Caesalpinimin A, a novel rearranged furanoditerpene with an unprecedented carbon skeleton from the seeds of Caesalpinia minax Hance

Abstract

Caesalpinimin A (1), a novel cassane furanoditerpene with an unprecedented carbon skeleton containing a spiro-A/B ring system was isolated from the seeds of Caesalpinia minax. The complete structure and absolute configuration were elucidated by spectroscopic methods and computational analysis.

---

Cassane furanoditerpenes are intriguing natural products in the context of structural diversity and broad biological activities. They are mainly distributed in the genera Caesalpinia and Pterodon, and characterized by a molecular skeleton constructed from fusion of three cyclohexane rings and a furan ring.¹ Some previously isolated cassane furanoditerpenes were reported to show significant antiviral,¹ anti-cancer,² anti-inflammatory, anti-analgesic,³ and plant growth regulation⁴ activities.

Caesalpinia minax Hance, a member of the genus Caesalpinia, is a prickly scandent shrub distributing widely in the tropics and subtropics. The seeds of this plant (called ‘ku-shi-lian’ in Chinese), have been used as Chinese folk medicine for the treatment of common cold, rheumatism, fever and prostatitis.^5,6 During our previous investigation on the active constituents of C. minax from Guangxi province of China, we identified some rearranged cassane diterpenoids with novel carbon skeletons, such as spirocaesalmin with a spiro-CD ring system⁷ and macrocaesalmin possessing a ten-membered macrocyclic 1,5-diketone ring.⁸ In our continuous search for structurally unique and biologically interesting cassane furanoditerpene, caesalpinimin A (1), a novel rearranged furanoditerpene (Fig. 1) with an unprecedented carbon skeleton featuring a spiro-A/B ring system, was isolated from this plant collected in Guangdong province (yield 0.000068%). Herein, we describe the isolation, structural elucidation and biological activities of 1.

Fig. 1 Chemical structure of 1.

The molecular formula of 1 was established as C₂₄H₃₂O₉ deduced from the pseudomolecular ion peak at m/z 487.1920 [M + Na]⁺ in the positive HRESIMS, indicating nine degrees of unsaturation. The UV spectrum showed the maximum absorptions at 213 and 258 nm, indicating the existence of conjugated system. The IR bands at 3573 and 1737 cm⁻¹ implied the existence of hydroxyl and carbonyl groups, respectively.

The ¹H-NMR spectrum showed two adjacent protons on a furan ring at δ_H 6.41 (1H, d, J = 1.8 Hz, H-15) and 7.21 (1H, d, J = 1.8 Hz, H-16), five signals at δ_H 0.91 (3H, s, H-19), 1.35 (3H, s, H-18), 1.21 (3H, s, H-20), 1.28 (3H, t, J = 7.1 Hz, 17-OCH₂CH₃), and 2.08 (3H, s, 7-OAc) ascribable to three tertiary methyls, one primary methyl and one acetoxy methyl, five oxygenated protons at δ_H 5.68 (1H, t, J = 5.9 Hz, H-2), 5.32 (1H, t, J = 10.8 Hz, H-7), 4.26 and 4.13 (1H each, m, 17-OCH₂CH₃), 3.80 (1H, dd, J = 11.6, 10.8 Hz, H-6), and an aldehyde proton at δ_H 9.85 (1H, s, H-1). The ¹³C NMR and DEPT spectra revealed that 1 possessed 24 carbons, comprising five methyls, three methylenes (including one oxygenated carbon at δ_C 61.2), three sp² methine carbons (including one aldehyde carbonyl at δ_C 203.5, and two olefinic carbons at δ_C 107.9 and 141.5), six sp³ methine carbons (including three oxygenated carbons at δ_C 96.8, 78.0, and 72.8), four sp² quaternary carbons (including two ester carbonyls at δ_C 173.3 and 172.2, and two olefinic carbons at δ_C 149.5 and 112.3), three sp³ quaternary carbons (including one oxygenated carbon at δ_C 92.4). These observations, together with molecular formula, indicated that 1 should be a cassane furanoditerpene.⁹

The more downfield nature of H-2 and C-2 relative to H-6 and C-6 indicates that C-2 is connected to two oxygen atoms to form a hemiketal structure. The full assignments and connectivities were determined by ¹H–¹H COSY, HSQC, and HMBC spectra.

Analysis of ¹H–¹H COSY spectrum revealed four spin systems (H-2 ↔ H-3, H-6 ↔ H-7 ↔ H-8 ↔ H-9 ↔ H-11, H-8 ↔ H-14, and H-15 ↔ H-16) which were shown in bold face in Fig. 2. The B, C and D rings were established by comparison with those of the known compound caesalpinin MG (2) (Scheme 1),¹⁰ which was also isolated from this plant along with 1. The ¹H–¹H COSY correlation between the methylene (δ_H 4.13 and 4.26) and the primary methyl (δ_H 1.28), along with the HMBC correlation between the methylene and C-17 suggested the formation of an ethyl ester functional group, and the HMBC correlation H-14 (δ_H 3.42) → C-17 (δ_C 173.3) indicated the location of the ethyl ester at C-14.

Fig. 2 ¹H–¹H COSY (bold), selected HMBC (arrows) and ROESY correlations (double arrows) correlations of 1.

Scheme 1 Plausible biogenetic pathway of 1.

The HSQC spectrum revealed that the signal at δ_H 5.32 (H-7) corresponds to the proton attached to the carbon at δ_C 78.0 (C-7), and the HMBC spectrum showed that H-7 was correlated to C-6, C-8 and C-21 (carbonyl group), which established that an acetoxyl group was attached at C-7. Similarly, the characteristic aldehyde proton (H-1) showed HSQC correlation to δ_C 203.5, and the HMBC spectrum showed that H-1 was correlated to C-10, C-20, C-5 and C-9, which established the location of the aldehyde group at C-10. The HMBC correlations of H-2, H-3, H-6 and H-7 to C-5 (δ_C 92.4), and the absence of correlation between H-2 and C-10 indicated a five membered furan ring A was connective to ring B via a spiro atom (C-5). Thus, the spiro-A/B ring system was established, and 1 was inferred to be an unprecedented cassane furanoditerpene with 5/6/6/5 ring system including a spiro-A/B rings.

The relative configuration of 1 was established by ROESY experiment. The ROESY spectrum showed NOE correlations (Fig. 2) between H-1 (δ_H 9.85) and H-9 (δ_H 2.38), H-9 and H-7 (δ_H 5.32), and H-7 and H-14 (δ_H 3.42), indicating these protons were in the same α-orientation. Furthermore, the NOE correlations between H₃-20 (δ_H 1.21) and H-6 (δ_H 3.80), and H-6 and H-8 (δ_H 2.53) suggested those protons were all β-oriented. In addition, NOE correlation between H-2 (δ_H 5.68) and H-1 revealed that H-2 approached to the aldehyde group in the same direction, and the ring A was nearly vertical to ring B. Based on the above observations, the relative configuration of 1 was established as 2R*, 5R*, 6S*, 7R*, 8R*, 9S*, 10R*, and 14S*.

Since 1 was established to have an unprecedented carbon skeleton, some further evidences were necessary to confirm its structure and absolute configuration. However, a single crystal could not be obtained despite many attempts. We decided to theoretically calculate the ¹H and ¹³C NMR of 1 by density functional theory (DFT) with Gaussian 09 package.¹¹

To obtain a reasonable starting structure and coarsely exclude energetic unreasonable conformation for 1, a conformational search using molecular mechanics calculations was performed with Discovery Studio 2.5 Client.¹² Then several minimum geometries were fully optimized at the B3LYP/6-31+G (d, p) level in the gas phase, and the most stable conformer for 1 was obtained by further frequency calculations. As recommended by a recent review paper,¹³ the NMR shielding constants for C/H atoms of 1 were computed using the GIAO technique at the mPW1PW91-SCRF/6-311+G (2d, p) level of theory using the PCM solvent continuum model and chloroform as the solvent. The calculated chemical shifts of 1 (Table 1) were obtained after corrections using the slope and intercept by linear regression analysis. The largest deviation and average error were only 4.3/1.4 and 0.51/0.12 ppm (similar to those of reported deviations¹⁴) for ¹³C and ¹H NMR data, respectively as compared with the experimental results, confirming that the proposed structure was reasonable and the NMR data were corrected assigned.

Table 1 Experimental NMR chemical shifts for 1 in CDCl₃ and its calculated NMR data (δ in ppm, J in Hz)^a

No.	Hexptl^b	Hcalcd	Cexptl^c	Ccalcd
a Largest deviation ΔδH = 0.51 ppm. Largest deviation ΔδC = 4.3 ppm (Δδ = |δcalcd − δexptl|). Linear fitting formula for ^13C NMR and ^1H NMR are δ = (185.14356 − σ)/1.04687 and δ = (31.4236 − σ)/1.03503 respectively, while the R values for two fitting are 0.9989 and 0.9819.b In 300 MHz.c In 75 MHz.
1	9.85 (1H, s)	10.00	203.5	205.1
2	5.68 (1H, t, 5.9)	5.55	96.8	97.0
3	2.01 (1H, m)	2.01	51.6	51.4
	2.25 (1H, m)	2.08
4			47.3	51.2
5			92.4	90.6
6	3.80 (1H, dd, 11.6, 10.8)	3.83	72.8	73.2
7	5.32 (1H, t, 10.8)	4.81	78.0	82.3
8	2.53 (1H, m)	2.71	38.2	40.5
9	2.38 (1H, td, 11.7, 4.9)	2.40	35.5	36.9
10			55.4	57.4
11	1.89 (1H, dd, 16.2, 4.9)	1.99	22.1	22.1
	2.24 (1H, m)	2.34
12			149.5	149.3
13			112.3	112.8
14	3.42 (1H, d, 7.8)	3.58	45.7	44.9
15	6.10 (1H, d, 1.8)	6.20	107.9	106.5
16	7.21 (1H, d, 1.8)	7.21	141.5	139.7
17			173.3	173.1
18	1.35 (3H, s)	1.31	24.2	22.6
19	0.91 (3H, s)	0.84	30.2	27.4
20	1.21 (3H, s)	1.20	9.45	8.8
7-OAc	2.08 (3H, s)	2.31	20.7	19.8
			172.2	170.2
17-OCH2CH3	4.13 (1H, m)	3.95	61.2	61.5
	4.26 (1H, m)	4.02
17-OCH2CH3	1.28 (3H, t, 7.1)	1.29	14.2	12.4
6-OH	4.00 (1H, d, 11.6)

---

The absolute configuration of 1 was further established by comparison of experimental and calculated ECD spectra (time-dependent DFT (TD-DFT)).¹⁵ The ECD spectra of the two possible enantiomers at the PBE0/6-311++G (2d, 2p) level were shown in Fig. 3. The calculated ECD spectrum of the enantiomer with configurations of 2R, 5R, 6S, 7R, 8R, 9S, 10R, and 14S (A, Fig. 3) showed a well match with the experimental spectrum, unambiguously proving that it was the most reasonable absolute configuration. Moreover, we analyzed the frontline molecule orbital of compound 1 (shown in Fig. 4). The bigger Cotton effect, which showed a negative peak at ∼225 nm for compound 1, was originated from the π(HOMO) → π*(LUMO) excitation because the theoretically predicted HOMO–LUMO gap is 5.42 eV that corresponds ∼229 nm for the predicted λ. Similarly, the smaller Cotton effect (the small negative peak around 300 nm) was assigned to the n → π* transitions.

Fig. 3 Determination of the absolute configuration of 1 by comparison of the calculated CD spectra with the experimental one (UV-shift = −4 nm, σ = 0.32 ev were assigned to match the experimental spectra band with SpecDis).¹⁶

Fig. 4 The HOMO–LUMO plot of compound 1.

Normally, cassane furanoditerpene are characterized by a molecular skeleton constructed from the fusion of three cyclohexane rings and a furan ring. Compound 1 is a novel cassane furanoditerpene with an unprecedented carbon skeleton containing a spiro-A/B ring. The possible biogenetic pathway was proposed as shown in Scheme 1. Compound 1 might be generated from caesalpinin MG (2),¹⁰ which could be converted to the key precursor a with ortho-dihydroxyl groups at C-1 and C-2 through a series of hydroxylation, deacetylation, and transesterification as shown in Scheme 1.¹⁷ After undergoing an oxidation procedure, the intermediate b with two aldehyde groups was formed.¹⁸ Finally, 1 would be yielded through a key intramolecular hemiacetal formation by condensation of hydroxyl and aldehyde groups,¹⁹ which was further verified by theoretical studies. The computational relative free energy profile for the last step of the plausible biogenetic pathway was shown in Fig. 5. All structures were optimized at B3LYP/6-31+G (d, p) level and the subsequent frequency calculation were carried out at four different theory levels. The results indicated that the reaction was thermodynamically and kinetically reasonable in the gas phase model, as it is about 12 k cal mol⁻¹ exothermicity and 36 k cal mol⁻¹ reaction barrier at M06-2X/6-311+G** level. Therefore the intramolecular hemiacetal formation was feasible to occur with enzyme catalysis in vivo.

Fig. 5 Energy profiles for the intra-molecular hemiacetal formation step of the biosynthesis pathway. R means reactant state, TS means transition state and P indicates product state. The imaginary frequency of the TS is −1808.2561 cm⁻¹ at M06-2X/6-311+G** level.

Because of the pronounced treatment effect of Caesalpinia minax on prostatitis,⁶ compound 1 was evaluated for COX-2 inhibitory activity using an enzyme immunoassay (EIA) kit (catalog no. 56131, Cayman Chemical, Ann Arbor, MI). It showed an IC₅₀ value of 4.0 μM. In addition, the cytotoxic activities of compound 1 were evaluated in DU145 and PC3 (both prostate) cancer cells using the reported MTT assay method.²⁰ Compound 1 only exhibited weak cytotoxic activities against these cells with IC₅₀ values over 50 μM. Thus, compound 1 might be a potential anti-inflammatory agent targeting the COX-2 enzyme with low toxicity.

In summary, compound 1 was the first example of cassane furanoditerpene with an unusual 5/6/6/5 tetracyclic carbon skeleton bearing a spiro-A/B rings. The isolation and structure elucidation of 1 has added an active entity to a diverse and complex array of cassane furanoditerpene family. Especially, the complete structure and absolute configurations have been assigned by both spectroscopic methods and computational analysis. Moreover, the plausible biogenetic pathway was reasonably proposed, and the key procedure of this pathway was also certified by computational method. Further investigations regarding compound 1's effects on COX-2 mediated functional gene expression and signaling pathway are in progress.

Acknowledgements

This work is dedicated to Prof. Xin-Sheng Yao on the occasion of his 80^th birthday, and was financially supported by the National Science Foundation of China (21203257 and 21272289) and Guangdong high level talent scheme (R.W.J.)

Notes and references

1. R. W. Jiang, S. C. Ma, P. P. H. But and T. C. W. Mak, J. Nat. Prod., 2001, 64, 1266–1272.
2. D. P. Ashok, J. F. Alan, W. Lee, Z. Gary, R. Rex, F. B. Mark, F. Leo and K. J. Randall, Tetrahedron, 1997, 53, 1583–1592.
3. M. M. M. Rubinger, D. Pilo-Veloso, G. M. Stefani, D. L. F. Alves and C. R. J. Maltha, J. Braz. Chem. Soc., 1991, 2, 124–127.
4. J. D. Antonio, C. A. B. Luiz, P. V. Dorila and L. F. A. Dalton, J. Nat. Prod., 1996, 59, 770–772.
5. Dictionary of Chinese Traditional Medicine, Shanghai People's Publishing House, Jiangsu New Medical College, 1977, p. 1289.
6. M. J. Gong, Y. Y. Xie and Z. J. Zou, World Journal of Integrated Traditional and Western Medicine, 2011, 6, 938–939.
7. R. W. Jiang, S. C. Ma, P. P. H. But and T. C. W. Mak, J. Chem. Soc., Perkin Trans. 1, 2001, 2920–2923.
8. R. W. Jiang, P. P. H. But, S. C. Ma, W. C. Ye, S. P. Chan and T. C. W. Mak, Tetrahedron Lett., 2002, 43, 2415–2418.
9. R. W. Jiang, S. C. Ma, Z. D. He, X. S. Huang, P. P. H. But, H. Wang, S. P. Chan, V. E. C. Ooi, H. X. Xu and T. C. W. Mak, Bioorg. Med. Chem., 2002, 10, 2161–2170.
10. S. K. Kalauni, S. Awale, Y. Tezuka, A. H. Banskota, T. Z. Linn and S. Kadota, Chem. Pharm. Bull., 2005, 53, 214–218.
11. Gaussian 09, revision A.1, Gaussian, Inc., http://www.gaussian.com.
12. Discovery Studio 2.5 Guide, Accelrys Inc., San Diego, 2009, http://www.accelrys.com.
13. M. W. Lodewyk, M. R. Siebert and D. J. Tantillo, Chem. Rev., 2012, 112, 1839–1862.
14. (a) A. Zheng, L. Chen, J. Yang, Y. Yue, C. Ye, X. Lu and F. Deng, Chem. Commun., 2005, 2474–2476; (b) J. X. Pu, S. X. Huang, J. Ren, W. L. Xiao, L. M. Li, R. T. Li, L. B. Li, T. G. Liao, L. G. Lou, H. J. Zhu and H. D. Sun, J. Nat. Prod., 2007, 70, 1706–1711; (c) M. W. Lodewyk and D. J. Tantillo, J. Nat. Prod., 2011, 74, 1339–1343.
15. L. B. Dong, J. Yang, J. He, H. R. Luo, X. D. Wu, X. Deng, L. Y. Peng, X. Cheng and Q. S. Zhao, Chem. Commun., 2012, 48, 9038–9040.
16. T. Bruhn, A. Schaumloffel, Y. Hemberger and G. Bringmann, Chirality, 2013, 25, 243–249.
17. (a) S. R. Peter, W. F. Tinto, S. Mclean, W. F. Reynolds and M. Yu, J. Nat. Prod., 1997, 60, 1219–1221; (b) T. Z. Linn, S. A. Awale, Y. Tezuka, A. H. Banskota, S. K. Kalauni, F. Attamimi, J. Ueda, P. B. S. Asih, D. Syafrudedin, K. Tanaka and S. Kadota, J. Nat. Prod., 2005, 68, 706–710.
18. (a) H. O. House, J. H. C. Lee, D. Van Derveer and J. E. Wissinger, J. Org. Chem., 1983, 48, 5285–5288; (b) E. P. Balskus, J. Me'ndez-Andino, R. M. Arbit and L. A. Paquette, J. Org. Chem., 2001, 66, 6695–6704.
19. L. Zhang, Y. Shen, H. J. Zhu, F. Wang, Y. Leng and J. K. Liu, J. Antibiot., 2009, 62, 239–242.
20. T. Mosmann, J. Immunol. Methods, 1983, 65, 55–63.

---

Footnotes

† Electronic supplementary information (ESI) available: 1D and 2D NMR, MS, UV, IR, and ECD spectra of caesalpinimin A (1) and detailed experimental procedures, as well as the optimized Z-matrixes for all computational models. See DOI: 10.1039/c3ra46502k

‡ These authors contributed equally to this work.

---