Freezing crystal

From: Theresa H. Hsu
Date: 5 February 2012 22:49

Hi all

Is there a list of conditions to be tried *first* for cryoprotectant? My crystals diffract at room temperature capillary but no in 30% PEG 400. Crystals are from 2 M ammonium sulfate.

Thank you.

Theresa

----------
From: Vineet Gaur


Hi Theresa,

Once I had crystals in 3.5 M Amm. Sulfate. I used Paraffin oil and Paraton-N-oil (in 1:1 ratio). I also used 30% Xylatol.  

Best,

Vineet

----------
From: Sean Seaver


Hi Theresa,

Try glycerol at 10 % with 5 % increments up to 30 % v/v.  You may also be able to attain protection by increasing the ammonium sulfate concentration along with incremental glycerol approach.

Crystallization of RNA/protein complexes by Garber and colleagues (Acta Cryst. (2002). D58, 1664-1669) reported using 25% - 30% (w/v) of glucose for cryoprotection with ethanol, MPD, PEG, ethylene glycol, sucrose not being found suitable with concentrated ammonium sulphate.

May also want to consider sodium malonate and DMSO along with the already suggested Paratone-N/Paraffin Oil blends.

DMSO: http://www.bio.brandeis.edu/publications/bi035849h.pdf
Sodium malonate: http://journals.iucr.org/d/issues/2003/12/00/fw5004/fw5004.pdf

Take Care,

Sean Seaver

----------
From: Eric Larson


Hi Theresa,

A good place to start when searching for suitable cryo conditions are the tables in these references:

Garman, et al. J. Appl. Cryst. (1996). 29, 584-587.

McFerrin, et al. J. Appl. Cryst. (2002). 35, 538-545.

hope they help and good luck.

Eric

__________________
Eric Larson
Boehringer Ingelheim
Ridgefield, CT




----------
From: Tommi Kajander

if you use oil do direct dry Paratone-N, with paraffin oil is not as good. Li-salts should

work also - i would almos imgaine you can freeze directly from so high (NH4)2SO4 conc.

but perhaps not. little bit (10%) glycerol probably does it also..

Tommi

----------
From: Mark J van Raaij


if your crystals are from 2 M AmSO4 without buffer, try to measure the pH in the drop, if possible.
Or if you have plenty of crystals, transfer to 2 M AmSO4 buffered at a wide range of different pHs to see where the crystals are stable, before adding cryoprotectant.
In the end, you may need to grow new crystals in the presence of buffer and, if possible, glycerol. The latter could allow direct freezing.
Mark

----------
From: Savvas Savvides

Dear Theresa

Cryo-cooling the crystals straight out of their drops or after brief incubations in crystal stabilization solutions containing >2M ammonium sulfate may be the way to go.

Here is a copy/paste piece from Kyndt et al (2007) Biochemistry 46, 95-105.

All the best

Savvas

Aliquots (0.5 í L) of the seed suspension (diluted 1:100 in stabilization buffer) were

introduced into a series of fresh hanging drops (containing

4 í L of protein sample and 4 í L of reservoir solution) that

had been equilibrated for 24 h over reservoirs containing 3

M ammonium sulfate and 20 mM sodium phosphate, pH

5.4- 6.2. Single crystals with bipyramidal morphology grew

after 1 week to a final size of 0.150 mm   0.150 mm

 0.100 mm.

To prepare crystals for data collection under cryogenic

conditions (100 K), crystals were flash-cooled by plunging

them directly from their native drops into liquid nitrogen. A

series of cryocooling conditions using a variety of cryoprotecting

reagents such as glycerol, sucrose, PEG 400, and

paratone indicated that only crystals flash-cooled by plunging

them directly from their native drops into liquid nitrogen

produced diffraction of acceptable quality.

----------
From: Ed Pozharski

<begin \personal_bias>

Sodium malonate is your friend

http://scripts.iucr.org/cgi-bin/paper?fw5004

<end>

--
Oh, suddenly throwing a giraffe into a volcano to make water is crazy?
                                               Julian, King of Lemurs

----------
From: Kris Tesh

When using oil on protein crystal mounts I suggest:

 

1.  once the crystal is under oil, remove as much adhered solution from the crystal surface.

2.  if there are any volatile components in your drop condition, presaturate the oil with that solvent.

3.  consider using perfluoropolyether since it has a much lower viscosity than ParatoneN, has a very low vapor pressure (used in turbo molecular and diffusion pumps) and is immiscible with just about everything.

 

And remember that some membrane bound crystals will dissolve in oil.

 

Kris

 

Kris F. Tesh, Ph. D.

----------
From: Enrico Stura


Theresa,

Several suggestions that have been given are excellent advice:
Li salts suggested by  Tommi Kajander is what I would use
in particular:
80% saturated lithium sulfate.
This should work. I would be very surprised if it does not.

Malonate as suggested by Sean Seaver is another great idea, but difficult to prepare
see ccp4bb by Doug Ohlendorf :

Malonic acid is dissolved in water and then pH adjusted to the desired value with NaOH. Caution: dissolving malonic acid is highly exothermic. Do it slowly, in a hood.

In general terms what needs to be considered is that
when one introduces a cryoprotectant it is important to
match the precipitating power of the precipitant.
If you just add glycerol, as suggested by some, you will tend to dissolve your
crystals with respect to 2M ammonium sulfate.

"Ideally in a balanced cryoprotectant the mild solubility
enhancing effect of the glycols should be counteracted by
the addition of MPD or DMSO that can act as precipitants.
Therefore, an approach using a complex mixture of
glycols and cryoprecipitants should yield improved
results."

I have assembled a kit initially for my own personal use
that creates mixtures respecting these principles.
CEA Saclay has licenced it to Molecular Dimensions.
The details of the components, the mixtures and their use
are freely available:
http://www.moleculardimensions.com/applications/upload/CryoProtX.pdf - for the product flyer and:
http://www.moleculardimensions.com/shopdisplayproducts.asp?id=201&cat=CryoKits - to get more info and order the product.

What I prefer about Lithium sulfate compared with malonate is that I just transfer the crystals in the 80% saturated Li2SO4
without bothering about the pH (no buffer) and it works. You may want to transfer with a capillary rather than a loop to avoid
shocking the crystals, but for the rest it should give good results.

Enrico.

-- 

----------
From: Prince, D Bryan 

Dear Theresa,

Gary Gilliland's paper on cryosalts would seem to be useful for your problem.

http://scripts.iucr.org/cgi-bin/paper?en0028

Also, I have used 15% glycerol in a synthetic mother liquor to effectively freeze crystals grown in 2M Ammonium Sulfate. Another method that Jim Pflugrath teaches is to use ordinary table sugar (sucrose) dissolved in the reservoir solution to cryoprotect your crystal. A webinar is located on the rigaku website here: http://www.msc.com/protein/webinar-001.html

Also, Hampton Research has a good list of tips in the Tech Support portion of their webpage. www.hamptonresearch.com


Whichever you choose, good luck with your crystals!

Bryan


----------
From: Theresa H. Hsu


Hi all

Thanks for all the suggestions which I will try soon.

How do the crystallization condition (PEG vs. salts like ammonium sulfate) affect the croyprotectant condition? Do factors like presence of low concentration of high molecular weight PEG (> 2000) mean PEG is better? Do buffers and salts in protein also important?

Trying to rationalize it :)

Theresa

----------
From: <Herman.Schreuder


Hi Theresa,

What works well for me, is to use like with like, e.g. to use salts or
salt-like cryoprotectants for salt conditions (e.g. malonate,
li-sulfate, sucrose/xylitol for ammonium sulfate conditions) and
glycerol, low mw PEG etc. for PEG and other alcohol conditions. I have
very bad experiences using glycerol with ammonium sulfate conditions. Of
course, the pH of the cryo-solution should be the same as of the
crystallization conditions to avoid a pH shock.

As mentioned, oils which do not mix at all with the reservoir solution
could be tried in both cases.

In many cases it helps to increase the salt/peg concentration in the
cryoprotectant solution, since under crystallization conditions there is
often an equilibrium between protein in solution and in the crystal. By
increasing the salt/peg concentration, the protein will no longer be
soluble and stay in the crystal. Also often some of the water is pulled
from the crystal lattice, resulting in a tighter packing, which is more
robust versus soaking and which may diffract better (dehydration
effect).

Good luck!
Herman

----------
From: Mark J van Raaij


Rationalising it completely may only be possible once you know the nature of the crystal contacts, i.e. when you have solved the structure. Until then it is mainly a matter of experimenting.

----------
From: Jacob Keller

One last thing--sometimes crystals can be frozen as is, particularly
if you use mitegen mounts and get nearly all of the mother liquor off
the crystals by dabbing the loop on the dry surface next to the drop
several times. So simple it is always worth a try....

JPK

--**

----------
From: Bosch, Juergen

Something to add into this discussion is also go fro the tiny crystals versus the big ones.

BIGGER is not always BETTER - in particular if you try to freeze directly out of your conditions without an additional cryo-protectant. And with small or tiny I mean 10 micron, whatever you are capable of mounting. It is also important to keep the amount of liquid volume around the crystal low, so rather use a loop in which you scoop the crystal up instead of having a large loop with lots of liquid.

Then one last remark, LN2 versus cryo-stream freeze. Dipping in LN2 leads to a quicker freeze of your material.

If you have the option to anneal your crystal after testing it in the beam try it out and assess the success or damage, this will be very different depending on what cryo-additives you have around.

Good luck,

Jürgen

......................
Jürgen Bosch

http://web.mac.com/bosch_lab/

----------
From: Enrico Stura


BIGGER is not always BETTER?

Theoretically it should be better because you have more scattering matter. If it is
not something has gone wrong in prior steps:
Purification: You were less selective and picked up more heterogeneous
protein.
Crystallization: The bigger crystals grew under conditions that were less
controlled because of changes in the state of protein supersaturation, precipitant or
temperature. These inhomogeneities contributed to give you bigger, but not better
crystals.
Cryo-soak: Bigger crystals are more prone to be shocked when transfered to a
less than optimal cryo-solution. This is a critical step, and crystals do not like too
much the cryo-chemicals. To test this I tried lower concentrations of various
cryo-compounds instead of a huge quantity of a single component and in most
cases the mixture was better tollerated than the single components.
Flash-freezing: Bigger crystals will cool more unevenly than small ones. A change from
liquid nitrogen to liquid ethane could achieve faster cooling because of the
greater heat capacity of the latter liquid.

Large crystals are more difficult to handle than small ones but after experimenting with
small ones we can build up a good experimental protocol so that the big crystals will give exceptionally
good results. I compared small crystals on high intensity beamlines at the ESRF against
large crystals on BM30 and the big crystals were statistically better. Unfortunately, it takes
a lot of time and a certain amount of expertize to optimize the conditions. So ...  although I
strongly disagree with Jürgen, I will also advise to start with small ones.

Enrico.>

----------
From: Bosch, Juergen

Hi Enrico,

I was just looking at non-optimal cryo-conditions and the original posters starting point.

Of course if you have a good cryo bigger is better for the reasons you write but if you have no clue how your crystals will perform then I'd rather go for small to be cautious and also have those around and not only the big ones which everybody mounts because they looks so nice. To be disappointed by big crystals is often not a surprise to me and if you have not tried small crystals from the same batch well then you missed 50% of your chances to solve s structure with the first light the crystals saw.

Jürgen

----------
From: Dirk Kostrewa

Dear Jürgen,

Am 07.02.12 16:58, schrieb Bosch, Juergen:
<snip> </snip>

Are you sure? There was a publication by Warkentin et al. [1] about a cold gas layer above liquid nitrogen that reduces the expected cooling rate a lot!
My very personal experience is, that cryo-cooling in the N2-stream worked better for me than in LN2 in a variety of projects - but the reason could just be me ;-)

Best regards,

Dirk.

[1] Matthew Warkentin, Viatcheslav Berejnov, Naji S Husseini, and Robert E Thorne: "Hyperquenching for protein cryocrystallography", J. Appl. Crystallogr., 39, 805-811 (2006)

--

*******************************************************
Dirk Kostrewa

----------
From: Bosch, Juergen

Hi Dirk,

I remember a neat paper don't recall who wrote it. I think it was in Acta D where the authors made a tiny probe the size of an elongated crystal glued to a [/Advertisement on] Hampton loop [/Advertisement off]. The probe was a temperature sensor and they recorded the cooling rate under different methods. The winner as far as I recall was freezing in liquid propane for the lack of the missing gas layer, but the second best method was LN2. Propane for whatever reason has gone extinct in certain areas of the world :-) . I'll try to find that reference but perhaps somebody else on this highly educated board knows which paper I'm referring to. I want to say it was published around 2004-2006.

Jürgen

----------
From: Jim Pflugrath

Just a thought for those that mentioned propane and ethane, I would like to suggest that they try carbon tetrafluoride (CF4) instead.  It certainly should be much safer.  It melts at 90 K and boils at 145 K, so you know you are below 145 K if you see it as a liquid.

----------
From: David Schuller


Yes, Warkentin, et al found the cold gas layer affects cooling rate - but they were able to overcome it by blowing air across the surface. Your point is taken that complications such as this affect the cooling rate much more than a simple liquid vs. stream choice.


-- 

----------
From: Enrico Stura

BIGGER is almost always BETTER if you spend the time to do your experiment well, and in particular
freezing (flash-cooling) your crystal after an appropriate soak in a good cryo-mixture that is respectful
of your crystal.

Thierry,

X-ray absorption, varies with the wavelength, and by going to shorter wavelengths you can reduce
absorption. That will allow you some extra margin before you have a crystal that is too big for
X-ray diffraction. At that point you can go to neutron diffraction and position all your hydrogens
too!!

BIGGER is almost always BETTER!

AND

Jürgen,

I never said that you need to look ONLY at large crystals. I just said that once you have extracted
as much information as possible from your small crystals, (if possible consolidated
a structure at any resolution you can get), by improving your experimental methodology, you should
aim towards BIGGER crystals, better Cryo-conditions and that WILL give you higher resolution.
Refining a structure at 1.6-1.2A will be much easier than doing it at 3A or lower. The time you spend
getting better crystals will save you a lot of time trying to get a decent structure out of bad crystals.

BIGGER is almost always BETTER!

AND

Dirk Kostrewa,

Does stream-cooling beat liquid ethane as well or not?


Enrico.

Enrico: what you state above is only true if you ignore absorption of the X-ray by the crystal.

Thierry

----------
From: Francis E Reyes


I'm going with Jurgen on this one.

http://img27.imageshack.us/img27/3232/pastedgraphic1.png


Sad was the day when I mounted this puppy and it shot to 8-10A. Room temperature. And messing around with cryos didn't help either.


Can't remember the size, but I think I had scooped it with a 0.8 mm loop.

I should've mounted it on a ring and given it to my wife. And that's not chromatic artifact, the ligand was red.

On a side note, I had a very small crystal embedded in a chunk of ice at the end of a 0.025 mm loop.  Couldn't even see it on the very nice on-axis cameras at the ALS. I shot blindly into the ice.. .it diffracted to about 1.8A (and the ice wasn't bad at all)


F
---------------------------------------------
Francis E. Reyes M.Sc.

----------
From: Jacob Keller

But if that puppy had been smaller, it might have diffracted even
worse, and just think if that little icy crystal had been bigger...

JPK
*******************************************

----------
From: Edward A. Berry


Not highly educated, but I remember hearing Haken Hope talk about
this experiment at a workshop at SSRL- cooling a thermocouple or
thermister in the cold stream vs in LN2.
Maybe described here:
Cryocrystallography of biological macromolecules: a generally applicable method.
Hope H. Acta Crystallogr B. 1988 Feb 1;44 ( Pt 1):22-6.

BTW I don't thing the greater cooling rate with ethane/propane is due
to greater heat capacity so much as the fact that LN2 is used at
it's boiling point: heat capacity is irrelevant since it can't absorb
any more heat as a liquid, latent heat of vaporization is the reelevant
parameter, but once it is vaporized the gas has low heat capacity and
thermal conductivity.
The liquid hydrocarbons are prepared by chilling to around their
freezing point (hydrocarbon slush) so can absorb a lot of heat before any gas forms.

----------
From: Theresa H. Hsu

A little off from the original question. Why don't small crystals dissolve to make a bigger crystal, especially when the small ones grow on top of each other? Can the clustered 3D crystals (I think it is called macroscopic twin) be used for full data collection?

Again, thank you.

Theresa