Microbial glycosylation of tanshinone IIA by Cunninghamella elegans AS 3.2028

Abstract

Tanshinone IIA (TIIA) is a natural product with significant anti-atherogenic activities. However, poor water-solubility and low bioavailability hinder its further exploitation as a drug candidate. In this study, microbial transformation of TIIA by Cunninghamella elegans AS 3.2028 was conducted to obtain two new glycosylated derivatives. Their structures were identified as hydroquinone TIIA 11-O-β-D-glucopyranoside (1) and hydroquinone TIIA 12-O-β-D-glucopyranoside (2) based on extensive NMR and MS spectral analyses. The solubility of 1 in 50% MeOH–H₂O solution was approximately 50-fold that of TIIA, and 1 showed remarkably improved oral absorption in mice. Furthermore, 1 and 2 exhibited similar Nrf2 activation activity to TIIA.

---

Tanshinone IIA (TIIA), with a phenanthrenequinone skeleton, is a major bioactive constituent of Salvia miltiorrhiza, a well-known traditional Chinese herbal medicine, Danshen.¹ Recently, Danshen has been recorded as a monograph in United States Pharmacopeia, and TIIA is one of the marker compounds.² A big array of experimental and clinical investigations have demonstrated that TIIA can be used to treat atherosclerosis, hyperlipidemia, endangiitis, and cancer.³ However, it has very poor water solubility and fairly low bioavailability (around 3% in rats after oral administration).¹ Thus, TIIA is usually prepared in sodium tanshinone IIA sulfonate injection (STS) in clinical use. However, the rapid elimination of STS led to reduced pharmacological functions,⁴ and serious side effects have been reported when STS was taken along with warfarin.⁵ In recent years, a number of groups have tried to improve solubility and pharmacokinetic properties of TIIA by structural modification, including sulfonation, bromination, and modification of the phenanthrenequinone skeleton.⁶ Some of these chemical reactions take place under harsh conditions, and toxic reagents are needed.

Microbial transformation is a powerful approach for structural modification of structurally complex natural products. It shows many advantages such as simple operation, mild reaction condition, and high selectivity. In some cases, the biotransformed metabolites may have enhanced biological activities.⁷ Furthermore, glycosylation, hydroxylation and other biocatalytic reactions could increase polarity and solubility of the substrates.⁸ However, the microbial transformation of TIIA has never been reported, thus far.

The Cunninghamella species have been reported to catalyze hydroxylation, glycosylation, demethylation, cyclization, and sulfation reactions.⁹ In this study, we report simultaneous hydrogenation and glucosylation of TIIA with >95% conversion rate by Cunninghamella elegans AS 3.2028. Two new glycosylated products (1 and 2) were obtained, and their structures were identified by extensive NMR and MS spectral analyses. Compound 1 exhibited 50-fold higher solubility in 50% MeOH–H₂O solution than TIIA. Moreover, 1 showed remarkably improved oral absorption in mice. The Nrf2 activation activities of TIIA, 1 and 2 were also evaluated.

We screened 15 fungal strains derived from Aspergillus, Penicillium, Syncephalastrum, Alternaria, Cunninghamella, Mucor and Botrytis species, and found that Cunninghamella elegans AS 3.2028 was capable of metabolizing TIIA into two major products (1 and 2) according to HPLC analysis. The ratio of 1 to 2 was approximately 4 : 1 based on their peak area at 254 nm (Fig. 1), and the substrates were almost completely metabolized. The UV spectra of 1 and 2 were very similar to those of TIIA, suggesting they had a similar skeleton. In the mass spectra, 1 and 2 exhibited the same [M − H]⁻ ion at m/z 457, 164 Da greater than TIIA, indicating that 1 and 2 should be hydrogenated (+2 Da) and glycosylated (+162 Da) derivatives of TIIA. This deduction was confirmed by the neutral loss of 162 Da (a glucosyl residue) in the MS/MS spectra (Fig. 2).

Fig. 1 HPLC chromatograms for microbial transformation of TIIA. (A) TIIA standard. (B) Biotransformed products of TIIA by Cunninghamella elegans. (C) Blank control of C. elegans.

Fig. 2 UV and MS/MS spectra of TIIA, 1 and 2.

For preparative-scale microbial transformation, a total of 32 mg of TIIA was added into C. elegans, and 8.0 mg of 1 (25.0% yield) and 1.5 mg of 2 (4.7% yield) were obtained by repeated column chromatography and semi-preparative liquid chromatography (Fig. 3). The purities were above 95% by HPLC analysis. Both 1 and 2 are new compounds.

Fig. 3 Preparative-scale microbial transformation of TIIA by Cunninghamella elegans.

Compound 1 had the molecular formula of C₂₅H₃₀O₈, as established by its HRESIMS spectrum (m/z 459.2009 [M + H]⁺, calcd 459.2013). The ¹³C NMR resonances at δ_C 107.1, 77.0, 76.1, 74.1, 69.8 and 61.2, as well as the anomeric proton resonance at δ_H 4.41 (1H, d, J = 8.4 Hz), indicated the presence of a β-D-glucose residue (Table 1). In order to confirm the sugar residue, 1 was hydrolyzed by trifluoroacetic acid, and the water-soluble fraction was analyzed by ion chromatography coupled with pulsed amperometric detection (IC-PAD) as we had previously reported.¹⁰ As expected, the chromatogram showed one peak at the same retention time as D-glucopyranose (Fig. 4). The other NMR resonances were very similar to those of TIIA, except for C-11 and C-12. The two carbonyl resonances at δ_C 183.8 (C-11) and 175.9 (C-12) in TIIA disappeared, and two oxygenated olefinic carbon signals appeared at δ_C 136.1 and 142.1 in 1, indicating that a double bond formed between C-11 and C-12. In addition, a new phenolic hydroxyl group resonated at δ_H 9.83, and showed HMBC correlations with C-11 (136.1), C-12 (142.1) and C-13 (115.3). These evidences suggested the phenolic hydroxyl group was connected to C-12. The HMBC correlation between the anomeric proton (δ_H 4.41) and C-11 (136.1) indicated the β-D-glucose residue was linked to C-11. Therefore, the structure of compound 1 was identified as hydroquinone TIIA 11-O-β-D-glucopyranoside (Fig. 3).

Table 1 NMR spectroscopic data for compounds 1 and 2 (DMSO-d₆)

Position	1 (600 MHz for ^1H)		2 (400 MHz for ^1H)
	δH (ppm, J in Hz)	δC type	δH (ppm, J in Hz)	δC type
1	3.98, m; 2.96, dt (4.2, 18.0)	31.0, CH2	3.49, d (6.4)	31.0, CH2
2	1.80, m; 1.50, m	20.0, CH2	1.74, d (6.4)	19.8, CH2
3	1.66, br s	38.8, CH2	1.66, br s	38.3, CH2
4		34.7, C		34.5, C
5		142.7, C		141.9, C
6	7.43, d (8.4)	123.4, CH	7.55, d (8.4)	125.8, CH
7	7.83, d (8.4)	116.7, CH	7.88, d (8.4)	118.1, CH
8		114.2, C		118.6, C
9		126.2, C		121.9, C
10		132.4, C		133.7, C
11		136.1, C		142.9, C
12		142.7, C		135.0, C
13		115.3, C		116.3, C
14		148.6, C		144.6, C
15		116.3, C		117.2, C
16	7.73, d (1.2)	141.1, CH	7.69, s	141.4, CH
17	2.38, d (1.2)	9.6, CH3	2.39, s	9.6, CH3
18	1.30, s	31.5,CH3	1.32, s	32.0, CH3
19	1.29, s	32.5, CH3	1.30, s	31.9, CH3
glc-1	4.41, d (8.4)	107.6, CH	4.57, d (8.0)	106.7, CH
glc-2	3.43, m	74.1, CH	3.40, br s	74.0, CH
glc-3	3.25, m	76.1, CH	3.30, br s	76.2, CH
glc-4	3.21, m	69.8, CH	3.17, br s	69.8, CH
glc-5	3.04, m	77.0, CH	3.17, br s	77.5, CH
glc-6	3.57, d (10.8); 3.44, m	61.2, CH2	3.68, m; 3.49, m	61.0, CH2
OH	9.83, s		9.21, s

---

Fig. 4 IC-PAD analysis for sugar residues of 1 and 2.

Compound 2 had the same molecular formula as 1, as deduced from the HRESIMS spectrum (m/z 459.2008 [M + H]⁺, calcd 459.2013). Its ¹H and ¹³C NMR spectra were very similar to those of 1 (Table 1). The carbon resonances at δ_C 106.7, 77.5, 76.2, 74.0, 69.8 and 61.0, as well as the anomeric proton resonance at δ_H 4.57 (1H, d, J = 8.0 Hz), indicated the presence of a β-D-glucose residue. This was confirmed by IC-PAD analysis with a reference standard (Fig. 4). By careful analysis of the 2D NMR spectrum, we found the phenolic hydroxyl group at δ_H 9.21 showed HMBC correlations with C-9 (δ_C 121.9), C-11 (δ_C 142.9) and C-12 (δ_C 135.0) in 2, suggesting it was connected to C-11. In addition, the HMBC correlation between the anomeric proton (δ_H 4.57) and C-12 (135.0) indicated the β-D-glucose residue was linked to C-12. Based on the above evidences, the structure of compound 2 was established as hydroquinone TIIA 12-O-β-D-glucopyranoside.

Hydroquinone TIIA 11-O-β-D-glucopyranoside (1): brick-red amorphous powder; [α]^D₂₅ + 0.087 (c 0.1, MeOH); UV (MeOH) λ_max (log ε) 260 (3.34) nm; IR (KBr) ν_max 3433, 2926, 1636, 1065 cm⁻¹; HRESIMS m/z 459.2009 for [M + H]⁺ (calcd 459.2013), corresponding to C₂₅H₃₀O₈; ¹H NMR (600 MHz, DMSO-d₆) and ¹³C NMR (150 MHz, DMSO-d₆) data, see Table 1.

Hydroquinone TIIA 12-O-β-D-glucopyranoside (2): brick-red amorphous powder; [α]^D₂₅ + 0.019 (c 0.1, MeOH); UV (MeOH) λ_max (log ε) 260 (3.34) nm; IR (KBr) ν_max 3433, 2927, 1674, 1636, 1068 cm⁻¹. HRESIMS m/z 459.2008 for [M + H]⁺ (calcd 459.2013), corresponding to C₂₅H₃₀O₈; ¹H NMR (400 MHz, DMSO-d₆) and ¹³C NMR (100 MHz, DMSO-d₆) data, see Table 1.

A time-course study showed TIIA could be almost completely metabolized within 10 h of co-incubation with C. elegans (Fig. 5). Compound 1 was the major product and its relative amount increased to 80% at 10 h. Interestingly, the relative amount of 2 maintained at around 20% during 24 h. Similarly, previous reports had revealed that only hydroquinone TIIA 11-O-β-D-glucuronic acid, but not hydroquinone TIIA 12-O-β-D-glucuronic acid, was detected as a major metabolite of TIIA in rats after oral administration.¹¹

Fig. 5 Time-course for the biotransformation of TIIA by Cunninghamella elegans (the relative amounts were calculated on the basis of peak areas in the HPLC chromatogram at 254 nm).

Glycosylation is a feasible way to prepare natural product derivatives with improved solubility and bioactivities.¹² Although a big number of TIIA derivatives have been synthesized by chemical means, this is the first time the glycosides of TIIA were prepared. Moreover, the biocatalytic reaction could take place efficiently under mild reaction conditions.

We tested the solubility of TIIA and 1 (2 was not tested due to its limited amount). Since both compounds were almost insoluble in H₂O, their solubility in 50% MeOH–H₂O solution was determined. As expected, glucosylation could significantly enhance the solubility of TIIA. TIIA was only slightly soluble in 50% MeOH–H₂O (1.7 μg mL⁻¹). In contrast, the solubility of 1 increased by almost 50-fold to 84.6 μg mL⁻¹ (Table S2†).

In order to compare the bioavailability of TIIA and 1, they were orally administered to mice at the same mole dose (0.17 mmol kg⁻¹). The plasma samples were collected at 1 h and 2 h after administration, because the time of occurrence for maximum drug concentration (T_max) for TIIA in rats was around 1 h.¹ The plasma samples were combined and analyzed by HPLC/DAD/ESI-MSⁿ. The results indicated both TIIA and 1 could be absorbed into circulation within 2 h, and only the prototypes were detected (Fig. S2†). According to LC/SRM-MS quantitative analysis, the concentrations of TIIA and 1 in the plasma samples were 0.10 nM and 1.39 nM, respectively, indicating that glycosylation could remarkably increase the oral absorption of TIIA (Fig. S3†).

The antioxidant activity of TIIA could contribute to its cardiovascular therapeutic effects.¹³ In this study, the Nrf2 transcription activation activities of TIIA, 1 and 2 were evaluated by using a luciferase reporter assay in HepG2 human hepatocellular carcinoma cells.¹⁴ The glycosylated products 1 and 2 had similar antioxidant activities to TIIA, and all of them could significantly activate Nrf2 transcription in a concentration-dependent manner (around 4-fold of the control at 20 μM) (Fig. 6).

Fig. 6 Evaluation of TIIA, 1 and 2 to activate Nrf2 transcription in HepG2 cells.

Conclusions

In summary, we obtained two new glycosylated products of TIIA (1 and 2) by microbial transformation of Cunninghamella elegans AS 3.2028. Their structures were established based on extensive NMR and MS spectroscopic analyses. The solubility of 1 in 50% MeOH–H₂O solution was approximately 50 times higher than that of TIIA, and the oral absorption in mice also increased remarkably when TIIA was glycosylated. Furthermore, 1 and 2 exhibited similar antioxidant activities to TIIA. This is the first report on the glycosylation of TIIA by microbial transformation. The glycosylated products show noticeable druggability, and warrant to be further evaluated in future work.

Acknowledgements

This work was supported by National Natural Science Foundation of China (No. 81222054), and the Program for New Century Excellent Talents in University from Chinese Ministry of Education (No. NCET-11-0019). We wish to thank Dr Zhenbang Jiang (Thermo-Dionex Inc.) for his technical help with IC-PAD analysis.

Notes and references

1. S. Gao, Z. P. Liu, H. Li, P. J. Little, P. Q. Liu and S. W. Xu, Atherosclerosis, 2012, 220, 3–10.
2. United States Pharmacopoeia <Salvia 1-3> Dietary Supplements, USP 37/NF, 32 edn, U. S. Pharmacopoeia Convention, Rockville, MD, 2014.
3. (a) Y. Zhang, P. X. Jiang, M. Ye, S. H. Kim, C. Jiang and J. X. Lü, Int. J. Mol. Sci., 2012, 13, 13621–13666; (b) F. T. Tang, X. Q. Wu, T. Q. Wang, P. Wang, R. F. Li, H. J. Zhang, J. Gao, S. R. Chen, L. P. Bao, H. Q. Huang and P. Q. Liu, Vasc. Pharmacol., 2007, 46, 427–438; (c) W. Y. Chen, F. T. Tang, S. R. Chen and P. Q. Liu, China Pharm., 2008, 19, 884–887; (d) B. Wei, W. W. Li, J. Ji, Q. H. Hu and H. Ji, Atherosclerosis, 2014, 235, 318–327; (e) Y. Zhang, S. H. Won, C. Jiang, H. J. Lee, S. J. Jeong, E. O. Lee, J. H. Zhang, M. Ye, S. H. Kim and J. X. Lü, Pharm. Res., 2012, 29, 1595–1608.
4. W. Z. Chen, Y. L. Dong, C. G. Wang and G. S. Ding, Acta Pharmacol. Sin., 1979, 14, 277–282.
5. (a) J. Liu, X. R. Wang, Z. W. Cai and F. S. C. Lee, J. Am. Soc. Mass Spectrom., 2008, 19, 1568–1575; (b) J. Liu and Z. P. Chen, Nurs. Prac. Res., 2009, 6, 63.
6. (a) W. G. Huang, J. Y. Li, Y. Luo, J. Li and W. Lu, Chin. Chem. Lett., 2009, 20, 1461–1464; (b) Y. F. Bi, H. W. Xu, X. Q. Liu, X. J. Zhang, Z. J. Wang and H. M. Liu, Bioorg. Med. Chem. Lett., 2010, 20, 4892–4894; (c) C. J. Sun and D. L. Bai, Acta Pharmacol. Sin., 1985, 20, 39–43; (d) X. Z. Bu, Z. S. Huang, M. Zhang, L. Ma, G. W. Xiao, Z. P. Zhong and L. Q. Gu, Chin. J. Org. Chem., 2001, 3, 223–230.
7. (a) V. Křen, J. Kubisch, P. Sedmera, P. Halada, V. Přikrylová, A. Jegorov, L. Cvak, R. Gebhardt, J. Ulrichová and V. Šimánek, J. Chem. Soc., Perkin Trans. 1, 1997, 2467–2474; (b) H. X. Ge, J. Zhang, L. Chen, J. P. Kou and B. Y. Yu, Bioorg. Med. Chem., 2013, 21, 62–69; (c) X. B. Yang, J. Hou, D. Liu, S. Deng, Z. B. Wang, H. X. Kuang, C. Y. Wang, J. H. Yao, K. X. Liu and X. C. Ma, J. Mol. Catal. B: Enzym., 2013, 88, 1–6; (d) K. Shimoda and H. Hamada, Nutrients, 2010, 2, 171–180.
8. J. R. Jiang, S. Yuan, J. F. Ding, S. C. Zhu, H. D. Xu, X. T. Chen, X. D. Cong, W. P. Xu, H. Y. Yi and Y. J. Dai, Appl. Microbiol. Biotechnol., 2008, 81, 647–657.
9. H. Cao, X. Q. Chen, A. R. Jassbi and J. B. Xiao, Biotechnol. Adv., 2015, 33, 214–223.
10. W. Song, L. Si, S. Ji, H. Wang, X. M. Fang, L. Y. Yu, R. Y. Li, L. N. Liang, D. Zhou and M. Ye, J. Nat. Prod., 2014, 77, 1632–1643.
11. P. Li, G. Wang, J. Li, H. Hao and C. Zheng, J. Mass Spectrom., 2006, 41, 670–684.
12. V. K. Kapoor and A. Kaur, Mini-Rev. Med. Chem., 2013, 13, 584–596.
13. H. S. Zhang and S. W. Wang, Biochem. Pharmacol., 2007, 73, 1358–1366.
14. (a) T. W. Kensler and N. Wakabayashi, Annu. Rev. Pharmacol. Toxicol., 2007, 47, 89–116; (b) J. D. Hayes and A. T. Dinkova-Kostova, Trends Biochem. Sci., 2014, 39, 199–218.

---

Footnote

† Electronic supplementary information (ESI) available: NMR and HRESIMS spectra for 1 and 2, and experimental details. See DOI: 10.1039/c5ra09745b

---