Materials and Apparatus:



GROUP NUMBER____T2______

TITLE____ HPLC Analysis of Goldenseal Root______

DATE SUBMITTED__12-20-01__

ROLE ASSIGNMENTS

ROLE GROUP MEMBER

FACILITATOR………………………..____ Ronit Morris ____

TIME & TASK KEEPER………………___Anoop Kowshik___

SCRIBE………………………………..____ Namrata Choudhury

PRESENTER………………………….____ Howard Lopez

ABSTRACT

In this experiment, the main objectives were to develop an HPLC method to identify berberine from Goldenseal root powder and compare the concentrations of berberine in GNC Fingerprinted Goldenseal Root and Bedrock Farm Goldenseal Powder. Berberine analysis in HPLC was conducted at a wavelength of 344 nm, which produced discernible absorbance peaks for both the berberine standard and the unknown solutions. The concentration-gain parameters used for the GNC and Bedrock Goldenseal powders were 40 and 20 times greater than that of the berberine standard, respectively. However, the berberine peak could not be isolated in the unknowns, as the retention times of the main peaks of the standard and unknowns did not agree. The retention times of the peaks for the GNC and Bedrock root were 13% higher and 8.6% higher than the standard retention time, respectively. These peaks were most likely combinations of many smaller peaks representing other alkaloids or impurity compounds. The most likely variables that contributed to this broad peak are the mobile phase (which was set at 50% water, 50% MeOH), or the lack of a solid phase extraction protocol to separate non-organic impurities. The berberine was assumed to be present in the unknown because of its high solubility in water, but possibly in very low concentrations. An ultra-high concentration unknown solution was tested and revealed another small peak within the broad unknown peak, signaling the presence of at least one other compound. Further manipulation of experimental variables was needed to isolate the berberine peak, and thus, the second objective of finding and comparing the berberine concentrations of the two root powders could not be achieved.

Objectives

• To develop a method to identify berberine from Goldenseal root powder through use of isocratic, reversed phase mode HPLC.

• To determine and compare the concentrations of berberine in GNC Fingerprinted Goldenseal Root 500 mg capsules and Bedrock Farm Goldenseal Powder.

Specific Aims:

• To determine the optimal wavelength and concentration-gain conditions at which to detect berberine so as to maximize the berberine sensitivity in unknown solutions.

• To determine the optimal mobile phase and the necessity of solid phase extraction for isolating the berberine peak in unknown solutions.

• To isolate berberine from both root powder and determine concentrations.

Hypotheses:

1) The optimal wavelength for berberine detection will be approximately 235 nm in the UV spectrum

2) The optimal concentration-gain parameters for the unknowns in HPLC will be 20 to 200 times more than the optimal concentration-gain product for the standards.

3) Isolation of berberine is possible without the use of Solid Phase Extraction.

4) Percent of recovered alkaloids for both unknowns shall range between 0.5-6% of sample.

Background

Hydrastasis Canadensis (Goldenseal) is a plant whose root contains alkaloids, mainly hydrastine and berberine, which have been used for medicinal purposes as an anti-inflammatory, oxytotic stimulant, anti-bacterial agent, and mild sedative. The most important principles of goldenseal are a group of isoquinoline alkaloids consisting mainly of hydrastine (1.5 - 4%) and berberine (0.5 - 6%), lesser amounts of canadine, candaline, and related alkaliods(Willard). Note that the numbers above correspond to fractional percentages of the total alkaloids. However, another source stated that berberine is 0.5 – 6% of the total root (Polaris).

The chemical composition of berberine chloride, a salt, is C20H18NO4Cl. It is a yellow powder, which is very soluble in water or methanol. Upon dissociation, it separates into the berberinium ion and chloride ion.

Hydrastine is the principle white alkaloid of Hydrastasis, and was discovered in 1850, by Mr. Alfred B. Durand (Amer. Jour. Of Pharmacology, Vol. 23, p. 13), who described it as being insoluble in water (Felter).

In 1873, Mr. A. K. Hale (Amer. Jour. Pharmacology, 1873, p. 247) noted that Canadine (C20H21NO4), another alkaloid found in the root which resembled berberine, but being darker in color, and behaving differently toward solvents (Felter). Canadine forms an almost insoluble nitrate by means of which the alkaloid was obtained from hydrastis. The free base nitrate form of this alkaloid is insoluble in water, soluble in alcohol, whereas the sulfate form, an exception, has been found to be soluble in water (Felter).

The berberine molecule is a positive ion (cation) due to the formal charge on the nitrogen of +1. It has a polarity due to the presence of electronegative atoms on oxygen molecules, along with the formal charge on nitrogen. The Nitrogen heteoroatom places it into the group of alkaloids. The chemical structure is shown below.

[pic]

Berberine is a yellow compound that occurs naturally as a berberinium salt of either chloride or sulfate, with both chloride and sulfate forms being very soluble in water (Seddon). It can also exist as berberine nitrate, although the nitrate does not naturally occur. The CRC Handbook of Physics and Chemistry rates berberine as highly soluble (a rating of 4/4, with 4 being very soluble) in diethyl ether and ethanol. Berberine is also mentioned as being soluble in water; one gram dissolves in 20ml of water (Merck). Hydrastine is given solubility ratings of 1 in water, and 3 in both acetone and benzene. Canadine is not listed as having any solubility rating within the web edition of the HBCP.

HPLC analysis is a powerful tool for the identification and quantification of active compounds of Herbal supplements, which in our case are the alkaloids of Goldenseal. Lazarowych, Etal described the use of HPLC for this purpose in their paper "Use of Fingerprinting and Marker Compounds for Identification and Standardization of Botanical Drugs: Strategies for Applying HPLC Analysis to Herbal Products. She examined several herbals such as Valerian Root and FeverFew using chemical "markers" which were basically pure standardized sources of the active compounds of the herbals. Different sources of herbals from different North American sources were analyzed for active compound concentrations, and many were found to contain little or none of these compounds (Lazarowych).

Solid Phase Extraction is a method by which aqueous samples can be processed in order to isolate and concentrate organic analytes from the sample matrix and provide a suitable sample extract for instrumental analysis. This process has a direct impact on accuracy, precision and quantitation limits and is often the rate-determining step for many analytical methods. SPE uses bonded alkyl bonded silica sorbents packed into disposable plastic or glass cartridges through which an aqueous solution of the analyte passes. Reverse phase SPE uses non-polar functional groups such as octadecyl (C18), which binds to the polar target molecule, while other compounds pass through. The solution will need to be highly non-polar to facilitate interactions of the polar analyte with the non-polar silica. The compound may then be removed fully from the tube by a washing step using a slightly more polar solvent. This procedure is highly accurate, reducing evaporation losses and exposure of analytes to organic solvents, which can be problems in traditional liquid-liquid extraction methods. (Beney, etal)

In previous research studies Goldenseal alkaloids, such as berberine and hydrastine, were run through HPLC at wavelengths corresponding to UV light. None of these studies noted using Solid Phase Extraction. In one Japanese Study, berberine standard and unknown were analyzed with a reverse phase chromatography tube at a flow rate of 0.6 mL/min, a 70 water/30 CH3CN (with 0.5 TFA) eluent, and a wavelength of 254nm (Shodex Webpage). Another study from Ansys Technologies and Metachem utilized a wavelength of 220nm with an eluent consisting of a changing mobile phase of 0.1%TFA in H2O to 0.1%TFA in MeCN (Polaris Webpage).

Materials and Apparatus:

• Reversed Phase HPLC , poly(styrenedivinylbenzene) (PSDVB) packed column

• Spectrophotometer at 344 nm

• Solid Phase extraction tube, silica gel, C18 Bonded phase

• HPLC Mobile phase: Methanol: 20mM KH2PO4. pH 2.6; 50% methanol, 50% water.

• Berberine Chloride Standard (Sigma)

• Goldenseal Powdered Root:

- Bedrock Farm Powdered Goldenseal Root

- GNC fingerprinted Goldenseal, 500 mg capsules

Methods:

A solution of 10 mg/ml berberine chloride (standard) was made in DI H2O and diluted to 1 mg/mL in 50:50 mobile phase ratio. Using the Genesis spectrophotometer, a survey scan, based against pure mobile phase, was run over a range of 200nm-500nm. Another scan was conducted in the range from 200-800 nm, which includes the visible range. No appreciable peaks were seen after 500nm in this scan. Initially the absorbance peaks went off the scale, so the solution was diluted again until a final concentration of 0.01 mg/mL was reached, which yielded absorbencies within the limitations of the spectrophotometer. The peaks in the survey scan corresponded to the wavelengths at which the detector absorbed the minimum amount of radiation that was passed through the berberine compound, corresponding to the maximum absorbance at that wavelength. A wavelength of 344 nm (UV) was chosen as the optimal wavelength and used for subsequent scans.

For the HPLC experiments, standard berberine solutions of 0.5 mg/mL were used because the detector on the built-in spectrophotometer is less sensitive at low concentrations. The sample loop only injects a volume of 20 μl in the column, which is further diluted with mobile phase. Therefore, the concentration injected must be higher in order for the detector to measure within its normal range.

As stated in the background, berberine makes up 0.5 and 6% of the total root, or of the total alkaloid content only. Thus, if berberine were 0.5 to 6% of the total root, the optimal concentration-gain product for the unknown solution would be anywhere from 20 to 200 times that of the standard (approximately). However, at that time, solubility was also a concern. At an unknown solution concentration of 100 mg/mL, the total solubility of berberine was questionable. Root powder solutions of 1mg/ml exhibited reasonable solubility, yielding golden yellow solutions, and were used in HPLC experiments. To adjust for this low concentration, the aspect of gain control was introduced. The HPLC Spectrophotometer has a gain control which amplifies the signal from solutions that contain the target compound in low concentrations. The gain control was adjusted for both the standard and unknown solutions so that HPLC experiments would result in graphs that contained the maximum possible peaks (for better sensitivity), but still lie within the detection limits of the spectrophotometer. Note that the concentration-gain parameters for the GNC root powder was 40 times that of the Berberine standard, while that of the Bedrock powder was 20 times that of the standard.

[pic]

Solid Phase Extraction:

Solid phase was not used to purify goldenseal root powder solutions prior to analysis with HPLC. One literature source that used HPLC to examine different herbal compounds used solid phase extraction in experiments involving other herbals, but not on Goldenseal Root (Alltech). The root was simply pulverized in mobile phase ratio, and analyzed directly. At least 2 other sources reported isolation of berberine from goldenseal root without the use of solid phase extraction. Further evidence came from experimental data. An HPLC trial of an unknown sample of root contained one clear, visible peak with no evidence of overlap. This was thought to represent only one compound at that retention time.

Results:

Figure 1 below, is the survey scan of .01 mg/mL berberine solution. The scan was based against the 50:50 methanol-KH2PO4 mobile phase.

Figure 1: Survey Scan of .01 mg/mL Berberine Solution

Peaks in the graph occurred at 420, 344, 263 and 228nm. 344nm was chosen as the optimal wavelength because it had the broadest peak and corresponded with literature that used wavelengths in the UV spectrum.

The following graph (Figure 2) shows the HPLC spectrophotometer data collected from the LabView software for the standard at 0.5 mg/mL.

[pic]

Figure 2: HPLC of berberine chloride standard at 0.5 mg/mL

The average area under the curve is 176.31 AU2. Considering a sample of 0.5 mg/mL solution in a 20 μL sample loop there is 0.01 mg of berberine. Therefore the concentration factor is 17637.13 AU2/mg. This value was used to calculate the amount of berberine in the unknown solutions. The retention times of each of the peaks agreed well, with the largest difference between times being 18.25 sec or 3.03%.

Figures 3 and 4 are the HPLC spectrophotometer graphs for the GNC and Bedrock Goldenseal Root.

[pic]

Figure 3: HPLC of GNC at 1 mg/mL

[pic]

Figure 4: HPLC of Bedrock at 0.5 mg/mL

Note that GNC trial 3 and Bedrock trial 1 are not aligned with the other peaks of the same unknown. These two trials were conducted in succession, and that the retention times of both trials were lower than the other trials of their respective sources (GNC, Bedrock).

The detector which converts absorbance data to electrical data on the apparatus was set at a gain of 0.01 for GNC and Bedrock roots compared to 0.2 for the standard solution. The areas obtained for the unknowns were therefore divided by 20 to agree with the rest of the HPLC results. Table 1 displays these results as well as fraction of alkaloids contained in each sample based on the concentration factor of 17637.13 AU2/mg.

|  |Standard |GNC |Bedrock |

|Avg. Area (adjusted to .2 gain) |176.3713 |10.819 |8.529 |

|Concentration of sample (mg/mL) |0.5 |1.0 |0.5 |

|Amount berberine/alkaloid input (mg) |0.1 |0.000613 |0.000484 |

|Fraction berberine/alkaloid in sample |  |3.07% |4.84% |

Table 1: Fraction Alkaloid in each Unknown

The unknowns contained alkaloid concentrations in the range given by literature values, so the data made sense. However, the data shows that the retention times of the unknowns did not match up with the standard; therefore, the calculated amounts of alkaloid are not necessarily berberine. The amounts given represent the total area under the graph of the peak, which may contain some berberine, but also other unknown compounds such as different types of alkaloids.

|Trial |Standard |GNC |Bedrock |

|1 |584 |686.5 |575 |

|2 |600 |683.5 |685 |

|3 |602.25 |650 |679.5 |

|AVG |595.42 |673.33 |646.5 |

|95% CI |24.719 |50.336 |153.972 |

Table 2: Retention Times

The GNC retention time is 13% higher than the standard retention time and over one minute longer. The Bedrock retention time is 8.6% higher and 51 seconds longer than the standard.

Additional trials were performed with the HPLC to determine the reason for the discrepancy in the retention times. A super high concentration of GNC was sampled (1000 mg/8 mL). This solution was used to test the theory that the berberine concentrations in the unknown solutions were too low to be detected as a noticeable peak in the detection range used.

[pic]

Figure 5: HPLC of GNC at 1000 mg/8 mL

At 659.5 seconds a second peak is present. The retention time for the larger peak is 493.5 seconds.

Also, a mixture of standard and GNC was tested. Equal volumes of 1 mg/mL GNC and 0.5 mg/mL standard were mixed together and injected into the HPLC at a gain of 0.2 and wavelength of 344nm.

[pic]

Figure 6: HPLC, Mixture of 1 mg/mL GNC and 0.5 mg/mL Standard

Unfortunately, the absorbance of the sample surpassed the limitations of the HPLC spectrophotometer. The actual composition of the peak present between 600 and 800 seconds is not known. Other peaks which had not shown up before at the lower gain setting are also clues that there are other interfering compounds in the unknown solution.

Discussion:

The final results of the experiment yielded no definite conclusion about the difference in berberine concentration between two sources of root powder. Berberine was never isolated using an HPLC method, however, there is some evidence of the existence of berberine in the sample, and many conclusions about what steps need to be taken in order to isolate the berberine peak at the correct retention time using HPLC. Much of the practical information on developing an HPLC method and refining the experiment is in the form of a troubleshooting guide, and the topics here are presented in a similar format.

Wavelength:

The purpose of the survey scan was to determine the best wavelength to use for the HPLC experiments, or the radiation wavelength at which berberine absorbs maximally. Based on a scan of berberine chloride standard, this was determined to be 344nm (UV).

Another option would have been to use a wavelength of 235 nm as quoted in literature sources (Alltech); however, one source used 344 nm with positive results. From literature, it is also known that both berberine and hydrastine absorb at 234nm because HPLC experiments resulted in a graph with a clear, sharp peak for each alkaloid (Alltech). However, from the survey scan data, it was also known that berberine absorbs strongly at 344 nm, so either one is an appropriate choice for the isolation of berberine. Any overlapping of peaks of other compounds would have been due to other factors, not to the choice of wavelength.

Solubility:

The solubility of the various alkaloids will dictate which, if any, are represented by the combined peak of the unknowns. Berberine and hydrastine are soluble in a 50:50 water/methanol ratio as demonstrated by the Alltech paper. In the Alltech experiment, root powder was pulverized in mobile phase ratio, and analyzed directly by HPLC. Both berberine and hydrastine peaks were evident in the results. The Handbook of Physics and Chemistry give berberine a ‘4’ (very soluble) rating in solubility in di-ethyl-ether and ethanol, but does not mention its solubility in water. Hydrastine has a rating of ‘1’ (very slightly soluble) in water. Therefore, the peak present for the unknown should not have contained hydrastine since the root was dissolved in water, but should have contained some amount of berberine. Another possibility is that this peak contained appreciable amounts of canadine, yet another alkaloid present in the goldenseal root with a chemical structure which differs from berberine by only one double bond. Canadine is also soluble in water. The only compound which would have appeared yellow is berberine. The yellow color is due to the chromophore in the berberinium cation. Hydrastine and canadine are colorless.

[pic] [pic]

Canadine Berberine

Concentration:

The concentration of berberine in the unknown samples is critical to the choice of concentration-gain factors, and the ability to separate berberine effectively using HPLC. The concentration was assumed to be between 20 and 200 times less than the standard, based on literature which claimed that berberine comprised 0.5-6% of the total mass of the root. This information may have been wrong, and the actual concentrations may have been much, much lower. Thus, the choice of concentration-gain parameters used to test the unknown solution may have been inappropriate. One example experiment below in figure 7 demonstrates the possibility of an extremely low berberine concentration compared to other alkaloids in the root.

[pic]

Figure 7: Berberine isolated from Goldenseal in a very low concentration (from SGE).

The Peak for berberine is tiny compared to the peak for hydrastine, demonstrating the plausibility of a low concentration, despite the differences in mobile phase (acetonitrile). The buffer used was sodium phosphate, which is equivalent to potassium phosphate for the purposes of HPLC, and the wavelength used was 235 nm, in the same (UV) range as 344 nm.

Solid Phase Extraction:

Solid phase extraction was not used in the preparation of unknown samples for HPLC analysis. This method is typically used to separate impurities from a target compound. Among these impurities would be the inorganic compounds in the goldenseal root which may have been present in solution. Because an HPLC scan of the unknown yielded one peak only, this was not considered a problem. There may have been other alkaloids present in the sample which showed up as part of the broad peak in the data, but these would not have been easily separated using solid phase extraction. The similarity in chemical composition between berberine and the other alkaloids would have made it very difficult to separate. Solid phase extraction involves using a solvent mixture of methanol:water of a specific polarity which will allow unwanted compounds to elute while berberine is left behind. The determination of this exact solvent ratio would have been a very lengthy process. Also, the preparation of a solvent ratio which is accurate to the correct number of significant digits may have been impossible given the equipment in the lab, and certainly not reproducible because of the volatility of methanol. Evaporation of methanol within the mixture during its preparation may have changed the polarity of the solvent by as much as 1%, and would render the process useless. Also, additional experiments would have been needed to guarantee the accuracy and reproducibility of the procedure. This procedure would have only been useful in separating impurities from the unknown solution, not in separating other alkaloids from berberine. Considering the amount of time allotted for the entire project, and the uncertainty it would have created, and the fact that it may not have helped in isolating berberine if the broad peak contained other alkaloids, and not impurities, solid phase extraction was not performed.

Some evidence that solid phase extraction may have been useful in separating impurities emerged during one of the last trials conducted. The HPLC of the addition sample of standard plus unknown revealed other minor peaks before the larger “berberine” peak. These peaks should have also emerged on the scans of the unknown, but did not. The unknowns were tested at a lower gain which increased the sensitivity of the detector 200 fold, so it is strange that they would not have been present in the original unknown solution scans. They were also not any other appreciably large peaks in the super concentrated solution of unknown root powder. These peaks may have represented new chemical compounds formed by the reaction of the Cl- ions in the standard solution with some soluble compounds in the unknown solution, and not impurities which were separable by solid phase extraction.

Mobile Phase Ratio:

The mobile phase ratio is the most important consideration in the analysis of this experiment. It controls the overlap of the individual peaks and the retention times of the compounds. A difference of as little as 1% in the mobile phase ratio can change the relative retention times of individual compounds enough so that the peaks either overlap, or they don’t. The two graphs below in figure 8 show a situation where an impurity was separated from the target compound by a change in mobile phase ratio of 1%. The wavelength (UV) and column (silica C18) used in the sample graph below are similar to those used in this experiment. The flow rate here is 1.0ml/min as opposed to 1.5 ml/min, and the compound is a polar organic, but not an alkaloid. The mobile phase used was acrylonitrile:water, not methanol:water, however the overall point is clearly demonstrated. A difference in 1% mobile phase changes the retention time of the “impurity” by 6 minutes. The difference in retention times in the peak expected for berberine and the peak in the resultant unknown data was only 1 minute. One literature source used 52:48 methanol:water as the mobile phase to separate berberine and hydrastine, however, this experiment was done using a 50:50 ratio. It was assumed that such a small difference could not have made a significant difference, but this was incorrect. Many variables were changed during the experiment, but a change in mobile phase ratio was overlooked.

[pic]

Figure 8: an example of one step in the development of an HPLC method to separate impurities from a target compound (from Bulletin 826B).

Yet another consideration is the use of a constant mobile phase, mixed immediately before reaching the end cap of the column by two separate pumps. The use of a variable mobile phase ratio may have been effective in isolating berberine, but unfeasible given the limitations in equipment and time. Air bubbles in the sample loop tube may have also contributed to the variable retention times. In addition, some of the experiments in the literature to isolate berberine used acetonitrile (MeCN) as a mobile phase, but this compound is toxic and was not an option in this experiment.

The trial which used an ultra-high concentration of root powder showed some evidence of another, emerging peak. This peak was not near the correct retention time for berberine, but may have represented another alkaloid, or an impurity. These data also indicate that there may have been many more peaks contained within the broad peak originally obtained, and further manipulation of the variables in the experiment was needed to isolate each individual peak, including berberine.

Conclusions

1) A peak absorption wavelength for berberine was found at 344 nm, giving discernible absorption peaks for the berberine standard and the unknowns in HPLC

2) The berberine peak could not be separated during the three weeks allotted for this experiment, due to several possible factors, the most likely of which were incorrect mobile phase or incorrect concentration-gain parameters for the given berberine concentration in the unknown.

3) Solid phase extraction would not have been useful in isolating berberine unless peaks corresponding to impurities were present in the broad, unknown peak.

Recommendations

This experiment suffered from a lack of an overall organizational plan. The data was not analyzed immediately after taking it, which delayed the discovery of the mismatched retention times (between the standard’s main peak and the two root powders’ main peaks) until the third week. Thus, it was not known that the peaks in the two powders did not necessarily represent berberine, which essentially wasted HPLC time with trials that would not tell us anything. It is recommended therefore that the first goal of future HPLC experiments be to isolate the peak of the compound in question in the unknowns. This must be done by manipulating the relevant variables (such as mobile phase and concentration) in a systematic way, until a peak emerges at a time which agrees with that of the standard, to within a margin of error. Only after the compound’s HPLC peak is identified and isolated can the determination of compound concentration be achieved.

References:

1. Beney, P.J. et al. Review, Evaluation, and Application of Solid Phase Extraction Methods. Hotline Vol 35, No 6: 1-5 December 1996

2. Budavari, Susan. The Merck Index: An Encyclopedia of Chemicals, Drugs, and Botanicals. Merck & Co., Inc: Whitehouse Station, NJ. 1996.

3. Felter, Harvey, and John Lloyd. King’s American Dispensatory. Copyright 1898.



4. Hennion, Marie-Claire. Solid-phase extraction: method development, sorbents, and coupling with liquid chromatography. [Review]. Journal of Chromotography 856(1-2):3-54, 1999 Sep 24.

5. Lazarwych, Natalie, and Pecos, Peter. Use of Fingerprinting and Marker Compounds for Identification and Standardization of Botanical Drugs: Strategies for Applying Pharmaceutical HPLC Analysis to Herbal Products. Drug Information Journal Vol 32: 497-512, 1998

6. Polaris Webpage



7. Seddon, K.R. “The Dunhuang Diamond Sutra: a challenging problem for scientific conservation techniques.” Dunhuang and Turfan: Contents and Conservation of Ancient Documents from Central Asia, The British Library Studies in Conservation Science, 1 (1996) 59-69.

8. Shodex Webpage



9. Willard, Terry. Herbs For Your Health- Goldenseal. Albert Food and Rural Development.



10. Technical Article: HPLC Method Development. An Example. .

11. HPLC Troubleshooting Guide: How to Identify, Isolate and Correct the Most Common Problems. Bulletin 826B Sigma Aldrich, 1997.

12. Natural Product Analyses with Alltima Rocket Columns Brochure #438.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download