In allosteric enzymes, the binding of a (effector) ligand at the
allosteric site, i.e., a site distant at least 1 nm from the functional
active site, regulates the enzymatic function, thus involving communication of a chemical signal from the allosteric to the active site.
This allosteric signal is expected to involve physico-chemical interactions between, generally conserved amino acid residues important for allostery [13–15], encompassing secondary structure
elements that define the “allosteric pathways” of the enzymatic
system [16]. Experimental characterization of allosteric pathways
is extremely challenging, and computational chemistry techniques,
such classical molecular dynamics (MD) simulations, could provide
unique information on protein dynamics at atomistic resolution
that significantly contributes to the elucidation of allosteric
mechanisms [17–19]. Standard MD simulations, in fact, are routinely used to monitor protein motions up to the μs timescale,
yielding MD trajectories that comprise the protein dynamics underpinning the allosteric mechanisms [20]. On the other hand, the
atomistic resolution of MD simulations and the large size of proteins (in the atomistic scale) make the recognition of allosteric
pathways (within the large network of physico-chemical interactions typical of proteic systems) a real challenge for standard analysis
of MD trajectories. Graph theory encompasses the appropriate
tools for modeling complex dynamical networks of chemical interactions resulting from atomistic MD simulations. Various graph
theory approaches could be exploited to represent a protein and
to decipher its allosteric mechanism, including the contact and the
elastic network models reported in details in this book series
[21, 22]. Here, we describe a method that combines the information on the correlated protein motions resulting from atomistic
MD simulations with a network analysis based on graph partitioning into mutually exclusive groups, named communities. The community network analysis (CNA) has been applied to several
biological systems, including allosteric enzymes, nuclear receptors,
and bacterial adaptive immune systems [23–30], providing elucidation of allosteric pathways and supporting rational discovery of
synthetic allosteric modulators [31, 32]. As exemplifying case, we
report here the CNA analysis of the imidazole glycerol phosphate
synthase (IGPS) enzyme from the thermophile Thermotoga maritima, an allosteric enzyme that represents a potential target for
allosteric drugs development [23, 33–35].
2 Materials
2.1 Initial Conditions
The proposed CNA method is based on the measure of protein
motion correlations and thus relies on the protein dynamics resulting from classical MD simulations. In order to set up MD simulations, initial conditions need to be defined. First, basic structural
138
Ivan Rivalta and Victor S. Batista
allosteric site, i.e., a site distant at least 1 nm from the functional
active site, regulates the enzymatic function, thus involving communication of a chemical signal from the allosteric to the active site.
This allosteric signal is expected to involve physico-chemical interactions between, generally conserved amino acid residues important for allostery [13–15], encompassing secondary structure
elements that define the “allosteric pathways” of the enzymatic
system [16]. Experimental characterization of allosteric pathways
is extremely challenging, and computational chemistry techniques,
such classical molecular dynamics (MD) simulations, could provide
unique information on protein dynamics at atomistic resolution
that significantly contributes to the elucidation of allosteric
mechanisms [17–19]. Standard MD simulations, in fact, are routinely used to monitor protein motions up to the μs timescale,
yielding MD trajectories that comprise the protein dynamics underpinning the allosteric mechanisms [20]. On the other hand, the
atomistic resolution of MD simulations and the large size of proteins (in the atomistic scale) make the recognition of allosteric
pathways (within the large network of physico-chemical interactions typical of proteic systems) a real challenge for standard analysis
of MD trajectories. Graph theory encompasses the appropriate
tools for modeling complex dynamical networks of chemical interactions resulting from atomistic MD simulations. Various graph
theory approaches could be exploited to represent a protein and
to decipher its allosteric mechanism, including the contact and the
elastic network models reported in details in this book series
[21, 22]. Here, we describe a method that combines the information on the correlated protein motions resulting from atomistic
MD simulations with a network analysis based on graph partitioning into mutually exclusive groups, named communities. The community network analysis (CNA) has been applied to several
biological systems, including allosteric enzymes, nuclear receptors,
and bacterial adaptive immune systems [23–30], providing elucidation of allosteric pathways and supporting rational discovery of
synthetic allosteric modulators [31, 32]. As exemplifying case, we
report here the CNA analysis of the imidazole glycerol phosphate
synthase (IGPS) enzyme from the thermophile Thermotoga maritima, an allosteric enzyme that represents a potential target for
allosteric drugs development [23, 33–35].
2 Materials
2.1 Initial Conditions
The proposed CNA method is based on the measure of protein
motion correlations and thus relies on the protein dynamics resulting from classical MD simulations. In order to set up MD simulations, initial conditions need to be defined. First, basic structural
138
Ivan Rivalta and Victor S. Batista