Sunday, 9 October 2011

Neutron data collection

From: REX PALMER
Date: 21 September 2011 10:52

Re Neutron Data Collection:
1. What are the limits to data set completeness imposed by a Laue experiment versus those of monochromatic data collection?
2. What problems are caused by flash freezing the larger protein crystals used for neutron data collection which do not occur for X-ray data collection ie because smaller crystals can be used.
Any help will be greatly appreciated.  


----------
From: <mjvdwoerd


Rex,

There are people more qualified to answer your question 1 than I am, so I am going to politely defer that answer. The answer depends on the unit cell dimensions, detector distance etc, and yes, there are more observations rejected due to overlap than would be the case in monochromatic data collection. As for 2, you should not freeze your crystals but mount them the old-fashioned way in capillaries. In practice neutron diffraction does not cause radiation damage to your crystals so you should not freeze and collect data as much as your time allotment allows for.

Hope this helps.

Mark van der Woerd

----------
From: Brad Bennett
Hello Dr. Palmer-
There may be other representatives in the literature by now but the one study I know of that examines the usefulness and limitations of determining "cryo"-neutron structures is "The 15-K neutron structure of saccharide-free
concanavalin A", Blakeley et al. (2004), PNAS, 47(101):16405-16410. There's a brief section there on the practicality of cooling such large crystals. Mark is right of course in that the vast majority of neutron data sets have been collected at RT, simply because neutrons are "soft" probes and non-damaging to the crystal, so there's no real need for cryocooling. However, the reason the authors give in the PNAS paper for collecting the data at cryo temps (mind you, this is helium and not nitrogen) is to help identify more solvent molecules in the sugar binding site. It also decfeased the B-factors for most of the atoms, too. This is severely paraphrasing the paper but hopefully you can take a look and see if it helps.

Best-
Brad 

----------
From: Sean Seaver
Dear Rex,
Laue allows for a greater number of Bragg reflections to be measured compared to monochromatic data collection over a give time period.  The limiting factor in neutron crystallography in regards to data completeness is predominately collection time and instrumentation (detector size).
In my hands, flash cooling becomes more difficult as the crystal size increases due to the volume that must achieve the glass transition state (must be fast and well centered in the cryogenic stream).

I would reach out to the beamline scientist, if you are serious considering collecting at a macromolecular neutron beamline using cryogenic temperatures.  Most data collections with neutrons are measured in days with some up to a month and don't believe any neutron beam lines have an automated refill system of liquid nitrogen.  I hope that helps.

Take Care,

Sean Seaver

P212121
http://store.p212121.com/

----------
From: David Schuller
With X-rays, Laue diffraction leads to some systematic overlap as reflections from different wavelengths fall on the same detector position, and this cuts into completeness.

With neutrons, it is possible to use a time-resolved detector such that all events are time-stamped, and the reflections from lower energy neutrons do not overlap with those of higher energy neutrons (neutrons having measurable mass, and thus noticable velocity differences).  I know that this is possible, I do not know whether it is commonplace.

See, for example:
Protein crystallography with spallation neutrons: the user facility at Los Alamos Neutron Science Center (2004) P. Langan, G. Greene & B.P. Schoenborn, J. Appl. Cryst. 37(1) 24-31.


--
=======================================================================
All Things Serve the Beam
=======================================================================
                              David J. Schuller
                              modern man in a post-modern world
                              MacCHESS, Cornell University
                             

----------
From: Jacob Keller
Wow, neutrons are pretty cool! No radiation damage--and time
resolution? I guess this is since they have much higher energy, and
are measurable individually? What are the numbers for fluxes
(neutrons/sec)? Are the neutrons all at one energy, or is there a
bandwidth?

JPK
--
*******************************************
Jacob Pearson Keller
Northwestern University
Medical Scientist Training Program
*******************************************

----------
From: Murray, James W
Actually, as calculated by Richard Henderson in 1995, there is non-negligible radiation damage from neutrons due to infrequent but energetic nuclear reactions. The reason that radiation damage by neutrons is not observed in practice is that neutron sources are so weak.

The potential and limitations of neutrons, electrons and X-rays for atomic resolution microscopy of unstained biological molecules.
Henderson R.
Q Rev Biophys. 1995 May;28(2):171-93.

best wishes

James

--
Dr. James W. Murray
David Phillips Research  Fellow
Division of Molecular Biosciences
Imperial College, LONDON

_________________________________


----------
From: Jacob Keller
Yes, I have read that paper (a seminal one and the source of the
"Henderson limit," no?), and saw that the best "deal" is electrons, I
think, but I was just delighted to learn that it doesn't happen in
practice. As I recall, x-rays are the worst deal?

JPK

----------
From: David Schuller
Maybe we should be using neutrinos, in hopes of getting some data _literally_ before there is any damage.

----------
From: Andreas Ostermann

The energy of neutrons is even lower when compared to X-rays.
A neutron with a wavelength of 1.8A has an energy of about 25 meV.
The flux at neutron sources compared to synchrotrons is unfortunately low:

Diffractometer "LADI III"  reactor ILL/France:
 3 x 10^7  neutrons/sec/cm^2  (quasi-Laue, delta L / L = 20%)

Diffractometer "BioDiff"  reactor FRM II / Germany:
 1 x 10^7  neutrons/sec/cm^2  (monochromatic, delta L / L =  2.5%)

Diffractometer "BIX4"  reactor JRR3M / Japan:
 4 x 10^6  neutrons/sec/cm^2  (monochromatic, delta L / L =  2.0%)

BUT you can detect hydrogen atoms even at a moderate resolution of
about 2A ! With neutrons the scattering power of hydrogen/deuterium
is "comparable" to the scattering power of carbon. You can even distinguish
between isotopes.  Since the nucleus is a point scatterer the "form factor"
-for neutrons called scattering length- is not scattering angle depended.
A typical measurement time is about 2-3 weeks for a crystal of 1 mm^3.
I know...of course not every protein can be crystallized up to 1 mm^3 but
if you have such a system and you are interested in the protonation states
of amino acids in the active centre for example, than neutrons are worth a try
for sure! If you fully deuterate your protein (which gets more and more routine
work for example at the D-LAB at  ILL/EMBL) you can even work with smaller
crystals.

Because of the relative low flux most reactor based neutron diffractometers for
proteins uses large cylindrical neutron image plate detector, which cover a solid angle
of about 2 Pi. At spallation sources (which are pulsed neutron sources) detectors
with time resolution are used. This instruments (PCS in Los Alamos; iBIX in Japan
and MANDI in Oak Ridge) are time of flight instruments. They uses the fact that
neutrons with different energy/wavelength show different velocities ( a 1.8A neutron
has a velocity of about 2200 m/s). They measure different wavelength neutrons at
different time at the detector.

Hope to see some of you as new "neutron users" in the future,
cheers,

Andreas

--
Dr. Andreas Ostermann
Technische Universität München
Research reactor FRM II
Instrument "BioDiff"
Lichtenbergstr. 1
D-85747 Garching

Web: http://www.frm2.tum.de/en/science/index.html

----------
From: Jacob Keller
That value, 2200m/s, is pretty slow--there are some bullets that go
faster than that, I think...

JPK


From: Leif Hanson

This thread caught my attention several days ago and I now have enough time to add my two cents worth. These are my own biases and probably do not reflect the views of my friends and colleagues at various neutron facilities.

With respect to the size of crystals for neutron diffraction, a good rule of thumb is that there should be at least 10exp24 uniformly ordered unit cells in a D2O exchanged crystal to have successful diffraction on par with rotating anode data measured on a crystal with a tenth the volume. Several data sets have been measured from smaller crystals, and perdeuteration lowers the volume needed to extract useful information. Most of the neutron data has a resolution cutoff of 1.8 to 2.0Å, which permits unambiguous placement of deuterons and solvent molecules, especially when completing dual refinement of X-ray and neutron data from the same crystal.

There have been a limited number of low temperature neutron diffraction experiments for several reasons. First, of the available neutron beamlines for macromolecular data measurement there are only one or two with open flow cryostats available, limiting the locations for standard macromolecular cryocrystallography. Second, there are a tremendous number of important structures that can be done at room temperature. It is difficult to justify the time needed for low temperature work to experimental review panels when crystals are available to resolve a knotty enzyme mechanism problem. Third, the size of crystal needed for successful neutron diffraction is right at the limit of the size of crystal that can be successfully flash-cooled without inducing excess mosaicity. Most neutron beamlines use some form of quasi-Laue data collection strategy. Mosaicities in excess of 0.5º render most crystals unusable for neutron data measurement. Remember that a lot of uniform unit cells are needed to get a usable diffraction signal from neutrons. Often a large flash-cooled cooled crystal appears to have low mosaicity when exposed to 0.5mm x-ray beam. However, when placed in the 3mm neutron beam, limited streaky low-resolution diffraction appears. It is difficult to judge the quality of flash-cooled neutron diffraction sized crystal without placing it in the neutron beam. Returning to point 2 it is difficult justify the time needed on fishing expedition. So far the only large crystals I have been able to flash-cool that met the demands of size and crystal perfection had very low solvent content or were grown in high levels of cryoprotectant. That said, several critical problems cry out for low temp neutron studies so there is every reason to persevere. I would be pleased to answer any questions off-line for those of you with more interest in neutron cryocrystallography.

Finally with respect to radiation damage, Benno Schoenborn has had a myoglobin crystal in sealed capillary that he has used as a “standard candle” for testing neutron beamlines. There has been no discernable degradation of the crystal in all the years he has used it. The neutrons used for neutron diffraction are ‘cold’ neutrons, usually with energies of 1 – 10 meV. Damage could come from activated nuclei, but these are usually very limited on a molar basis within the crystal. As can be seen with Benno’s myoglobin crystal, 30 years of iron activation has yet to produce a measurable defect.  

Leif Hanson
University of Toledo

----------
From: Patrick Shaw Stewart
Can't the crystal be flash-cooled at high pressure? The inventors of the Crystal Harp in Zurich use a machine that does this automatically for cryo e.m., which many universities already have.




No comments:

Post a Comment