I have a strong affinity for protein purification, and an avid interest in streptavidin. I love its spectacular ability to bind biotin and how we can harness the couple to get proteins stuck. But if you want to de-stick them – good luck! When sticking reversibly is what you have in mind you’ll want modified versions like Strep-Tactin & “Strep Tags” that less-strongly bind. I love protein purifications, but this is probably my last couple preps for a while as we prepare for a transition to working from home to help flatten the curve, so I thought today I’d take you behind the scenes with me as I do “strep tag” affinity chromatography.
Chromatography is just a fancy dancy way of saying we take a mixture of things in a “mobile phase” (like a mix of proteins in a buffer (pH-stabilized salt water) and get it to travel through some “stationary phase” (like little beads called resin packed into a column). As the things in the mix travel through, they interact differently with the mobile phase (e.g. some proteins bind the beads, others don’t), so they get separated (either some things get stuck and don’t come out, or they come out, but with a delay). The word chromatography comes from it initial being used to separate colored (chrom-) pigments.
In affinity chromatography, the proteins you want selectively stick to the resin, while the proteins you don’t want flow through. And then you push off the proteins you do want with a competitor to get them all by themselves. But of course, you can’t just will the proteins to stick or not stick, you’ve got to stick to the biochemistry! The protein you want has to have something that none of the other proteins have, and which the column likes to bind specifically. Thankfully, with recombinant protein expression this is “easy” to achieve – since we’re providing the genetic instructions for making the protein (recombined into a piece of DNA called a vector which we stick into cells in a flask to make it for us), we can add a bit on to those instructions to add a bit onto the end of the protein that can serve as an “affinity tag” – and one of my go-to tags is the “strep tag”
This particular strategy is based on biotin-avidin/streptavidin interactions. Those are some of the strongest known non-covalent and thus “easily reversible” interactions and we can harness their power to help us purify proteins by sticking a biotin mimic as a tag on our protein and binding it to little beads coated with a streptavidin mimic. And don’t worry – that’s not blood in my column – I’m just “leeching” out the desthiobiotin with the HABA dye vampire!
Streptavidin is a naturally-occurring protein that was isolated from a bacterium called Streptomyces avidinii. It’s part part of a 2-component antibiotic complex these bacteria use to protect themselves from other bacteria by binding small peptides called stravidins that the bacteria need for making biotin.
Why would they need it? Biotin is aka vitamin B7 and it’s an important co-factor (helper molecule) for some enzymes. Bacteria can synthesize biotin, but we can’t – but that doesn’t mean we can’t use it. There are enzymes in our cells that use it to help carboxylate things (add CO₂ to them).
Streptavidin is tetrameric, which means it hangs out in groups of 4 (with each individual protein on its own being a “monomer”). And since each of the monomers in the tetramer is the same we call it a homotetramer.
It looks (and acts) A LOT like another protein called avidin even though sequence-wise they’re only 30% identical. They both bind biotin: avidin even binds a little stronger – but avidin also has properties that make it less appealing – it’s glycosylated (has sugar chains attached), positively-charged (cationic), tends to aggregate (clump up) etc. There’s also NeutrAvaidin, which is deglycosylated avidin, which tries to get around some of those problems. But, although avidin binds free biotin better than streptavidin, streptavidin is better at binding biotin-tagged things.
Avidin was discovered in egg whites after it was found that animals that were fed an egg-white diet were having health problems. They called it “egg white injury” and it turned out that what was happening was that avidin in the egg whites was binding & sequestering biotin (potentially to prevent bacteria from having it).
The binding of biotin to avadin is the strongest known non-covalent interaction (covalent bond are bonds like those linking the generic parts of amino acids – they’re strong and involve actual electron sharing instead of non-covalent interactions which are just attractions without electron co-habitation)
When speaking of binding strength or “affinity” we usually use a term called Kd which can seem kinda counter-intuitive because the LOWER the Kd the STRONGER the binding. Kd is the “dissociation constant” (it’s only constant for a single protein under certain conditions, it’s not like a generic constant). And it’s the inverse of the affinity constant – so Kd = 1/Ka
More on Kd here: http://bit.ly/2JyBbuj
But basically if you think of binding as a reversible process A + B <> AB, Kd says how much A do I need in order for half of B to be bound.
If B really likes A (high affinity) it’ll latch on almost anytime they happen to meet, so it won’t take much A to get all the B stuck, and you’ll have a lower Kd. But if B has a lower affinity it won’t always stick when they meet, so you’ll need to have more chances of chance encounters, so you need more A to get the same amount of B stuck, so you’ll have a higher Kd.
Biotin has a Kd for streptavadin of ~10^-14 M (femtimolar range!) So if we want to get a protein stuck to something and STAY stuck, biotinylating it’s great. This is good for things like microfluidic experiments where you bind things to chips and flow different liquids over them containing things you want to see if will interact with the thing bound to the chip.
But sometimes strong isn’t a good thing. When you’re purifying a protein with an affinity tag you need the tag to bind to the resin strong enough that it’ll stay bound when you wash the column to remove non-specific binders. But the interaction has to be weak enough that you can get your protein unstuck without harming it. Biotin binds streptavidin so strongly that the conditions needed to get it off are so harsh they’ll probably hurt your protein
So, instead of sticking “real biotin” onto our proteins and binding it to “real” streptavidin, we use modified forms.
For the biotin mimicking part, I use a “Strep Tag.” The Strep-tag is an 8 amino acid (protein letter) sequence (Trp-Ser-His-Pro-Gln-Phe-Glu-Lys) (WSHPQFEK). In terms of amino acid qualities, this sequence is pretty well-balanced so it doesn’t affect the properties of the protein like its overall charge.
I use a Twin-Strep-tag (WSHPQFEKGGGSGGGSGGSSAWSHPQFEK) which has that bindable sequence twice with a linker in between them. Since glycine’s really small and flexible (it’s side chain is just an H) the tag can contort itself so that both sites can bind a single bead (kinda like the His tag we looked at last week did). But unlike with the His tag, where each bead only had 2 spots open, the Strep-Tactin resin has 4, so it can bind more. More on the this His tag: http://bit.ly/2lzYxVE
Instead of sticking it on after the protein’s been made like what happens when proteins get biotinylated inside of cells, it gets put on as part of the protein because we add the instructions for it in front of or at the end of the instructions for our protein that we stick into cells to make it for us.
And speaking of sticking things places, we need the other half of our “velcro” – the streptavidin mimic. The resin I use is a modified form of streptavidin called Strep-Tactin that binds less tightly (so more reversibly). Like always, this is NOT an endorsement, it’s just what our lab uses. The Kd of the Strep-Tag II to Strep-Tactin is about 1μM (10^-6) compared to that Kd of ~10^-14 of biotin to streptavidin.
You have to take into account volumes – usually we have a LOT more than 1mL of lysate (I often have more like 100-200mL) and my protein’s spread out all throughout it – so concentrations are low and I need to take into accounts kinetics (speed of binding)
So I need to make sure that all of my protein has a chance to find resin. So I usually use about 5mL resin (10 or so for large preps like the 2 I did today) and bind it in “batch mode” – add lysate to resin & let it spin on a hotdog spinner in the cold room for ~30 min before column-izing it by pouring it into an empty plastic/glass cylinder with a filter at the bottom and letting the liquid flow through (trapping the resin and the bound protein)
Then I wash it with the same buffer the protein’s in and the resin was equilibrated in. This washes off anything that’s just kinda randomly stuck on. When I do insect cell expression, I co-express my protein with YFP that tells me that things are working – more on this here: http://bit.ly/2YQtPFJ
So that glowing yellow’s not the protein I want and it gets washed off.
You know how I said proteins can get biotinylated in cells? That can cause problems when it comes to purification with a biotin-binder…
There’s a lot of biotin & biotinylated stuff in bacteria, so I don’t use Strep Tags for bacterial expression. But there’s less endogenous biotin & biotinylated proteins in non-bacterial cells like the insect cells I use to express proteins. So there’s less gunk to randomly stick. Which is important because that’s the real biotin so it binds stronger – so we’re gonna have to use some harsher conditions later to force the stuff that’s really stuck off & regenerate the resin (but AFTER we give our proteins a polite shove…)
So at this point I have my protein mostly by itself, but stuck to the beads and I need to do the shoving. Biotin could easily compete off our protein but then we still have the problem of needing to get the biotin off the column so we can use the resin again. So we turn to a slightly less tight binder that’s still a great competitor – desthiobiotin. As the “desthio” part of the name indicates, desthiobiotin is biotin minus the sulfur. Biotin has a Kd for streptavadin of ~10^-14 M (femtimolar range!) whereas desthiobiotin has a Kd of ~10^-11M (100uM) (remember lower Kd is stronger affinity).
Then I add the elution buffer (which has the competitor, desthiobiotin). I let ~ 1/2 a CV (column volume) flow through (e.g. if I had 10mL resin, I’d let out ~5mL which is the liquid that was in the column before the new liquid got there). I collect this as my elution 0 (E0) – it shouldn’t have much of my protein but it’s good to keep & check.
After letting that through, I close the stopcock and give the desthiobiotin a chance to find & bind the strep-tactin (pushing my protein off). Then, after ~15 min I open the stopcock to let the liquid out (E1). Then I do more elutions with the elution buffer, collecting fractions I can test for the presence of protein (e.g with a Bradford assay or UV-spectrophotometer). Once protein stops coming out I can stop.
So now I have my protein largely by itself and unstuck but the resin’s now stuck to the desthiobiotin & any biotinylated (the real deal) proteins, biotin, etc. that was naturally in the cells. So I need to not-so-politely shove that stuff off.
Thankfully, although it’s a protein, streptavadin – and its Strep-Tactin derivative – is REALLY sturdy. It can hold up to harsh conditions that’d kill most proteins (8M urea or guanidine, 0.5 M NaOH, 50% form amide, proteases, etc.). And the tetramers only start turning into monomers if you raise the temp above 60°C.
This stability is great because it allows us to clean the resin with harsh conditions (after we get our protein off with mild conditions) without damaging the resin itself.
To regenerate the resin we need to remove the desthiobiotin to open the spots back up. Can do this with NaOH or a dye called HABA (2- [4’-hydroxy-benzeneazo] benzoic acid) – which displaces the biotin and turns the resin red. (if the resin doesn’t turn really red it’s probably time to use some new resin – but the resin can be reused many times). more here: http://bit.ly/2ROUIHW