Lipid Leaders: Gary Siuzdak

Tell us a little bit about yourself.
My current role at Scripps is Professor and Director of the Scripps Center for Metabolomics; I’ve been at Scripps since 1990 (33 years). I received my B.A. (Mathematics) and B.S. (Chemistry) at Rhode Island College and my Ph.D. (Physical Chemistry 1990) at Dartmouth College where to complete my thesis research I had to build a mass spectrometer. That effort positioned me well for the subsequent transition to Scripps. Beyond that, I live in Cardiff, California where I spend my time biking, swimming, and somewhat surprisingly (to me) a fair amount of time doing volunteer landscaping on an orchard for a local hermitage.
Did you always want to be a scientist or did someone in your life play a major role in influencing your decision to become a scientist?
As it is for most scientists, science was a natural inclination. I can’t say I ever regretted it.
What do you consider to be the greatest breakthrough in “omics” technology in the last decade?
While less of a breakthrough, and rather a slow evolution, I believe the cumulative developments of mass spectrometry technologies on a whole have addressed (and will continue to address) many of the challenges in proteomics, metabolomics and lipidomics.
Could you give us a general overview of drug-initiated activity metabolomics (DIAM)?
Drug-initiated activity metabolomics (DIAM) is a relatively simple concept, we use a known drug that acts on a particular disease, and then use metabolomics to investigate how the drug perturbs the metabolome. We then screen the perturbed metabolome for activity in the same disease. Simple, and it makes sense how DIAM could help identify endogenous metabolites related to a disease, and even identify metabolites that could treat the disease.
Your group recently discovered myristoylglycine as an endogenous metabolite for human brown fat differentiation. Could you describe how DIAM was used to make the discovery?
The myristoylglycine discovery is a great example of how DIAM works (Figure 1). Myristoylglycine was discovered using metabolomics and a known drug (zafirlukast) that promotes human brown fat differentiation. Zafirlukast perturbs the metabolome giving us finite number of metabolites to investigate for activity. Our activity screens on these metabolites determined that myristoylglycine was the most active and, at the same time, much less toxic than the drug, zafirlukast. A win-win.

Could you tell us about METLIN – how it came about, the goal of the database, any new or exciting things to be on the horizon?
METLIN is a database populated with experimental MS/MS data on over 875,000 molecular standards (Figure 2) and was key to the DIAM work in that METLIN was used to identify myristoylglycine and all the other dysregulated metabolites. METLIN came about from our early efforts on a sleep study in the 1990s, when I realized the difficulty in identifying metabolites could be improved with a tandem mass spectrometry database. We started building METLIN in the early 2000s.

What to look out for with respect to METLIN? Several developments are underway, 1) continued growth in MS/MS data acquisition on standards, 2) new scoring mechanisms incorporating neutral loss data, 3) the addition of our experimental ion mobility data (CCS values) on METLIN molecules, and 4) well, that’s a surprise.
Most people are probably familiar with MS/MS, but the METLIN database also includes neutral loss data as well. Could you explain neutral loss and how it can be used in molecular species analysis?
Unlike MS/MS databases, neutral loss spectra are created by subtracting the fragment ions’ m/z from the precursor ion’s m/z (Figure 3), this subtraction represents the neutral loss that occurs from the intact molecular ion to create the subsequent fragment (or product ion). For example, for a precursor ion with a m/z of 200 and a fragment ion with m/z 182, the corresponding neutral loss peak would be 18 (Dm/z). As shown in Figure 3, the NL spectrum is essentially the mirror image of the MS/MS spectrum, where MS/MS spectra are represented by fragment ion intensity versus m/z (Fragint vs m/z), while neutral loss spectra are neutral loss intensity versus Dm/z (NLint vs Dm/z). We are all familiar with MS/MS spectra and their Fragint vs m/z axes, however NLint vs Dm/z is serving to provide another dimension of molecular characterization, an especially useful technique for identifying unknown molecules.

We created what is currently the only neutral loss database (METLIN-NL) for small molecules. The orthogonal component of METLIN-NL to its METLIN-MS2 database is a powerful combination in identifying unknowns, this extra dimension helps identify previously uncharacterized chemical entities, including lipids and other metabolites.
If you had to give one piece of advice to someone in the research field, what would you tell them?
Don’t take advice.