Carl-Mikael Suomivuori

The mission of the Suomivuori lab is to uncover atomic‑level mechanisms in biology for the development of safer, more effective drugs. The lab uses state-of-the-art computational methods, including all-atom molecular dynamics (MD) simulations, quantum chemical calculations, machine learning–based structure prediction, as well as large-scale virtual screening and other in silico drug discovery tools, to drive complementary experimental work—the results of which, in turn, closely guide the computational work.

Enabling safer, more effective drugs through the rational modulation of binding kinetics. Compared to conventional measures, such as a drug’s binding affinity for its receptor, a drug’s binding kinetics—the rates at which the drug binds to and unbinds from its receptor—can better predict how well that drug works in a living system. However, despite the promise of safer, more effective drugs, rationally optimizing a drug’s binding kinetics remains an unsolved challenge, mainly because doing so requires characterizing not only the bound and unbound states (which determine affinity), but also the binding pathway. To achieve the rational design of ligands with desired binding kinetics, we combine advanced MD simulations—which can uniquely reveal binding pathways at atomic-level resolution—with efforts in medicinal chemistry (to synthesize designed ligands) and pharmacology (to experimentally measure the kinetics). We focus on G protein–coupled receptors (GPCRs)—the single largest class of drug targets—such as serotonin, dopamine, and opioid receptors (the targets of, e.g., psychedelics, antipsychotics, and opioid painkillers, respectively).

Uncovering the molecular mechanisms of neurotransmitter loading into synaptic vesicles. Nerve signaling relies on vesicular neurotransmitter transporters (VNTs), which harness a proton electrochemical gradient to pump large numbers of neurotransmitters into synaptic vesicles, whose contents are released into the synapse for nerve signal propagation. Despite the crucial roles VNTs play in neurophysiology, and despite the current utility and future promise of VNTs as drug targets, the mechanisms by which VNTs pump neurotransmitters have long remained unclear. To determine these mechanisms, we use a multiscale computational approach, combining molecular simulations based on classical physics (to access conformational dynamics involving large distances or long timescales) with simulations based on quantum mechanics (to get at proton transfer reactions). We have a particular focus on the VNT known as VMAT2, which transports all monoamine neurotransmitters. In addition to its central role in regulating our brain chemistry, VMAT2 is an important drug target, as it is acted on by drugs that treat movement disorders resulting from, e.g., long-term use of certain antipsychotics. Interestingly, VMAT2 is also acted on—through mechanisms that remain unclear—by amphetamines, including MDMA (“ecstasy”), which is showing promise in treating, e.g., post-traumatic stress disorder.

Toward next-generation cancer drugs by elucidating how ligands induce distinct signaling outcomes through the same receptor tyrosine kinase. Receptor tyrosine kinases (RTKs) control cell proliferation and development in a broad range of biological and pathological contexts, and they represent a pivotal class of cancer drug targets. Although RTKs have traditionally been thought of as binary switches that are either in a state that signals or one that does not, it is now clear that different ligands can trigger distinct signaling outcomes through the same RTK. How such “functional selectivity” occurs at a molecular level remains unclear, despite its importance both from a fundamental science perspective and for developing new therapeutics. We use large-scale MD simulations to uncover these mechanisms, leveraging lessons learned from previously published work on successfully uncovering corresponding mechanisms for GPCRs (Suomivuori et al., Science, 2020).