Re: What thermodynamic changes accompany the generation of a prion?

Hi,

This is no softball question; by that, I mean you are probing an area that
is still being researched very actively and thus it is difficult to provide
you with a simple answer at the moment.  Significant progress has been made
in understanding the mechanism through which normal, soluble proteins are
transformed into insoluble fibrils, as is the case for prions and for the
process of amyloidogenesis in general.

At the heart of your question is the process through which proteins
normally fold; the model currently accepted (and believe me, there is
significant controversy within the field) describes protein folding as more
of a funnel, rather than a single “linear” pathway.  Whether folded or not,
proteins exhibit a great degree of conformational flexibility, due
primarily to the thermal resonant energy that causes individual bonds to
vibrate.  At absolute zero temperature, all motion is thought to be frozen
out; in physiological conditions (310 K, or 37C for mammals, slightly more
or less for bacteria), there is substantial amount of thermally activated
motion present in all systems.  For proteins, the result is that native
proteins rattle around a preferred conformation represented by the lowest
free energy state; more motion is typically observed for side chain
fluctuations at the surface of the protein than in the folded core, but at
some frequency all atoms are in motion and all bonds are vibrating.  In the
unfolded “state”, there are no stable bonds holding the protein together
and so there is great deal more conformational freedom for bonds to rotate
and for the chain to writhe around.  Just considering the 6 degrees of
freedom allowable for the backbone atoms of a single amino acid of a
protein, for a smallish 100 amino acid protein there are 100^6 or
1,000,000,000,000 possible conformations allowable; in principle, all
possible conformations are approximately equal in free energy for the
unfolded protein while only a small number (on the order of 10 to 1000) are
available for the native protein, all with very similar free energies.

In short, the unfolded state is really an ensemble of states, while the
native state is a very small number of possible conformations around a
dominant low-energy minimum.  The folding funnel model says that any of the
10^12 conformations of our hypothetical protein can collapse to a native
state, each one described by a unique pathway.  Experimental data indicates
that, consistent with the two-state model of protein folding, there is for
most proteins a well-worn path through which the majority of unfolded
proteins conformations will pass through in attaining the native-like fold.
 Think of a ski-slope that has a very wide starting line, and the
ski-course traverses a valley ending at a very narrow finish line.  The
presence of intermediates on the folding pathway are represented in a
folding funnel by areas that have local stability; these can be shallow
bowls in our ski-slope or they can be deep crevasses.  The significance for
us between a shallow local minimum and a deep one is the energy required to
climb back uphill to get back on the path towards the finish line, the
native conformation.  For real proteins, the consequences are both
energetic and importantly kinetic; a deep local minimum (a crevasse in our
ski slope) can trap proteins in a partially folded state.  Still, folding
is overall a downhill event and generally occurs on the order of
milliseconds for the majority of proteins.

The phenomenon of the transformation of proteins into amyloid-fibrils is of
great scientific interest, firstly, because it is closely connected to the
so-called conformational diseases, many of which have proven to be
incurable, and secondly, because it remains to be fully explained in
physical terms (energetically and structurally).  You’ve asked specifically
about prions, but we are also able to draw on research from studies from a
number of diseases, including Alzheimer's disease, Huntington’s disease,
senile systemic amyloidosis, and type II diabetes.  Deposition of the
protein fragment, A-beta, as insoluble fibrillar aggregates is one of the
defining pathological features of Alzheimer's disease, and a number of
epidemiological, genetic-linkage, and transgenic mouse studies indicate
that A-beta deposition is causally linked to neurodegeneration.  The
process leads to fibrous aggregates in the form of extracellular amyloid
plaques, neuro-fibrillary tangles and other intracytoplasmic or
intranuclear inclusions.

For the transmissible prion diseases typified by Mad Cow Disease, Variant
Creutzfeldt-Jakob Disease, and the Transmissible Spongiform
Encephalopathies, the origins of the disease have been linked to the
transfer of a prion-forming protein from one organism to another; the
transfer of the prion-forming protein is often insufficient in and of
itself to cause amyloidogenesis; rather, host proteins that have
amyloidogenic proteins are thought (and in one case I know of have been
proven) to be recruited into amyloid fibrils.  However, the majority of
prion diseases have been shown to have de novo origins.  In some diseases
(typified by early onset versions of the more generic disease), mutations
in an amyloidogenic protein make it more prone to forming the pathogenic
fibrils.  In these familial forms of amyloidogenic diseases, the rate of
onset is fast but the pathophysiology is often very similar if not
identical to the more generic form of the disease, suggesting a kinetic
phenomenon.

The unifying mechanism through which most amyloidogenic diseases operate is
one of a protein-misfolding or refolding of a normal native protein.  If on
our protein funnel we find ourselves in a local minimum where part of the
protein unfolds and perhaps refolds into a different, non-native state, we
have the opportunity to elaborate on that new, non-native structure.  This
seems to be the case for most prions; it is a small segment of the whole
protein that is disease causing, not the whole protein.  When this small
segment occasional unfolds with some LOW probability (1 in 10^6 to 10^9),
it can adopt a dimeric conformation with a neighboring protein that is in a
similar state of undress.  This new non-native dimer can act as a
nucleation point for elaborate b-sheet structures that grow into
amyloidogenic fibers.  That the initial nucleation occurs with a low
probability tells us that it is slow kinetically and initially an uphill
process in terms of free energy.  I mentioned proteins tend to fold on the
millisecond time scale; Alzheimer’s tends to develop over the course of
50-80 years, so though amyloidogenesis is a low probability event, it tells
us that events with small though finite probability can have dire consequences.

Many properties of amyloidogenic fibrils have emerged: a common structure
for filaments and fibrils, nucleation dependent kinetics, the role of
oligomeric intermediates and the existence of at least two protein
conformations separated by a high energetic barrier, which behave as two
macroscopic states. 

Some fundamental questions regarding the mechanism of fibril formation (the
details of which are beyond the scope of this forum) still facing
researches include:

   (a) Can domain-swapping be a mechanism for fibrillization of globular
proteins?

   (b) What is the role of {alpha} helical parts of proteins? Do they
remain helical in the fibrils?.

   (c) What is the role of {alpha} helical intermediates observed in
folding and fibrillization studies where temporarily non-native {alpha}
helices appear?

I hope this general, though broad answer gives you fuel for thought and
energy for filling in this area as best fits your interests.  For more
advanced information, I would suggest a visit to the National Library of
Medicine’s pub-med websites:
 http://www.ncbi.nlm.nih.gov/

Be sure to enter “review” as one keyword to help refine the scope of your
literature searches.

Regards,

Dr. James Kranz
Johnson & Johnson PRD