I’m sure you’re dye-ing to know how strings of DNA letters turn into colorful peaks on a screen. Don’t FRET – actually do – because Forster Resonance Energy Transfer (FRET) is how the ENERGY TRANSFER FLUORESCENT DYES used for DNA SEQUENCING work! While some 20-somethings are preparing for their “big days” 👰🤵💒 I’m preparing posts about BigDyes 🤓 I’m in love with biochemistry! My “something blue” is ddG-EO-5CFB-dR110. And that “something borrowed”? Some energy from light will do…
Yesterday we saw how we can use DNA sequencing to “read” DNA. The DNA alphabet consists of 4 letters: A, T, C, and G and a molecule called DNA Polymerase (DNA Pol) helps link them together like a long string of cursive letters using another strand as a template. We can get DNA Pol to do this in vitro “in a tube” using polymerase chain reaction (PCR) to make lots of copies. Theoretically, you could link together an infinite number of letters because each letter comes with “ticket in hand.”
This “ticket” is energy in the form of phosphate groups. These groups are negatively charged and the “left arm” (5’ site) of nucleotide triphosphates (NTPs) have 3 of them scrunched next to each other, their opposite charges repelling each other, desperate to escape but they don’t have enough electrons to go it alone.
This is where the “ticket taker” comes in. The “left leg” (3’ site) of the end of the growing letter strand has a hydroxyl (OH group) which attacks the phosphate group closest to the new letter’s “body” (the α phosphate), kicking the other 2 off as inorganic pyrophosphate (PPi).
This changes the chain’s “seniority” – you know how some companies make the “new guy” do the busy work, DNA’s like that. The newly added nucleotide now becomes the ticket taker. In DYE-TERMINATOR sequencing, you trick the old ticket maker into “hiring” an incompetent new ticket taker -> a letter without the 3’ OH needed to grab onto another nucleotide (a dideoxynulceotide triphosphate (ddNTP). So the company stops growing.
With SANGER SEQUENCING (more here: http://bit.ly/2Tul9SJ ) you set up the reaction so that most of the hiring pool is competent, but eventually each company will hire one of those incompetent guys and have their “assets frozen.” And they’ll have different lengths when they do. You can then separate the different pieces of DNA by their length by using their natural negative charge to attract them through a gel column towards a positive charge – the bigger ones will get slowed down more because they keep getting tangled up in the gel’s meshy net as they go. more here: http://bit.ly/2SDKE8I
But you don’t just want to know the length, you want to know the culprit – what was the last nucleotide added that caused the stall? Was it an A, G, T, or C?
Say you have a sequence that ends in ????. With terminator nucleotides (ddNTPs)(?*), you can get pieces like ?*, ??*, ???*, and ????*. But that doesn’t tell you anything about what ? is. But if you label the terminator nucleotides according to their letter (i.e. ddATP is distinguishable from ddGTP is distinguishable from ddCTP is distinguishable from ddTTP) you start to get hints – you get pieces like A*, ?T*, ??G*, ???A* -> since they all start from the same place, you can just line them up and read them out
The sequence is ATGA
So that’s what you want to happen. But you need to make sure you “blame the right guy.” You need to be able to accurately tell the different terminators apart. So we want to label the ddNTPs – and we want to label each letter with a different label. And this isn’t as easy as it sounds.
The labels we use are fluorescent – FLUORESCENCE is where a molecule absorbs light at one wavelength (excitation wavelength) and releases it as light of a different wavelength (emission wavelength)
Light is ElectroMagnetic Radiation (EMR) – when people think of “light” they usually think of “visible light,” but that’s just a teeny little piece of the EMR spectrum squeezed in between infrared to the left (lower energy, longer wavelength) and ultraviolet to the right (higher energy, shorter wavelength).
Light can be thought of as little packets of energy (photons) traveling as waves through space. The only thing different between the light you see and the light you don’t see is the amount of energy in their photons. All light travels at the same speed (the speed of light), but the photons with more energy want to go “faster” but they can’t go faster forward so they just bounce more along the way – they have a higher frequency (think of a surfer getting hit by more waves in the same amount of time) and, to keep the overall forward speed at the speed of light, the peaks of their waves have to be closer together to compensate. So the more energy in a photon, the higher the frequency (f) and the shorter the wavelength (λ).
From the perspective of a molecule in light’s path, it’s like being bombarded with “coins” of “energy money” with different wavelengths corresponding to different coin denominations. Molecules can only grab the coins if they have the right size “coin slots” which are determined by where their electrons are housed and how much it costs for those electrons to move to a “luxury” house (higher energy orbital). They can’t maintain the high life for long, so they fall back down and give back “money” as light, but with lower energy because they used some of it “crying and stuff” (lost as heat, etc.)
So the emission wavelength is lower energy, longer wavelength (red-shifted) compared to the excitation wavelength and the difference between the 2 is called the Stokes shift. http://bit.ly/2Or1J3q
Different molecules have different sized “coin-slots” so they absorb different wavelengths and they give off different wavelengths of light.
If we only needed to label 1 thing, we could “just” attach a fluorescent dye to a ddNTP & shine laser light on it as it runs through the gel -> it would absorb that light and spit it back out at a different wavelength -> measure this is and voila! We’d only have a few things to worry about
The emission wavelength has to be far from the excitation wavelength so that the laser light doesn’t get confused for emission light. It’d also be great if that optimal absorbance matches the emission wavelength of common lasers.
The signal has to be strong enough to detect. The signal strength depends on the Extinction coefficient – how much do you absorb & the quantum yield – how much of what you absorb do you give back as light
The dye has to be “nonobtrusive” in terms of not messing up the thing you’re trying to measure. In our case, the dyes can’t interfere with Pol’s ability to add them to the chain (nothing can be added after it, but it still needs to be added itself (like sticking a cap on it). So we stick the dye sticking off the base pointed away from where the action’s occurring – the bases can still base pair and get added.
We also need to make sure the dyes don’t mess up how the pieces run through the gel. The gel separates DNA fragments by size. If you change the size too much (or the charge) you’ll alter how it runs, so the smaller DNA piece could overlap with a bigger piece, etc. (you might confuse ?T* for ??T* for example). So you want the mobility shift to be small.
We need to label 4 different things (ddATP, ddTTP, ddCTP, and ddGTP) and be able to tell them apart, which offers additional challenges.
Their emission wavelengths have to be far enough apart that you can tell them apart (have good spectral resolution). And not just the “peak” – fluorophores have a maximum emission (and absorbance), but there’s also “fuzziness” around it – the absorbance and emission spectra are like bell curves. And we want the curves to be “sharp” so they don’t overlap – we don’t want to mistake a T from an A, etc.
To help with this we can use energy-transfer conjugated dyes. These take advantage of something called Forster Resonance Energy Transfer (FRET), which involves a donor absorbing, light, handing energy (but not light) to an acceptor, and that acceptor then giving off that energy as light. (kinda like a money wire transfer vs. exchange of coins)
We label the ddNTPs with dyes in which a donor and an acceptor are physically linked. The laser light excites the donor electron, which hops up to a higher energy level. When the electron falls, it’s gonna give off energy. But instead of giving off that energy as light, it gives it off as “nonradiative energy” (not-light) that the acceptor molecule snatches up -> excites the acceptor molecule’s electron to a higher level -> now, when this electron falls it does give off its energy in light form (fluoresces)
We can use the same donor for each of the 4 letters so they absorb the same wavelength – so you can excite them all with a single laser as they run through the gel.But the donors are conjugated to different acceptors, which have different emission wavelengths, and we want their emission spectra to be well-separated.
Because the signal strength is proportional to the absorbance, you want the absorbance wavelength to be optimal, which is why these donor-acceptor dyes are great because they have the same optimum absorbance, so we don’t have to optimize for one at the expense of the others or use a wavelength that’s just “meh” for each of therm. And, ideally, the signal strengths for each of the letters should be comparable so that the quiet guys don’t get drowned out.
In addition to preventing overlapping peaks in “wavelength space” we also need to prevent overlapping peaks in “physical space” – we want the mobility shift as they go through the gel to be small and comparable for the different dyes.
The “classic” dyes used are “BigDye” terminators. They use 5-carboxy-dichlororhodamine dyes as acceptor dyes & 5- or 6-carboxy isomers of 4′-aminomethylfluorescein as donor dyes. The 5-carboxy-dichlororhodamine dyes are slightly different for each letter so they give off different colored light. Linking the linked donor-acceptor to the nucleotide is a propargylamino (PA) or propargyl ethoxyamino (EO) linker.
If you want more details on FRET – http://bit.ly/2FKtInv