOH–H₂O, 50% EtOH–H₂O and 70% EtOH–H₂O, to yield six fractions (D1–D3, and E1–E3). Fractions D1 (60 g) and D2 (57 g) were separated through a polyamide column and eluted with H₂O–EtOH (100 : 0, 30 : 70, 50 : 50, 70 : 30, v/v), to yield eight fractions (D1a–D1d and D2a–D2d). Fractions E1 (20 g) and E2 (35 g) were separated through a polyamide column and eluted with a gradient of H₂O–EtOH (100 : 0, 30 : 70, 60 : 40, v/v), to yield six fractions (E1a–E1c and E2a–E2c). Fraction D1a (4 g) was separated via an MCI column and eluted with a gradient of H₂O–MeOH (1 : 9–9 : 1, v/v), to obtain four subfractions D1a1–D1a4. Subfraction D1a1 (0.9 g) was purified by semipreparative HPLC with MeCN–H₂O (23 : 77, v/v, 4.5 ml min⁻¹), to yield 13 (4.5 mg, t_R = 37.5 min), 14 (1.0 mg, t_R = 34.5 min) and 15 (0.6 mg, t_R = 32.5 min). Fraction D2a (4 g) was separated through a MCI column and eluted with a gradient of H₂O–MeOH (1 : 9–9 : 1, v/v), to obtain four subfractions D2a1–D2a4. Subfraction D2a1 (1.5 g) was purified further using semipreparative HPLC with MeCN–H₂O (30 : 70, v/v, 4.0 ml min⁻¹), to yield 1 (26.0 mg, t_R = 70.0 min), 18 (11.0 mg, t_R = 72.5 min), 20 (83.0 mg, t_R = 57.5 min), and 21 (180.0 mg, t_R = 55.5 min). Subfraction D2a2 (0.5 g) was purified further using semipreparative HPLC with MeCN–H₂O (22.5 : 77.5, v/v, 4.0 ml min⁻¹), to yield 3 (1.2 mg, t_R = 75.0 min), 8 (2.5 mg, t_R = 79.0 min), 10 (3.5 mg, t_R = 80.0 min), and 19 (14.0 mg, t_R = 91.0 min). Subfraction D2a3 (1.3 g) was purified further using semipreparative HPLC with MeCN–H₂O (24 : 76, v/v, 4.0 ml min⁻¹), to yield 11 (36.0 mg, t_R = 103.0 min), 12 (2.5 mg, t_R = 70.0 min), 23 (16.0 mg, t_R = 61.0 min), and 24 (71.0 mg, t_R = 67.0 min). Fraction D2b (8 g) was separated through an ODS column and eluted with 10% CH₃OH–H₂O, 30% CH₃OH–H₂O, 50% CH₃OH–H₂O, and 70% CH₃OH–H₂O, successively, to yield four fractions (D2b1–D2b4). Subfraction D2b3 (1.5 g) was purified by reversed-phase semipreparative HPLC with MeCN–H₂O (23 : 77, v/v, 4.0 ml min⁻¹), to yield 2 (5.7 mg, t_R = 115.5 min), 7 (8.0 mg, t_R = 151.5 min), 9 (1.0 mg, t_R = 104.5 min), and 22 (7.6 mg, t_R = 114.5 min). Fraction E1a (3 g) was chromatographed on a YMC RP-C18 column (17% CH₃CN–H₂O, v/v, 5 ml min⁻¹) to yield seven fractions (E1a1–E1a7). Subfraction E1a2 (0.6 g) was purified further using semipreparative HPLC with MeCN–H₂O (8 : 92, v/v, 4.5 ml min⁻¹), to yield 4 (6.5 mg, t_R = 21.5 min) and 6 (5.0 mg, t_R = 38.0 min). Fraction E1b (5 g) was separated through an ODS column and eluted with 10% CH₃OH–H₂O, 30% CH₃OH–H₂O, 50% CH₃OH–H₂O, and 70% CH₃OH–H₂O, successively, to yield four fractions (E1b1–E1b4). Subfraction E1b1 (2 g) was chromatographed on a YMC RP-C18 column (14% CH₃CN–H₂O, v/v, 5 ml min⁻¹) to yield four fractions (E1b1a–E1b1d). Subfraction E1b1a (0.3 g) was purified by semipreparative HPLC with MeCN–H₂O (12 : 88, v/v, 5.0 ml min⁻¹), to yield 5 (5.0 mg, t_R = 25.5 min). Fraction E2a (15 g) was separated via an MCI column and eluted with a gradient of H₂O–MeOH (1 : 9–9 : 1, v/v), to obtain five subfractions E2a1–E2a5. Subfraction E2a1 (1 g) was purified further using a semipreparative HPLC with MeCN–H₂O (6 : 94, v/v, 5.0 ml min⁻¹), to yield 16 (6.0 mg, t_R = 32.0 min) and 17 (6.0 mg, t_R = 18.5 min).

Pieristoxinlactone A (1). Colorless crystals; mp 268–269 °C; [α]²⁰_D −9.9 (c 0.29, MeOH); IR (KBr) ν_max 3472, 2986, 2941, 1790, 1741, 1368, 1223, 1087 cm⁻¹; ¹H and ¹³C NMR data, see Tables 1 and 2; HRESIMS m/z 703.2568 [M + Na]⁺ (calcd for C₃₃H₄₄O₁₅Na, 703.2572).

Pieristoxinlactone B (2). White, amorphous powder; [α]²⁰_D −5.2 (c 0.23, MeOH); IR (KBr) ν_max 3505, 2921, 2850, 1744, 1646, 1374, 1233, 1053 cm⁻¹; ¹H and ¹³C NMR data, see Tables 1 and 2; HRESIMS m/z 661.2477 [M + Na]⁺ (calcd for C₃₁H₄₂O₁₄Na, 661.2467).

Pieristoxinlactone C (3). White, amorphous powder; [α]²⁰_D −46.2 (c 0.03, MeOH); IR (KBr) ν_max 3500, 2924, 2853, 1785, 1745, 1376, 1231, 1067, 925 cm⁻¹; ¹H and ¹³C NMR data, see Tables 1 and 2; HRESIMS m/z 661.2467 [M + Na]⁺ (calcd for C₃₁H₄₂O₁₄Na, 661.2467).

Pieristoxinlactone D (4). White, amorphous powder; [α]²⁰_D −11.9 (c 0.07, MeOH); IR (KBr) ν_max 3395, 2976, 2936, 1759, 1676, 1202, 1139, 1076, 928 cm⁻¹; ¹H and ¹³C NMR data, see Tables 1 and 2; HRESIMS m/z 415.1984 [M + H]⁺ (calcd for C₂₀H₃₁O₉, 415.1963).

Pieristoxinlactone E (5). White, amorphous powder; [α]²⁰_D −25.3 (c 0.17, MeOH); IR (KBr) ν_max 3390, 2965, 2944, 1777, 1741, 1232, 1075, 923 cm⁻¹; ¹H and ¹³C NMR data, see Tables 1 and 2; HRESIMS m/z 457.2075 [M + H]⁺ (calcd for C₂₂H₃₃O₁₀, 457.2068).

Pieristoxinlactone F (6). White, amorphous powder; [α]²⁰_D −7.9 (c 0.19, MeOH); IR (KBr) ν_max 3377, 2965, 2939, 1766, 1677, 1202, 1140, 911 cm⁻¹; ¹H and ¹³C NMR data, see Tables 3 and 2; HRESIMS m/z 421.1831 [M + Na]⁺ (calcd for C₂₀H₃₀O₈Na, 421.1833).

Pieristoxinlactone G (7). Colorless crystals; mp 270–272 °C; [α]²⁰_D +57.0 (c 0.26, MeOH); IR (KBr) ν_max 3505, 2982, 2920, 1773, 1731, 1373, 1238, 1059, 929 cm⁻¹; ¹H and ¹³C NMR data, see Tables 3 and 2; HRESIMS m/z 661.2478 [M + Na]⁺ (calcd for C₃₁H₄₂O₁₄Na, 661.2467).

Pieristoxinlactone H (8). White, amorphous powder; [α]²⁰_D +49.5 (c 0.10, MeOH); IR (KBr) ν_max 3536, 2937, 1747, 1370, 1235, 1045 cm⁻¹; ¹H and ¹³C NMR data, see Tables 3 and 2; HRESIMS m/z 603.2398 [M + Na]⁺ (calcd for C₂₉H₄₀O₁₂Na, 603.2412).

Pieristoxinlactone I (9). White, amorphous powder; [α]²⁰_D +7.5 (c 0.04, MeOH); IR (KBr) ν_max 3536, 2939, 1746, 1372, 1233, 1044 cm⁻¹; ¹H and ¹³C NMR data, see Tables 3 and 2; HRESIMS m/z 603.2417 [M + Na]⁺ (calcd for C₂₉H₄₀O₁₂Na, 603.2412).

Pieristoxinlactone J (10). White, amorphous powder; [α]²⁰_D +23.2 (c 0.19, MeOH); IR (KBr) ν_max 3440, 2978, 2942, 1741, 1373, 1235, 1049 cm⁻¹; ¹H and ¹³C NMR data, see Tables 3 and 2; HRESIMS m/z 663.2618 [M + Na]⁺ (calcd for C₃₁H₄₄O₁₄Na, 663.2623).

Pieristoxinlactone K (11). Colorless crystals; mp 231–233 °C; [α]²⁰_D +67.3 (c 0.10, MeOH); IR (KBr) ν_max 3469, 2942, 1786, 1731, 1371, 1245, 1054 cm⁻¹; ¹H and ¹³C NMR data, see Tables 4 and 2; HRESIMS m/z 661.2469 [M + Na]⁺ (calcd for C₃₁H₄₂O₁₄Na, 661.2467).

Pieristoxinlactone L (12). White, amorphous powder; [α]²⁰_D +52.1 (c 0.07, MeOH); IR (KBr) ν_max 3472, 2983, 2942, 1732, 1372, 1222, 1053 cm⁻¹; ¹H and ¹³C NMR data, see Tables 4 and 2; HRESIMS m/z 647.2307 [M + Na]⁺ (calcd for C₃₀H₄₀O₁₄Na, 647.231).

Pieristoxinlactone M (13). White, amorphous powder; [α]²⁰_D +25.0 (c 0.10, MeOH); IR (KBr) ν_max 3423, 2978, 2941, 1785, 1730, 1375, 1265, 1239, 1214, 1048, 913 cm⁻¹; ¹H and ¹³C NMR data, see Tables 4 and 2; HRESIMS m/z 619.2355 [M + Na]⁺ (calcd for C₂₉H₄₀O₁₃Na, 619.2361).

Pieristoxinlactone N (14). Colorless crystals; mp 267–268 °C; [α]²⁰_D −62.0 (c 0.05, MeOH); IR (KBr) ν_max 3368, 2919, 1738, 1374, 1242, 1049 cm⁻¹; ¹H and ¹³C NMR data, see Tables 4 and 2; HRESIMS m/z 679.2572 [M + Na]⁺ (calcd for C₃₁H₄₄O₁₅Na, 679.2572).

Pieristoxinlactone O (15). White, amorphous powder; [α]²⁰_D −53.2 (c 0.04, MeOH); IR (KBr) ν_max 3490, 2986, 2941, 1739, 1373, 1239, 1061 cm⁻¹; ¹H and ¹³C NMR data, see Tables 4 and 2; HRESIMS m/z 679.2572 [M + Na]⁺ (calcd for C₃₁H₄₄O₁₅Na, 679.2572).

Pieristoxinlactone P (16). White, amorphous powder; [α]²⁰_D −17.3 (c 0.30, MeOH); IR (KBr) ν_max 3480, 3391, 2975, 2945, 1761, 1445, 1232, 1117, 1077 cm⁻¹; ¹H and ¹³C NMR data, see Table 5; HRESIMS m/z 417.2123 [M + H]⁺ (calcd for C₂₀H₃₃O₉, 417.2119).

Pieristoxinlactone Q (17). White, amorphous powder; [α]²⁰_D −21.6 (c 0.19, MeOH); IR (KBr) ν_max 3355, 2977, 2937, 1761, 1676, 1201, 1184, 1140 cm⁻¹; ¹H and ¹³C NMR data, see Table 5; HRESIMS m/z 433.2065 [M + H]⁺ (calcd for C₂₀H₃₃O₁₀, 433.2068).

X-ray crystallographic analysis

The crystals of compounds 1, 7, 11, and 14 were obtained from recrystallization in a mixture of solvents (MeOH/H₂O, 10 : 1) at room temperature. The X-ray crystallography studies of the four compounds were conducted according to the usual procedure (for details, see Tables S1–S4 in the ESI†). The crystallographic data for compounds 1, 7, 11, and 14 reported in this paper have been deposited at the Cambridge Crystallographic Data Centre (deposition No. CCDC 1548366, 1548369, 1548367, and 1548368† for 1, 7, 11, and 14, respectively).

Animals

All animal care and experimental procedures were in accordance with the current guidelines of the National Institutes of Health (NIH). Animal experiments were approved by the Ethics Committee of Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College.

Acetic acid-induced writhing test

Groups of 8 Kunming mice were used as control and test mice. The mice were given an intraperitoneal injection of 1% v/v acetic acid solution 15 minutes after the administration of the compounds (blank control received the vehicle, 0.9% NaCl, 5 ml kg⁻¹, ip; positive control received the morphine, 0.5 mg kg⁻¹, ip). The mice were placed individually in glass boxes. The numbers of writhes produced in these mice were counted for 20 minutes. For scoring purposes, a writhe was indicated by a stretching of the abdomen with a simultaneous stretching of at least one hind limb.

Conclusions

In summary, herein we have reported seventeen new 3,4-secograyanoids (1–17), together with seven known compounds (18–24), which were isolated from the roots of Pieris formosa. Their structures with absolute configurations were characterized by a series of spectroscopic methods and X-ray diffraction. Compounds 1, 2, 4–8, 10–13, and 16–24 exhibited significant analgesic activity at 5.0 mg kg⁻¹ (ip) compared to vehicle-injected mice (p < 0.05). In particular, compounds 16 and 17 showed highly potent activities with writhe inhibition rates of 63.5% and 69.9%, respectively.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by National Natural Science Foundation of China (No. 81673314) and CAMS Innovation Fund for Medical Sciences (No. 2016-I2M-1-010). The authors are grateful to the Department of Instrumental Analysis at our institute for the spectroscopic measurements.

References

1. Y. Li, Y. B. Liu and S. S. Yu, Phytochem. Rev., 2013, 12, 305–325.
2. W. H. Tallent, M. L. Riethof and E. C. Horning, J. Am. Chem. Soc., 1954, 79, 4548–4554.
3. P. Narayanan, M. Rohrl, K. Zechmeister and W. Hoppe, Tetrahedron Lett., 1970, 11, 3943–3944.
4. S. Fushiya, H. Hikino and T. Takemoto, Tetrahedron Lett., 1974, 15, 183–186.
5. S. J. Wang, S. Lin, C. G. Zhu, Y. C. Yang, S. Li, J. J. Zhang, X. G. Chen and J. G. Shi, Org. Lett., 2010, 12, 1560–1563.
6. Z. Y. Wu, H. Z. Li and W. G. Wang, Phytochem. Lett., 2012, 5, 87–90.
7. Y. Li, Y. B. Liu, J. J. Zhang, Y. H. Li, J. D. Jiang, S. S. Yu, S. G. Ma, J. Qu and H. N. Lv, Org. Lett., 2013, 15, 3074–3077.
8. J. W. Burke, R. W. Doskotch, C. Z. Ni and J. Clardy, J. Am. Chem. Soc., 1989, 111, 5831–5833.
9. S. Z. Zhou, S. Yao, C. P. Tang, C. Q. Ke, L. Li, G. Lin and Y. Ye, J. Nat. Prod., 2014, 77, 1185–1192.
10. A. Furusaki, N. Hamanaka and H. Miyakoshi, Chem. Lett., 1972, 1, 783–786.
11. M. K. Zhang, Y. Zhu, G. Q. Zhan, P. H. Shu, R. J. Sa, L. Lei, M. Xiang, Y. B. Xue, Z. W. Luo, Q. Wan, G. M. Yao and Y. H. Zhang, Org. Lett., 2013, 15, 3094–3097.
12. Y. Li, Y. B. Liu, Y. L. Liu, C. Wang, L. Q. Wu, L. Li, S. G. Ma, J. Qu and S. S. Yu, Org. Lett., 2014, 16, 4320–4323.
13. Y. Li, Y. B. Liu, H. M. Yan, Y. L. Liu, Y. H. Li, H. N. Lv, S. G. Ma, J. Qu and S. S. Yu, Sci. Rep., 2016, 6, 36572.
14. J. S. Chen and S. Zheng, Chinese Poisonous Plants; Science, Beijing, 1987, pp. 226–227.
15. C. S. Niu, Y. Li, Y. B. Liu, S. G. Ma, F. Liu, L. Li, S. Xu, X. J. Wang, R. B. Wang, J. Qu and S. S. Yu, Org. Lett., 2017, 19, 906–909.
16. C. S. Niu, Y. Li, Y. B. Liu, S. G. Ma, L. Li, J. Qu and S. S. Yu, Tetrahedron, 2016, 72, 44–49.
17. C. H. Li, X. M. Niu and Q. Luo, Org. Lett., 2010, 12, 2426–2429.
18. Z. Y. Wu, Y. D. Li and G. S. Wu, Chem. Pharm. Bull., 2011, 59, 492–495.
19. W. G. Wang, Z. Y. Wu, R. Chen, H. Z. Li, H. M. Li, Y. D. Li, R. T. Li and H. R. Luo, Chem. Biodiversity, 2013, 10, 1061–1071.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed 1D and 2D NMR, HRESIMS, IR, and X-ray crystallography data. CCDC 1548366–1548369. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7ra08635k

---