Unable to reproduce robot tray hits in hand trays

From: Matthew Lalonde
Date: 26 March 2012 18:57

Dear All,

I have searched the archives and would like more information about reproducing robot tray hits using 24-well hand trays. I reproducibly get crystals when I use small volumes (0.5 ul) in 3-well intelliplates but only precipitate in 1-2 ul sitting drops in 24-well hand plates.

What parameters should I vary to reproduce crystals in hand plates? References that discuss this procedure exhaustively would also be appreciated.

Thanks,
Matt

----------
From: Nian Huang


I have seen people only use robot to optimize their crystal and get
good diffraction (~2 A). If you keep having trouble, you can try this
method instead, even though generally the case is the bigger the drop
the bigger crystal.

I remember the archive suggest us to use higher concentration of
protein to reproduce the result when using bigger volume.

Best,

Nian

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

On Mon, 2012-03-26 at 11:57 -0600, Matthew Lalonde wrote:
> What parameters should I vary to reproduce crystals in hand plates?

First of all, protein concentration.  It also does not hurt diluting
your reservoir since you are getting precipitates.  If your goal is to
get bigger crystals (which is not always better), consider using
crystals that grow on intelliplates for seeding.

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

----------
From: David Briggs


Hi Matt,

This doesn't really answer your question directly, but sidesteps
around the issue -
I wrote a little something on this exact subject not so long ago -
http://xtaldave.wordpress.com/2012/02/23/on-protein-crystallisation/
(You can ignore the first 5 paragraphs of intro - it was written with
a lay audience in mind)

Basically, I setup fine gradient screens in deep well blocks by
titrating stock solutions across a row - which I can then put straight
into use with our Phoenix crystallisation robot. In most situations I
would hope to get decent diffracting crystals from a standard 3-well
intelliplate setup this way.

HTH

Dave

----------
From: Richardson, Brian C.

Most of the labs sharing our Phoenix have had enough trouble with exactly this that our standard procedure is to now to use 1 µl drops (.5/.5 protein/well) in our initial screens – those scale up much more reliably to the 24-well format and seem less finicky in general, reducing the chances of an unusable hit.  We often get a somewhat different set of hits than we get at the lower volume, as well, so it might be worth rescreening while you work on the existing hit. 

 

Of course we also try the usual tricks with optimizing anyway (my first crystallized protein precipitated in sitting drops but not hanging drops, etc). J

 

 

------------------------------------------------

Dr. Brian C. Richardson

 

----------
From: Patrick Shaw Stewart

Hi Matt

Rajesh asked a similar question last week, below

Essentially, you have to reduce the protein concentration when you scale up because you lose proportionally more protein from smaller drops.

This usually works very well and we see no reason to use more than 0.3 + 0.3 for initial screening.

There's a page on our web site which explains this in much more detail, see http://www.douglas.co.uk/Scaling_Up.htm

Best wishes

Patrick

Rajesh

If you set up the volumes you suggest you will probably get precipitation.  This is counterintuitive until you realize that (as Ed says) you will be losing a lot of protein with those small drops.  When you scale up the surface area to volume ratio is lower, so a smaller proportion of the protein is lost.  Therefore you go up on the phase diagram and get precipitation or very small crystals.

Normally halving the amount of protein for the hits from 200 nl drops works (suggesting that half the protein is lost from such small drops).  Try say 500+1000+500 (don't reduce the volume of seed stock because the solution that you suspended the crystals in may be important).  Or dilute the protein and use 1000+1000+500.

For the hits from the 450 nl drops you could reduce or dilute the protein by say 25.%.

Or make plenty of seed-stock and try seeding into a random screen again with larger drops, say 1.5+1+0.5 ul

Those tiny crystals should be good for seeding, don't worry about that (provided they are protein of course).

Streak seeding may work but bear in mind that roughly a third of the precipitant comes from the seed stock in your 250 nl drops.

You can add the seed stock with a syringe and needle if you don't have suitable robot ;)

Experience and data-mining suggests that increasing the salt precipitant (in high-salt drops) or salt additive (in PEG drops) by around 50% may be helpful too when scaling up - I'm not sure why this works.

Good luck

Patrick






For the hits in the 250 nl drops you are probably losing 

On 19 March 2012 20:31, Rajesh kumar  wrote:

Dear All,

I have few papers in hand which  explain me about microseeding, matrix microseeding, and cross seeding.

I have also read few earlier threads and some more literature in google.

Using Phoenix robot, I did a matrix micro-seeding and matrix cross seeding. I have few hits with this.

In 96 well I used 100+100+50 nL and 200+200+50 nl (protein+screen+seed) in separate expts.

I have hard time to plan to translate this 96 sitting drop well plate to 24 well plate to refine the conditions to get better crystals. only 1-2 hits are small crystals and they are tiny. 

 I wonder in 24 well plate, if I should do-

1)  for Example 500+500+50nl (I am sure I cant add less that 500 nL precisely)

2) to a drop of 500+500 nL do microseeding/streaking with a hair

I appreciate if you could advise and share some practical ways to further my experiment.

Thanks in advance

Regards,

Rajesh