Crystal Manipulation for Data Collection at Low Temperature

1
Crystal Manipulation for Data Collection at Low
Temperature

• Sean Parkin - Department of Chemistry, University
  of Kentucky

2

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

3
Benefits of Low Temperature
Reduced radiation damage i) Primary dose
dependent ii) Secondary time dependent iii)
Thermal damage Decreased thermal motion (
disorder) iv) Improved resolution limit v)
Possibility of disorder resolution vi)
Sharper electron density Increased crystal
lifetime Full data collection for most
structures can be done on one single crystal.
4
Intensity decay from radiation damage
Haas Rossmann (1970) Acta Cryst. B26, 9981004
5
Visible damage and reduced resolution
Lysozyme diffraction at 100K before and after 22
minutes on beamline 14-BM-C at APS
Teng Moffat (2000) J. Synch. Rad. 7, 313-317
6
Primary damage is dose dependent
J J J J gt J J K L J J J J x-rays L L J J J
J J J gt J J J L J J J J gt J J K J
Some molecules become chemically altered so the
average electron density gets smeared. Primary
radiation damage to H2O produces OH radicals
that can initiate further damage via secondary
events Low temperature has little or no effect
on primary damage.
7
Secondary damage is time (and dose) dependent
J J K L gt J ) K µ L L J J x-rays L L J
J J J J L time J J L J J K J gt J J K J
A cascade of free-radical initiated reactions
destroys long-range order even more, further
smearing the electron density which increasingly
destroys the high resolution data Low
temperature can inhibit secondary damage
8
Thermal damage from very intense sources
J J K L gt J ) K µ L L J J intense L L
J J J J J L x-rays J J L J J K J gt J
N K J
X-ray absorption dumps energy into the
crystal Heating accelerates secondary
damage Non-uniform heating causes temperature
gradients and stress-induced damage A
consideration on insertion device beamlines at
third generation sources. All modes of damage
will compound each other
e.g. Teng Moffat (2000) J. Synch. Rad. 7,
313-317
9
Improvement of resolution
I µ Ioexp-2B(sinq/l)2 T1 300 K T2 100 K,
Bo 5 Å2 b 0.05 Å2K-1 assume B(T) Bo bT
(Bo bT1)/r12 (Bo bT2)/r22 r2 r1 (Bo
bT2)/(Bo bT1)1/2
Roughly three times as much data at 100K vs 300K
Hope, H. (1988) Acta Cryst. B44, 22-26.
10
Reduced thermal motion
General reduction in refined B values. Dramatic
reduction may indicate resolved disorder.
BPTI, main chain
BPTI, side chains
Parkin Hope (1996) Acta Cryst. D52, 18-29.
11
Resolved disorder in favourable cases
BPTI C-terminus, 298K
BPTI C-terminus, 125K
Parkin Hope (1996) Acta Cryst. D52, 18-29.
12
Other benefits
Full datasets are usually obtained from just one
crystal. For MAD data especially, systematic
errors are minimized. Crystals can be harvested
and stored. Important if crystals degrade, e.g.
oxidization of Se-Met. Crystal mounting can be
much less damaging to crystals because of
reduced amount of manipulation.
13

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

14
Prior to cooling the crystal
i) Find a suitable cryoprotectant
Paratone oil, mineral oil, polyfluoroethers 60
success rate Remove surface water
oils
PEG lt 4K increase PEG, add other small PEGs PEG
gt 4K add small PEGs MPD increase MPD
concentration Salt add MPD/glycerol or even more
salt ? exchange salt e.g. to sodium formate n.b.
low salt requires more help than high
salt Modify surface water
antifreezes
Small organics can bind at active sites - not
good !
15
Finding a suitable cryoprotectant
Elspeth Garmans table of minimum amounts of
glycerol needed to prevent ice formation in
Hampton Screen I.
These were arrived at by dilution rather than by
replacement of water, so the numbers should be
used with caution.
Mitchell Garman (1996) J. Appl. Cryst. 29,
584585
16
Finding a suitable cryoprotectant
Structures in Acta Cryst. D57, 2001.
Garman Doublié (2003) Meth. Enzymol. 368,
188-216.
17
Finding a suitable cryoprotectant
Structures in Acta Cryst D56-57, 2000-2001.
Garman Doublié (2003) Meth. Enzymol. 368,
188-216.
18
Finding a suitable cryoprotectant
Garman (1999) Acta Cryst. D55, 16411653
19
Prior to cooling the crystal
ii) Optimize concentration of antifreeze
The minimum required to suppress ice is not
necessarily the optimum amount
Mitchell Garman (1994) J. Appl. Cryst. 27,
10701074
20
Prior to cooling the crystal
iii) Introduction of antifreeze Single or
multi-step soak / wash Sequential partial
exchange of mother liquor Dialysis iv) Length
of soak / wash Seconds to wash, minutes (hours
?) to soak
a) Move crystal between solutions
b) Solution pipetted onto crystal
40
40
30
30
cryo-solution
cryo-solution
20
20
10
10
0
0
0
1
2
3
4
5
0
1
2
3
4
5
Time (minutes)
Time (minutes)
The goal has been to minimize the shock to the
crystal Generally little optimization done unless
there are real problems
Garman (1999) Acta Cryst D55, 1641-1653
21

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

22
Cooling methods - cryogen characteristics
Variable temperature is easy Very
controllable Single step EASY
Stream cool
Fixed temp., 77 K Very controllable Two steps EASY
LN2 dunk
Harder Several steps Variable temp. possible but
tricky Not easy to control without practice
Propane dunk
23
Cooling methods - cryogen cooling rates
N2 stream
liquid N2
liquid propane
N2 stream
liquid N2
liquid propane
Walker, Moreno Hope (1998) J. Appl. Cryst. 31,
954956
Teng Moffat (1998) J. Appl. Cryst. 31, 252257
24
Cooling methods points to consider

• Cooling rate is proportional to DT
• ii) Liquid propane can be dangerous around
  potential sources of ignition
• iii) Due to inherent complexity, liquid propane
  methods are the hardest to make reproducible
• iv) Leidenfrost gas layer insulation of large
  objects is insignificant with ordinary-sized
  crystals
• v) Liquid propane has a large liquid range -
  constant stirring is required for
  reproducibility
• Make the process as simple as possible

25

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

26
Crystal mounting tools
loops - homemade
arcs
tongs
loops - bought
special vials
vials and holders
Pictures Hampton Research Bruker-Nonius Sean
Parkin
27
Crystal mounting aqueous film removal under oil
Aqueous films clinging to the crystal can often
be teased away with a needle point. Or they may
be wicked away with a wedge of pre-moistened
filter paper. Most dry oils will accept a little
water so small amounts will diffuse into the
oil. This may be good or it may be bad.
28
Crystal mounting crystal pick up
29
Crystal mounting manipulations in the dewar
1) Pre-cool tongs, plunge crystal.
2) Clasp the mounting pin.
3) Remove the pin holder.
4) Carry it to diffractometer.
Parkin Hope (1998) J. Appl. Cryst. 31, 945-953.
30
Crystal mounting - Tongs
transfer to diffractometer takes a couple of
seconds
open the tongs so that the cold stream blows in
the gap.
31
Crystal mounting - Cryovials
Note the inverted f axis
means no cryogen spillage.
pictures courtesy of MSC
32
Crystal mounting - robots
SSRL
Automated Fast Reproducible Expensive
BruNo
MSC - Actor
pictures courtesy of MSC, SSRL, Bruker-AXS
33

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

34
Potential problems Control and reproducibility
Crystal environment should be controlled so that
it is reproducible.
Temperature versus time for a "crystal" held in
stainless steel block tongs. Warming rate is
about 0.5 per second (depends on tongs)
Parkin Hope (1998) J. Appl. Cryst. 31, 945-953
35
Potential problems Control and reproducibility
Ensure the crystal temperature is controlled
throughout mounting and that it is reproducible.
mount temp.
dismount temp.
Temperature vs time during mount / dismount
Parkin Hope (1998) J. Appl. Cryst. 31, 945-953
36
Potential problems Ice
Ice caused by inadequate cryoprotectant. Solution
optimize concentration
Ice caused by snow from slushy liquid nitrogen
sticking to the drop. Solution carefully
remount from fresh cryogen, gently tease off the
snow etc.
Pictures Elspeth Garman (Oxford University)
37
More ice problems
1
2
3
A poorly positioned nozzle or draughts will cause
snow to grow on the pin end. This can get
serious if left too long.
Pictures (1) Sean Parkin (2,3) Elspeth Garman
(Oxford University)
38
Potential problems Ice prevention
Even in humid environments ice can be prevented
without elaborate contraptions.

• The important point is a well-defined geometric
  relationship between cold stream, mounting pin
  and goniometer head and to rigorously exclude
  draughts
• Pin design
• Cold stream geometry
• Turbulence
• Exclude draughts

Parkin Hope (1998) J. Appl. Cryst. 31, 945-953.
39
Potential problems Mosaic spread
Minimize mosaic spread to optimise data quality.
It should prove possible to approximate the
mosaicity of crystals at room temperature.
Therefore it helps to know what this is !
Garman, E. (1999) Acta. Cryst. D55, 1641-1653.
Dauter, Z. (1999) Acta. Cryst. D55, 1703-1717.
40
Potential problems maximize cooling rate
1) Keep cryogen close at hand 2) Go for a large
surface area to volume ratio - so small crystals
have an advantage. Generally, S/V gt
12mm-1 e.g. 0.4mm x 0.4mm x 0.4mm block, S/V
15mm-1 0.5mm x 0.5mm x 0.5mm block, S/V
10mm-1 0.4mm x 0.5mm x 0.2mm block, S/V 19mm-1
Thus a plate should cool faster than a rod or a
block.
41
Things to consider
Be in control throughout the experiment The bare
minimum antifreeze concentration needed to
suppress ice formation is probably not the
optimum amount for minimizing mosaic spread and
maximizing resolution. Minimize crystal
handling. Smaller crystals are easier to cool
evenly. Attempt some sort of annealing (next).
Make the mounting and retrieval process as simple
as possible, but not simpler Simplicity leads to
reproducibility
42

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

43
Annealing of macromolecular crystals
Quick and (hopefully not so) dirty approaches
Macromolecular Crystal Annealing
Flash Annealing
Systematic approaches
Controlled slow annealing
Controlled flash annealing
44
1) "Macromolecular Crystal Annealing" - procedure
Crystal is quickly removed
placed in cryoprotectant
and then re-flash cooled.
for three minutes
Harp, Timm, Bunick, (1998) Acta Cryst. D54
622-628
45
Macromolecular Crystal Annealing
Increased resolution and better mosaicity for
nucleosome core particle crystal after 3 minute
anneal in antifreeze or oil
Harp, Timm, Bunick, (1998) Acta Cryst. D54
622-628
46
Macromolecular Crystal Annealing
A flash-cooled crystal of Patatin gave
diffraction to 3.7 Å clearly showing multiple
lattices. On annealing it broke into two pieces.
On remounting, the larger piece diffracted to
2.3 Å.
Hansen, Harp, Bunick, (2003) Meth. Enzymol.
368, 217-235
47
Macromolecular Crystal Annealing
An otherwise trashed nucleosome core particle
crystal resurrected after a 3 minute anneal
Harp, Timm, Bunick, (1998) Acta Cryst. D54
622-628
48
2) "Flash Annealing" - procedure
Flash cooled
stream diverted
re-flash cooled.
The cold stream blowing over the flash cooled
crystal is blocked for a short period of time
(seconds) until it has thawed. Then the
obstruction is released to re-cool the crystal.
Yeh Hol (1998) Acta Cryst. D54, 479-480
49
Flash annealing - results
Glycerol kinase, resolution limit 4 Å, poor
mosaic spread. Flash annealing by blocking the
cold stream for 1.5 - 2 seconds three times
gives 2.8 Å resolution and better mosaic spread.
Yeh Hol (1998) Acta Cryst. D54, 479-480
50
Controlled annealing without thawing
Accomplished in two ways Slow warming using a
controllable stream heater. Rapid warming to
some pre-determined temperature by either
dynamic mixing of cold and warm streams or by
rapid switching of two cold gas
streams. Questions to be answered Is it a
protein or a bulk water phenomenon ? How does
annealing work ? Why does annealing work
? What are the mechanisms of protein crystal
annealing ?
51
3) Slow annealing Cell volume vs temperature
Concanavalin A1
TcAChE (trigonal)2
DV gt 1200 Å3
Crystals with channels show an abrupt volume jump
at some well defined temperature. Implies that
effect is in the bulk water and that there is a
surface effect via the connection to the crystal
surface.
TcAChE (orthorhombic)2
2) Weik et al. (2001) Acta Cryst. D57, 566-573
1) Parkin (1993) Ph.D. Thesis, UC Davis
52
Cell dimension changes on annealing
a axis
b axis
c axis
change relative to 95 K
temperature (K)
There is an abrupt jump in both the b and c axes
in concanavalin A on warming from 160 to 165 K.
There is no corresponding jump in a.
Parkin Hope (2003) Acta Cryst. D59, 2228-2236
53
Origin of annealing effects
(view down c ) c axis expands protein moves
waters move
(view down b ) b axis expands protein moves
waters move
(view down a ) a axis constant - waters move
Parkin Hope (2003) Acta Cryst. D59, 2228-2236
54
Annealing affects diffuse solvent diffraction
Overall background for concanavalin A diffraction
is reduced after annealing. It is also a bit
smoother. Still dont know what the likely
mechanism of annealing is.
Parkin Hope (2003) Acta Cryst. D59, 2228-2236
55
4) Flash annealing without thawing
Rapid adjustment of warm and cold gas flows onto
the crystal. Can be tricky
With two low-temperature machines we can rapidly
switch cold streams. Much easier
Kriminski, Caylor, Nonato, Finkelstein, Thorne
Acta Cryst. (2002), D58, 459-471
Parkin (2002) unpublished
56
Flash annealing seen by in-situ X-ray imaging
At room temp. the whole crystal is in the
diffracting position over a very small angular
range.
After flash cooling, the mosaic spread is much
worse and the resolution limit was severely
degraded to 4.3 Å.
After 25 s controlled anneal at 250 K, resolution
limit is 2.4 Å, mosaic spread generally not as
good as pre-cool value.
Kriminski, Caylor, Nonato, Finkelstein, Thorne
Acta Cryst. (2002), D58, 459-471
57
What happens to the mosaic structure ?
On flash cooling, individual domains have a
mosaic spread similar to that of the whole
crystal. After annealing, small domains have much
narrower mosaic spread, but are themselves
somewhat mis-aligned.
Kriminski, Caylor, Nonato, Finkelstein, Thorne
Acta Cryst. (2002), D58, 459-471
58
What happens to the water ?
a) The distribution of water is fairly uniform in
solvent regions of fresh crystals at room
temperature. b) During flash cooling, water is
squeezed out of small domains and collects in
the regions surrounding the domains. c) Which
leaves the domains somewhat more
mis-aligned. Annealing likely gives a partial
fix by increasing order within domains and by
reducing the spread of lattice spacings within a
crystal.
Kriminski, Caylor, Nonato, Finkelstein, Thorne
Acta Cryst. (2002), D58, 459-471
59
Annealing - General comments

• Consider annealing if diffraction is
  uncharacteristically poor after flash cooling.
• ii) Try the quick methods first. MCA appears to
  be more general.
• iii) For MCA, the crystal must be stable in its
  cryoprotectant.
• iv) Size may be a less important factor than
  shape. Thin crystals may be better suited to
  flash annealing.
• v) In MCA, a full three minutes may not be
  needed. Longer times appear less likely to yield
  optimum results.
• Multiple cycles of MCA are not necessary and may
  be undesirable. For flash annealing, multiple
  cycles may be required.

Hansen, Harp, Bunick, (2003) Meth. Enzymol.
368, 217-235
60

• Benefits of low temperature
• Crystal treatment prior to cooling
• Cooling methods and cryogens
• Crystal mounting etc.
• Potential problems
• Annealing methods
• When nothing seems to work

61
When nothing seems to work (in no particular
order)
Does it diffract at room temperature ? Try other
cryoprotectants Try more than once. Vary time
and temperature of crystal handling steps Match
antifreeze to the system Exchange buffers
etc. Attempt annealing.
62
Does the crystal diffract at room temperature ?
Photograph Elspeth Garman (Oxford University)
Capillary scheme lifted from Practical Protein
Crystallography by Duncan McRee
63
Try more than once.
Picture courtesy Elspeth Garman, after Schneider,
Bravo Hansen
64
Match antifreeze to the system - Osmotic
Pressures

• Find osmotic pressure of mother liquor in the CRC
  Handbook of Chemistry and Physics (section D232)
  11th column, O (Os/Kg)
• Find osmotic pressure of your antifreeze.
• Modify the concentration in mother liquor to
  minimize the change in osmotic pressure.
• Osmotic shock Water will either be pumped into
  the crystal or sucked out of the crystal. Either
  one may cause cracks, increase mosaicity, lower
  resolution etc. All bad.
• Rapid transfers between solutions will give
  greatest shock but will minimize the time over
  which damage could occur.
• Experiment !

65
Osmotic pressure matching - Elspeths worked
example.
Mother Liquor Osmolarity (Os/Kg) 2.0 M
NaCl 50mM pH 7.8 Tris HCl 3.95 Cryoprotectant
20 glycerol 2.90 Difference here is
1.05, so the plan is to alter the cryoprotectant
so that it matches. From the CRC Handbook, 0.55M
NaCl exerts an osmotic pressure of 1.05
Os/Kg. Hence try 0.55M NaCl, 20 glycerol in
50mM pH 7.8 Tris HCl
66
Advantages and disadvantages of low temperature
work.
FOR Reduced radiation damage Gentler
mounting Lower backgrounds Higher
resolution Fewer crystals Transportation is
easy Harvest crystals at their peak.
AGAINST Expense (money) Expense (time) Mosaic
spread increase