Thank you. Now I understand the difference. I thought there was separation.
Maia Xuewu Zhang wrote:
Hi Maia,I have seen your post regarding this before and I just want to point out that you may have confused "AUC" (analytical ultracentrifugation) with gradient-based ultra-centrifugation methods for separating macromolecules. AUC does not involve separation of different species in the sample. There are two types of AUCs: sedimentation velocity and sedimentation equilibrium. In sedimentation equilibrium experiments, the system reaches the equilibrium at the end, and the monomer/dimer ratio, Kd, etc parameters can be worked out by fitting the data to a model globally. The shape of the molecule does not matter. For starters:http://en.wikipedia.org/wiki/Ultracentrifuge Xuewu ZhangOn Wed, Aug 11, 2010 at 10:37 AM, chern <ch...@ualberta.ca <mailto:ch...@ualberta.ca>> wrote:Hi Anastassis,We are back to the same argument that AUC is not a good method. Aseveryone knows, it's a dynamic equilibrium between monomers and dimers that exists before separation. Once you started separation in any method, the equilibrium is disturbed now in each separated band. That will cause re-equilibration and constant migration of newly formed dimers from the monomer band and newly formed monomers from the dimer band. The t(eq) is the re-equilibration time. Your method of separation of monomers and dimers should be quick enough before any re-equilibration occurs (t(sep)<<t(eq)). Otherwise, you get a mess and smearing of bands. Also, most conventional methods depend on shape etc. I find SEC is most convenient.Maia----- Original Message ----- *From:* Anastassis Perrakis <mailto:a.perra...@nki.nl> *To:* chern <mailto:ch...@ualberta.ca> *Sent:* Monday, July 05, 2010 2:38 PM *Subject:* Re: [ccp4bb] monomeric coiled coil--updated On 5 Jul 2010, at 22:04, chern wrote:Hi, Anastassis If you had just a monomer at the start time then t(eq) is the time to get to equilibrium with the dimer and vice versa.sorry to say but the definition of that time in a biophysical sense, is in my opinion equal to infinity and cannot be defined. I am being a bit pedantic here, but I am just saying that t(eq) cannot be defined, it can be approximated, and thus t(eq) is wrong to define. Why not talk about kD and kON and kOFF that have robust definitions based on kinetic properties and a physical meaning?When you separated the two bands (monomers and dimers) in AUC, and then the equilibrium is quickly established in each band again what's the point? So, to be successful in this method, you need to have t(eq) much lower than the separation run. Ideally, if you could separate monomers and dimers instantly and freeze them in the separated state, then you can have good estimate of the both fractions. I think this is clear.But, I disagree and I think what you say is wrong. The equilibrium is dynamic. Why do you insist there is a point in 'separation'? The monomer changes to a dimer and vise versa in a continuous fashion. All you can say is that in a given concentration the equilibrium is shifted towards one or the other form. But its a dynamic one. Even at a concentration which is 50-50 between two states, the molecules that are in one state or another are changing according to kinetic parameters that are characteristic for the complex. Even at 100% - lets say of a dimer - by your definition, (100% cannot exist since its reached asymptotically by any derivation about equilibriums) molecules will fall to monomer and will reassemble to a dimer rapidly. To be honest I think that talking about t(eq) is largely wrong in biophysical terms, since it does not exist. A.That's what I meant. Maia----- Original Message ----- *From:* Anastassis Perrakis <mailto:a.perra...@nki.nl> *To:* chern <mailto:ch...@ualberta.ca> *Sent:* Monday, July 05, 2010 11:45 AM *Subject:* Re: [ccp4bb] monomeric coiled coil--updated On 5 Jul 2010, at 19:30, chern wrote:Thank you for reply.1.It will be nice to have mass-spec method fornon-covalent complexes.Carol Robinson is doing these routinely ... so does Albert Heck2. 10nM solution is very dilute to catch mostcomplexes. So methods that work only for diluted solutions are not reliable to determine the multimeric state. What about > 100 mkM solutions for Kd`10^(-4)M?But its all matter of kD!! If my kD is 1nM, I will see it at 10 nM. If you have 1 mM kD, then you need 100mM sample! Most methods you can measure what you want, its a sample preparation issue.3. For study of multimers in AUC. If you have twoseparate bands, one of a monomer and one of a dimer, there will be continuous association in the monomer band and continuous dissociation in the dimer band even while they are separating, so there will be continuous migration between two bands and smearing. That's what I meant. AUC never worked for determination of multimeric state for me. I got many times uninterpretable results. I know that it works sometimes, most likely when teq>>tsep (who knows what t(eq) is for different proteins), even withsome broadening of the bands.I really do not agree with that. Are you talking about equlibrium or sedimentation runs? I agree it does not always work well, but that is mostly misbehaving samples over large times in eg the wrong temperature, rather than anything else. A complex with kOFF of about a minute or two can be measured over weeks in AUC, the one is unrelated to the other. btw, I still dont get what you mean t(eq). The only definition of the time you need to reach equilibrium in a system is infinity. A.Maia----- Original Message ----- *From:* Anastassis Perrakis <mailto:a.perra...@nki.nl> *To:* CCP4BB@JISCMAIL.AC.UK <mailto:CCP4BB@JISCMAIL.AC.UK> *Sent:* Monday, July 05, 2010 9:36 AM *Subject:* Re: [ccp4bb] monomeric coiled coil--updated On Jul 4, 2010, at 20:35, chern wrote:It's an interesting discussion.1.Usually it's not possible to use mass-specfor non-covalent complexes.Native mass spec is well developed. I would hope to give you my favorite reference, but its for now pending, courtesy of some reviewers in Mol. Cell. ;-) Seriously now, native mass spec is a very nice option and the work on non-covalent complexes by Carol Robinson and others is good proof for that.2. Most methods depend on macromolecule shape and concentration.Yes, but we need to be specific, for example, MALLS is shape independent, unlike SEC which is size dependent. I would add that most methods depend on having a good estimate of concentration. The result you get for your system, not the method itself, is concentration dependent: eg MALLS is perfectly fine for a 10nM complex, but utterly useless for a 10mM "complex". AUC has the advantage to work across a wide range of expected kD's although some tuning is needed.3. SAXS method looks limited to me. It uses diluted monodisperse solutions. That excludes complexes that can associate/dissociate. How can you calculate Kd from it?Yes, although its not the best way. Since you do measure at different concentrations, if you are above-below the kD so in the low end you have monomer and in the higher end dimer,you can estimate well the kD.4. All methods for determination of multimeric state using separation technique depend on three different cases: time of equilibrium (teq)>> time of separation (tsep), or teq << tsep or teq ~ tsep. Even without equilibrium, you cannot have only one component. For a successful separation, you would want teq >> tsep, which is less likely in AUC method.I guess there is a confusion of two different issues: 1. Separation of multimers: To separate two things, indeed as you put it, you want teq>>tsep, and I would add that you want that for a the concentration(s) that your separation method works. In general, to separate two things, you want to be at a concentration where one state is clearly preferred over the other. 2. Study of multimers: In AUC for example you study the presence of multimers as a function of time, so you can observe all kinds of multimers and from their relative abundance you can study the association characteristics.5. There are papers on capillary electrophoresis methods where they study exactly these effects (time of equilibrium vs time of separation).A.Maia ----- Original Message -----From: "aidong" <a...@xmu.edu.cn <mailto:a...@xmu.edu.cn>> To: <CCP4BB@JISCMAIL.AC.UK <mailto:CCP4BB@JISCMAIL.AC.UK>> Sent: Sunday, July 04, 2010 3:01 AM Subject: Re: [ccp4bb] monomeric coiled coil--updated > In light of several wonderful responses,Iwould like to provide an > update for this question: > > 1. I would agree that SEC might not be able to identify monomer vs > multimer forms for this likely rod-shaped protein. > > 2. It is extremely low kd for dimer. AUC and SAXS experiments have > measured its kd at ~0.1 mM. > > 3. MALS might not be able to pick up dimer form since it might be only > a few percent when the concentration is low. We might overcome > concentration effect by direct injection to dawn heleos and refraction > index. > > 4. Mass spec has found both monomer and dimer forms although the > abundance of each one is not known. > > 5. Intramolecular coiled coil is quite possible since intermolecular > dimer is unstable. We hope our structure mightprovide an answer.> > Many thanks for your time and ideas > > Cheers > > Aidong > > > On Jul 4, 2010, at 1:09 AM, AnastassisPerrakis wrote:> >> A few thoughts on these, since I do not fullyagree. >> >> 1. Detection by light scattering is a methodthat can be used either >> without separation, or while separating.>> If you have a scattering detector, you canstick in a cuvette, or >> stick it to the end of a column, your choice.>> >> 2. Sec is not a good method to show ifespecially a coiled coil is >> monomer-multimer. A long coil, will>> have a hydrodynamic radius bigger than itsMW, thus any prediction >> based on SEC will be misleading,>> especially for this class of proteins. >> >> 3. In AUC (although I am not an expert at itat all) I cant see the >> connection between the disassociation time>> and the run time. In sedimentation orequilibrium runs, depending on >> what you want to see, I think you can look>> at monomer-multimer equilibrium over a widerange of kD and >> combinations of k(on) and k(off).>> >> 4. The physiological concentration is a bitmisleading. First, its >> clear now that cells have microenvironments,>> and 'physiological' concentrations are hardto define. Also, in a >> cell, I think (and I think others tend to agree)>> that kD plays little role at the end. kD is acombination of k(on) - >> which is concentration dependent but in a cell>> very likely diffusion limited - and of k(off)which I think is what >> matters most in the cell.>> >> Going to Aidong's question, I think thatMALLS was a good >> experiment. The fact that these constructs dono associate, >> can mean that >> >> a. the prediction is wrong - likely withthese scores, but not >> necessary>> b. the kD in solution is indeed higher thatthe concentration you >> used for MALLS>> c. The constructs are not well chosen for some reason >> >> You could use AUC to detect kD as high as~100uM, depending on the >> concentration of the start sample of course.>> The next question will anyway be if that kDhas any sort of >> physiological significance - which you cannottell by magnitude - >> so you are back at the drawing board formutants. Three years later >> the referees will still not believe it ... sorry, now it gets >> personal,>> so I stop here. >> >> My two cents. >> >> A. >> >> >> On 3 Jul 2010, at 18:10, chern wrote: >> >>> The multimeric state depends on a proteinconcentration. You can >>> get any>>> multimer to dissociate if you dilute it tolow enough >>> concentration. If>>> your complex is a homodimer, thenKdiss=[complex]/[monomer]^2. >>> Let's say>>> your Kdiss~10^(-3)M, and your proteinconcentration is ~10^(-4)M, >>> then>>> [complex]=Kdiss/[monomer]^2=10^(-3)/10^(-4)^2=10^(-5),that means, >>> the dimer>>> concentration is approximately ~10 timesless then the monomer >>> concentration>>> at this particular protein concentration.Let's say, the mol weight >>> is 50>>> kDa, then at 5mg/ml you will have only about~10% of the dimer. Of >>> course,>>> if your Kdiss~10^(-4)M, then you will have approximately similar >>> concentrations of monomers and dimers at 10^(-4). >>> Because this is a dynamic equlibrium betweenmultimers and >>> monomers, some>>> methods are not good for the determinationof a multimeric state. >>> Some>>> reviewers demand to prove the multimeric state by size-exclusion >>> chromatography (SEC) or analytical centrifugation. The analytical >>> ultracentrifugation method will not work, asthe characteristic >>> time of the>>> dissociation/association is much lower thanthe centrifugation time >>> (`24>>> hours). The separated monomer will startassociation and the >>> separated dimer>>> will start dissociation according to Kdissand the bands will be >>> smeared.>>> SEC is faster, like half an hour, it givesyou a better chance. The >>> methods>>> without separation are the best Like lightscattering), just make >>> protein>>> concentration high. Here comes the otherquestion. What is the >>> physiological>>> concentration. You want to be close to it. Iread some literature >>> on this>>> and it looks like it is between 10^-(4) to10^-(6) for majority of >>> proteins.>>> >>> >>> >>> >>> >>> >>> >>> ----- Original Message ----- >>> From: "aidong" <a...@xmu.edu.cn <mailto:a...@xmu.edu.cn>> >>> To: <CCP4BB@JISCMAIL.AC.UK <mailto:CCP4BB@JISCMAIL.AC.UK>> >>> Sent: Saturday, July 03, 2010 6:26 AM >>> Subject: [ccp4bb] monomeric coiled coil >>> >>> >>>> Sorry for this ccp4 unrelated question. >>>> >>>> We recently have a protein that a multicoil program >>>> (http://groups.csail.mit.edu/cb/multicoil/cgi-bin/multicoil.cgi/cgi-bin/multicoil >>>> ) predicts to have very high probabilityfor dimer and trimer. >>>> Their>>>> scores are close to 0.4 and 0.6 for lengthsof more than 60 amino >>>> acids.>>>> However, two constructs that cover thisregion have demonstrated >>>> monomers>>>> in solutions by Multiangle lightscattering?! For the same >>>> question, we>>>> could not get any response from thisprogram manager therefore we >>>> turn to >>>> ccp4 for help. We wonder whether some of you might have similar>>>> experience. Thank you in advance. >>>> >>>> Sincerely, >>>> >>>> Aidong >>>> >*P** **please don't print this e-mail unless you really need to* Anastassis (Tassos) Perrakis, Principal Investigator / Staff Member Department of Biochemistry (B8)Netherlands Cancer Institute, Dept. B8, 1066 CX Amsterdam, The NetherlandsTel: +31 20 512 1951 Fax: +31 20 512 1954 Mobile / SMS: +31 6 28 597791
