crystallization methods and screening · 0 5 10 15 20 25 30 35 40 45 50 0 2 4 6 8 10 12 14 16 18 20...

Juan Manuel García-Ruiz

Laboratorio de Estudios CristalográficosCSIC-Universidad de Granada, Spain

[email protected]

Crystallization Methods and Screening

Crystallization methods

water removal, either by: Evaporation. Dialysis.

solubility change driven by: Temperature pH Dielectric constant Ionic strength Polymers

Because of lack of material protein crystallization techniques are microscale or nanoscale design

A. McPherson, Crystallisation of Biological Macromolecules, Cold Spring Harbor Laboratory Press, 1999.A.Ducruix and R.Giege (eds) Crystallisation of Nucleic Acids and Proteins: A Practical Approach, IRL Press at Oxford University 1999Methods in Enzimology, volume 368 (2003)Journal of Structural Biology, volume 142 (2003)Proceedings of the ICCBMs at Acta Crystallographica (2001 and 2003) and Journal of Crystal Growth (1998)

Solvent is water (plus detergent in membrane protein crystallisation)

Low reproducibilityLack of understanding

Crystallizability: the propensity of a compound to crystallize, i.e. to enter the a7rac8ve regime to make ordered arrays of their molecules, both with orienta8onal and transla8onal symmetry.

Crystallizability depends on the proper8es of the macromolecule itself and on the chemical cocktail used to bring there into the a7rac8ve regime, including ligands that are assumed to help crystallizability.

However, the way to mix the macromolecules and the molecules of the precipita8ng cocktail, the rate at which they mix and how fast supersatura8on is achieved, depends on the crystalliza,on technique.

Thus, the output from any crystallization experiment depends, in some extent, on the crystallization technique used. This is particularly true for screening

Vapour diffusion technique

Advantages

Drawbacks

• Easy to understand by the users• Simple to implement• Inexpensive• Robo8c available for H-‐T structural genomics• Crystals can be handled for cryoXtallography

• The solu8on moves from equilibrium to far from equilibrium• Ambiguous interpreta8on of drop’s crystalliza8on phenomena • Problems of reproducibility and scale-‐up• Each drop experiment screens a narrow crystalliza8on space (measured by visited values of supersatura8on and supersatura8on rate)

Juan Ma. Garcia-‐Ruiz XII InternaBonal Conference on the CrystallizaBon of Biological Macromolecules Cancun, 6-‐9 of May, 2008

Batch method

Direct mixing of protein and precipitant

Protein Precipitant

Trial is blind• Super(under)saturation is immediately achieved• Screening performed by repeating a number N of experiments using different initial conditions • Higher the number N of experiments higher the possibility of success• Number of visited crystallization conditions per trial: One supersaturation value, One (inLinite) rate of

supersaturation

Oil thin layer

Slow evaporation

0 5 10 15 20 25 30 35 40 45 500

2

4

6

8

10

12

14

16

18

20

E'

0'

Understurated zone

Metastable zone

Labile zone

0

E

b) Vapour diffusion experiment

Mac

rom

olec

ule

conc

entra

tion

Precipitating agent concentration

Evaporation of water increases the concentration of precipitating agent and protein at the same rate.

and the concentration of protein decreases

As a result protein precipitates hopefully as crystals …

Vapour diffusionHanging/Si1ng drop

Transport of water molecules

Crystallisationin the drop

Protein solution + precipitant at low concentration

Precipitant at high concentration(usually twice the one in the drop)


Note that the soluBon moves from equilibrium to out of equilibrium:It means iniBal condiBons and/or rate of evaporaBon need to be tuned

Experimental set-up

A Mach-‐Zehnder interferometer study

mirror 0

mirror 1

mirror 2

sample

633 nm

Laser

filter

collimator

Beam spliVer1

Beam spliV

er2

3 mm

0.4

mm

The drop was sandwiched between two glass plates with a gap of 0.4 mm.

Precipitant solu7on

T = 20ºC


Low concentration =Low density solution

Dynamics of a crystallising dropA demonstration of the role of gravity

Vapour diffusionHanging/Si1ng drop

Growing crystal

Mach-Zehnder interferometry

Crystals sedimentate

Crystals nucleate at the surface of the drop

Convective flow motion


Nucleation frequency J is defined as the number of stable nuclei forming per unit of time and unit of volumen. It is given by:

Scale-‐up problems

The most common precipitants used in macromolecular crystallization

Salts Volatile organic compounds

Polymers Non-volatile organic solvents

1. Propanol and isopropanol

2. 1,3-propanediol3. Dioxan4. Acetone5. Methanol6. Ethanol

1. Poly(ethylene glycol) 1000, 3350, 6000, 8000, 20000

2. Jeffamine3. Polyamine

1. 2-methyl-2,4-pentanediol2. Ethylene glycol 400

1. Ammonium phosphate and sulfate

2. Lithium sulfate3. Sodium or potassium or

ammonium chloride, phosphate, acetate, tartrate

4. Magnesium or calcium sulfate, chloride

5. Sodium or magnesium formate

The Yirst step of a Protein Crystallization experiment is always mixing The way to mix the macromolecules and the molecules of the precipitating cocktail, the rate at which they mix and how fast supersaturation is achieved, depends on the crystallization technique.

The chemical cocktail is in most cases Protein + water +

buffer + precipitant

Problem: Low reproducibility; Same chemistry yields different results within a single drop withing chemically identical drops: Lack of understanding

Precipitant

Protein

Density of protein soluBon = density precipitant soluBon

Protein heavier

Protein lighter

RelaBve viscosiBes also mater

¿Precipitant over protein or protein over precipitant?

Problems of Reproducibility

Case 1: An interferometric video corresponding to the experiment of pouring a lighter NaCl solution (1%) over a denser Lysozyme solution (30 mg/ml). The diffusivity of the protein molecules towards the upper salt solution is slower than the diffusivity of the salt molecules towards the lower protein solution. This triggers a convective mechanism that enhances mixing. Time scale, 2 frame per second.

Case 2: An interferometric video corresponding to the experiment consisting of pouring a NaCl solution (2%) over a lighter Lysozyme solution (30 mg/ml). Notice that the system becomes turbulent as a result of the development of Rayleigh-‐Taylor instability and the cocktail becomes homogeneous in less than one minute. Time scale, 2 frame per second.

Case 3: An interferometric video recording an experiment consisting of pouring a lighter lysozyme protein solution (30 mg/ml) over a NaCl solution (3%). It is clearly observed the phenomenon of double-‐diffusive salt Ringering instability. The salt molecules diffuse faster into the pockets of proteins than the protein molecules into the pockets of salts. That simple mechanism creates density differences between the pockets and the surrounding solution that provokes the formation of upwards salt-‐rich Ringers and downwards protein-‐rich Ringering. Time scale, 2 frame per second.

Case 4: An interferometric video corresponding to an experiment consisting of pouring a Lysozyme protein solution (60 mg/ml) over a NaCl solution (9%). The interferometric mode switch to normal optical illumination between 27 and 34 seconds. Then, it can be clearly noticed that due to high concentration of the solutions, the Rluid is mixed by drag of the precipitated protein. Time scale, 2 frame per second.

t = 25.0 s t = 34.0 s t = 64.0 s t = 96.0 s t = 153.0 s

t = 0.0 s t = 3.0 s t = 10.0 s t = 16.0 s t = 20.0 s

Time sequence of interferograms. A solu,on of Lysozyme 30 mg/ml poured on top of a solu,on of NaCl 2% m/v.

Protein lighter more viscousNaCl heavier less viscous

As a pracBcal corollary of the results, it is recommended to perform mixing when seeking for reproducibility while avoid mixing when seeking for a more complex and extensive screening of the crystallisaBon space.

To Mix or not to Mix. Is that the ques7on?

From: On the mixing of protein crystallisa,on cocktailsEduardo I. Howard, José Miguel Fernandez and Juan Manuel Garcia-‐RuizSubmi]ed for publica,on

Batch Method

Batch Method no mixing = liquid-‐liquid diffusion

Capillary counter diffusion


Requirements : Convection must be removed One reactant must flow faster than the other one It must be room to display the spatial pattern Initial conditions far from equilibrium

A precipitant chamber A long protein chamber

A physical buffer (optional)

The counter-‐diffusion technique

Based on diffusion–reaction patterns forming when two reactants are allowed to diffuse one against the other under initial conditions very far from equilibrium*

How do counter-‐diffusion works?

* J.M. Garcia-Ruiz, Counterdiffusion method for protein crystallisation. Methods in Enzimology 368 (2003) 130-154

tDdxerfcccppt

pptppt 221 0 +

= tDxerfcccp

pp 221 0 −

=The diffusion step: with Dppt >> Dp

The precipitation step is governed by supersaturation with respect to the solubility curve of the protein


Counter-‐diffusion technique:How do it works?

Precipitant

Typical counterdiffusion crystallization patterns

Protein concentration

Precipitant concentration

Precipitant

Precipitant


A supersatura8on wave for protein crystallisa8on

Numerical simulaBon of the counter-‐diffusion technique

J.M. García-Ruiz, F. Otálora, M.L. Novella, J.A. Gavira, C. Sauter and O. Vidal. J. Crystal Growth 232 (2001) 149-155

F. Otálora and J.M. García-‐Ruiz. Computer simulaBon of the diffusion/reacBon interplay in the gel acupuncture method. Journal of Crystal Growth 169 (1996) 361

F. Otálora and J.M. García-‐Ruiz. Computer simulaBon of the diffusion/reacBon interplay in the gel acupuncture method. Journal of Crystal Growth 169 (1996) 361

Requirements : Convection must be removed One reactant must flow faster than the other one It must be room to display the spatial pattern Initial conditions far from equilibrium

A precipitant chamber A long protein chamber

A physical buffer (optional)

The counter-‐diffusion technique

Based on diffusion–reaction patterns forming when two reactants are allowed to diffuse one against the other under initial conditions very far from equilibrium*

tDdxerfcccppt

pptppt 221 0 +

= tDxerfcccp

pp 221 0 −

=The diffusion step: with Dppt >> Dp

The precipitation step is governed by supersaturation with respect to the solubility curve of the protein


NaCl

35 m

m

A demonstra8on of the dynamics of pa7ern forma8on in counterdiffusion technique

5 mm

[NaCl] 20% p/v[Agarose] 0.5% p/v[Lysozyme] 100 mg/ml[Agarose] 0.25% p/v

Advantage of CCD : Diffusive mixing assures reproducibility


Precipitant concentration

Protein concentration

In batch + evaporation and vapor diffusion techniques the system moves towards far from equilibrium

In counter-diffusion technique the system moves from far from equilibrium to equilibrium:

The technique self-search the best crystallization conditions


How to perform counter-diffusion experiments?

23

forces drag viscousforcesbuoyancy −⋅⋅Δ⋅⋅== να gcLGrN

Microgravity981-9.8 x10-4 cm2 s-1

Gels (10-7-10-6 cm) and capillaries (2 x 10-2 cm)

Viscosity 0.015 g cm-1 s-1

Density gradient≈ 10-3 gr cm-3 mol-1

Looking for diffusion mass transport scenarios

J.M. García-‐Ruiz, J. Drenth, M. Ries-‐KauV and A. Tardieu. A world without gravity -‐ Research in space for health and industrial processes. Edited by G. Seibert et al., ESA SP 1251 June (2001)

At high GrN convecBon dominate

At low GrN convecBon dominate


Volume of protein solution required to fill a capillary as a function of its

For capillaries of 0.1 mm internal diameter the required volume of protein solution is 314 nanoliters for 4 cm long capillaries and 235 nanoliters for 3 cm long capillaries. For a capillary diameter of 0,75 mm (the limit for a good visualization) the volume of protein solution is just 166 and 133 nanoliters respectively.

20 40 60 80 100 120 140 160 180 200

0

200

400

600

800

1000

1200

1400

1600

Volu

men

in n

anol

iters

Capillary diameter (microns)

capillary length = 3 cm capillary length = 4 cm capillary length = 5 cm

20 40 60 80 1000

50

100

150

200

250

300

350

400


Apoferri7n

A) Hfq, B) AP-‐Sm1, C) AF-‐Sm1 associated to a 14 bases RNA target RNA, D) EndoVII, E) EndoVII in complex with a DNA cruciform juncBon, F) EndoVII in complex with a mismatched DNA, G) lysozyme.

From: Sauter, Biertümpfel et al., Acta Cryst. (2002). D58, 1657

photosystem II core complex of Pisum sativumIvana Kuta et al., Photosynth Reserach 2007

Or by counterdiffusion techniques, since you hace already seen a good collec7on of crystal, here there are some more from other laboratory.

Structure of 21 simple mutants of tiorredoxine de E. Coli

native

Residuos cargados superficiales

Residuos hydrofobos enterrados

PDB I.D. 2FCH, 1ZZY, 2FD3, 2H6X, 2H6Y, 2H6Z, 2H70, 2H71, 2H72, 2H73, 2H74, 2H75 y 2H76

D EE DI VV I

Residue number20 40 60 80 100

ΔΔ

G (kJ/m

ol)

-5

-4

-3

-2

-1

0

1

2

85

60

41

38

23

55

45

91

7572

13

2

4347

9

10

4

5 16

25

3061

44

48

101

10486

A par7r de un análisis de secuencia se han determinado las mutaciones mas conservadas evolu7vamente. Para entender el por que de esta conservación se llevan a cabo estudios termodinámicos y estructurales que nos permitan en un futuro diseñar proteínas funcionales de mayor robustez.

SER38

Crystal structure of dihydropyrimidinase de S. meliloti...

…and mutants C76A y C181A of the hidantoin racemase of the

C76A C181A C76A-hidantoinaC76A-D,L-IPH C181A-D,L-MTEH

Dpto. Química, Área de Bioquímica y Biología Molecular. UAL. Dr. Felipe Rodríguez-Vico y Dr. Sergio Martínez-Rodríguez

N114A Abl-SH3 domain mutant complexed with a high affinity peptide ligand

Table 1.-‐ X-‐ray data collec7on sta7s7cs

Space group P212121

Unit cell dimensions 48.170 50.093 56.431 90.00 90.00 90.00

Resolution range (Å) 50-1.75

Number of observations 84702

Unique reflections 13236 (1638)

Data completeness (%) 92.2 (75.8)

Rmergeb (%) 0.07 (0.24)

I/σ(I) 16.4 (4.4)

a The values in parentheses are for the highest resolu7on bin

Camara-Artigas et al., Acta Cryst. (2007). D63, 646–652

A demonstration of the use of the method to solve structures at room temperature directly from 0.1 mm capillary screening.

In situ data collection and structure determination at room temperature

Protein Lysozyme Thaumatin HLFBPase FerritinMW 14,3 KDa 22 KDa 147 KDa 456 KDaIPtheoretical

9.3 8.5 6.6 5.4Precipitant NaCl Na/K Tart. PEG 4000 CdSO4pH 4.5 7.0 9.0 5.0

Protein and low concentration agarose

Precipitant + glycerol (cryo solution)

Crystal

Gavira, J.A. Et al.. (2002). Ab ini4o crystallographic structure determina,on of insulin form protein to electron density without crystal handling. Acta Crystallogr D58 :1085-‐1254Ng, J.D et al. (2003). Protein crystalliza,on by capillary counter-‐diffusion for applied crystallographic structure determina,on. Journal of Structural Biology 142:218-‐231.

Each protein was concurrently grown, deriva7zed with a halide, treated with a cryogenic solu7on in a single capillary tube and used directly for data collec7on using synchrotron radia7on source without ever handling the crystals.The structures were determined ab ini7o by Itera7ve Single Anomalous Sca[ering (ISAS) yielding to de novo crystallographic models. This procedure is proposed as a direct applica7on towards high thorough-‐put screening and structure determina7on for proteins in general.

Cryocrystallography with crystals grown by counterdiffusion techniqueKeeping crystals in the capillary (Method 1) In situ data collection and structure determination

XTALVIEW(Duncan E. McRee)

ISAS((Bi-Cheng Wang)

O(Alwyn Jones)

PHASES(W.Furey & S. Swaminathan)

XTALVIEW(Duncan E. McRee)

CCP4 (UK Science and Engineering Research Council)

+

J.A. Gavira, D. Toh, F.J. López-Jaramillo, J.M. García-Ruiz and J.D. Ng. Ab Initio crystallographic structure determination of insulin: from protein to electron density without crystal handling. Acta Cryst. D58 (2002) 1147-1154

SH3 Alpha-‐Spectrin pH 9.0mixPEGs

SH3 Alpha-‐Spectrin pH 5.0PEG8K

Example of capillary monted crystal for cryo-‐cooling

Spectrin-‐SH3Spectrin-‐SH3

RT 100K

Wavelength (Å) 0.977 0.977

Space group P212121 P212121

Cell parameters (Å) 34.46, 42.47, 50.04 33.66, 42.24, 49.73

Resolu7on limit (Å) 1.55 1.55

T data collec7on (K) RT 100

Precipitant mixPEGs mixPEGs

Crystalliza7on Technique CD Capillary CD Capillary

Crystalliza7on pH 9.0 9.0

Mosaicity 0.3-‐0.4 0.8-‐1.0

B factor 19.7 21.0

Example of a SH3 Alpha-‐Spectrin crystal collected at room temperature and then cryo-‐cooled at 100K.

SH3 Alpha-‐Spectrin pH 9.0mixPEGs

Precipitant/buffer cocktail

Gel plug

capillaries

cover


Dip one capillary into the protein solution. The protein solution will rise by capillarity and the capillaries will be filled. Seal the upper end with the putty.

Dip the filled capillary into the GCB-Domino. Just punch the unsealed end of the capillary across the gel located on top of the precipitant.

J.D. Ng, R.C. Stevens and P.Kuhn, submiVed

Juan Ma. Garcia-Ruiz Tokyo 2nd International Symposium on Diffraction Structural Biology 2007

Low-cost plastic constructs fabricated from cyclic olefin polymers shown non-birrefringent and compatible materials for microfluidic crystallization.

Tested with bacteriorhodpsin with lysozyme, bacteriorodopsin and thaumatin

New technologies applied to counterdiffusion

In situ data collection and structure determination

Joseph D. Ng, Peter J. Clark, Raymond C. Stevens and Peter Kuhn. In situ X-‐ray analysis of protein crystals in low-‐birefringent and X-‐ray transmissive plasBc microchannels. Acta Cryst. (2008). D64, 189–197

Joseph D. Ng, Peter J. Clark, Raymond C. Stevens and Peter Kuhn. In situ X-‐ray analysis of protein crystals in low-‐birefringent and X-‐ray transmissive plasBc microchannels. Acta Cryst. (2008). D64, 189–197

Sauter et al., From macrofluidics to microfluidics for the crystallizaBon of biological macromolecules CGD 7 2007.

Song, H.; Tice, J. D.; Ismagilov, R. F. Angew. Chem., Int. Ed. 2003, 42, 768-‐772.

Using microfluidics to decouple nucleaBon and growth of protein crystals. J.U. Shim, G. Cristobal G, D.R. Link, T. Thorsen, S. Fraden, Crystal Growth & Design 7, 2192-‐2194 (2007).

Our aim is to reduce the screening for crystallizaBon condiBons as much as possible based on understanding

Could a mixture of PEGs do as well as each PEG separately?Is it the pH of the crystallizaBon experiment important when used salts as well as PEG as precipitaBng agent?

Study base on 20 proteins along one year:

PEG 400K (30%) ; six pH values PEG 4000 (PEG 30%); six pH values PEG 8000 (30 %); six pH values PEG 400K (20%) ; PEG 4000 (PEG 15%); PEG 8000 (10 %); six pH values

Ammonium sulphate 3M at six different pH values.

pH 4 pH 5 pH 6 pH 7 pH 8 pH 9

The precipitant cocktail contains:

a) A buffer that keep the solution within one unit of pHb) A low molecular weight polyethylene glycol c) A high molecular weight polyethylene glycold) An additive that may help to precipitate the protein.

)(TRTD B

ηπκ

≅According to Einstein-Stokes relation the diffusivity depends on the radius of the molecule. Therefore:

Buffer molecules diffuses faster and change the pH of the protein solutionLow MW PEG molecules go later onFinally high MW PEG molecules scan the capillary

Several effects in one single experiment !All concentrations tested in one experiment !


Crystallizability is very sensiBve to pH –as expected. The bad new is that there is not A clear correlaBon with pI. When crystals form at any pH, precipitaBon occurs in any other pH

Crystallizability is sensiBve to pH. There is not a clear correlaBon with pI.The 3PEG mixture does as well as each of them. However , it is required to increase the concentraBon of PEG 8K

Some additional results observed :

1. It is confirmed that, for the same screening matrix, counterdiffusion yields at least as many hits as vapor diffusion, within the first three weeks after set-up of the experiments.

2. It is confirmed that for similar crystallization conditions, nucleation is retarded in the case of counterdiffusion with respect to drop techniques.

3. It has been also proven that counterdiffusion yields crystals under screening conditions that drop techniques never produced. Noticeably the number of new crystallization conditions increases with time, sometime with waiting periods of month(S)

crystal form changes with crystallisation technique The way in which supersaturation is achieved has an effect on the crystal form. Protein crystallisation crystal form technique

Cablys3*lysozyme hanging drop PEG P212121 67 69 113 90 90 90 hanging drop formate C2 133 73 39 90 105 90 APCF C2 129 73 38 90 107 90 counterdiffusion formate P212121 68 69 113 90 90 90

TIM hanging drop AS P3221 125 104 counterdiffusion AS P3x12 215 106

Parvalbumin sitting drop P212121 51 50 35 counterdiffusion P21212 59 59 26 counterdiffusion P1 26 32 53 85 83, 79

From I. Zeggers, J.M. García-Ruiz and coworkers, unpublished data

PDB ID Variant S G Condition Method1KEB P40S. P 1 pH: 3.8

Ethanol: 25% CuAc2:10 mMNaAc:100 mM

Diffusion-Vd

1SRX Forma oxidada C 2 pH: 3.8Ethanol: 25%CuAc2:10 mMNaAc: 100 mM

Diffusion-Vd

2TRX Silvestre C 2 T: 4° CpH: 3.7-4.1Butane 0.001-0MMPD: 48-50%NaAc: 0.01-0MCuAc2:1 mM

Micro-dialysis

2FCH G74S P 21 21 21 T: 4° CpH: 5.4 MPD: 60%NaAc: 15 mMCuAc2: 1 mM

Counter-diffusion

1ZZY L7V P1 pH: 3.8Ethanol: 25%CuAc2: 10 mMNaAc: 100 mM

Counter-diffusion

2FD3 P34H P1 21 1 T: 4° CpH: 8.0 MPD: 60%Hepes: 15 mMCuAc2:1 mM

Counter-diffusion

2H6X2H6Y a Z2H70 a 76

WildE48D, E44DD9E,D47E, E85DD43E, D2E, D13ED10E

P61 pH: 6.9pH:3.5, 7.0 pH: 7.0, 7.0, 5.4pH:7.0, 3.5, 7.0pH:4.1T: 4° CMPD: 60% NaAc/Hepes:15mMCuAc2:1 mM

Counter-diffusion

Counterdiffusion produces new polymorphs

Vrije Universiteit Brussel, Université de Liege, Univesrité Catolique de Louvaine

Screening with counter-diffusion technique

Simple and fast preparation of the experiments

No tools required. Just fill the capillary and punch it into the precipitation cocktail

Extensive search of the crystallization space for pH and precipitant concentrations

Actual grid screen can be made possible and are recomended

Requires minimal amount of protein

Even lest than 250 nanoliters per capillary.

No pictures and image analysis required.

The whole growth history is stored in the capillaries.

Unambiguous interpretation of the results

The technique does not yield single results but a trend

The number of conditions for effective screening can be reduced

Mix of PEGs can be used to replace singular MW PEGS.

Coauthors in the work presented in this talk:

Luis Antonio Gonzalez: CrystallizaBon experiment implementaBonJosé Luis Gavira: Handling of crystals and X-‐ray diffracBon analysisEduardo Howards: Work on mixing experimentsJoe Ng: First screening and in situ X-‐ray diffracBon analysisFermín Otalora: Computer simulaBons

Financial support from and help from:

European Community VI Framework ProgramSpanish Government (MEC)Triana Science and TechnologyThis is a product of La Factoría de Cristalización

Thank you

This is a product of La Factoría de Cristalización

crystallization methods and screening · 0 5 10 15 20 25 30 35 40 45 50 0 2 4 6 8 10 12 14 16 18 20...

Documents