Hunting the Unknown

The human body is teeming with amino acids, carbohydrates, and other small molecules. Collectively referred to as metabolites, these molecules are the intermediates and end products of cellular processes. Analyzing metabolites and lipids is a critical step in human health and disease research – from studying how tumors grow to predicting the severity of COVID-19.

Mass spectrometry (MS)-based methods are the industry standard for measuring metabolites and lipids in biological samples. Over the last few decades, advances in MS have led to the emergence of large-scale approaches to study these molecules: metabolomics and lipidomics – but sample complexity and the limitations of current technologies lead to challenges in identifying and elucidating the structure of unknown metabolites and lipids.

In each cell or tissue sample, thousands of different molecules may be present – challenging enough, but there is yet more complexity. Some molecules are far more abundant than others, pushing the limits of dynamic range; the levels of others may vary by several orders of magnitude in time and space; and, finally, some compound classes – particularly lipids – may include many molecules with similar structures. In short, finding unknown metabolites and lipids in biological samples and identifying their molecular structures is a task akin to searching for needles in haystacks.

So how can MS help? First, a little background for those less familiar with the technology. In simple terms, MS ionizes compounds and separates them by charge-to-mass (m/z) ratio, generating plots of ion signal against m/z – known as spectra – which researchers can use to identify compounds. To tackle the complexity of determining metabolites and lipids in biological samples, researchers use tandem (MS²) or sequential (MSⁿ) mass spectrometry. In these techniques, a first MS step ionizes and separates all the compounds so that users can select ions of a specific m/z to be broken into fragment ions; subsequent MS steps separate and detect fragment ions by m/z until structures can be assigned to all molecules present.

The challenge of identifying unknown molecules using MS is further complicated by the presence of irrelevant (from the chemical background) or redundant (from adducts, isotopes and in-source fragment ions) spectra. Existing instruments do not have sufficiently high-resolution to consistently prevent these spectra from interfering with detection of the compounds of interest, making MS data less reliable. Furthermore, low-resolution MS technologies cannot unambiguously identify elements within molecules, making it difficult to spot isotopologues (molecules that have identical molecular structures except that one or more atoms has a different number of neutrons).

In summary, the complex nature of biological samples and the limitations of current technologies prevent researchers from effectively annotating acquired spectra in many cases; sometimes only known metabolites and lipids can be identified.

But things are no less important for being difficult to find – particularly in the context of human health. To that end, new technologies are emerging to allow researchers to embrace the unknown. One such example is real-time spectral library matching, which uses spectra from known molecules to refine MSⁿ workflows in the search of unknown (but closely related) molecules, including metabolites of a drug. It can simplify data analysis and speed up identification because it only triggers MS³ scans on spectra similar to the compound(s) of interest, focusing those scans on fragments unique to the unknown molecules. It also increases the effective MS² scan rate, allowing the sampling of more MS precursors, and it can be combined with other workflows for automated background exclusion. As a result, real-time spectral library matching enables researchers to discover more unknown molecules related to compounds of interest. For example, a recent study of the drug amprenavir – a treatment for HIV – in human liver microsomes found 17 metabolites using real-time spectral library matching, only 11 of which emerged using a conventional MS³ approach (1).

Real-time spectral library matching can also help identify lipids and other molecules that produce fewer fragment ions. For example, by refining the search criteria to use a narrow mass tolerance and only advancing fragment ions with high match scores, researchers at the University of Wisconsin-Madison were able to adapt the real-time library search workflow to identify molecular structures of lipids, including the acyl chain composition of phosphatidylinositol (2).

Other new technologies are also helping to elucidate molecular structures by providing additional fragmentation pathways for ions not amenable to traditional collision-induced dissociation (CID). Ultraviolet photodissociation (UVPD), for example, uses high-energy photons to break apart the ions. An angiotensin II receptor blocker called telmisartan – commonly used to treat high blood pressure and heart failure – produces few fragment ions when exposed to CID. Researchers recently demonstrated that using UVPD on telmisartan produced more fragment ions than a traditional CID approach (1). Furthermore, the fragment ions produced using UVPD were of different sizes, allowing researchers to identify the drug in the sample with more confidence.

Finally, some researchers use a technique called stable isotope labeling to track small molecules and their metabolites through downstream biochemical pathways, but these traditional approaches are only suitable for tracking known compounds and their metabolites because they require prior knowledge of the compounds’ molecular structures. However, the introduction of ultra-high resolution accurate mass MS systems – coupled with new software solutions – give researchers the opportunity to use stable isotope labeling as a discovery tool for unknown metabolites.

Here, I’ve provided a few examples of advanced MS technologies that help researchers identify unknown metabolites and lipids, but there are others – and this exciting field is constantly evolving. Untargeted methods based on modern MS technology can produce more reliable data than ever before, enabling researchers to discover unknown compounds, elucidate their molecular structures, and understand their fate in the body. With this knowledge, we can dig even deeper into the details of disease and better understand the impact of treatment – allowing us to provide better care now and in the future.