Skip to main content

Analytical Chemistry

Tie Me Proteins All Down, Sport

There are a lot of slick ideas in molecular and chemical biology that depend on immobilizing proteins or small molecules onto solid supports. Consider affinity chromatography: if you can tether a “bait” onto some solid matrix, you can then flow all sorts of mixtures over it (gorp from freshly lysed cells, for example) and let the tightest binding partners stick to the column material while you merrily wash everything else off. Surface plasmon resonance is another example: there are funky electromagnetic effects (those “surface plasmons”) that take place at a thin layer of gold on a chip, and these are altered depending on the size/weight of molecules that are bound to the surface. So if you can tether a small molecule, a protein, a piece of DNA (whatever) to that surface, you can then flow analyte solutions over the chip (drug screening candidate molecules, say) and actually watch their binding events as they stick to the immobilized target and then leave.

All great things, and there are plenty more where those came from. . .if you can get your desired species tethered to the solid support. And if you can get enough of it on there. And especially if you can do these things in a way that leaves your bait species in a useful state – that is, exposing the same binding surfaces in the same way that it would in solution. That ain’t trivial. If your bait is a small molecule, the situation is especially fraught, because you’re going to have to significantly modify its structure to introduce a linker that can tether to the solid support. You’ll probably want some reasonably long spacer group in that linker – make it too short, and your species has its face jammed up against the solid surface. Too long is not so great, either, because those linkers tend to ball up on themselves like an old coiled telephone handset cord, which can sometimes cause odd behavior, too.

You can buy a lot of beads, powders, chips, etc. that are already functionalized with reacting partners (carboxylic acids or activated esters, primary amines, click-triazole-precursor azides or alkynes, and so on). But you’re going to have to build in the complementary reacting functional group into your small molecule of interest, and you’re going to have to do that from some non-crucial part of its structure. You may well not know where the non-crucial parts are, though, and even when you identify them, the synthetic routes they leave you with may not be friendly ones. If your bait is a protein, the situation is improved, but not as much as you would probably like. Proteins generally have a number of different side chain residues sticking out from their surfaces, and that gives you a variety of handles to choose from. Problem is, you might not be able to choose: you can end up with a heterogeneous mix of different-immobilized proteins, each of which might have somewhat different behavior (and some of which are now inactivated).

That inactivation problem is a constant pain, actually. It’s almost always the case that a tethered-down protein performs less vigorously than it does under free-living conditions, but it’s too often the case that it dies completely. Here’s a new paper that tells us some more about the problem. The team (groups from Oregon State and Colorado) are looking at everyone’s favorite prototype enzyme, carbonic anhydrase, as it get immobilized through several different chemical reactions. CA really gets a brutal workout in chemical biology, it needs to be said. It’s a sturdy enzyme, extremely well characterized, and it’s cheaper than some forms of dirt. So it gets wheeled out every time someone things of something weird to do to an enzyme!

These groups have been working on this problem for a while, and they’ve already found that a big problem is that proteins unfold and denature in contact with some surfaces. They seem to have a separate sort of diffusion along the solid interface, and as they bump and roll along the surface they can encounter small regions that favor an interaction with some sort of inactive, partially unfolded protein state. That led immediately to the thought that the less of this wandering around, the better, and this current study looks at several reactions with a wide range of rate constants. Other variables were minimized by building the ligation site into the carbonic anhydrase via an unnatural amino acid at the same residue in each system (126, well away from the active site).

Looking at the activity of the enzyme after ligation, and at its unfolded state (via tryptophan fluorescence) did indeed show a strong correlation between the speed of the ligation reaction and the amount of well-folded, active immobilized protein. The faster it can find a home without encountering some denaturing region on the surface, the better. You really do wonder what those regions are like, but as the authors mention, these things are probably nanometer-sized and are going to be hard to characterize. But since the surface in this case was silica, you could imagine some local concentration of free Si-OH silanols or something of that sort.

The paper even shows that the wandering-through-the-desert process can be tracked by making a fluorescent variant of the CA enzyme. The resulting FRET experiment confirmed that the faster ligation reactions do have the expected shorter trajectories along the surface. There were all sorts of behaviors, though: some of the CA enzyme molecules remained well-folded the whole time and found a home after only a brief search. The longer searches, as mentioned, tended to lead to more unfolding – and in many cases, the protein was unfolded by some surface contact and then went on to wander along the surface some more in that state before finally binding. Overall, there is a strong correlation of proper folding to immobilization as close as possible to the initial point of contact to the surface. Meanwhile, some of these unfolded proteins just wandered out into solution again before binding to anything. That makes a person wonder how many proteins get denatured (to some extent or another) every time they’re passed over a solid support.

As the paper notes, these results also explain some past reports that nonspecific tethering strategies led to better protein behavior than trying to set everything up perfectly. The nonspecific ones have a greater chance of immobilizing more quickly, though (more residues can be involved), whereas the specific-setup ones will likely have to roll around on the surface for longer before the right match is found. And that increases the chance of unfolded trouble. So in general, you want to set up your ligations to take place as quickly as possible, such as the use of the tetrazine/trans-cyclooctene system, which is red-hot compared to many of the more traditional ones. Chemical biologists take note!


12 comments on “Tie Me Proteins All Down, Sport”

  1. Ezra Abrams says:

    I spent some time putting Antibodies onto latex
    And the answer always was, take your plate of mAbs and try a bunch till you find one that works

    As for “denatured on a surface” this has been known for a long time – pretty sure people studying BSA (which is, iirc, sort of like a coiled spring, on a surface it unfolds to reveal the phobic inside) knew this a while ago


    1. Aaron says:

      I worked on an immunoassay development project for a time in grad school. The specific, oriented attachment chemistries never worked as well as I wanted them to, and it’s bothered me ever since. Thanks for writing this up Derek.

  2. Ezra Abrams says:

    One time, for a sandwhich assay of some sort, we were having trouble finding an antibody that worked when put on latex
    I forget all the details, but my boss said, Ezra, why don’t you see what alternatives there are

    So I find this company in the UK that makes Rabbit mAbs, and I have a phone call with the guy

    he says
    our fee is a flat US $50,000 dollars to start (2005 dollars)
    we only do biz with companies like Merck, and they come to us when they can’t find a mouse mAb that works

  3. Another Guy says:

    “God made the bulk; the surface was invented by the devil” – Wolfgang Pauli

  4. theasdgamer says:

    This was an interesting article, Derek.

    I have a suggestion for a future article. Could you ask someone–maybe Dr. Young–to do a guest post about the model of the covid disease and how it progresses? That might be very helpful in resolving some things.

  5. Wat says:

    Are you referring to Dr. Robert O. Young? Because this is on his webpage.

    “In 1994, Dr. Young discovered the biological transformation of red blood cells into bacteria and bacteria to red blood cells.”

    That would be like a human being able to turn into a snake and a snake to turn back into a human. There are orders of magnitude of difference between structure in bacterial cells vs erythrocytes, which are eukaryotic.

  6. David E. Young, MD says:

    So it gets wheeled out every time someone *thinks* of something weird to do to an enzyme!

    (minor spelling error)


    1. theasdgamer says:

      If you would be so kind as to discuss doing a post with Derek. Please discuss things like mean time to viral load and mean time to viral clearance as well as the clinical progress. Thank you.

      Paging Derek…

  7. Siddharth Dasgupta says:

    I wonder what the effect is on DEL Technologies

    1. sgcox says:

      DEL is not with surface immobilised targets . Enrichment is after ligand binding event happened in solution (usually). Might have some effect on AC-MS screening but do not think it would be that much.
      More practical example here is probably the amine coupling in SPR. After pH scan people often choose the value with the highest response which is basically a measure of “stickiness” and might be not very wise choice, for protein might be already damaged in such condition.

      1. polyalanine says:

        The protein bait is immobilized. The DNA encoded small molecule is in solution.

        1. sgcox says:

          Is it always ? I thought immobilisation, say, strep beads are added after binding of ligand to bait. Most protein pulldowns, especially in cells are done this way. Should work fine for DEL too, why pre immobilise unless for particular technical assay setup ?

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.