Keep your ION the prize – pure protein! Lost? Use the isoelectric point (pI) to guide you and your protein of interest on your Ion EXchange chromatography (IEX) journey! IEX is a common protein purification technique in which we separate proteins based on charge. Proteins are made up of amino acid letters, some of which are sometimes charged (depending on pH). Different proteins have different amino acid spellings and thus different charges and we can exploit these different charges to separate them using columns filled with charged resin (basically little beads).
Unlike affinity chromatography (Ni-NTA, streptactin, etc.), which recognizes a specific feature of a specific protein (often a tag you genetically-engineer onto the protein when you express it) and thus is only useful if the protein you’re trying to purify has that feature, IEX can be used for “any” protein without relying on a tag. Therefore, you can use it on never-tagged proteins or proteins that were tagged but then you cleaved the tag off after the affinity chromatography step and now you want to purify the tag (and other lingering contaminants) away. That’s what I’m doing. And here’s how it works. First an overview and then more details on protein charge and practical advice on IEX
Today is day 3 of epic protein purification week (5 days, 6 protein purifications, >60L worth of insect cells, not enough time for coffee…) and I’m purifying a protein using (as one step in a multistep workflow) ion exchange chromatography. It’s a technique in which you separate proteins by their charge by getting them to bind to oppositely-charged resin (little beads) in a column, washing the other stuff off, and then, once all that stuff’s far away, pushing your protein off with salt. Since a salt (such as table salt, NaCl) is made up of a positive part (e.g. Na⁺ in table salt) and a negative part (e.g. Cl⁻ in table salt) that come apart when you dissolve them (it both dissolves and dissociates), it provides a source of both positively-charged & negatively-charged competitors. The more opposite the protein & bead charges are, the stronger the charge-opposite charge interactions, and thus the more salt it’ll take to get pushed off. So if you gradually increase the salt you can push proteins off based on their charge, separating them in the process.
Because of that whole “positive part AND negative part,” a salt can be used to unstick either a negatively-charged (anionic) OR a positively-charged (cationic) protein (e.g. the Na⁺ can compete off, thus exchanging for, a positive protein and the Cl⁻ can compete off a negative protein). But in order for the protein to stick to the resin in the first place, that resin needs to be oppositely-charged compared to the overall protein. So if your protein’s +-charged (we call this cationic) you’ll want – charged resin (like an S column) but if your protein’s -charged (we call this anionic) you’ll want + charged resin (like a Q column).
I always had a problem remembering the terminology, but when it comes to ion exchange chromatography – the flavors refer to the type of ion that’s being exchanged – NOT the one that the resin – so anion exchange chromatography uses positively-charged resin to bind negatively-charged (anionic) proteins – and cation exchange chromatography uses negatively-charged resin to bind positively-charged (cationic) proteins.
How do you know what to use?
Whether, in which direction, and to what extent, a protein is charged depends on “how it’s spelled” (what protein letters (amino acids) does it have in what order?) and what the pH is. And the 2 are directly related – pH is a measure of how many protons (H⁺) are in a solution – it’s an inverse log, so the more H⁺ (more acidic), the lower the pH. Some amino acids are able to give protons (act as acids) or take protons, (act as bases) depending on the proton availability, and since those protons are positively-charged, give/take-ing them changes the protein’s charge. Gaining a proton (protonation) can positivize a neutral or neutralize a negative whereas losing a proton can negativize a neutral or neutralize a positive.
I like to visually think of solutions as being a sort of “proton cookie jar” and the pH reflects the fullness of that jar. The fuller the jar (more free proton “cookies” floating around) the lower the pH and vice versa. Acids and bases are like friends donating (in the case of acids) or taking (in the case of bases) proton “cookies,” so the number of acids & bases in a solution influences the pH. But not all acids and bases are created equal – some acids are more generous and some bases are more greedy.
Proteins are like a big linked-up group of friends where some of them are cookie givers or takers. The more “proton-greedy” a protein is, the more positively-charged it will be, and it will remain charged even when there’s not a huge excess of protons (super acidic solution)(cookie jar not very full). Conversely, a “proton-generous” protein will tend to be negatively-charged unless the pH gets really low (cookie jar overflowing).
No matter how greedy or generous your protein is, there’s some “tipping point” where the protein is neutral overall. I stress “overall” because the individual letters are the places where the charges are coming and going – and they act “independently” – so the overall charge of the protein is the sum of all those pluses and minuses. When the number of pluses = the number of minuses the protein has no overall charge (even if parts of it remain charged) and we call the pH at which this global neutrality is reached the isoelectric point, pI.
Interactions between charged chains & water contribute to keeping proteins dissolved. & like charges repel so when protein molecules are highly charged the same say they can repel each other, so usually a protein is the LEAST soluble at its pI, but this isn’t always the case.
For any protein you know the sequence of, you can calculate the theoretical pI using a pI predictor (free online software like ExPASY ProtParam can do this for you). What it gives you is just an estimate of the pI, since the letters can be locally influenced by their neighbors. So the pI its algorithm spits out is a “best guess” about how proton-greedy your protein is. It tells you the “theoretical pI” – above this pH, there aren’t enough protons for all the takers to take them*, so your protein is likely negatively charged (anionic, and thus suitable for anion exchange chromatography). Below the pI, there are plenty of protons for even the reluctant takers to take them, so your protein is likely positively-charged (cationic, and thus you would choose to use cation chromatography).
*note: it’s not really that there are no free protons – there always are lots of them, but if you think about protons coming and going frequently, when there are fewer protons around it’s harder to grab another one if you let go. So I don’t want to confuse anyone with my oversimplified analogy, but I personally find it helpful when trying to figure out whether something will be charged or not
If you need a review of pH, check out: http://bit.ly/phacidbase
But here’s the gist and how it relates to protein charge. Molecules are groups of atoms held together by strong, “covalent” bonds – the type of bond that doesn’t break easily. Those atoms (individual hydrogens, carbons, oxygens, etc.) are made up of even smaller parts called “subatomic particles” which include protons (which are positively-charged), neutrons (which are neutral), and electrons (which are negatively-charged). ⠀
Even though they’re attached by covalent bonds (strong bonds that involve the sharing of pairs of electrons), bonds to H are weaker than bonds to other atoms. H’s are really small & don’t offer much latchbility so they’re relatively easily removed. Just how easily depends on what they’re attached to. Hydrogen atoms only have 1 proton & 1 electron & they often leave that electron behind when they go, so all they’re left with is a H⁺, so we often call H⁺ a proton.
These protons can come from water itself, which is mostly H₂O, but also contains O ⁻ & H⁺. Or they can come from other molecules, including proteins. Protons can easily find a water to latch onto (forming hydronium ion (H₃O⁺) because the vast vast majority of water is in water form (not ionized). The ions are spread out throughout the solution, but for the sake of visualization you can think of all the protons as being in a jar, and for the sake of fun analogy-ness you can think of these protons as cookies! So, water is just cookies waiting to be made!
You can think of a solution kinda like a middle-school cafeteria with a cookie jar. pH is a description of how full the jar is and it is a property of the cafeteria as a whole – it comes from contributions from all the students (molecules), some of whom share cookies and some of whom take them, and proteins are like “cliques” of students whose contributions we can view at the individual student (amino acid) or clique (protein) levels.
Like any molecules, proteins are made up of atoms (of elements like hydrogen (H), carbon (C), oxygen (O), nitrogen (N), & sulfur (S)) joined together through covalent bonds (strong bonds that involve atoms sharing pairs of electrons). In any solution you have a fixed # of each type of atom – that is, you can’t do alchemy & voila suddenly have gold. You can’t change the total # of hydrogens, but you can change how they’re distributed.
The atoms in proteins are arranged into amino acid building blocks that are linked up to form proteins. You can think of the individual amino acids as “students” that link up to from large “cliques.” There are 20 (common) amino acids, & proteins can be made up of “any” # of any combination of them (e.g. you can have 15 “Janes” in one clique).
We can talk about “cookie-having” at the clique (protein) level or the friend (amino acid level). At the clique (protein) level we talk pI (ISOELECTRIC POINT) and it comes from what’s happening at the friend (amino acid) level, which we measure as pKa.
Some amino acids have more H⁺s than other amino acids. If they “look at the jar” & see that there are plenty of cookies, they say “why should I give up my cookie if there are already cookies out there?” So they keep it. Others would like a cookie but they have better “self-control” – not until the jar’s pretty full will they take some for themselves.
Different amino acid “friends” have different levels of greediness, which we describe using a # called pKa. The pKa is the pH at which 1/2 of a thing will be protonated & half won’t be. Below the pKa, there are more free H⁺ (fuller cookie jar), so the thing’s more likely to be protonated. Above the pKa there are less free H⁺ (less cookies in the jar) so the thing’s more likely to be deprotonated.
But not all things are protonatable/deprotonatable. They have to have an H for one thing that H can’t be too tightly held. Each location that is protonatable/deprotonatable has its own pKa. So let’s look where on proteins we can find these sites.
Amino acids all have the same generic backbone that they use to link together & then they have unique “side chains” aka “side groups” aka “R groups” The generic backbone part has an amino (-NH₂) group on 1 end & a carboxyl -(C=O)-OH group on the other end. They link up amino to carboxyl in an amide aka peptide bond —NH-(C=O)— , shed a water in the process & the part that’s left we often refer to as the “residue” – so when alanine is by itself, we call it a free amino acid. But when it’s in a chain (peptide) we call it a residue.
When you have free amino acids, they all have at least 2 groups that can give or take hydrogens. The amino end can exist as -NH₂ (neutral) or -NH₃⁺ (+ charged) & the carboxyl group can exist as -(C=O)-OH (neutral) or -(C=O)-O⁻ (- charged). So a free amino acid can be +, -, or neutral. The pKas of these groups depend on the amino acid but are typically ~10 & ~2 respectively.
Physiological pH (the pH inside your body) is about 7.4, which is above 2, so most of the carboxyl groups are protonated & thus in the negatively-charged carboxylate form. But below 10, so most of the amino ends are in the +-charged NH₃⁺ form. So the charges cancel out – we call this a ZWITTERION
When you link amino acids together, you’re only left with 1 amino end & one carboxyl end. So you only have to take those group’s protonation state into account once. But some of the side chains in the residues in between those ends can get protonated or deprotonated as well & when we talk about pKa’s for amino acids in the context of proteins or peptides, we usually are talking about the pKa of the side chain (if there is one). Sometimes this is referred to as the pKR or pK3.
There are 2 side chains that are frequently deprotonated at physiological pH – we call these ACIDIC because they donate H⁺s & when they do they become negatively charged & now capable of accepting H⁺s (acting as a base) so we call them “conjugate bases.” It can seem kinda confusing because “acidic” residues often play important roles by acting as bases in their deprotonated form. The “acidic” refers to its NEUTRAL form being acidic.
These side chains are: Aspartic acid (Asp, D)(pKa ~3.65), which gives up a H⁺ to become aspartate & Glutamic acid (Glu, E)(pKa ~4.25) which gives up a H⁺ to become glutamate. Cysteine (Cys, C) & tyrosine (Tyr, Y) can can also deprotonate to give negatively charged chains, but they do so much less readily – pKa of 8.4 for Cys & 10.5 for Tyr.
There are 3 side chains that are frequently protonated at physiological pH – we call these “basic.” Lysine (Lys, K) (pKa ~ 10.28) & Arginine (Arg, R) (pKa ~13.2) are predominantly protonated at cellular pH, but Histidine (His, H) (5.97) is more “iffy” (remember that pKa tells you when HALF the groups are deprotonated on average so it’s not like you hit the pKa & bam they’re all deprotonated – you have a mix. Also, an important caveat is that pKas are context-dependent so the pKa you get from a table is likely close to but not exactly the “real” pKa in the situation you’re looking at.
So this is all happening at the “friend” level, but what if we want to know about the clique’s “cookie-ness”? We need to take into account that clique’s unique combination of friends -> We calculate the pI (isoelectric point), that pH at which the protein is neutral overall. It doesn’t mean that each individual amino acid is neutral, just that the non-neutral ones balance out.
So proteins with high pIs are like cookie-greedy cliques. You have to get them to see that the jar’s really low & others are desperately cookie-deficient before they’ll donate (at the net level)
The higher the pI, the more H⁺-rich the protein is likely to be -> you have to take it to a higher pH (more basic meaning less free H⁺) before it will shed H⁺s. We often refer to such proteins that are + charged at neutral pH as “basic” & they get that “basicity” because they have lots of basic residues (His, Lys, &/or Arg). They’re great for binding negatively-charged things like DNA or RNA.
Proteins with a low pI are like generous cliques. They don’t need cookies to be happy and will only take it once they know there’s plenty for everyone!
Proteins with pI’s below neutral are “acidic” and they usually have lots of Glu’s and Asp’s.
in summary for this background part before I get into the more practical tips: Key points to keep in mind
🔹 pH is a measure of free protons (H⁺) which are positively charged, but pH is on a negative log scale so LOWER pH means more positively-charged protons floating around
🔹some parts of proteins can grab onto them – the more there are (lower the pH) the more likely this is -> increases charge (an make neutral side chains ➕ charged or ➖ charged side chains neutral)
🔹when pH is high, kiss those H⁺ goodbye! at high pH there aren’t many protons so this is less likely and some protein parts can actually donate them to the cause -> decreases charge (➖ side chains can’t get neutralized (they stay ➖) And ➕ side chains have to “give up” their H⁺ (they get neutralized)
🔹proteins have different combinations of parts that give & parts that take
🔹as a result, a protein has an overall charge that depends on the pH
Each protein has a specific point at which it at which the protein OVERALL is NEUTRAL – there can be ➕ & ➖ charged chains, but they perfectly cancel each other out & we call this charged but neutral condition one of my favorite words: ZWITTERIONIC. The pH at which it occurs the pI or ISOELECTRIC POINT
🔹 Above the pI, the protein is ➖
🔹 Below the pI, the protein is ➕
Now that we understand the molecular magic behind ion exchange chromatography, let’s look a bit closer at how we can but this knowledge to practical use.
As I hinted at earlier, there are 2 “flavors” of ion exchange chromatography: ANION exchange and CATION EXCHANGE. Ions are charged things and basically you “exchange” ions from salts (like the Na⁺ or Cl⁻ of NaCl (table salt) with protein ions. Then you can gradually increase the salt concentrations so that those salt ions outcompete the protein and you get another exchange. Or you can change the pH to change the protein’s overall charge (the lower the pH, the more free H⁺ for the protein to latch onto -> become more positive & vice versa)
In CATION EXCHANGE chromatography you have negatively charged resin & you’re binding & exchanging positively-charged (cationic) proteins & salt ions.
ANION EXCHANGE chromatography is the opposite – you have positively charged resin & you’re binding & exchanging negatively-charged (anionic) proteins & salt ions.
The more oppositely-charged something is, the more salt it will take to compete it off. And since different proteins have different charges, they’ll come off at different salt levels.
I’m currently purifying a protein that is “basic” – it has a predicted pI ~9. My buffer is at 8.0, which is below the pI. Lower pH means more protons, means fuller cookie jar, so the protein will take – and when it takes those H⁺ it gains positive charge. So at the pH it’s in, it should be cationic.
So I want to exchange it for other cationic things, which called for CATION EXCHANGE CHROMATOGRAPHY – the “-“ thing stays stationary (negatively-charged groups permanently stuck to the little beads in the column) while I swap out the “+” thing the beads are bound to – first I exchange salt (like the cation Cl⁻ coming from KCl) for protein – and then I exchange that for salt again (but I have to really flood it with salt at this point in order to push it off)
Normally with ion exchange I set up an automated method where I equilibrate the column with a buffer that’s the same as the one my protein is in, load the sample using a sample pump that uses a tube like a straw to suck up the protein solution until it senses air, wash the column with the same salt concentration it’s already at, and then go straight to a gradient elution – gradually increasing the salt concentration and collecting what comes out in ~1mL portions in a deep-well block.
Sometimes you have to do some troubleshooting: If no protein comes off, your protein could have precipitated on the column (another indication of this problem is if the column pressure increases because the clumped up protein blocks flow). One reason this could have happened is the initial low-saltiness – In the equilibration step you wash the column with a lot of buffer at the salt concentration you want to start at. This is the same salt concentration your protein is currently in, which is usually diluted from the salt concentration you do the initial purification steps at – proteins often aren’t happy in low salt environments – and they’re more likely to bind non-specifically because there’s less charge-based “competition” – so I normally do the affinity chromatography steps at a higher salt concentration. But with IEX you want to start low before you can go high – because you really want to be able to get a good gradient – and you need your protein to stick – so you don’t want too much competition in the beginning -If the salt’s too high, it can “win” from the start. So I dilute the protein first. If you’re worried about precipitation you can take a small test sample of the protein and dilute it to various salt concentrations to check for “crashing out” (precipitation) before you risk diluting it all.
When you lower the salt, In addition to lowering the competition, you “unshield” the proteins from the sea of charge around them (lower the ionic strength) so it’s easier for opposite charges to find one another – even if those charges aren’t that strong. http://bit.ly/phacidbase
note: if you’re doing it after tag cleavage, cleave the tag before you dilute it so the cleaver (protease) can find the protein to cleave more easily)
Another thing to keep in mind – at high salt concentrations, because the liquid’s so “unattractive” for water-avoiders you get enhanced hydrophobic interactions (basically the charge-avoiders would rather be with other charge avoiders so they stick together.
Remember that the charge of the protein is pH dependent! The further the pH of the buffer is from the protein’s pI, the more charged the protein will be. So if you want your protein to get the protein to bind tighter, run farther away! Raise the pH for anion exchange and lower it for cation exchange – but don’t stray too far or the protein won’t like you.
For more practical protein-purification posts (and background/theory), check out the new page on my blog where I’ve collected some of my protein purification posts. http://bit.ly/proteinpurificationtech ⠀