Hi Jackie,

there are still plenty things you may try. First of all, you've mentioned you got heavy precipitant during your optimisation. You may try to reduce your protein concentration and see what happens. You can also try different protein:precipitant ratio and different drop sizes. You may explore different setups as well - switch between hanging and sitting drop, try microbatch under oil (with paraffin and volatile oils in parallel), capillary crystallisation. You may also try to run the additive screen (sold by Hampton Research). Last but not least, if you get the hit in your initial screening, try to fish that crystal(s) and put it in the X-ray beam - may be you will have good diffraction. Also, have you verified that the hit you got is indeed a protein crystal? Also it never hurts to setup more screens if there are available in your lab - you may get different condition(s) which may produce better quality crystals. Good luck!

Hi Jackie,

One area to consider is your protein sample. This is an important variable that you should consider altering too. You probably have two areas here that you could look at. Firstly, you could try setting up drops with a different ratio of protein and precipitant (I assume that you are using vapour diffusion, although the outcome will likely be the same with other methods). I have seen others try 1+1, 2+1 and 1+2 drops as standard in optimisation.

Secondly, you could look at altering the conditions of your protein sample. Changing the concentration, as Albert suggested, is an obvious thing to do. Other areas include the buffer that the protein is in - is it optimal? If you haven't done so, you could try buffer optimisation, for example using differential scanning fluorimetry (see Niesen et al. (2007) Nature Protocols 2:2212-21 for details). Also, consider the salt concentration of your protein. Phase separation can indicate that the protein would prefer more NaCl in the protein sample. I had one protein where changing the salt concentration upwards changed the crystallisation from phase separation to nice crystals.

Good luck!

Agreed that probably your protein concentration is too high, hence altering protein/precipitant ratio might be very beneficial.

I would also attempt adding some water to the droplet, but not to the well(e.g. 1 ul protein + 1 ul water + 1 ul precipitant). This changes the initial concentrations of all species and the kinetics to get to the final concentration, but does not affec these last ones.

Another suggestion (which I have heard from Terese Bergfors) is to use the "oily" phase separation as a source of seeds for streak seeding. Just dip your "cat's whisker" into the phase separation and seed a fresh droplet in which you have reduced the starting concentrations of both protein and precipitsant.

Hope this helps.

Marco

Apart from the other suggestions (lowering protein conc, adding water to you your drop etc.) you may try to vary the pH of the crystallization in very small steps (0.1) at a time, to explore say +/- 0.3 around your condition [7.2 to 7.8 in your case]. Sometimes it can change the dynamics of crystallization in your favour. But be very precise in your pH measurements for better reprodcucibility. Bye the way how quick is the phase separation after setting up the expriment?

Hi Jackie,

The crystallization proceeding and set the conditions take time. As everybody said try change your protein conditions, first the buffer and salt concentrations to avoid phases. For precipitates, reduce first protein concentration leaving precipitant constant, and then do it in the other sense. Another important thing is the tempeture, set drops at low temperatures, even 4 degrees.

Good luck!

Marisa

Jackie

Your results are actually typical. Very often, if you scale up, you will get precipitate. The reason is simple, although not many people seem to know this. With small drops, you lose roughly half of your protein, through denaturation on the plastic of the plate and at the air-aqueous interface. Therefore when you scale up from say 100+100 nl to 1+1 microlitres, you normally need to roughly halve your protein conc (as others have said above).

The other important point is that you will almost certainly get a great benefit from microseeding, using the crystals that you have to make a seed stock. This will (probably) allow you to get new hit conditions, obtain better-diffracting crystals right out of the screen and also - arguably most important - allow you to control the number of crystals per drop by diluting the seed stock. If you are trying to soak ligands into the crystals you can also have fun with seed stocks with different unit cells.

See

http://www.douglas.co.uk/Scaling_Up.htm

http://www.douglas.co.uk/mms.htm

http://www.douglas.co.uk/MMS_proc.htm

I hope that all helps!

Patrick

Most HT throughput screens are done as sitting drops. Optimize your hit in sitting drop plates too.

Also, try adding a NaCl gradient to your screen to reduce precipitation.

You got great suggestions above. Keep in mind that there is always more to try. E.g. optimize your sample and screen from scratch. Some constucts never crystallize but after slight modifications they crystallize in many conditions.

Good luck!

To add to the points on concentration and seeding, remember that crystallization involves both nucleation and growth. In my view, tiny volume HT crystallizations use very high protein concentrations to increase nucleation, which seems limiting in the tiny volumes. When we looked at samples from Protein Structure Initiative efforts that were using such HT conditions, we found that all samples contained aggregated proteins by x-ray scattering in solution (GL Hura et al., Nature Methods. 2009 Aug;6(8):606-12). So it you have a way such as seeding to provide nucleation then you can usually use much lower protein concentrations.

You do not give conditions, but I assume you are using some combination of salt and PEG. These conditions often work by driving the protein into the PEG phase creating super high protein concentrations that provide nucleation often at the PEG-salt interface. If this is your situation, then I would try varying temperature 10 degrees higher and lower plus PEG and salt concentrations. When you can actually see phase separation then you have likely over shot your conditions as you want microscopic rather than macroscopic phase separation. So you will want to do a grid that reduces precipitant concentrations to points below which you can see the gross phase separation. You can even put protein and precipitant into separate drops and pull a liquid bridge between them that will form a gradient if you wish to get fancier. By the way, you should also check to see if your protein precipitate will go back into solution or not– i.e. is it an equilibrium precipitate or is in unhappy protein?

Hello everyone! I am very grateful to see so many suggestions coming and I am pretty excited to try them all. Thank you so much for the quick replies - I have been quite anxious and at a loss at what to do.

Misc replies:

@Umesh Singh: The phase separation comes after 1-2 days!

@Patrick Shaw Stewart: I know you (and your immensely helpful Douglas Instruments webpage)! The idea of halving of concentration (from 19mg/mL to 9mg/mL) came from what you mentioned (nice tartan drop btw).

Hi Jackie,

Phase separation may indicate protein heterogeneity and i often see it if protein is truncated.

I am not quite sure about your protein. Is it a soluble protein? Maybe remove of cysteins can work. Or just lower you protein concentration. Sometimes protein concentration does no directly relate to the crystalization.

Good news! I tried the various suggestions that all of you have made, and right now I have...spherulites! Which is a little better above phase separation I suppose.

The condition change to get to spherulite from phase separation was slightly more drastic than I thought:

Different protein conc: 2.5 mg/mL vs 9 mg/mL

Same salt

Same buffer

Different PEG%: 3% vs 10%

I hate to impose on you guys again, but are there any good protocols out there on additive screens? I have been searching for protocols in the Hampton website but to no avail - they only have protocols for sitting drops and I'm looking for hanging drops. Thanks again guys, this thread will be quite helpful for not just myself but for many others as well.

Dear Jackie, the protocol for the hanging drop setup with the additive screen is straightforward - there are two possibilities - 1) if you do it by hand, you add each additive into a corresponding well of a plate with your condition in 1+9 manner, i.e. if you want to setup 500µL reservoirs, you should add 50µL of additive to 450 µL of your condition. And then you just setup normal hanging drop (1:1 of your protein to the reservoir solution or any other ratio you prefer) 2) If you have a robot then you may add additives only to the drops on the cover foil, except for the conditions with volatile additives, which you should still add into reservoirs as well. You may also try to do the same in manual setup, but then it is vey tricky to pipette extremely low volumes, unless you follow Hampton Research reccomendations and make gigantic drops of 10µL. In my personal opinion, if you want to save your sample and the additive screen and you have access to the robot - do the sitting drop setup with the robot.

Good luck!