My name is Grace McDermott, and I am starting my second year of study in biomedical sciences at the University of Ottawa. I’m very excited to be dipping my toes into the field of scientific research by volunteering at the John L. Holmes Mass Spectrometry Laboratory! I have acquired invaluable practical knowledge and experience here. Thanks to the mentorship of Dr. Sharon Barden and Dr. Paul Mayer I have had to opportunity to study the popular medicinal herb, Withania somnifera, more commonly referred to as ashwagandha. Ashwagandha is known for its antioxidant, anti-anxiety, anti-inflammatory, and cognitive-enhancing effects, along with many more medical benefits. Interestingly, the plant’s roots contain an abundance of naturally occurring steroidal lactones called withanolides, which are believed to be the main active component. In order to identify the intriguing compounds that are present in ashwagandha root, I have employed a specific GC-MS method, TMS derivatization, and Kovats indices for a proper analysis.

There are many compounds expected to be present in the roots of ashwagandha, like triterpenes, alkaloids, flavonoids, and fatty acids. Here I have an example of each of the main classes of compounds reported to be found in the plant’s roots. I was originally hoping to be focusing on the class of compounds that received their name from their discovery in the plant: withanolides. Withanolides are polyoxygenated steroidal lactones categorized by their 28-carbon ergostane skeletons and they are the primary active component in ashwagandha. However, after conducting many trials using multiple analytical techniques, withanolides remained undetected in the ground-up root extract. This could be due to the possibility that the abundance of withanolides was very low, and/or the limitations of the GC-MS in terms of oven temperature and molecule size since withanolides are larger molecules with high boiling points.

For the extraction, the roots of ashwagandha from Mountain Rose Herbs were ground up and then placed in vials with ethanol, methanol, ethyl acetate, and pentane respectively and then centrifuged and stored in refrigeration. These samples were filtered and run though the GC-MS with little to no detection of meaningful data besides the presence of linoleic acid and palmitic acid. The use of the GC-MS was avoided initially due to the nature of the compounds I was interested in analyzing.

 It was then decided to derivatized using TMS. More specifically, the oily extraction was mixed with 100 ul of BSTFA + TMCS 99:1 and warmed to speed up the reaction. In the hope of making the sterols more volatile and more ionisable.

The GC-MS method was changed to:

The inlet was kept at a hot 250℃ due to the higher boiling points of fatty acids and sterols. The temperature started at 150℃ for 1 minute, then went up 15℃/min to 320℃ which was held for 5 minutes to hopefully pull out as many fatty acids and sterols as possible.
The result was rather surprising!

Here we have a chromatogram of the TMS derivatized CO₂ extracted ashwagandha root. As you can see, there is a staggering pattern that displays each of the saturated fatty acids including odd-numbered ones! Just to specify, for any double bonds present like in oleic and linoleic acid, I cannot confirm nor be confident of their placement. However, I want to point out the rarity of finding odd-numbered fatty acid chains in a non-microbial species. Animals mostly only produce even-numbered fatty acid chains due to their biological constraints in fatty acid synthesis. I won’t get into the details of that, but I will mention that some plant species do have the capacity to produce odd-numbered fatty acids which are more specifically found in the plant and seed oil. This is important since the CO₂ extraction produced an oily substance due to the non-polar nature of CO₂, giving us an oil from the dried root. However, I have yet to find any documentation of this being recorded in the roots of ashwagandha. Interestingly, there are multiple biological applications that odd-numbered fatty acid chains pose in the human body. For example, supplementary C₁₅ and C₁₇ in the bodily systems have been associated with a lower risk of type 2 diabetes, lower incidence of obesity, and lower risk of coronary heart disease. These advancements are relatively new, but the positive implications of odd-chain fatty acids for human health may be an emerging field of study! As you can see, we also have two phytosterols present: campesterol and β-sitosterol.

We can confirm the presence of the TMS derivatives of all the chromatographic peaks using the mass spectra they produce. The first step is to compare the mass spectrum of the peak of interest with the literature mass spectrum to see if they line up. Each molecule has its own unique spectrum, so the two should be very similar, if not the exact same. I’m using my experimental data of TMS derivatized palmitic acid for an example here; the two spectra line up perfectly, showing us that the peak of interest is in fact TMS derivatized palmitic acid. The loss of CH₃ from the parent ion produces a very stable fragment ion, this is seen in this spectra at 313da, this loss of CH₃ (15da) we use to identify all TMS possible peaks. The addition of TMS add 72da to the molecular weight of the compound in question. This TMS is used to derivatize alcohols and acids.

The math:

Palmitic acid 256da.............-1 255 + 73 or (256=72) TMS = 328 m+ ion - CH₃ 15 = 313da the intense fragment ion. This theory works with all TMSed compounds.

All Electron Impact (EI) Spectra unique to the compound you are analyzing but are practically identical between instruments.

The spectra of TMS derivatized even numbered fatty acids, are compared to a database used in the instrument NIST MS search 2.0 and this is integrated into the GC-MS computer system and it was able to identify them. The database did not however contain spectra for the odd fatty acids. Here the NIST Chemistry WebBook was used which is formatted slightly differently but it’s helpful for comparison nonetheless. i was able to match spectra for derivatized heptadecanoic acid (C17) and heneicosanoic (C21) acid. It should be noted that NIST does not have data for TMS derivatized C23 and C25 saturated fatty acids so I could not compare my data with the mass spectra. 

You may still be wondering how we’re so sure of our results; the answer is the Kovats retention index! The Kovats retention index is essentially a dimensionless quantity that standardizes the variable retention times of a given molecule. With millions of different combinations of methods that can be programmed and many different columns that can be installed on the GC-MS, retention time loses meaning. Using the equation for the temperature programmed Kovats index we standardize the quantity which gives us a retention index that we can compare with published literature values.

The CO₂ extraction of aswagandha, was repeated in order to replicate the data, and in doing so, we used a more concentrated solution of TMS derivatized and CO₂ extracted ashwagandha which allowed us to see the presence of long chain alkanes in the sample, as well! When the chromatogram of the alkanes standard was enhanced at the same magnitude as the sample chromatogram and lined up, the sample’s long chain alkanes matched up perfectly which verifies the accuracy of the Kovats index. Now if we look at the equation again the variables make more sense, with it being the retention time of the peak of interest, t_n+1 being the retention time of the alkane after the peak of interest and t_i being the retention time of the alkane before the peak of interest. After plugging all the lovely data in, we receive the dimensionless quantity that we can compare with literature. Also, it should be noted that the scales for the chromatograms are different. This also confirmeded the presence of the odd fatty acids again!

The CO₂ extract of Ashwagandha root, showed the presence of phytosterols, long chain alkanes, fatty acids and very surprisingly odd numbered fatty acids.

Using the NIST Chemistry WebBook we search for the compound of interest and find the data for the gas chromatography that contains the retention indices. We must use the data for the same column as our GC-MS which is the DB-5 MS. While the first two letters simply identify the company that made the column, the number is the most important variable as it represents the material the column is lined with along with the molecular interactions that couple that. The MS at the end is the more specific; it means it was developed for a mass spectrometer to have less column breakdown. The MS would be unnecessary for a flame ionization detector since it would not matter if the column bled some in that case. Here we have palmitic acid with a calculated retention index of 2045 and the two literature values listed prove the validity of our retention index. As you can see, it’s not always 100% precise, but within approximately 20 units. Moving forward I will only acknowledge the odd-numbered fatty acids as they are the most interesting, but the other even-numbered fatty acids will be on screen for those interested in comparing the values. For margaric acid we received almost the exact same value with ours at 2140 and literature at 2137.4.

Our retention index for heneicosanoic acid is 2535 and the literature value is 2533.9. The C19 in our sample appears at 2da lower than it should, suggesting a double bond 19:1 is how fatty acid nomenclature describes double bonds, the retention index for the 19 would be different than the 19:1, which is not in the data base. 

Then I have the remaining even-numbered fatty acids because NIST did not have sufficient data on the TMS derivatives of C23 or C25 saturated fatty acids. I also included campesterol and β-sitosterol for those wondering!

In conclusion, I originally set out with a quest to find withanolides and along the journey I found something arguably more compelling! As you will see with others from the John L. Holmes Lab, you don’t always get what you want, but sometimes that’s what you need.

As a student who just completed her first year, this experience has really enlightened me and introduced me to the crazy world of chemistry. With the discovery of some odd-carbon fatty acids in the roots of ashwagandha, I’m not entirely certain of their applications. It may be that these fatty acids provide supplementary explanations for the adaptogenic effects of ashwagandha, especially concerning the treatment of metabolic diseases like diabetes in folklore! The answer may never be known for certain, but I’m more than happy to contribute to the combined efforts of scientists from all over in finding additional treatments for ailments like diabetes.