When refining the manufacturing process for a nutraceutical ingredient or dietary supplement, whether it be at a small or large scale, there are many aspects that must be considered, including impurities that may be present in the formulation.
Determining what these contaminants are and how they may impact both the efficacy of an ingredient — as well as its safety profile — is highly important, as these types of molecules can cause significant regulatory and legal hurdles for formulators and manufactures alike.
The current standard protocols for identifying impurities, such as mass spectrometry, liquid chromatography (LC) and nuclear magnetic resonance (NMR) can prove costly and time-consuming when used independently, and may not offer certainty when identifying sample impurities.
To find out more about how nutraceutical developers can optimise their LC-MS procedures, Annabel Kartal-Allen (AKA) spoke to Ahmed Ben Faleh (ABF), co-founder and CEO of Isospec Analytics.
AKA: What are the current limitations when identifying unknown molecules?
ABF: If you’re using basic technologies like mass spec or liquid chromatography (LC), you are generally able to identify what you’re looking at, though you often run into issues when identifying something that’s not yet in a database.
When identifying unknowns, there may be many molecules of the same mass, which makes identification with mass spec somewhat tenuous. It can result in analysts making educated guesses into what they’ve found, which inevitably leads to mistakes. This can be dangerous from a regulatory standpoint, as well as for the safety of consumers.
There are currently few analytical options for identifying unknown molecules that don’t take weeks, or even months
Also, there are currently few analytical options for identifying unknown molecules that don’t take weeks, or even months. When you encounter an unknown impurity, there are three possible options: ignore it, purify the sample and then identify it with NMR, or use mass spectrometry. The first is clearly not ideal, as you may be leaving something harmful in the formulation, which is particularly bad when discussing ingestible products, such as nutraceuticals. The second can take weeks and requires high concentrations of a sample to be accurate. The third requires guesswork, followed by a trial and error process involving chemical synthesis of the suspected impurity, which is both costly and time consuming.
AKA: How does your technology work?
ABF: To identify unknown molecules in a sample, we generate an additional dimension of data by measuring its infrared ‘fingerprint’. We first separate the components of a sample using LC and then measure the infrared spectrum of each component directly inside a mass spectrometer. This allows us to measure an infrared fingerprint from the specific molecule you’re interested in with no interference from any other components of the sample, giving you information that’s entirely unique to the impurity in question.
If we don’t already have a molecule’s infrared fingerprint in our database, we have methods of determining the molecular structure, either from the fingerprints of its fragments, or from quantum chemical calculations.
An infrared fingerprint provides information on the molecule’s structure, which can elucidate the role it plays in one’s metabolism. This may ultimately allow one to optimise a product, making it more efficient and removing risk. 
Ahmed Ben Faleh, co-founder and CEO of Isospec Analytics
AKA: How does this technology expedite the identification process?
ABF: Instead of waiting weeks — or even months — to achieve impurity identification, we can determine unknowns in timescales as fast as ten minutes. This is because we eliminate lengthy purification steps as well as the need to synthesise guessed structures. We basically obtain this new data dimension, an IR fingerprint, without introducing any additional time to a standard LC-MS workflow.
If one has a match to the infrared fingerprint, identification is 100% certain
AKA: How would your platform reduce business costs?
ABF: The field of nutrition is highly competitive, and whilst you’re developing your product, someone else is eating up the market shares you could be taking. Moreover, reducing the need for chemical synthesis not only speeds up the identification process, but reduces labour costs.
AKA: How does your technology compare to the standard identification protocols in accuracy and repeatability?
ABF: The most important advantage of including an infrared dimension into current LC-MS workflows is that if one has a match to the fingerprint, identification is 100% certain, since an infrared fingerprint is an intrinsic property that is highly sensitive to molecular structure. In addition, these measurements are insensitive to small changes in experimental conditions, allowing readings to be consistent wherever you do them; a difficulty often faced when implementing standard protocols.
Identifying the structure of an impurity with certainty can offer peace of mind for ingredient developers, allowing them to adjust their processes to remove it if necessary.
AKA: How are your databases laid out?
ABF: We have various databases to complement a range of project types. Our largest database contains most molecules commonly encountered in nature – such as sugars and lipids – which can be used as a centralised platform for a large percentage of customers.
However, when we’re working with a company that produces novel synthetic molecules which do not naturally occur, we are able to create a customised database, allowing us to refer back to it when they want their samples analysed. From this, one can determine if a sample impurity is something already discovered or another molecule that may have occurred from a change in manufacturing or formulation procedures.
AKA: How can this technology help scale up manufacturing?
ABF: When scaling up manufacturing procedures, adjustments to the production process are usually necessary, since the system biology will be different at larger scales. Our technology can be useful in this context because it allows manufacturers to identify novel impurities that weren’t present before scale-up.
For example, if you wanted to formulate a new glycan additive to a nutraceutical product, information about the linkages between monosaccharides can be used to alter the bacterial strains to favour the production of useful molecules over other products and impurities.
When developing a bacterial strain to make a specific molecule through fermentation, obtaining information on the structure of the products and by-products is also important to highlight any modifications needed for both the bacteria and the reactor conditions to optimise the overall output. Optimisation of workflows based on data rather than trial and error is an important factor for saving time, money, and resources.
