Seems to me that if you really do have a little magnet inside your
ferritin you would expect the protein crystal to behave like a "spin
glass". Spin glasses are a classic homework problem in statistical
thermodynamics that makes most students question their choice of major.
However, invoking them is also a good way to impress physicists and get
you beam time. The most bizzare thing about spin glasses is that they
can actually have negative temperatures, but I digress.
In a nutshell: Any "ordering up" of the little magnetite nanocrystal
orientations must not just work against "entropy", but also against the
magnetic fields of all its neighbors. If they were all aligned
(ordered) then each magnet is oriented the "wrong way" in a massive
field generated by all the other magnets, so there is a very strong
incentive for each of them to turn over. But then once a lot of them
have randomly turned over the situation gets a lot more complicated.
Ideally, the lowest-energy state is where the magnets exactly alternate:
up, down, up, down, up, down all the way through the crystal. However,
in reality there is essentially no pathway for the magnets to "find"
this orientation, so they never do. The system is "frustrated" (and
yes, that is the official term).
So, in a way, you have a lot of reasons to expect that the iron core of
the ferritin will be "disordered", but you should be able to perturb
this "disorder" with a strong enough magnetic field. Not sure how
strong you need, but there will also be a temperature where the magnets
cannot rotate anymore, so perhaps plunge-cooling into the liquid helium
inside a superconducting magnet is what you want to "shoot for". But,
then again, it could be as simple as exposing a crystal to a relatively
"weak" magnetic field at a temperature just above the glass transition
of whatever shares the lumen of the ferritin sphere with the
nanocrystal. Probably somewhere around 130 to 160 K. Warkentin &
Thorne (2009) presented a potentially general way of holding a protein
crystal at any temperature you want between 100 and 300K:
http://dx.doi.org/10.1107/S0021889809023553
-James Holton
MAD Scientist
On 5/8/2012 7:16 AM, R. M. Garavito wrote:
Dear Anna,
I know that you already have gotten replies from some top experts, but
your intriguing problem brought up some issues I have run across in
the past.
First, from you experience with single crystal diffraction, your
results are not that much different from those seen in virus
structures where the nucleic acid structure is averaged out. As the
nucleic acid doesn't (and mostly can't) adopt the symmetry of the
protein shell, the crystallization process alone does the "averaging."
Just because that ferritin and magnetite have cubic symmetry
elements, if they don't line up, the magnetite structure can be
"averaged out" upon crystallization. So, working at lower symmetry
may not help, unless there is some directional correlation of
the magnetite symmetry and position with the crystal axes. But try P1
and see what happens.
A second comment is why not try neutron scattering (SANS or single
crystal neutron diffraction), particularly as you can match out the
protein with D2O and see just the magnetite. While the same concerns
apply for single crystal neutron diffraction, you see more clearly
regions of higher average density inside the protein shell.
And lastly, have you tried crystallizing your ferritin/nanoparticle
complexes in the presence of a magnetic field? It would be a neat
trick, and people have tried such things in the past, such as for
orienting biomolecules. Some even used old NMR magnets. Would be
wild, if it worked.
Good luck,
Michael
/****************************************************************/
/R. Michael Garavito, Ph.D./
/Professor of Biochemistry & Molecular Biology/
/603 Wilson Rd., Rm. 513///
/Michigan State University/
/East Lansing, MI 48824-1319/
/Office:////(517) 355-9724Lab:(517) 353-9125/
/FAX:(517) 353-9334Email: rmgarav...@gmail.com
<mailto:garav...@gmail.com>/
/****************************************************************/
On May 7, 2012, at 12:30 PM, anna anna wrote:
Dear all,
I'd like some suggestions/opinions about the sense of an experiment
proposed by a collaborator expert in saxs.
In few words, he wants to collect SAXS data on a suspension of
protein xtals to investigate "low resolution periodicity" of the xtal
(more details below).
The experiment requires a very huge number of xtals to obtain the
circles typical of saxs and it is very time-consuming to me (I know
nothing about saxs, I have only to prepare the sample). I proposed to
measure a single rotating xtal (like in XRD) but he told they don't
have a goniometer on saxs beamline.
Here is my concern: does it make sense to measure many xtals
together? Don't we lose information with respect to single xtal? And,
most of all, what can I see by /s/axs that I can't see by/w/axs??
Sorry for the almost off-topic question but I think that only someone
who knows both the techniques can help me!!
Some detail for who is intrigued by my story:
we prepared doped magnetite nanoparticles using ferritin as
bioreactor. I crystallized this spheres filled with metal and solved
the structure at 3.7A but I can see only the protein shell while
there is no density inside, even if I know that the nanoparticles are
there. A simple explanation is that the particles are free to move in
the cavity(note that the diameter of the nanoparticle is shorter then
the inner diameter of the protein shell), ie are disordered, and do
not contribute to diffraction, in fact, to my knowledge, nobody have
ever seen the metal core inside ferritin or dps proteins. However,
since they are magnetic particles they must "see" each other through
the protein wall, ie they can't be completely free to move in the
cavity. Maybe, but this is just my opinion, I don't see the particle
because the "period of the particle" in the xtal is different/longer
than the period of the protein shell.
Anyway, we are interested in the relative distance between the
magnetic particles in the xtal to study the effects of magnetostatic
interactions in nanoparticles 3D arrays. We are going to do this by
saxs since, they told me, lower resolution is useful in studying this
long range periodicity (the diameter of ferritin is about 120A) but
it seems fool to me using a suspension of so many xtals to obtain a
scattering curve while I could collect diffraction images from a
single xtal!!! I know that saxs is used when you don't have xtals but
if you have xtals, ie your system is ordered, xtallography is much
more powerful!!
Another question: how can I handle my diffraction data at 3.7A
resolution to "look for" nanoparticles? Should I try a lower
symmetry? Maybe the anomalous signal? Have you any reference for a
similar case?
Thank you very much!!
anna