IGEM:IMPERIAL/2006/The Talk Page
Purpose of this Page
Tom is away revising for exams and John is going away to the US. This page is a place where questions are raised and answered and a place to keep everyone aware of what is going on. Also, if anyone external to the project wishes to raise a point, contribute to a discussion or simply to ask a question please feel free!
Major Issues (Everyone please read and contribute)
Non-functional AiiA (Unresolved)
- John C: Data showing aiiA does not work
- Tom 13:55, 26 August 2006 (EDT): I'd be a little hesitant to condem AiiA with that data alone, for the following reasons:
- The control of having non-induced cells I don't think is adequate. Having non-induced cells as a control we are assuming the operator sequence is functioning and intact and we're also making the assumption there is enough native LacI repressor to effectively shut off AiiA production. We need a control where we have AHL innoculated with non-AiiA/LuxI/Flourescence producing cells.
- Secondly, the method for working out whether the difference between induced and non-induced cells is statistically significant is not given. Just because a difference is between 3-4% doesn't make it statistically insignificant by default. Especially if you're consistently getting those results. I'd suggest a T test on repeated samples to determine significance.
- Thirdly, the dilution of cells and small AHL innoculation time may be having an effect. I'd suggest growing a culture overnight in the prescence of AHL and IPTG, along with appropriate controls, and taking the supernatent directly from that for the T9002 assay.
- Johnsy 06:49, 27 August 2006 (EDT): I agree with Tom in that the AiiA may work. There are several reasons for non-expression (or maybe non-detection?) I have not tried to increased the IPTG amount, so maybe there is not enough IPTG to induce expression. Moreover, as our T9002 has been acting up within the past few days, it might be our T9002 cell activity which is the problem. However, having run an SDS gel with no results might signify that there is not expression of AiiA in the first place (or that there is not enough expression to produce any significant results). We might want to try to increase the IPTG concentration so that we are sure to induce AiiA expression and retry the SDS gel? just a suggestion...Then again, it could be the problem with the part itself in the registry. We can wait for another few days until our J37022 test construct finishes and try to run the tests using that construct.
- Tom 19:14, 27 August 2006 (EDT): I reckon first thing to do would be to induce overnight in the prescence of AHL. You can then do the T9002 assay straight from the o/n culture, and use some of the cells for the SDS page.
- JohnChattaway 07:53, 28 August 2006 (EDT) I think you may be right, the western blot expriment is flawed because DH5a can't express protein at high levels, to do that you need to use a different strain so the western blot may have failed for that reason. The new strain kirstin wants to use is in the next safety catagory above DH5a but that just means it has more forms to be fille in. Secondly I think that seeing as the activity of aiia is so low an hour may not be a long enough time period to get mesurable results I will try an assay which runs overnight and look into small chemostats to run it in.
- Tom 10:00, 28 August 2006 (EDT): I don't think a chemostat would be necessary. Just whack in some IPTG and AHL in a standard overnight culture. Just to act as a diagnostic as to whether AiiA is being produced and is working. You shouldn't need a chemostat until assembley of the final system is complete.
- JohnChattaway 04:48, 29 August 2006 (EDT) The reason no aiia is being produced is that we were testing a contaminant. The part should grow in kanomycin media only but we were testing in ampacillin so it was an amp-resistant contaminant growing, not the real part. will retest in kan medium
- Tom 06:27, 29 August 2006 (EDT): The AHL assay test data should be okay though, as I'm pretty sure John S tested in Kanamycin (I set up a culture with Deepti, defiantley used Kan). And as we were only doing the western blot because we didn't think that was working, doing the overnight innoculation/induction is still probably the best move.
- Johnsy 18:00, 29 August 2006 (EDT): Yes indeed I have been using Kanamycin for S01656 testing. Furthermore, I believe that we grew up the S01656 culture for SDS gel in Kanamycin, so there should still be some expression of AiiA. However, the culture might be contaminated already, since when we were growing up the cultures, we accidentally grew up in Amp, but we did have substantial bacterial growth :S...not really a good sign...and might be a contaminant of something else...but definitely not good. I would susupect that maybe the AiiA genome might have been mutated or as John C suspects, that the AiiA isn't being produced enough in the medium, so we are not getting enough expression to be measured.
Non-functional Prey Cells (Unresolved)
- John C: The prey cells don't flouressce! This is a majour problem!!! what are we doing about it?
- This data suggests that the prey cells are flourescing however there is no propper controll so the experiment needs to be repeated. I am pretty sure the increaced flourescence is due to the presence of cells not producing GFP. There needs to be a positive controll.
- My experiment yesterday showed that the prey cells DO NOT PRODUCE AHL! this is a big problem.
- How can we fix this?
- Tom 14:10, 26 August 2006 (EDT): When I did the initial testing of J37015 I initally got a AHL and flourescence flatline. However, repeating the experiment and missing out the dilution steps from the o/n culture I got this output for AHL (although I forgot to measure J37015 direct flouresence, oops!). The data does show that AHL is being produced.
- Tom 10:01, 28 August 2006 (EDT): Any updates on this?
- JohnChattaway 04:40, 30 August 2006 (EDT) I have decided that the prey cells are producing AHL based on tom's data. Here's how I think we can test the prey cell positive feedback loop, uising a chemostat.The rate of AHL production for a given OD and AHL concentration will be constant. The chemostat will wash out cells and AHL keeping AHLconcentration and OD constant therefore you will reach a steady state where AHLproduction = AHL washout. So by mesuring the concentration of AHL in the medium and knowing the volume of washed out medium we will know the rate of AHL washout which will equil AHL production thereby giving us the intrinsic growth rate of AHL.JohnChattaway 04:40, 30 August 2006 (EDT)
anyone thought of using an auxostat rather than a chemostat? (just to make it even more complicated)
- Tom 18:55, 26 August 2006 (EDT): You'll all be pleased to know it is possible to assess AHL levels via HPLC. They did it here. Dunno whether you can incorporate this into testing, but I defiantley think it's a more accurate way to assess AHL levels.
- Johnsy 07:11, 27 August 2006 (EDT): Again, the problem with using HPLC is that we will not have a direct measurement of the AHL levels in the actual oscillator. This will be okay for our testing phase, and actually might be helpful if we begin by testing the concentrations of the AHL in the stock AHL, so we are sure that we are using correct AHL concentrations. Degradation of the AHL (or incorrect concentrations) might account for some of the bloopers we have had in our T9002/J37016/J37020 testing.
- Tom 09:16, 27 August 2006 (EDT): There is facility in the bioreactor suite to take samples from the chemostat and run them through a HPLC within the room. There is no reason we can't use this to test AHL levels in the oscilator, although I think the GFP flourescence would be a nicer reporter. It could, however, prove useful if things don't go according to plan and we have to troubleshoot. And if you think the AHL is messed up, it might be an idea to make up a new stock when Vincent gets back.
- Farah 18:27, 28 August 2006 (EDT): John, you've landed!! And you're already typing on the wiki! Just have to say - thanks for leaving your lab book...As for the HPLC - I think we should go ahead and do it so that we can check our T9002 transfer curve but to save manpower and time I'd rather give it to the ABC to sort out than do it for ourselves.
- Johnsy 18:02, 29 August 2006 (EDT): Will the other stuff in the LB medium and the bacteria cell affect the HPLC readings, or do we have to purify AHL somehow?
- Tom 12:04, 24 August 2006 (EDT): Hi guys, something to note which I don't think has been that clear. In a growing cell culture cells grow and divide, increasing the volume of cells. If the production of a stable protein was shut off then the effective half-life of that protein would be the division time of the cells. IE, as the cell becomes two, the protein is shared between a volume twice as large as the original. Thus halving concentration. This needs to be taken into account when assessing protein half-lives.
- Although note, this does not effect AHL half-life as its volume is not constrained by the cell, but rather the fixed volume of the medium.
- Christin 12:51, 24 August 2006 (EDT): After cell division, the protein concentration would half - that means that that protein concentration is not in steady state any more but after a certain time, the steady state concentration would be reestablished again. How can this be taken into account?
Also, this time delay might have effects on the dynamics of the system.
- Tom 10:16, 25 August 2006 (EDT): You would still get establishment of steady states when protein is being produced. The only thing cell division is effecting is protein half-lives when we consider the volume of cells in the chemostat is always the same. IE, the washing out rate is the same as the growth rate. And thinking about it, in the chemostat AHL half-life would also be equivalent to the doubling time of Bacteria, IE ~30 minutes, as this is being washed out at the same rate. PS: Is the full model on the wiki yet?
- Johnsy 07:18, 27 August 2006 (EDT): Something else to note regarding the cell division: Once the cell divides, yes the concentration of AiiA is halved WITHIN each cell, but eventually, the AiiA concentration will again reach a steady state and return to the level before, which means that in time, there would be an actual increase in the amount of AiiA within the entire system. However, if we consider the prey cells, they too are dividing. Thus, also increasing the concentration of AHL (doubling it every doubling time). Intuitively, this would suggest that our amplitude would double every time the cells divide (which is not necessarily a bad thing?), but more modelling needs to be done on this to confirm.
- Tom 11:23, 27 August 2006 (EDT): The changing of amplitudes won't happen within the chemostat, as the number of cells remains constant. And the actual division of cells themselves doesn't effect protein levels in the way that there is a sudden drop and increase back to steady state. The act of division itself does not effect concentrations as the resulting daughter cells will still have the combined volume of the single parental cell. What is effecting concentrations is growth, which is continuous.
- Christin 19:11, 29 August 2006 (EDT): Wait a second ... the concentration within the cell as it divides does not change at all. As concentration is number of moles per volume, the concentration stays the same. However, are you saying that the growth of the cell before it divides will affect the concentrations? Yes, it does, but is there such a large change in cell size that it will have a significant effect on the concentrations?
- Tom 07:35, 21 August 2006 (EDT): Hi everyone, just been reading through this. Big drawback of the 3 antibiotic way of monitoring cell population that you've identified is the extra ligation steps. What extra ligations steps are involved? If you just mean grafting your parts into a plasmid with appropriate antibiotic markers, can't you just co-transform a seperate plasmid with the marker on, thus saving a ligation step? I'm pretty sure those kind of plasmids would be easy to get hold of as they are the basis of a lot of gene cloning.
- Tom 10:03, 21 August 2006 (EDT): Actually, on second thoughts, that wouldn't work, as you'll end up losing the plasmid containing the antibiotic you're not screening for while the cells are growing
- Johnsy 11:30, 21 August 2006 (EDT): From what I understand from the team, we will use the antibiotic way of monitoring as a backup plan, in case we decide to monitor cell populations. For now, we have not decided to monitor cell populations, but we will go ahead and electroporate & Maxi the parts we might need. Just in case...
- Deepti 15.24, 22 August 2006 (EDT) -Talking to Dr. Jensen yesterday morning, it seems that the 3 antibiotic method is not very accurate an requires extra ligations steps which can be avoided.
- Waiting for plated samples to grow will require atleast a day before we can assess the differences in population sizes.
- The number of bacteria in the plated samples does not necessarily accurately reflect the relative proportion in the common broth culture.
- Even if the plating method were to be an accurate measure of the colony sizes in the broth culture, we could still only record, and not control population sizes
- According to Dr. Jensen, the best way to do it would be the OD method.
- By growing up the bacteria separately, we can measure their separate growth rates which, hopefully, should not be too disimilar.
- When mixing the two cultures, the ODs should be equal, and preferably, between 0.6 and 0.8.
- Tom 13:05, 22 August 2006 (EDT): Would it be possible to check the population ratios after they've been mixed together using the OD method?
- Deepti 09.10, 23 August 2006 - I'm afraid not. Once the cultures are mixed there is no way to tell them apart using the OD method. Best we can do, I reckon, is ensure the OD of both individual cultures is between (0.6-0.8) right before mixing and then just hope they stay more or less equal (we will have a rough estimate of the individual growth rates of each population from the OD measurement of the individual cultures done beforehand).
- Tom 05:27, 23 August 2006 (EDT): If we're gonna be using the chemostat for the final assembley (which I assume we are), Dr. Leak made it clear that it amplifies the smallest difference in compeitive growth ability. Meaning that once innoculated, there's a good chance the ratios are gonna change. I strongly recommend looking more into this, and looking into obtaining LacZ plasmids.
- Farah 12:22, 21 August 2006 (EDT): What are we doing about luxI in J37015? Are we going ahead with measuring the GFP levels in that part? Tom mentioned talking to Vincent about it...I can model the luxI but from the assumptions we've made for our current model for J37015 we ought to get a similar curve as for the AHL levels. To Tom: I remember you saying you've talked to Vincent about the luxI - what was the conclusion?
- Tom 15:50, 21 August 2006 (EDT): Yeah, we decided it might be easier and more accurate to measure the GFP from the pLux transcript as opposed to AHL. We'd still like to measure AHL to see the relationship between LuxI and AHL, however the relationship may be easier to characterise and more accurate than the T9002 assay.
- Deepti 23 August 2006: What would be the point of measuring LuxI if we cannot characterise the relationship between LuxI concentration and AHL concentration? The only way we could do that is to measure both compounds (using both protocols) and come up with a curve that describes their relationship. Correct me if I am wrong, but as far as I can gather, the reason for using this alternative protocol is to achieve greater accuracy than with the AHL assay. Using the alternative protocol we would gain a more accurate measurememnt of LuxI concentration but the AHL measurement would still depend on the original assay whether we use the curve produced by the alternative or not!! (Since the curve is extracted by comparing data from both assays).
Should we not stick with the original assay then, inaccurate as this may be? Unless ofcourse we do want to measure the concentration of LuxI in the system, for which I'm not sure I see the reason for.
LuxR in the Predator Cell
- Tom 15:50, 21 August 2006 (EDT): I remember when Farah and Christin assumed LuxR in the predator was constant we still got oscillations. If this is the case what is the point of LuxR being under the control of pLux as opposed to a constitutive promoter? Also, if this is the case and LuxR variance isn't important, than our design is exactly the same as the preivious MIT design, except spread over two cells. Maybe someone can look at the effects on the system of having LuxR vairable and LuxR constant?
- Farah 11:29, 22 August 2006 (EDT): When Christin and I assumed LuxR to be constant we were just 'hiding' the fact that LuxR is actually varying - as was pointed out by Vincent. I don't see the poing in modelling a simplified version of the system when our more 'accurate' model is working.
- Tom 13:39, 22 August 2006 (EDT): I think it would just be interesting to see the differences. We've made LuxR vary, and if we still get oscillations when it isn't varied people may ask why we bothered to make LuxR variable in the first place.
- Farah 18:20, 28 August 2006 (EDT): Sorry I didn't answer you sooner about this - just noticed your comments buried in the mass of text on this page. We don't really have the option of choosing whether levels of LuxR vary or are constant. Our predator (J37016 or J37020) just simply has varying levels of LuxR, so assuming constant LuxR levels in the modelling would just hide what was actually going on in our system (we tried it but this point was pointed out to us by Vincent). If we adapted the predator to produce constant LuxR then we could make the approximation but this could only be done with J37020 which has two promoters - doing so for part J37016 would mean producing constant amounts of AiiA too.
- Johnsy 07:08, 27 August 2006 (EDT): What is going on with the Cre/Lox part? Have we started to ligate it yet?
- Jonny 12:59, 27 August 2006 (EDT) This is being done by Ms. Jensen and now I think Dr. Mann also. Basically, we have to do two seperate PCR reactions and then fuse them together. One of the PCR reactions appears to be fine, however the other has failed four times. The reason for the failures is thought to be the very long overhanging fragment for one of the primers - around 60 base pairs. They have been playing around with the different temperatures, but there has been no success so far. If it continues to fail, then an option is to order two seperate sequences that anneal to each other which should then make the PCR simpler. Unfortunately none of the other ligations steps are possible without the PCRed fragment.
- Johnsy 18:04, 29 August 2006 (EDT): Ok...cool...but if it doesn't work, is it a very necessary part to have for our prey cell? Have we finished testing the J37015? or rather..how is the testing going for the J37015?
- JohnChattaway 05:55, 22 August 2006 (EDT): I am curious about what data this test will actually give us. It seems like you are measuring the time taken to reach steady state rather than characterising the positive feedback loop. I think that a much better test would be;
- Make a fresh day culture of J37015 (OD 0.4 (guess))
- Make a T9002 AHL assay
- then every 10 mins for 2hrs
- Take 1 1ml sample
- Record OD
- Centrifuge down cells, collect supernatant
- Put supernatant on ice
- After 2 hours this should leave you with 12 samples of different AHL conc in the ice bucket you can test these with the T9002 assay.
- We would then normalise the AHL reading against OD
- Tom 07:07, 22 August 2006 (EDT): That's pretty much exactly what the protocol says to do. Although it needs to be adapted to look at GFP output over time also. Also it'll have to be made more clear if you had trouble working out what was going on, any volunteers? :D
- Tom 18:43, 26 August 2006 (EDT): Hi guys, just thought I'd pop in with an idea. We were saying how we wouldn't be able to have a live demonstration at the jamboree because our period is too long. However, what about a simulated live demonstration? So we have a computer generated visual representation of the culture running off our model, and we can change parameters like cell ratios, growth rates and chemostat ODs. That way people can come along and fiddle with the paramters and be like 'wow, I can chage frequency/amplitude really easily!'. Plus, as it'll be based off a mathematical model, it isn't gonna go wrong in front of them. What do you reckon? I personally think audience participation is always a big plus. Also, just generally and not thinking about the jamboree, it'll be interesting and potentially very useful to see how the model responds to these new input parameters that we can easily change.
Working Bank Holiday Weekend
- Farah 18:37, 27 August 2006 (EDT): This is to John. Can you come in tomorrow afternoon to read the plates and culture up stuff for tuesday? I can call/mail you to let you know when you need to come in. We've worked it out so that then you, Christin, Jimmy and I will only need to contribute a half day...Our fate lies in your hands....
- Farah 17:51, 28 August 2006 (EDT): We've gathered a lot of data this weekend. Could everyone process whatever data they have (using the spreadsheet)and update the results page?
- Johnsy 18:05, 29 August 2006 (EDT): Pity they didn't have a holiday here in the US...but we do have one next Monday 4 Sept! :) Good to see you guys still dedicated!