Unmasking traditional Chinese medicine

When it comes to analysing the components in traditional Chinese medicine – vital if science is to understand how they work, if at all - rapid resolution and high throughput are a must 

JUJUBE, or spine date, has been used in Chinese foods and medicine for at least 2,500 years.  Although jujube is not used in Western medical practice, Europeans have recognised its usefulness as a tonic for all parts of the body since the 17th century1. Part of the nutritional value of jujube comes from saponins. These compounds are present in many plants such as Ginseng and Notoginseng and have been the subject of research for a long time. Spine date seeds have two main saponin compounds, jujubosides A and B, which, like other saponin compounds have a triterpenes structure (Figure 1). Other triterpenes have been proven to have therapeutic effects. Scientific studies with animals have shown jujubosides to have anti-anxiety and hypnotic effects, causing a reduction in the speed of conditioned reflexes, a reduction in hyperactivity, and a lowering of blood pressure1. Therefore, they have the potential to be used as a remedy for irritability, insomnia, anxiety, oedema, congestive heart failure, asthma and throat diseases.

With the potential for interactions between the chemical compounds found within drugs, food, and alternative medicines, it is increasingly important to accurately identify all known components in a sample, including those which may be structurally similar. However, many analysis and separation techniques used in laboratories can often incur significant costs.  For HPLC analyses one of these costs can be in the use of large quantities of solvent, as well as reduced throughput due to long analysis times. This article describes the use of a Rapid Resolution High Throughput (RRHT) column for the separation of jujubosides from spine date seed by high-performance liquid chromatography (HPLC). These columns are made with small particles, 1.8µm, and have shorter lengths resulting in dramatically lower consumption of solvent and faster analysis compared to standard columns.

Figure 1. Structures of Jujubosides

Spine date seeds have a complex sample matrix, with many compounds having absorption below 210nm. Further, jujubosides are lacking a chromophore (Figure 1) and are unable to produce a sufficient signal with ultraviolet (UV) detection. The lack of signal and the high background detection around 210nm makes it difficult to get a clean baseline, increasing the potential for error in quantitative results. To determine the relative amounts of jujuboside analogues A and B and to compensate for this potential source of error, UV detection was replaced by detection with an evaporating light scattering detector (ELSD), giving higher sensitivity compared to UV detection and a cleaner baseline (Figure 2).

Figure 2. Standards and sample chromatograms by UV 210nm/ELSD on 4.6mm × 150mm, 5μm traditional column

Comparative analyses were conducted using a 3.0mm × 50mm, 1.8μm RRHT column (ZORBAX Solvent Saver HT Eclipse Plus C18, Agilent Technologies) compared to a traditional 4.6mm × 150mm, 5μm column (ZORBAX Eclipse Plus C18, Agilent Technologies)2. Jujubosides A and B were found to separate well on the traditional column as shown in Figure 2, with excellent peak shape using ELSD detection, but the two target peaks detected at UV 210nm were not well resolved from matrix compounds (Figure 2).

The conventional method was conducted in approximately 12 minutes using 12ml of solvent consumption in one injection. Changing from a traditional column to a RRHT column resulted in a saving of 87% of the solvent and 67% of the time (Figure 3). This is a result of changing both the column internal diameter from 4.6mm to 3.0mm ID and reducing the column length from 150 to 50mm. Both changes were needed to make such a dramatic solvent savings. At the same time, the peak 1 and 2 in Figure 3 (representing the jujobosides in actual sample) shows excellent resolution on both the 4.6mm and 3.0mm ID column. The resolution is much greater than baseline and is an efficient use of time on the RRHT column without risking co-eluting peaks from additional compounds in the sample. This turns the method into a high throughput LC method at lower cost.

Figure 3. Comparison of the methods developed on the 3.0mm × 50mm, 1.8μm RRHT column and 4.6mm × 150mm, 5μm traditional column

This method can be run on a standard LC instrument with an RRHT column3 primarily because a 3.0mm ID column was selected over a 2.1mm ID column, but optimisation may be required to minimise extra column volume. Optimisation of the LC is a simple procedure and consists of replacing typical existing tubing with tubing that may be shorter and with a smaller internal diameter. Also the flow cell may be replaced with one with a smaller volume. The goal of optimisation is to minimise extra column volume such that high resolution is achieved with the RRHT columns in a short time. The ELSD detector parameters must also be optimised. In this method, when the column was first switched to an RRHT column all the parameters of the ELSD remained the same. However, the heights of the two target peaks on the RRHT column were not consistent with the peak height obtained on the traditional column. This resulted in lower theoretical plates of over 2,000 for peak 1 and over 4,000 for peak 2.

To correct the inconsistency in peak height between the traditional column and the RRHT column, two parameters of the ELSD were optimised. By adjusting the filter value between OFF, 2, 5, and 8, the theoretical plates increased dramatically as the filter value decreased. Baseline noise also increased simultaneously. In order to determine the ideal signal to noise value, the filter value was set to 2, which provided theoretical plates of over 5,000 for peak 1 and over 8,000 for peak 2. The height of the two peaks was then consistent between the two columns, as shown in Figure 3.

This method was then successfully transferred to the short column with sub 2-micron particles and substantial solvent savings were realised. Of concern to many analysts when attempting this would be whether additional sample preparation is required to make this successful. The sample preparation for the spine date seeds was not changed and was the same regardless of which column was used. The purchased powder was first degreased with ethyl ether using a Soxlet extraction technique. The residues from this step were extracted with methanol and dissolved in water and then back extracted with butanol. This extract was dried down, redissolved in methanol and filtered through a 0.45µm regenerated cellulose membrane filter to complete the preparation for HPLC analysis. Therefore no additional costs were incurred in the laboratory to apply this method to the RRHT column.   

The results of this study show that jujubosides A and B can easily be separated, with high efficiency and symmetrical peaks. The method developed on an RRHT column dramatically reduced the solvent used and the analysis time, increasing sample throughput and lowering laboratory operating costs. Detection by ELSD gave a cleaner baseline allowing better specificity and had higher sensitivity for these compounds lacking chromophores. This study demonstrates that when separating compounds from complex matrices such as traditional Chinese medicines, separation techniques can be fast and cost-effective without adding sample preparation steps or compromising on accurate quantification.