Photobleaching Model

From Course Wiki
Jump to: navigation, search

Trying to figure out a better model for the photobleaching of Sybr Green I in the DNA melting lab, because our current model fails at long times.

What is the current model?

The Matlab DNA melting simulation writeup has a fixed percentage of the active fluorophores being destroyed every time step. The bleaching coefficient B is set such that half the fluorophores will have been destroyed by the end of the simulation, if you had a constant input.

I think this makes at least some sense, because if you have the fluorescence be low and then go up at the end, then the bleached is more than half of the unbleached, because there were inactive fluorophores not being destroyed... argh I am confused.

How is it done in EstimateDnaMeltingParameters? 

bleachingCorrection = (1 - bleachingCoefficient .* cumsum(thermalQuenchingCorrection .* dnaFractionVersusTime)).^bleachingPower .* ...
                      (1 - bleachingConstant .* (0:(length(dnaFractionVersusTime) -1))');

What are bleachingCoefficient and bleachingConstant?

How does Sybr Green actually photobleach?

Photobleaching is "a very poorly understood phenomenon" [1]. Remember, the other word for it is photodegradation.

To a first approximation, every fluorescence event has an equal chance of causing the destruction of that one fluorophor (Wikipedia). Also fluorophor lifetimes given in terms of avg number of photons they can emit before dying.

  • Call Sigma Aldrich? Call Invitrogen/Life Technologies? Who all sells Sybr Green I? (Life Tech owns it and licenses to other companies. Molecular Probes is a sub-entity of Invitrogen.)
  • Zipper et al [2] is the classic sneaky structure-finding paper. Also talks about the importance of the stoichiometric ratio (dbprs). Need to reread this carefully because last time I was paying attention to a different aspect.
  • Henary & Mojzych [3] book chapter. Says Chen et al (ref 7) found first order photobleaching kinetics.
  • Chen et al [4] says first order kinetics for a cyanine-family dye. Full text not available grrrr. So I don't know how long they treated the dye, or whether they really mean pseudo first order.
  • Epling and Lin [5] find first-order kinetics in bleaching some dye using TiO2 as a catalyst, but no bleaching without catalyst. Unclear if this is applicable to us. Also, they seem to have used constant light input.
  • Wikipedia on cyanine dye applications notes that the dye to basepairs ratio matters; above 1 dye molecule per 60 bases, you can start to see quenching. Oh, fuck. Do we have to account for this too? What dbprs have we been using? Does it matter if the DNA is ss or ds? (i.e. if you melt the DNA does the dbprs in the tube increase?) The Zipper paper talked about this...
  • Ward & Marples [6]

"DNA-bound photobleached SYBR Green I is reported to lose its DNA-binding affinity and dissociate; free dye can then re-bind to DNA thereby restoring the fluorescence (Molecular Probes, Inc. personal communication). Consequently, it is possible to rescore images stained with SYBR Green I following photobleaching by storage in the dark and subsequent re-hydration." WHAT WHAT WHAT is this also definitely calling Molecular Probes tomorrow :) and maybe Ward & Marples too :) (Also, looking at their figure 1A, looks like they see first order kinetics too.)

  • Widengren & Schwille [7] check out equation 5, it looks pretty much like what we had in the matlab page. But... their timescales are different, and their derivation rests on assumptions about timescales. In particular, because they're looking at a small volume, they assume that photobleaching and diffusion have the same time scale, which is not true for us. (Well, is it? What the hell is the timescale for photobleaching anyway?) Also, they're using an autocorrelation equation, which I don't really understand how it works.
  • Patel et al [8] maybe we should use Sybr Gold? Does it have more endurance? See top of 2nd column on 1st page.
  • Aside: this paper [9] supports the "linear thermal quenching" idea
  • [10] and [11] talk about bleaching of organic chromophores. Not sure it's exactly relevant but I should skim them later.

Need to understand fluorescence correlation spectroscopy, then read this and this

Prior Photobleaching Experiments

It looks like the general procedure is to point light at it and fry it.

Fluorescein

Frequency-Domain Measurement

Gaigalas et al [12]: frequency-domain measurement using lock-in amplifier. A cool-sounding procedure, well suited to the existing DNA melter, and apparently also beneficial because it allows you to figure shit out without knowing lots of fiddly molecular details.

Second paper by Gaigalas et al [13] ugh clearly I have not understood the first paper well enough

NOTE: we should also do some constant-illumination experiments that completely burn out the SG, to check the frequency-domain results on the long time scale. Try different illumination intensities.

Understand How It Works: Why does the frequency response look the way it does? What exactly is the meaning of the quadrature/in-phase ratio? (Get some context from the use of this technique on other problems.) Follow through the equations and analysis. Do we need to introduce flow, or not? Try rederiving model without flow.

Cyanine Dyes

  • Silva et al [14]
    • apparently the deal with cyanine dyes is that "fluorescence quantum yields can be very low in fluid solution but are significantly enhanced in conformationally restricted environments". Huh. That explains a lot.
    • ...and thus if you put them in something viscous like glycerol, you can get a hundredfold higher quantum yield. Really??? whoa.
      • "(The low fluorescence of the dyes in methanol necessitated the use of 90% glycerol in water to observe fluorescence in the absence of DNA.)" We could do this. They seem to have gotten about 1/4 as much fluorescence in 90% glycerol as in aqueous solution with excess DNA (Fig 5). WARNING: my instinct says that the higher yield might be due to some molecules being stuck in the correct state and some in the incorrect state, so we might expect photobleaching behavior to differ somewhat. Therefore do not use glycerol for photobleaching experiments.
      • Does glycerol get less viscous with heat? Yes. Good question. Pending.
    • "The absorption maxima for all five dyes shift to the red upon binding to DNA, indicative of the lower dielectric environment of the stacked base pairs (Table 1)." Oh, that's what the Invitrogen guy meant.
      • Also, in glycerol, their tri-whatever orange derivatives seem to emit about 10-20nm redder than in DNA. I dunno if that's a problem with our filters.
    • "Dyes that bind with higher affinity to double-stranded versus single- stranded DNA induce a stabilization of the former, leading to an elevated melting temperature (Tm). For example, [Poly(dA-dT)2], a double-helical DNA polymer consisting of alternating A-T base pairs dissociates into single strands at Tm = 42.4 °C under the conditions described in the supporting information, but in the presence of TO at a ratio of one TO per 2 DNA base pairs, the melting temperature increases by 22 °C." FUCK
    • How they did it: irradiate 1min, take spectrum, repeat; 10min total irradiation. Control: same but block irradiating beam. Very high powered irradiation (150W with water cooling) -- that must be why such a short time.
  • Talhavini [15] studied fluorescein in polyvinyl alcohol -- possible polymer for us to use? (Kaduk says that a nonpolar polymer gel-like thing might work a lot better than trying to find glycerol-plus.) Of course, diffusion is absent and that might be bad.
    • Also, "temperature dependence of photobleaching" OH GOD.
  • Renikuntla et al [16]
    • cyanine dyes tend to aggregate (nonfluorescently) at high concentrations. But this goes really weird with DNA... the "nonfluorescently" part might go away, but it also affects the wavelength... whattttt
    • Test singlet oxygen mechanism: "This suggests that the fluorinated dye should be less susceptible to reaction with singlet oxygen, a prediction we tested using the photosensitizer eosin Y. This dye has a high quantum yield for production of singlet oxygen (0.57 in water)12 and allows singlet oxygen to be introduced to the system without excitation of the cyanine dyes." (As opposed to Fenton-like hydroxyl radical mechanism and others.)
  • Also uh these people seem to be using 150W Hg(Xe) arc lamps, not LEDs, which is why they can get results so fast.

Experiments

How to investigate photobleaching? Just shine light at samples for an hour?

I wonder if there's anything we can add to the solution to reduce photobleaching. Of course, it'll also affect melting, so I don't know if it's a good idea.

Grumble. This isn't a problem in RT-PCR because you measure at the same point of every cycle, not continuously. But it is possible to take melting curves with the LightCycler -- look that up.

A thought about the glycerol thing -- why doesn't this work with gels? Or does it? Viscosity is not the same as gelification, I suppose.

NB: photobleaching kinetics should depend on photobleaching mechanism, which may depend on strength or frequency of illumination. Results from radically different setups may not apply to ours.

To Do

Fluorescein and quantitative photobleaching models:

LightCycler:

Media:

  • Glycerol
  • Whatever glycol
  • Polyvinyl alcohol
  •  ?

Look at the SNR measurement VM for "two lock-in amplifiers at once". That's kind of like what it does. Play with frequencies.

Retrieved from "http://engineerbiology.org/w/index.php?title=Photobleaching_Model&oldid=4021"