Dear
James,
In
your most comprehensive review of in situ diffraction you
mention the
PX
Scanner: accordingly I am pleased just to make one or two
points.
Whilst
the PX Scanner has been used for ‘data collection’ in one or
two instances,
the
principal application is for determining the unit cells and
diffraction properties
(resolution
limit and mosaicity) of protein crystals and – of course,
for differentiating
between
salt and protein crystals at the earliest stages of
crystallisation optimisation.
In
the PX Scanner, the crystallisation plate is rotated only by
+/- 3° about the
horizontal:
therefore – as you mention, the crystal movement problems
that you refer
to
are avoided. There is, certainly, a ‘lot of stuff’ problem:
but, actually, not ‘‘too much’’.
Thus,
the scattering from the crystallisation plate and the mother
liquor is relatively
non-changing
between each of the slices / images that are collected
across the 6° range.
With
this, the sharp Bragg reflections really stand out very
well. As you comment,
ever-improving
‘X-ray friendly’ plates are being introduced. Inevitably,
though, with the background ‘issue’ the image quality from
the PX Scanner can never be as bright or
speedy
as from a mainstream diffractometer, or synchrotron !!
However, there is a
monotonic
relationship between the resolution limits as observed on
the PX Scanner
compared
to those observed at synchrotrons, etc. (Perhaps just a 1 Å
difference in
some
‘real’ examples.) Therefore, with experience, one can
easily judge if one’s crystals
are
‘synchrotron ready’ – and that the cryoprotectant is
compatible, etc ... from these
convenient
‘at home’, in situ, and largely automated PX Scanner
measurements.
Very Best Regards,
Marcus Winter
(Agilent Technologies)
From: CCP4 bulletin board [mailto:[log in to unmask]]
On Behalf Of James Holton
Sent: 13 June 2012 08:35
To: [log in to unmask]
Subject: Re: [ccp4bb] Are these xtal conditions
worth optimizing?
Having some experience with in-situ diffraction I feel like I
should mention the two major technical challenges you will
most likely encounter if you attempt it:
1) your tray cannot be rotated 90 degrees without "disturbing"
your crystals
2) there is no path through your tray that doesn't put WAY too
much "stuff" in the beam
Note that neither of these things is a property of your
crystal or the x-ray beam, but rather a "feature" of your
tray. What I mean by "disturbing your crystals" is that your
reservoir solution sloshes over them if you turn the tray 90
deg, or the drop deforms with gravity, dislodging the
crystals, and sometimes even the drop itself will slide out of
place.
What I mean by "too much stuff" is that the path the x-ray
beam must take in order to get to your crystal, diffract, and
then get out the other side of the tray and find the detector
is almost always much thicker than the path through the
crystal itself. Why does this matter? Because plastic,
water, oil, buffer, etc. all generate about the same amount of
x-ray background per micron of beam-traversed thickness as
your protein crystal does. Absorption per unit thickness is
also similar for these substances, but silica (aka glass)
scatters and absorbs about 10x more than a protein crystal of
the same thickness. The long and short of all this x-ray
physics is that once the thing you have "mounted" your crystal
in starts to scatter and absorb more more than the crystal
itself you start to notice your apparent "resolution"
degrade. Or at least degrade relative to a nice thin loop.
Yes, there is a reason why cryo loop mounts are so popular!
The loop and drop within it generally introduce about the same
thickness of "stuff" to the beam as the protein crystal, and
once the thickness of "stuff" gets thinner than the protein
crystal you see no more improvement in "resolution". This is
because the diffuse scatter from the protein crystal itself
starts to dominate the background. Funny how we tend to
converge on an optimal solution without really understanding
why, but we have.
True, you don't need an optimal system to see relative
differences between crystals or just to check if something is
"salt or protein", but when the tray is a hundred times
thicker than the crystal (say 1 mm of tray vs a 10 micron
thick plate or needle), then it is quite possible you just
won't see any spots at all. This is especially true if the
x-ray beam is wider than the crystal because, with trays, the
background increases with the AREA of the beam (square of the
beam size). That is, as long as your 10x10 micron crystal is
inside a "slab" of plastic or water, a 100x100 micron beam
will generate 100x more background than a 10x10 micron beam
after putting the same number of photons on the crystal (same
spot intensity, damage, etc.). For those keeping score, your
10x10x10 micron crystal in a 1 mm thick tray hit with a
100x100 micron beam will have to compete with 10,000x the
usual background. You can try going to smaller beams, but for
even moderately thick trays parallax effects make it very hard
to hit small crystals with a small beam, especially if the
tray rotates. This is because visible light is refracted by
the plastic, but x-rays are not. In fact, you even see
parallax effects with loop mounts! The drop containing your
crystal is, after all, a lens. Ever noticed how the crystal
never seems to be "centered" at every angle? That's
parallax. You might think you can get away with not rotating
the tray, but then you have to deal with the fact that a
"still" of a salt crystal can also give no spots. This is
because there is a good chance none of the hkls from the small
unit cell of a salt will be on the Ewald sphere.
So, it is more challenging than you might think! But the
advantages of in-situ diffraction are great enough that many
people (including myself) have set out to try and solve these
problems. In my case the result was the Fluidigm "Topaz(R)
DC" model chip:
http://www.fluidigm.com/home/fluidigm/docs/Datasheet_1.96DC.pdf
which has the same background as a 200 micron loop mount (100
microns of sample and 100 microns of plastic, (after you
remove the cover!). The next-lowest x-ray path tray I am
aware of is the Nat X-ray "CrystalQuick X" model with 300
microns of plastic plus the drop. Then again, the MiTeGen
"MicroRT(TM)" system also has the same background as a loop,
but does involve a bit more manipulation than a "true" in-situ
mount. Please note that I do not get kick-backs from any of
these companies and I don't make any money if you buy their
products, but I have spent a good deal of time building an
in-situ diffraction setup at my beamline and it is frustrating
to me that almost nobody uses trays that are good for in-situ.
I think the reason for this is not just because "diffraction
capable" trays tend to be more expensive than thicker trays,
but because thin layers of plastic are alarmingly
water-permeable. So, people tend to get better results with
(and therefore prefer) trays that "hold their water".
Unfortunately, after getting hits many people find themselves
in the situation of the original poster, and would like to try
in-situ, only to find:
3) you cannot reproduce your crystals in a new tray format
I think #3 is the main reason in-situ is not tremendously
more popular than it currently is. As I understand the PX
scanner, it was designed to address problem #1 (above), but
there is nothing it can do about problems #2 and #3. You can
try to "subtract the background", and that can help you see
spots that are not buried in the noise, but no amount of
computer trickery can remove the noise itself. That would
violate the second law of thermodynamics. The only way to get
rid of photon-counting noise is to not make it in the first
place, and that means thinner trays.
Yes, thinner. There really is no other way around it. I have
seen a few claims of "x-ray transparent" or "low scattering"
plastics, and I have evaluated dozens of them myself, but you
have to be careful to normalize for x-ray absorption when you
do this. If you just go by "counts on the detector", then you
might convince yourself that the best material to make trays
out of is lead. Turns out that if you do normalize for
absorption then pretty much everything made of carbon,
nitrogen, oxygen etc. at densities near 1 scatter about the
same. The choice of material allows you to push those
scattered photons around in reciprocal space, but they do all
have to go somewhere.
I have searched high and low for an ultra-low water
permeability material that is not made of high atomic number
elements, but near as I can tell there just isn't one. Except
perhaps diamond, but I don't think anyone would buy those
trays.
-James Holton
MAD Scientist
On 6/8/2012 7:15 AM, Prince, D Bryan wrote:
If
you’re looking for in situ screening for your
crystallization experiment plates, I would direct you to the
following link http://rigaku.com/products/protein/platemate
where you can mount your 96-well screening plate directly on
your home source beam and quickly tell if the drop in
question is worth pursuing. With respect to full disclosure,
a colleague at AstraZeneca in Alderley Park designed the
original prototype.
Hope
this helps!
Regards,
Bryan
Confidentiality
Notice: This
message is private and may contain confidential and
proprietary information. If you have received this message
in error, please notify us and remove it from your system
and note that you must not copy, distribute or take any
action in reliance on it. Any unauthorized use or disclosure
of the contents of this message is not permitted and may be
unlawful.
From:
CCP4 bulletin board [mailto:[log in to unmask]]
On Behalf Of Marcus Winter
Sent: Friday, June 08, 2012 9:18 AM
To: [log in to unmask]
Subject: [ccp4bb] Are these xtal conditions worth
optimizing?
Dear
Artem,
Thanks
for your reply. However, with respect to the CCP4BB, I’ll
reply to
you
offline on this second occasion...
Very
Best Regards,
Marcus.
From:
Artem Evdokimov [mailto:[log in to unmask]]
Sent: 08 June 2012 13:43
To: WINTER,MARCUS (A-UnitedKingdom,ex1)
Cc: [log in to unmask]
Subject: Re: [ccp4bb] Are these xtal conditions
worth optimizing?
Hi Marcus :-)
As per our previous exchange - the scanner is a cute machine
but in my opinion the cost/benefit ratio is not right.
On a more philosophical note I honestly think that what "we"
(that is macromoleclar crystallography community) really need
is a radically new way to grow crystals. The existing spectrum
of tools still fails too much and is still too reliant on
chance. If you have a new tool that can reliably and
significantly improve odds of crystallization for a given
protein - let me know. Yes, we have all the current best
practices going and also are trying a number of the more
exotic techniques too - but it is not enough.
Artem
On Jun 8, 2012 1:39 AM, <[log in to unmask]>
wrote:
Dear
Artem,
Thanks
for your reply. Indeed, as you describe there’s much
value in the
approach
that you describe. The principle of the PX Scanner is
that the
crystal
is left entirely undisturbed in the mother liquor: the
seal is left intact.
Yes:
PX Scanner users have noted many cases where addition
of the
prospective
cryoprotectant immediately diminishes the diffraction
properties
of
the unfrozen crystal.
-
If you’d like to evaluate the PX Scanner for your
purposes, then of course
you’d
be most welcome...
Best
Regards,
Marcus.
From:
Artem Evdokimov [mailto:[log in to unmask]]
Sent: 08 June 2012 03:19
To: WINTER,MARCUS (A-UnitedKingdom,ex1)
Cc: [log in to unmask]
Subject: Re: [ccp4bb] Are these xtal
conditions worth optimizing?
An
easy alternative is to scoop 'hopefully crystals' into
a loop via a layer of oil (but keep the mother liquor
around the crystal) and shoot them without
cryo-cooling. This ensures that cryo is not a factor
(in a very nasty recent case in our lab, it was) -
while also allowing you to avoid capillary mounting
(which is an awesome and fun thing to do, but not when
you want to test more than a few crystals). You only
need a few shots anyway (since it's really unlikely
that the stuff would actually diffract) and the
live-oil mount allows you to re-close the drop really
quickly thus hopefully saving the remainder of the
crystal-like artifacts to be studied later should they
prove to be interesting...
Artem
On
Thu, Jun 7, 2012 at 12:46 AM, Marcus Winter <[log in to unmask]>
wrote:
Dear
Eddie, Christine,
In
determining the diffraction qualities of
(putative) protein crystals directly
– in
situ,
in the crystallisation plate, and for the
optimisation of crystallisation conditions,
the
(home laboratory) Agilent Technologies PX
Scanner instrument can be extremely
powerful,
as you may be aware. Any SBS format plate is
compatible, along with other
technologies.
Yours
sincerely,
Marcus
Winter (Agilent Technologies)
From:
CCP4 bulletin board [mailto:[log in to unmask]]
On Behalf Of Edward Snell
Sent: 06 June 2012 23:02
To: [log in to unmask]
Subject: Re: [ccp4bb] Are these xtal
conditions worth optimizing?
Hi
Christine,
Not
to blow one’s own trumpet but check out
What’s
in a Drop? Correlating Observations and
Outcomes to Guide Macromolecular
Crystallization Experiments in Cryst. Growth
and Design, 2011, 11(3) 651-663. Luft,
Wolfley and Snell. http://pubs.acs.org/doi/abs/10.1021/cg1013945
I’ll
send a copy to you but cannot send it to the
whole ccp4 list. Basically, if the phase
separation (left image) is protein (check
with UV if you have Tryptophan) then these
are very close to crystallization conditions
and temperature can be used to get nice
crystals. Be careful not to shift
temperature too much otherwise you will
bring the Ammonium Sulfate out of solution.
Right image may also benefit from
temperature. There is a discussion about
which way to go in the paper.
Cheers,
Eddie
Edward
Snell Ph.D.
Assistant Prof. Department of Structural
Biology, SUNY Buffalo,
Senior Scientist, Hauptman-Woodward
Medical Research Institute
700 Ellicott Street, Buffalo, NY
14203-1102
Phone: (716) 898 8631
Fax: (716) 898 8660
Skype:
eddie.snell Email: [log in to unmask]
Telepathy:
42.2 GHz
Heisenberg was probably here!
From:
CCP4 bulletin board [mailto:[log in to unmask]]
On Behalf Of Harman, Christine
Sent: Wednesday, June 06, 2012
5:53 PM
To: [log in to unmask]
Subject: [ccp4bb] Are these xtal
conditions worth optimizing?
Hi
All,
I
have these very weird drops that I found
from screening (please find pictures
attached). I am not sure if they are
worth optimizing. I am very interesting
to know your opinions of what you think of
these drops could be (are these what you
call "spherulites"?). If you think these
conditions are worth optimizing, I welcome
any ideas on how to optimize. I have done
some optimization and still get the same
result which is many of these weird things
growing throughout the drop and with no
sharp edges, and sometimes a skin forms
after 2weeks (with the sodium malonate
conditions only). I have also opened the
drop and poked around to see if it is
phase separation and this things are
definitely solid and slightly-very mushy.
I haven't had a chance to check for
diffraction, but will be very soon. Both
of these drops contain the same protein
preparation of a Fab/peptide complex @
~5mg/mL in buffer containing 0.1M Sodium
Acetate pH 5, 150mM NaCl. I appreciate any
advice, thoughts or comments that you
could provide.
Peace,
Christine