I’ve used Ni-affinity chromatography to purify His-tagged proteins a lot, but until today, try out Cobalt I had not! This resin was once blue so what did I do? I stripped the Nickel off it and charged it with cobalt to give it this pretty hue – and hopefully purer protein too!
Why the NTA wardrobe change? It’s not that cobalt’s the “fall fashion line” for IMAC (Immobiolized Metal Affinity Chromatography). Instead I’m trying it out because my Ni was leaching out and taking my protein with it en route! So let’s see if cobalt’s as resin-sticky but protein-picky as cobalt proponents like to tout!
IMAC is a way to help purify proteins that we’ve added a “His tag” onto to make them specifically sticky to resin (little beads) that are coated in a metal such as nickel (Ni) or cobalt (Co). So when we express a protein with a His tag then break open the cells (lyse them) and flow all that stuff that was inside the cells (lysate) through it, our protein will stick, other proteins (without the tag) won’t, then we can get our protein to unstick – all alone!
The sticking works (hopefully) by the metal acting as a go-between connecting the resin and the protein (specifically the chain of Histidine (His) amino acids we’ve attached to the end), unlike some other resins where the protein interacts directly with the beads. So with IMAC you have bead:metal:His-tag where those : are something called coordinate covalent bonds, which are where an atom with a lone pair (like that N in the His chain) shares a pair of electrons with a metal.
And these bonds are reversible – which is good for your protein being able to come off – but not good if it comes off to easily – and not good if the metal comes off the beads. So instead of just binding in one place, you want something that “bites” the metal multiple times – and we call such molecules that can form multiple coordinate covalent bonds CHELATORS.
So you have a chelator (like NDA or NTA (more below)) “PERMANENTLY” stuck to the beads. And then that chelator clamps down around the metal. And then the string of Hises you’ve attached to your protein acts as a chelator that bites down around the metal too.
The resin we use for IMAC is usually beads made of agarose coated with such chelators – the one I’m using is NTA – nitrilotriacetic acid (NTA), which is a tetradentate chelator (4 sites). You might have seen cobalt resin advertised as TALON – that’s just a brand name for a slightly different linker – a patented carboxymethyl aspartate. Nickel and Cobalt offer 6 places to bite, so you still have 2 spots left for your protein to bite.
So that His-metal bite is weaker than the resin-metal bite (2 bites vs. 4), so you can get your protein to come off (elute) without stripping the metal off, just by adding a His mimic called imidazole that acts as a competitor. Imidazole mimics a single His so can only take a single bite – it’s not a chelator so you have to add a lot of it.
If you add a stronger competitor like EDTA, which *is* a chelator and can bite down on all 6 of the metal’s bite spots. you can get the metal to come off too. So you can turn your blue resin un-blue. So you can strip the metal off and give it a wardrobe change into a different metal – like cobalt.
Co & Ni are both “transition metals” – you find these in the center of the periodic table and they’re good at giving and taking electrons (can have multiple oxidation states – more here: http://bit.ly/31MPb9w )
They’re really similar in a lot of ways, and both bind the beads and your protein. But Ni has 1 more proton (giving it an atomic # of 28 to Co’s 27). And they have slightly different “tastes” when it comes to molecule dating.
It’s kinda like every time a protein meets the metal the metal has the choice of “swiping left or right.” The chances depend on how attractive the metal finds the protein. The higher the affinity, the more attractive. But if it finds everyone attractive, you have low specificity.
If your protein’s less attractive, you’ll need more metal-coated beads so you have more of those encounters – even though each encounter still has a lower probability of sticking, if you have a lot of them your chances improve. And as long as its affinity’s a lot better than the contaminating proteins – even if it’s still not great – you’ll get the binding to be almost all your stuff
Nickel finds His more attractive than cobalt does. So, cobalt has a lower affinity but higher specificity because it’s “pickier” and its pickier because it has stricter spatial requirements – the Hises have to be next-door neighbors like you have in a His tag or at least specially spatially positioned
Ni-NTA has a higher capacity – according to QIAGEN, it can bind up to 60mg 6X His-tagged protein (there’s that mg thing again… see yesterday’s rant on how you shouldn’t use protein weight without defining the protein since proteins all have different weights and I want to know if the resin’s binding 1 giant protein or lots of little ones). Co-NTA has a capacity of ~10
And what about the leaching stuff? Leaching is when the metal comes off the beads, not just your protein, and when you don’t want it to – this is more of a wardrobe malfunction! Nickel is more likely to unintentionally flash someone because it can actually bind the resin in a couple different ways – it can form 2 different coordination complexes. One of these complexes forms a nice little pocket that’s ideal for tag-binding with Ni bound to 3 carboxyl groups & 1 N. But the other complex has a flat (planar) structure, where nickel’s bound to 2 carboxyl groups & 1 N. This doesn’t bind the Ni very tightly, so the Ni can leach off.
With cobalt, the linker bites the Co harder, but the protein bites it more weakly. So the Co is less likely to leave the column, but your protein is more likely to leave the Co. So – Ni & Co both will bind the beads and your protein, but Co binds the beads tighter – so it’s less likely to leach off the beads (this is what I was aiming for). Co may bind beads tighter but proteins, not so much. It has a lower affinity than Ni for His, so, unless you use more resin, it might yield a lower yield. But, since it binds non-specific stuff worse too and that stuff bind worse to begin with, you end up with a purer protein product!
In today’s TIL… When I was looking online to see if it’s a-okay to strip & swap the metal from the Ni-NTA resin we buy, I saw some references to cobalt being blue. And isn’t “cobalt blue” even like a crayon color? But the cobalt I found on our lab shelf was cranberry-colored. That’s because cobalt’s color depends on who it hangs out with.
I prepared the cobalt solution from Cobalt Chloride hexahydrate CoCl2·6H2O- the “cobalt chloride (CoCl2) part” tells me that, in this salt form, each cobalt atom hangs out with 2 chlorine atoms (the Co has a +2 charge that’s neutralized by the -1 charge of 2 Cl’s). And the “hexahydrate part” tells me that each of these trios is surrounded by 6 water molecules.
And apparently when cobalt chloride is “wet” it’s pink or purple but when it’s “dry” (anhydrous) it’s blue. When you get the “anhydrous” form wet it first forms a dihydrate CoCl2·2H2O which is purple. And when you add even more water it turns pink. So cobalt can actually be used to indicate the presence of water and they make like water test strips out of it to test pipes for leakage, use it to indicate desiccator capacity, etc.
When I added imidazole-containing buffer, the color changed to a purpler form, but that’s just because when the imidazole coordinates with the cobalt it changes the absorption a little bit.
Note: Another affinity tag I use is a Strep tag which binds streptactin resin. I use the Strep tag when I do insect cell expression, but the His-tag with the bacterially-expressed stuff. His is underrepresented among the protein letters (amino acids) – there are 20 (common) ones yet His only makes up ~2% of amino acids you actually find in proteins. And, you’re more likely to find adjacent ones by chance in eukaryotes than prokaryotes – less contaminating proteins binding non-specifically when used for bacterially-expressed proteins
more on IMAC: http://bit.ly/2Fs8taY
more on column-stripping & regeneration: http://bit.ly/2XGQiZ6