Vaheh, I don't recall precipitation at all, but I do remember that we were prepared to change the crystallization recipes (i.e. adjust the recipes from previous 'large volume' crystallization to make more nuclei). For example, our first tests were with lysozyme (sorry, hardly a representative protein, I know) and it is known (paper by RA Judge et al) that buffer pH is a major determinant in nucleation (for that protein). Our initial concern was: is it possible to grow crystals in exceedingly small volumes? So we changed the pH to increase the nucleation rate. I think I recall that in case of lysozyme the nucleation rate is roughly inversely proportional to acid concentration, i.e. increase pH by 1 unit (10X less acid), gives ~10 times more nuclei. In general (for other proteins) we did make some minor adjustments in crystallization, but generally only in protein/precipitant concentrations. 'Minor' means adjustments of about 10% or so - going from 10% precipitant to 9% or 11%.
When you go to very small volumes there is another consideration that you have to think about: mixing. If you assume that mixing takes place by diffusion (no stirring, it is very difficult to stir a very small volume reliably, i.e. to the same extent every time, unless you use microfluidic flow, in which case complete mixing to homogeneous mixtures is trivial - note: the Fluidigm system is 'microstatic', not microfluidic), so - if you assume mixing to take place by diffusion only, then the diffusion length in 100nL drops is very small and you can calculate in the worst case scenario how long it will take to accomplish full mixing by diffusion only. The time scale will be in the order of seconds. If you do the same calculation for a large drop (1 uL), the time scale is MUCH longer. So in case of batch crystallization, the 'end point' is reached very quickly in small volumes, while it takes much longer in large drops. If the end point causes protein precipitation, then this will happen very quickly after starting the experiment. In the larger scale experiment it will take much longer to reach precipitation and your system may go slowly through a process of enabling crystal growth before you reach drastic insolubility that causes precipitation. So as someone else said, the kinetics are completely different, the end point is not, and this can significantly affect the outcome of the experiment. Mark -----Original Message----- From: Oganesyan, Vaheh <[EMAIL PROTECTED]> To: [email protected] Sent: Thu, 17 Jan 2008 7:40 am Subject: Re: [ccp4bb] crystallisation robot Mark, What was the state of the larger drops when tiny counterparts had crystals? My guess - they all precipitated. I’m trying to understand why some proteins or some conditions require change in protein concentration while others do not when migrating from smaller drops to larger ones. If it is protein dependent then I’m afraid there might be no one answer; if it is not then there should be a trend and explanation of phenomena. Vaheh From: CCP4 bulletin board [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Wednesday, January 16, 2008 8:31 PM To: [email protected] Subject: Re: [ccp4bb] crystallisation robot Once upon a time I worked in a group that was interested in developing crystallization in microfluidics. This was before the time that Fluidigm existed and we had not heard of crystallization with the aid of microfluidics at the time. We had good reason to try to make a system that was as small and light as possible - it had something to do with the cost of shipping proteins and precipitants - less was better. And we also wanted all protein drops to be fully enclosed, out of safety considerations. Like Tassos, we were very worried what would happen if you scaled back drops along the lines of this discussion - several uL downto tens of nanoliters. If the stochastic process had a major influence over this process, we thought that we would never get any crystals. So we set up side-by-side experiments at larger volumes and smaller volumes - basically scanning several orders of magnitude - expecting a decrease of the number of crystals when volumes decrease. To our great surprise the outcome was that smaller volumes almost always gave MORE (I almost want to say 'dramatically more') crystals, more nucleation, and indeed in various cases the crystals grew much faster also. Indeed, it was trivial to observe that the surface-to-volume ratio was the primary driver for the nucleation process. We had control over geometry to some extent and were able to observe surfaces while crystals grow. The crystals would most commonly nucleate on a surface. So although there probably is something to stochastic aspects, it is clear that other aspects can be more important and "overrule" the stochastic considerations. The somewhat unpleasant consquence is of course that results acquired in very small volumes (with larger surface-to-volume ratio) cannot necessarily be repeated in larger volumes (smaller surface-to-volume ratio). This is not a flame, even if heat might be a good thing on a night with temperatures predicted far below 0F. :-) Mark -----Original Message----- From: Anastassis Perrakis <[EMAIL PROTECTED]> To: [email protected] Sent: Wed, 16 Jan 2008 6:17 am Subject: Re: [ccp4bb] crystallisation robot > Oryxnano 50+50 nL > > Demetres > Which, indirectly, brings up an interesting (but not relevant to the Oryx) question. Nucleation is a process that does have a stochastic aspect. Thus, one could argue that compromising to 200-300 nl might be better than either extremes of 50nl (too small volume and less chance for nucleation) or 1000 nl (too much sample). any comments ? (let the flames begin). A. PS1 another interesting issue that has has been hardly touched in these emails is the real sample loss: left in wells and not easy to recover, lost because of contamination with system liquid, etc ... PS2 I see lots of people with new robots. please do have a look at the www.BIOXHIT.org page and if you have a few minutes to assemble a table we will be happy to add your specs to our pages. it can be a nice resource and it has already enough things and already one response to my last email ;-) To make life easier to potential contributors we can provide an Excel sheet to fill up with your specs - just ask. On Jan 16, 2008, at 12:46, Demetres D. Leonidas wrote: > > David Briggs wrote: >> I'll defend the honour of the phoenix... (again) >> >> Bernhard Rupp 100+100 nl >> Dave Briggs (and all users at Univ of Manchester, UK) 100+100nl >> Others.. >> >> Only time we have ANY problems is when the nano dispensing tip >> gets clogged. Often a good wash whilst still on the machine will >> clear the blockage. >> >> Dave >> >> >> >> >> -->> ============================ >> David C. Briggs PhD >> Father & Crystallographer >> http://www.dbriggs.talktalk.net <http://www.dbriggs.talktalk.net> >> AIM ID: dbassophile >> ============================ > > --> Demetres D. Leonidas, Ph.D. > Structural Biology & Chemistry Group > Institute of Organic and Pharmaceutical Chemistry > The National Hellenic Research Foundation > 48, Vassileos Constantinou Avenue > Athens 116 35, Greece > ================================================== > Tel. +30 210 7273841 (office) > +30 210 7273895 (lab) Fax. +30 210 7273831 > E-mail: [EMAIL PROTECTED] > URL: http://athena.eie.gr > ================================================== size=2 width="100%" align=center> More new features than ever. Check out the new AIM(R) Mail! ________________________________________________________________________ More new features than ever. Check out the new AIM(R) Mail ! - http://webmail.aim.com
