Origin of a Species: History and observations of one high throughput crystallization laboratory J. R. Luft, R. J. Collins, S. M. Gulde, A.M. Lauricella,

Origin of a Species: History and observations of one high throughput

crystallization laboratory

J. R. Luft, R. J. Collins, S. M. Gulde, A.M. Lauricella, C. A. Mancuso, J. L. Smith,

C. K. Veatch, and G. T. DeTitta

Recent Advances in Macromolecular CrystallizationLe Bischenberg, France

May 8-11, 2005

Hypothesis

By executing a large number of precipitation reactions for a large number of proteins we could predict crystallization conditions for a previously un-crystallized protein by comparison of precipitation behavior.

Protein Binary Code PSSUnknown 1011010010110010 --

Protein B 0111010111010010 11Protein A 0001100101010101 5

What is the most efficient way to execute precipitation reactions?

• Batch, in particular microbatch in a capillary– Minimize solution volume requirements for

precipitation cocktails

Good, but not Great• Maximum throughput (manual)

50 experiments/technician/day• Crystals were a ‘by-product’.

Microbatch Under Oil

Chayen, N.E., Stewart, P.D.S., Blow, D.M., Journal of Crystal Growth, 122 (1992) 176-180.

• Minimal solution volumes

• Plates for storage and handling

Mothers and Daughters

• Mother (source) / Daughter (destination) plates• Efficient parallel setups

…

source destination

Solution Delivery

• Pump

• Slot pins

• Pipettes

• Syringes

• Faster Syringes

http://www.vp-scientific.com/index.htm

Screening for Crystallization Leads

* 15361. Oil2. Crystallization Cocktail3. Macromolecular Solution

Quantity• How Many Experiments?

– 20 plates per day (~ 200/month)– 200(plates) x 1536(experiments) x 6(images)

• Generates 1.8 million images/month

As of 3/17/2005:4877 (plates) x 1536 (experiments)

7.5 million experiments

45 million images

469 active users

Setups per Month

HT Infrastructure at HWI

Variables

1. Plate

2. [Cocktail]

3. Oil

4. Sample [salt]

5. Sample State (solid or liquid)[M

]

[C]

1. The Plate

• February 2004– Square to circular cross-section well

• Better visibility

• Now in production at Greiner BioOne– LBR plates

• “significantly higher water absorption”

11

22

Image Quality

2000 2005

1.8 mm 0.7 mm

LBR Plate Imaged with Cross-Polarized Light

2. Cocktail Dehydration

• Plates prepared with oil

• Cocktail added to plates (200nl)

• Plates are stored at 4oC for 1 day – 2 weeks

[M]

[C]

Dehydrate C drop

Decrease Vc drop

Increase [C]Decrease dilution of [M]Increase [M]exp

??

3. The Oil

• Paraffin Oil – (IR Spec grad)

• Emulsion formed with some batches

• Mineral Oil – (USP grade)

• December 2004

4. [Salt] in the Samples

• Compared salt concentrations for 392 samples– ([KCl] + [NaCl]) mM

• All samples were set up using the ‘standard’ 1536 screen and outcomes reviewed

• Compared 158 samples shipped:

– 76 Frozen (dry ice)– 82 Liquid (wet ice)

• All samples were set up using the ‘standard’ 1536 screen and outcomes reviewed

5. State of the Samples

Frozen Samples versus Liquid Samples

frozenliquid

Samples with hits Samples with no hits

Num

ber

of S

ampl

es

Distribution # of Hits

45% had 8 or fewer leads from 1536 experiments

In Conclusion…

• Work supported in part by the John R. Oishei Foundation, the Cummings Foundation

• NIH RR016924, NIH P50 GM-62413 and NIH P50 GM-64655• Special thanks to Bob Cudney• Greiner Bio-One

• And ….

Acknowledgements

http://www.hwi.buffalo.edu