Summary/Abstract
Lithium is a first-line therapy for millions of people suffering from bipolar disorder, and is promising for
inhibiting development of dementia. Experiments show that a primary mode by which Li+ alters physiological
processes is by reducing activities of a surprisingly limited number of Mg2+-dependent phosphoryl-transferring
enzymes, including phosphomonoesterases and protein kinases. While the (Li-independent) catalytic
mechanisms of these enzymes are quite well-understood, much about the mechanistic details underlying their
Li-susceptibility remain unknown. Not surprisingly, it remains a major challenge to design enzyme variants that
are Li-resistant, and use them to disentangle signaling pathways associated with Li-susceptibilities of individual
enzymes. Here we focus on Li+'s action on kinases, and address the following problem central to alleviating the
issues raised above. Experiments on 71 human kinases show a wide range of Li-susceptibility — many are
unaffected and others are affected to varying degrees. But there is no explanation for these variations.
We address this gap in our understanding of Li-action by using state-of-the-art molecular mechanics
(MM), quantum mechanics (QM) and QM/MM simulations, as well as mutagenesis experiments guided by
bioinformatics and natural selection. Supported by experiments, we explore the overarching hypothesis that Li+
affects kinase activity by interacting directly with their catalytic sites. In Aim 1, simulations will examine how Li+
binds kinases, and how Li+ binding reduces kinase activity. Additionally, simulations will provide insights into
potential allosteric effects that regulate catalytic site activity. Our biochemical, cellular and in vivo experiments
in Aim2 are designed to (i) systematically examine effects of sequence differences between Li-sensitive and Li-
resistant kinases, with the goal of making a Li-sensitive enzyme, GSK-3, resistant to Li+; and (ii) discover key
residues that make certain kinases Li-sensitive. Experiments will also validate findings from simulations, and at
the same time, simulations will provide molecular insights to interpret results from mutational experiments.
Combined analysis of results from simulations and experiments will yield a Li-resistant GSK-3, which is
significant because it will, for the first time, enable us to disentangle GSK-3-driven physiological effects of Li+
from those of other Li-sensitive enzymes. This study will also provide a physical basis to explain observed
variations of Li-sensitivity across kinases, and these biophysical findings will serve as foundations for future
efforts to make other Li-sensitive kinases resistant to Li+, and map their specific phenotypes. We expect that
such efforts will improve understanding and predictions of patient responses to Li-treatments and dosages,
which remains a difficult task. This will both expedite therapy and avoid exposure to side effects. Finally, this
study will explore new advancements in modeling enzyme reactions and yield a validated polarizable force
field for describing Li+/Mg2+ interactions with proteins. This will enable future reliable studies of Li-action on
proteins not considered in this project and broaden exploration of the full range of Mg-binding proteins.