motions have been found to be much conserved in evolution and
often repurposed [2, 3]. Proteins also tend to evolve through
modular combination of domains with particular functions
[4]. Hence, rigid and dynamic or disordered proteins can be com-
bined in a myriad of ways. Also, and importantly, proteins often
react to stimuli such as pH, phosphorylation, changes in membrane
composition, and molecular recognition of other molecules.
Proteins have many rotatable bonds and their conformational
space is hyperdimensional, which makes it intrinsically hard to
understand. However, those rigid or machine like proteins, men-
tioned above, are characterized by small numbers of connected
spaces which have relatively low dimensionality, evolved so they
can carry out their function efficiently and structured by their native
contacts/interactions. The scientific community has started to
move towards a view of protein folding defined by intermediates
or network hubs with progressively lower dimensionality, rather
than the previous folding funnel theories [5–8]. This is a pertinent
example of one of the major problems we have in trying to under-
stand the high-dimensional space of proteins. High dimensionality
clouds our understanding of protein dynamics; we often project
important aspects of conformational space onto a small number
[2, 3] of important dimensions. However, many important pro-
cesses take place in more than three dimensions and this can be
difficult to understand and visualize.
Technology to design proteins already exists but thus far only
superstable examples have been generated and not the delicately
balanced dynamic ensembles found in most evolved proteins
[9]. This mastery of protein dynamic ensembles is not close
owing to the high complexity. Currently we are still developing
the ability to efficiently understand the dynamic ensembles of
evolved proteins we may come across in nature.1.2 What Is
a Conformational
Transition?
The phrase “conformational transition” is not always entirely clear
and can cause confusion. Hence, a clear definition of what we mean
by “conformational transition” is important. This requires the
definition of what is and what is not a separate conformation.
These discussions can, in some situations, revolve around the avail-
able structural data for a given protein. For those interested in
biomolecular simulations these discussions are more and more
related to the “slowest motions” of a system. The slowest motions
are, in general, most likely to be important functional motions. Of
course, the motions themselves are not necessarily slow, and they
may be rapid but rare as events, yielding a slow kinetic rate. How-
ever, the so-called faster motions can also be correlated and func-
tionally important. Fast motions can mediate allosteric signals over
surprisingly long distances across protein molecules with a confor-
mational change which cannot be detected easily with modern
experimental techniques [10]. Given this complexity, there is340 Benjamin P. Cossins et al.