Simultaneous Identification of Characteristic Components in HPLC-PDA-ELSD Fingerprint Profile of Ginkgo biloba Leaves Extract

Abstract

This is an updated fingerprint profile for the simultaneous identification of 29 compounds, mainly including characteristic -terpene trilactones and flavonoid glycosides from Ginkgo biloba leaves (EGb) by employing a high-performance liquid chromatography-photodiode array-evaporative light scattering detector method. Compounds 1 and 6 had not been previously either detected or described in any EGb samples. In general, these 29 compounds were distributed into 5 regions in the fingerprint according to their different structural properties. This efficient analytical method could be generally applied to the quality control of EGb761 and other commercially available EGb products in the Chinese active pharmaceutical ingredients market.

Nowadays, the extract of Ginkgo biloba leaves (EGb) is one of the best-selling herbal supplements in the world and is used extensively for treating cardiovascular and neurological disorders. ¹ EGb’s pharmacological activities are commonly attributed to the synergism of 2 major classes of components, flavonoids and terpene trilactones (TTLs). In general, the commercial EGb preparations (eg the international standardized EGb761) are formulated to contain 6% TTLs and 24% flavonoid glycosides (FGs). ² Quercetin, kaempferol, and isorhamnetin are principally the aglycones of FGs. The TTLs are mainly composed of bilobalide, and ginkgolides A, B, C, and J.

EGb fingerprinting typically employs high-performance liquid chromatography (HPLC) separation. Ultraviolet (UV) detection is useful for the flavonoids, ^3,4 but the TTLs lack a useful chromophore and are only detectable using either an evaporative light scattering detector (ELSD) ^5,6 or refractive index detector (RID). ⁷ Therefore, these 2 classes of components in EGb are usually detected by separate chromatographic analyses, which is both inconvenient and time-consuming. HPLC-mass spectrometry has proved to be a powerful and generally applicable method due to its sensitivity and specificity, especially for identifying minor components. However, these analytical methods require complex sample preparations and exorbitant operation and maintenance costs, making it impractical for any routine analyses. ^8,9 Moreover, FGs in EGb are quite difficult to distinguish by any mass spectrometry method since structural analogs usually produce the same parent ions and major mass fragments. ¹⁰ Thus, it is desirable to establish a rapid, simple, and informative method to identify simultaneously FGs and TTLs in the quality control of EGb. In this study, a comprehensive fingerprint profile using HPLC combined with both photodiode array (PDA) and ELSD has been developed to detect simultaneously the FGs and TTLs in EGb.

The LC system, including column selection, mobile phase, and gradient elution, were first optimized to achieve good resolution and short analysis time for a commercially available EGb powder (provided by a Chinese manufacturer, Shanghai Sine Promod Pharmaceutical Co., Ltd). Two analytical columns: YMC C₁₈ column (250 mm × 4.6 mm, 5 µm) and Waters Sunfire C₁₈ column (150 mm × 4.6 mm, 5 µm), were compared, and the latter provided better resolution and shorter run time. Mobile phases consisting of acetonitrile-water and methanol-water gradients with and without 0.05% trifluoroacetic acid (TFA) (v/v) were compared. ^5–7,10,11 The results indicated that TFA improved resolution of the analyte and that the acetonitrile-water (containing 0.05% TFA, v/v) and methanol-water (containing 0.05% TFA, v/v) gradients afforded similar chromatography. The latter was chosen due to its lower cost.

The optimized gradient elution consists of a linear gradient from 30% to 70% methanol over 40 minutes, followed by an isocratic elution with 100% methanol for 10 minutes. The flow rate was set at 1.0 mL/min and injection volume was 10 µL. Equilibration time was 15 minutes and the column temperature was maintained at 30°C. The PDA and ELSD fingerprints were recorded online at the same time using a precision flow splitter (PDA: ELSD, 3:1, v/v). Retention times (t _R) of the corresponding peaks in the PDA and ELSD fingerprints were considered to be equivalent as the maximum time deviation was less than 0.1 minutes.

The HPLC-ELSD fingerprint profile of the aforementioned EGb powder produced by Sine Promod shows circa 30 peaks, and the main marker peaks are well resolved (Figure 1). About 7 additional peaks (indicated by the downward vertical arrows in Figure 1) were observed in the ELSD fingerprint, but not in the PDA fingerprint; as described below, 5 of these were identified as TTLs. To identify the target peaks accurately (Figure S1, see Supplementary Material), the same HPLC method was used to collect and identify 29 compounds (1–29, Figure 2) from the Sine Promod’s EGb product ¹² ; these 29 compounds included 13 FGs (1, 7, 8, 10, 11, 13–20), 9 FG cyclodimers (21–29), 5 TTLs (3–5, 9, 12), 1 phenolic acid (2), and 1 lignan (6). Their structures were determined by comparing their MS, nuclear magnetic resonance (NMR) (¹H and ¹³C NMR), electronic circular dichroism spectroscopic data, and physicochemical properties with those reported in the literature (Figures S2–S7, see Supplementary Material). In fact, compounds 1 and 6 are reported herein for the first time from the EGb samples.

Figure 1

HPLC-PDA-ELSD fingerprint profiles of EGb separated on an ODS column (150 × 4.6 mm, 5 µm) with a linear gradient of methanol in water (0.05% TFA, v/v) from 30% to 70% over 40 minutes, followed by an isocratic elution with 100% methanol for 10 minutes. Twenty-nine peaks (marked according to the numbering in Table S1) in the HPLC-ELSD fingerprint were assigned. Section A: TTLs (compounds 3-5, 9, and 12); section B: FGs containing tri-glycosides (compounds 7, 8, 10, and 11); section C: FGs containing mono-glycoside (compounds 13 and 15); section D: FGs containing di-glycosides (compounds 14 and 16-20); section E: FGs containing tetra-glycosides (compounds 21-29).

Figure 2

Chemical structures of compounds 1 to 29.

Figure 3

Comparative fingerprint profiles of the standardized EGb761 and the analytical EGb powders using the proposed HPLC-PDA-ELSD method.

Figure 4

Comparative HPLC-PDA-ELSD chromatograms of 2 commercially available EGb powders (Sine Promod’s and Bescon’s) in China by a modified analytical method. The modified chromatographic systems were the same as those for Figure 1, except for the gradient eluting conditions. The new gradient ran from 25% to 40% methanol over 8 minutes then increased to 65% methanol in 32 minutes, followed by an isocratic elution with 100% methanol for 10 minutes. The peaks in the ELSD profile of the Sine Promod’s EGb powder were assigned by comparison with those in Figure 1.

To correlate the isolated compounds with those specific peaks in the fingerprint, the 29 isolated compounds were individually analyzed using the same HPLC conditions used for the fingerprint chromatogram (Figure S8, see Supplementary Material). The peaks in the EGb fingerprint were assigned based on their t _R and UV absorptions. Table S1 (see Supplementary Material) summarizes the comparative t _R and UV absorption data obtained from the characteristic peaks in the fingerprints and the corresponding isolated reference standards (IRSs).

Five TTLs (peaks 3, 4, 5, 9, and 12), with well-resolved t _R of 7.96, 9.11, 9.98, 15.38, and 16.44 minutes, were definitively identified as bilobalide, and ginkgolides J, C, A, and B, respectively, based on their t _R. As the 5 TTLs were well resolved and eluted within a short period of 17 minutes, this method also has potential for application to the quantitative analysis of TTLs in EGb products. However, because of the structural similarity of ginkgo FGs (mainly mono-, di-, or tri-glycosides of the aglycones quercetin, kaempferol, and isorhamnetin), separation of these derivatives using chromatographic methods with a simple binary elution system in a short runtime is always challenging. ¹³ In this work, some pairs of peaks (7/8, 10/11, 13/14, and 18/19) with poor separations in the fingerprints were difficult to identify just based on their t _R. Nevertheless, they could be distinguished by comparing their different UV absorption maxima (λ_max) in the fingerprints with those of the IRSs (Table S1), such as 7 (t _R: 13.20 minutes, UV λ_max: 255.5 and 356.8 nm)/8 (t _R: 13.88 minutes, UV λ_max: 257.9 and 317.4 nm), and 10 (t _R: 16.00 minutes, UV λ_max: 265.0 and 349.7 nm)/11 (t _R: 16.11 minutes, UV λ_max: 254.3 and 354.9 nm). Peaks 13/14 co-eluted, as did 18/19. The intensities of peaks 21-29 were relatively weak and their identifications would be more tentative.

The peak assignments in the HPLC-ELSD fingerprint profile of EGb are summarized in Figure 1. Interestingly, the elution ranges of the 29 isolated compounds in the fingerprint could be roughly divided into 5 sections A–E (Figure 1), and the distributions of FGs were based on the number of the glycosides in the structures. TTLs (3-5, 9, and 12, section A) were located in the range 7.96-16.44 minutes. The FGs containing tri-glycosides (7, 8, 10, and 11, section B) have an elution range from 13.20 to 16.11 minutes. Mono-glycoside compounds (13 and 15, section C) showed t _R of 18.36 and 19.66 minutes, while compound 1 (t _R: 4.84 minutes) eluted ahead due to the dihydroflavonol structure with a free 3-hydroxyl. Diglycoside-containing structures (14 and 16-20, section D) were located in the range from 18.92 to 26.95 minutes, whereas the compounds eluting in the range 25.88-33.09 minutes were FG cyclodimers (21-29, section E) with tetra-glycosides. Therefore, the proposed fingerprint chromatographic method could also provide valuable guidance for the separation of specific compound classes, especially for the FGs.

To validate the proposed HPLC-PDA-ELSD method, the chromatographic analysis of the international standardized EGb761 sample was performed, and the resulting fingerprint was compared with those of the standard EGb powder provided by Sine Promod. As shown in Figure 3, the chromatograms of these 2 EGb samples were very similar, and the amounts and t _R of peaks in each fingerprint were nearly equivalent by alignment analysis. Therefore, the peaks identified in the EGb fingerprint developed above can definitely be applied to EGb761. In fact, since the elution order and abundances of peaks in the developed fingerprint were relatively stable when analyzed using the same analytical octadecylsilyl (ODS) column, the chemically resolved EGb fingerprint in different elution conditions could also be rapidly assigned and used to estimate the EGb products conveniently by parallel analysis. Herein, the HPLC-UV-ELSD fingerprint proposed with a modified elution condition (Figure 4) was then established and applied for the evaluation of another commercialized EGb sample provided by Jiangsu Bescon Pharmaceutical Co., Ltd (http://www.jsbskyy.com). As shown in Figure 4, the overall profiles of the Bescon’s EGb powders were consistent with those of Sine Promod’s EGb, and the marker peaks in the ELSD chromatogram of Sine Promod’s EGb sample were assigned by comparisons with those in Figure 1. Five characteristic TTL peaks (3, 4, 5, 9, and 12) were also found in the Bescon’s EGb fingerprint, and their relative contents were in accordance with those in Sine Promod’s EGb sample. The amounts and t _R of the peaks ascribed to FGs (7, 8, 10, 11, and 13-29) in these 2 fingerprints were nearly identical, the difference being the lower intensities of peaks I and II (16 and 17) and higher intensity of area III (FGs cyclodimers, 21-29) in the Bescon’s EGb sample, which may still satisfy the requirement of at least 24% total FGs contents for such a standardized EGb powder.

This is the first report of the simultaneous identification of FGs and TTLs in EGb with a HPLC-PDA-ELSD analytical method. A total of 29 marker peaks in the fingerprint were accurately identified, which include 13 monomeric FGs, 9 cyclodimeric FGs, 5 TTLs, 1 phenolic acid, and 1 lignan. The commercial EGb products are formulated to contain 24% FGs. As shown in Figure S9 (see Supplementary Material 1), the common aglycones (eg quercetin, kaempferol, and isorhamnetin) in EGb are extremely minor components. In fact, they are generally either reported or detected as hydrolysis products of the corresponding FGs in EGb. The proposed fingerprint profile could rapidly and accurately reflect the original chemical information, especially the genuine flavonoids in EGb products, which could preclude the possibility of artificial adulteration of pure aglycones (quercetin, kaempferol, and isohamnetin) or rutin. ^14,15 The EGb samples from both Sine Promod and Bescon could satisfy the requirement of at least 24% total FGs and 6% total TTLs. To our knowledge, there are more than 90 manufacturers that produce EGb in the Chinese market nowadays, but their quality control is worrisome. This method could be viewed as complementary to specific quantitative analysis in the quality control of the ginkgo leaf extract products and also enables the application in guided isolation of specific compounds (eg FGs) in EGb. Additionally, the newly established fingerprint could also allow a rapid evaluation of the geographical/seasonal variations of FGs and TTLs contents in G. biloba leaves, which would provide valuable guidance for the cultivation of G. biloba and for the harvest of superior EGb products.

Experimental

The details of this section can be found in Supplementary data.

Supplemental Material

Supplementary Data - Supplemental material for Simultaneous Identification of Characteristic Components in HPLC-PDA-ELSD Fingerprint Profile of Ginkgo biloba Leaves Extract

Supplemental material, Supplementary Data, for Simultaneous Identification of Characteristic Components in HPLC-PDA-ELSD Fingerprint Profile of Ginkgo biloba Leaves Extract by Guang-Lei Ma, Jiang Wan, Juan Xiong, Guo-Xun Yang, and Jin-Feng Hu in Natural Product Communications