By comparing the NMR, MS data and optical rotations with the literature values (Table 1 and Table 2), compounds 1–11 (Figure 1) were identified as (±)-cis-khellactone (1), (±)-praeruptorin A [(±)cis-3′-angeloyl-4′-acetylkhellactone, 2], praeruptorin B [cis-3′,4′-diangeloylkhellactone, 3], (+)-praeruptorin E [(+)-(3′S,4′S)-3′-angeloyl-4′-isovalerylkellactone, 4], cis-3′-acetyl-4′-angeloylkhellactone (5), cis-3′-isovaleryl-4′-acetylkhellactone (6), (+)-cis-(3′S,4′S)-3′-angeloyl-4′-senecioylkhellactone (7), (+)-cis-(3′S,4′S)-3′,4′-diisovalerylkhellactone (8), (−)-cis-(3′R,4′R)-3′-angeloylkhellactone (9), (−)-cis-(3′R,4′R)-4′-angeloylkhellactone (10) and (+)-trans-(3′S,4′R)-3′-angeloylkhellactone (11), respectively.

Under the present analytical conditions, the enantiomers of (±)-cis-khellactone (1) achieved good separation on a CHIRALPAK AD-RH semi-preparative column (Figure 2A). The two compounds were eluted at the retention times of 6.7 min (1a) and 9.7 min (1b), respectively. Their optical rotations were +57° (c, 1.0) and −55° (c, 1.0), which were in good agreement with those reported previously for (+)-cis-khellactone and (−)-cis-khellactone, respectively [20]. Thus, 1a was assigned as (+)-cis-(3′S,4′S)-khellactone and 1b as (−)-cis-(3′R,4′R)-khellactone (Figure 1 and Figure 2A).

Enantioseparation of (±)-praeruptorin B (cis-3′,4′-diangeloylkhellactone, 3) was achieved on the same chiral semi-preparative column using a mobile phase with a different acetonitrile/H₂O composition. The two enantiomers were eluted at 19.8 min (3a) and 22.9 min (3b), respectively. The optical rotations measured [3a, +37° (c, 1.0); 3b, −36° (c, 1.0)] for the enantiomers were similar to the data reported by Okuyama and co-workers [7]. Thus the two enantiomers were identified as (+)-praeruptorin B [(+)-cis-(3′S,4′S)-3′,4′-diangeloylkhellactone, 3a] and (−)-praeruptoin B [(−)-cis-(3′R, 4′R)-3′,4′-diangeloylkhellactone, 3b] (Figure 1 and Figure 2B).

Our previous studies [14,21] and Ruan’s report [22] on the metabolisms of (±)-praeruptorin A (2), (+)-praeruptorin A (2a), (−)-praeruptorin A (2b), (+)-praeruptorin B (3a) and (+)-praeruptorin E (4) revealed that all these angular-type pyranocoumarins underwent stepwise hydrolysis and generated cis-khellactone without alteration of the absolute configuration when they were incubated with liver microsomal proteins from rats or humans in the presence of a NADPH-regenerating system. This finding suggests that the absolute configurations of this type of coumarins can be determined from the configurations of C-3′ and C-4′ of khellactone produced from hydrolysis of the parent coumarins by rat or human hepatic phase I isozymes.

To confirm this speculation, (+)-praeruptorin A [ +59° (c, 1.0), 2a], (−)-praeruptorin A [ −61° (c, 1.0), 2b], (+)-praeruptorin E (4), and the two metabolites of 2b formed in rat plasma, (−)-cis-(3′R,4′R)-3′-angeloyl khellactone [ −57° (c, 1.0), 9] and (−)-cis-(3′R,4′R)-4′-angeloyl-khellactone [ −39° (c, 1.0), 10], were incubated with rat liver microsomes. The chemical structures and absolute configurations of 2a, 2b, 9 and 10 have been unambiguously identified using NMR analysis and optical rotation data in our laboratory [14], while 4 was commercially available and its chemical structure was double checked by comparing its mass spectral profile and optical rotation data [ +35° (c, 1.0)] with the literature report [8].

When incubated with rat liver microsomes (RLMs) in presence of the NADPH-regenerating system, all these five pyranocoumarins generated cis-khellactone which was eluted at 12.0 min under present achiral HPLC conditions and exhibited identical UV and MS spectra to those of the authentic (±)-cis-khellactone (Figure 3). When the cis-khellactone produced from each reaction was collected and analyzed using chiral LC-MS/MS, the products formed from (+)-praeruptorin E (4) and (+)-praeruptorin A (2a) exhibited the same retention time at 9.0 min as that of (−)-cis-khellactone (1b, Figure 4A), while those metabolites of (−)-praeruptorin A (2b), (−)-(3′R, 4′R)-3′-angeloylkhellactone (9) and (−)-(3′R,4′R)-4′-angeloylkhellactone (10) shared the same retention time (10.6 min) with (+)-cis-khellactone (1a, Figure 4B).

These results together with the optical rotation data of these parent pyranocoumarins (Table 1) support that their respective hydrolytic metabolite cis-khellactone keeps the absolute configurations during hydrolysis of the parent compounds in rat liver microsomal proteins.

In contrast, when 2a [(+)-cis-(3′S,4′S)-3′-angeloyl-4′-acetylkhellactone] was subjected to basic hydrolysis [20], the formation of compound 11 [(+)-trans-(3′S,4′R)-3′-angeloylkhellactone] was evidenced by NMR and LC-MS/MS analysis and its optical rotation (Table 1 and Table 2). Peucedanocoumarins I, II and III, which are also trans-khellactone type coumarins, yielded cis-khellactone after alkaline hydrolysis [23]. Thus, epimerization might occur during alkaline treatment of pyranocoumarins and finally may cause misidentification of the absolute configurations of parent compounds.

This method was further applied to the determination of the absolute configuration of another four pyranocoumarin compounds 5a, 6a, 7 and 8 isolated from Qian-hu. cis-Khellactone obtained from the enzymatic hydrolysis of compound 5a was eluted at 12.0 min on the achiral column (Figure 3) and only appeared at 10.6 min on the chiral LC-MS/MS (Figure 4B), and identified as (+)-cis-khellactone by comparison with the reference compound. Thus a (3′R,4′R) configuration of compound 5a was suggested. Furthermore, the optical rotation of 5ain chloroform was +3.5° (c, 1.0) and the NMR data supported the structure as cis-3′-acetyl-4′-angeloylkhellactone (Table 2). These findings demonstrated that the absolute configuration of 5a predicted from the absolute configuration of its metabolite cis-khellactone corresponded well with the structure identified based on LC-MS/MS, NMR and polarimetric analysis (Table 1 and Table 2). These data of 5a were in good agreement with those of pteryxin [(3′R, 4′R)-3′-acetyl-4′-angeloylkhellactone], which has been isolated from P. praeruptorum [23], thus, compound 5a was unambiguously identified as pteryxin [(+)-cis-(3′R,4′R)-3′-acetyl-4′-angeloyl-khellactone].

Similarly, the end products obtained from hydrolysis of compounds 6a, 7, 8 by RLMs (Figure 3) showed the same retention time as that of (-)-cis-khellactone on the chiral LC-MS/MS system (Figure 4A), supporting the existence of cis-(3′S,4′S) configurations in these compounds. Correspondingly, the NMR data (Table 2) of compound 6a supported it as cis-3′-isovaleryl-4′-acetyl khellactone and polarimetric measurement revealed a positive optical rotation [ +33° (c, 1.0)]. Compound 6a was thereby identified as (+)-cis-(3′S,4′S)-3′-isovaleryl-4′-acetylkhellactone. NMR analysis (Table 2) of compounds 7 and 8 identified their structures as cis-3′-angeloyl-4′-senecioylkhellactoneand cis-3′,4′-diisovalerylkhellactone, respectively. Their optical rotations were +31° (c, 1.0) and +35° (c, 1.0). Taken together, the structures were assigned as (+)-cis-(3′S,4′S)-3′-angeloyl-4′-senecioylkhellactone (7) and (+)-cis-(3′S, 4′S)-3′,4′-diisovalerylkhellactone (8). So far, (±)-dihydrosamidin, the racemic form of compound 6a, has been isolated from P. Turgeniifolium [24,25], while neither the racemate nor any of the single enantiomers has been reported for Qian-hu. Both 3′-angeloyl-4′-senecioylkhellactone and 3′,4′-diisovalerylkhellactone have been reported in P. japonicum,but not in P. praeruptorum (Qian-hu) [26] and their absolute configurations were not determined in the previous report [27]. Again, determination of the absolute configurations of pyranocoumarin compounds 5a, 6a, 7 and 8 based on chiral LC-MS/MS analysis of their respective hydrolytic metabolite cis-khellactone formed by RLMs were proved to be feasible and should be applicable to other pyranocoumarins that have a hidden C-3' and/or C-4' hydroxy functionality that can be released by enzymatic transformations.

It’s interesting to note that compounds 5 and 6, which were obtained from Qian-hu to prepare 5a and 6a through crystallization, exhibited similar positive optical rotation and identical NMR data as compounds 5a and 6a, respectively. However, when subjected to enzymatic hydrolysis, both compounds formed the two optical isomers of cis-khellactone (Figure 4C,D), indicating that compounds 5 and 6 were isolated from Qian-hu as enantiomerically enriched compounds. This was further confirmed by two peaks observed in chromatograms obtained from enantioseperation of compounds 5 and 6 (data not shown) and the enantiometric ratios (±) of cis-khellactone formed from 5 (7/1) and 6 (1/8) by RLMs (Table 1). Thus, the NMR data plus net optical rotation could not distinguish the enantiomerically enriched angular-type pyranocoumarin from its pure enantiomer.

In the present study, (±)-praeruptorin B (3) was enantioseparated, for the first time, to obtain the two enantiomers. The two optical isomers (3a and 3b) were also subjected to enzymatic hydrolysis followed by chiral LC-MS/MS analysis individually. Similar to those observed with the two isomers from (±)-praeruptorin A (2), cis-khellactone from (+)-praeruptorin B (3a) shared the same retention time (9.0 min) with (−)-cis-khellactone (Figure 4A), while cis-khellactone from (−)-praeruptorin B (3b) was observed at 10.6 min, the same as (+)-cis-khellactone (Figure 4B), suggesting that the metabolite cis-khellactone keep the absolute configuration as their respective parent optical isomer. Again, this result agreed well with that obtained from a combination of NMR and polarimetric analysis (Table 1 and Table 2) in the present study and the previous report [7]. Thus, (+)-praeruptorin B (3a) was identified as (+)-cis-(3′S,4′S)-3′,4′-diangeloylkhellactone and (−)-praeruptorin B (3b) as (−)-cis-(3′R,4′R)-3′,4′-diangeloylkhellactone.

The stereoselectivity in pharmacological actions [13], metabolism [14] and absorption [28] of praeruptorin A have been revealed by previous studies, indicating that it is of great importance to determine the absolute configurations of the angular-type pyranocoumarins when correlate their structures with the resultant biological interactions during pharmacological/toxicological evaluation and/or ADME screening. Furthermore, because most pyranocoumarins from Qian-hu naturally exist as enantiomers and/or diastereoisomers, misidentification of the identity of the compound studied has occurred [29,30,31]. Thus, it is crucial to establish a method which enables rapid determination of the absolute configurations of the pyranocoumarin compounds to be worked on. These findings obtained in the present study demonstrated that it is feasible to determine the absolute configurations of pyranocoumarins from that of the hydrolytic product khellactone using enzymatic hydrolysis coupled with chiral LC-MS/MS analysis. Because of the high sensitivity and selectivity afforded by LC-MS/MS techniques, this method can be used to determine the configuration of compounds of this chemical type with advantages in lower quantity of the tested compounds required for in vitro incubation.

Qian-hu was obtained from Ningguo, Anhui Province, China and the crude drug was authenticated as the dried roots of P. praeruptorum DUNN by Professor Pengfei Tu from Department of Natural Medicines, Peking University (Beijing, China). The specimens were stored at the State Key Laboratory of Quality Research in Chinese Medicine, University of Macau.

The crude drug (6 kg) was crushed into a powder and extracted with 95% ethanol (50 L × 3, 80 °C) to afford an extract (1.83 kg). An aliquot of the extract (50 g) was loaded onto a silica gel column (200–300 mesh, 800 g) and fractionated with petroleum ether-ethyl acetate (20:1→1:2, v/v) to get 11 fractions (Fractions 1–11). Fraction 4 (4.2 g) was loaded onto another silica gel column (200–300 mesh, 100 g) was and compound 2 (432 mg) was eluted with petroleum ether-ethyl acetate (8:1→1:1, v/v). Compounds 3 (195 mg) and 4 (52 mg) were obtained by recrystallization of Fraction 5 (5.3 g) and Fraction 8 (7.1 g) from MeOH, respectively. Compound 6 (12 mg) was isolated from the remaining solution of Fraction 5 using a semi-preparative COSMOSIL 5C18-AR-II column (250 mm × 10.0 mm i.d., particle size 5 µm, Nacalai Tesque, Kyoto, Japan) that eluted with MeOH–H₂O (65:35, v/v) at a flow rate of 2.5 mL/min. The remaining solution of Fraction 8 was introduced to the same semi-preparative HPLC system and eluted with MeOH-H₂O (69:31, v/v) to afford 7 (11 mg) and 8 (17 mg). Compound 5 (13 mg) was prepared from Fraction 3 by semi-preparative HPLC with an isocratic elution utilizing MeOH-H₂O (77:23, v/v). Compounds 5a [(+)-cis-(3′R,4′R)-3′-acetyl-4′-angeloyl-khellactone] and 6a [(+)-cis-(3′S,4′S)-3′-isovaleryl-4′-acetylkhellactone] were obtained as white needles by recrystallization of 5 and 6 from methanol.

A basic hydrolysis of 2 (70 mg) was carried out to obtain 1 (23 mg) following a procedure described previously by our group [14]. (+)-Praeruptorin A [(+)-cis-(3′S,4′S)-3′-angeloyl-4′-acetyl- khellactone, 2a] and (−)-praeruptorin A [(−)-cis-(3′R,4′R)-3′-angeloyl-4′-acetylkhellactone, 2b] were obtained by enantioseperation of compound 2 according to our previous report [14]. (−)-cis-(3′R,4′R)-3′-angeloylkhellactone (9) and (−)-cis-(3′R,4′R)-4′-angeloylkhellactone (10) were prepared from an incubation of compound 2b with rat plasma and unambiguously identified using NMR and LC-MS/MS analysis [14]. Compound 11 was obtained from basic hydrolysis of compound 2a, according to a method reported by Wu and co-workers [20]. Additionally, authentic compounds 2, 2a and 4 were purchased from Shanghai Traditional Chinese Medicine Research Center (Shanghai, China) and their identities were confirmed using LC-MS/MS and polarimetric analysis. The purity of all the compounds was all above 98% as determined by achiral HPLC-UV. The identities of all pyranocoumarin constituents 1–11 isolated from Qian-hu were identified using polarimetric analsyis, LC-UV-MS/MS and NMR analysis and confirmed by comparison with previous reports. The spectrometric and spectroscopic data were summarized in Table 1 and Table 2.

Pooled rat liver microsomes (RLMs) were prepared at the School of Biomedical Sciences, the Chinese University of Hong Kong by differential centrifugation according to a standard procedure reported previously [32]. The content of microsomal proteins was determined using Lowry’s method [33] and stored at −80 °C until use.

Glucose 6-phosphate (G-6-P), glucose-6-phosphate dehydrogenase (G-6-PD), and β-nicotinamide adenine dinucleotide phosphate (β-NADP⁺) were obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA). Formic acid, acetonitrile and methanol of HPLC grade were purchased from Merck (Darmstadt. Germany). Ultra-pure water was obtained in house using a Milli-Q plus water purification system (Millipore, Bedford, MA, USA).

Enantioseparation of the enantiometric mixtures (±)-cis-khellactone (1) and (±)-praeruptorin B (3) were carried out on a semi-preparative CHIRALPAK AD-RH column (10 × 250 mm I.D., particle size 5 µm, Daicel, Tokyo, Japan). The mobile phases were acetonitrile-H₂O with a volumetric ratio of 30/70 for 1 and 45/55 for 3. The flow rates were 2.5 mL/min for elution of both analytes. Each analyte eluted was collected separately and the solvent was removed under reduced pressure to yield 1a (9.5 mg) and 1b (8.5 mg), 3a (23.6 mg) and 3b (19.9 mg). The purity of all the enantiomers was more than 98% as determined using chiral LC-UV/MS/MS.

The polarimetric analysis was performed on a Perkin-Elmer 243B digital polarimeter (PerkinElmer, Netherlands) in chloroform at 589.3 nm, 20 °C. Compounds 1–11 were dissolved in CDCl₃ containing 0.3% trimethylsilane (TMS) and their ¹H NMR spectra were acquired on a Bruker Avance 600 spectrometer (600 MHz, Bruker GmbH, Bremen, Germany). Chemical shifts were reported in δ scale in units of part per million (ppm) with TMS as the reference, and coupling constants (J) were expressed in units of Hertz (Hz).

Enzymatic hydrolysis of each of the angular-type pyranocoumarins (final concentration 25 µM) was carried out in a 200 µL reaction system containing 1 mg/mL of rat liver microsomal proteins in 0.1 M potassium phosphate buffer (pH 7.4) at 37 °C. The reactions were initiated by adding the NADPH-regenerating system (4 mM MgCl₂, 1 mM β-NADP⁺, 1 mM G-6-P and 1 U/mL of G-6-PD). Reactions were stopped at 60 min by adding an equal volume of ice-cold methanol and the samples centrifuged to remove the proteins before subjected to achiral HPLC-UV analysis.

The achiral HPLC-UV analysis was performed on an Agilent 1200 series liquid chromatographic system (Agilent Technologies, Palo Alto, CA, USA) that equipped with a vacuum degasser, a quaternary pump, an autosampler and a diode array detector (DAD) system. Data acquisition was controlled by an Agilent ChemStation B 3.0 software. Sample separation was performed on an ODS reversed-phase C₁₈ column (250 mm × 4.6 mm I.D., particle size 5 µm, Agilent) at 35 °C. The mobile phase consisted of water containing 0.1% formic acid (A) and methanol (B). A gradient elution was carried out at 1.0 mL/min as follows: 0–4 min, 25–50% B; 4–16 min, 50–53% B; 16–17 min, 53–56% B; 17–20 min, 56% B; 20–23 min, 56–69% B; 23–40 min, 69–76% B; 40–41 min, 76–100% B; 41–52 min, 100% B. All reference compounds and samples from reactions were monitored at 323 nm and UV spectra were recorded over 200–400 nm. The injection volume was 70 µL. cis-Khellactone, the end product yielded from the aforementioned reactions of all pyranocoumarins, was eluted at the retention time of 12.0 min under the present analytical conditions and collected separately for further chiral LC-MS/MS analysis.

Chiral LC-MS/MS analysis was performed on an Agilent 1200SL series (Agilent) liquid chromatography coupled online with an API 4000 Q-trap^@ mass spectrometer (Applied Biosystems, Foster City, CA, USA). The LC system consisted of a vacuum degasser, a binary pump, and an autosampler. Chromatographic separation was performed on an analytical CHIRALPAK AD-RH column (5 × 150 mm I.D., particle size 5 µm, Daicel, Tokyo, Japan), which was eluted with acetonitrile-H₂O (30:70, v/v) at a flow rate of 0.65 mL/min. The mass spectrometer was equipped with a Turbo V^TM source and a Turbo Ion Spray probe (500 °C) and operated in positive ion multiple reactions monitoring (MRM) mode. Ion optics was tuned using polypropylene glycol (PPG) standard dilution solvents. Nitrogen was used as the nebulizer, heater, curtain and collision gases. Optimum ion source parameters for cis-khellactone were as follows: nebulizer (GS1), heater (GS2) and curtain gas flow rates 50, 50, 10 instrument units, respectively; ionspray voltage 4500 V; heater gas temperature 550 °C; declustering potential (DP) 100 V. The ion pairs used for monitoring cis-khellactone were 263.1 > 245.1 and 263.1 > 203.1. Collision energies (CE) were set as 25 eV and 30 eV for the two ion pairs, respectively. The Applied Biosystems Analyst Software package (Version 1.5) was used for instrument control, data acquisition and processing.