Skip to Content

Chemical Biology

Methionine Is Now Your Friend

I’ve been meaning to mention this new paper, from a mixed Berkeley/UCSF team, which could open up a new area for chemical biology. If you want to label a protein, you have to have a handle to do it. “Label”, in this sense, can mean almost anything. You could be sticking a fluorescent group of some sort on there, something with an odd isotope pattern for mass spec analysis, a spin label for NMR experiments, even a tritiated or other radioactive side chain, what have you. There are no end of experiments and assays to be run once you’ve got some sort of detection technique in place, but you’ve got to get it in there first.

If you want these things covalently attached (and you often do) there are limited options, which often start and end with cysteine residues. Those SH groups are reactive enough to pick up various electrophiles, and the techniques that depend on them are legion. But there are problems. Not every protein has a Cys in a place that’s useful for your experiment, and not all Cys groups are reactive in quite the way you might need. Some proteins are ungrateful enough not to have one out on the surface at all, and other have them, but they can’t be modified without messing up some key part of their structure or function. There are other side chains that can also be modified with electrophiles, of course, but after Cys these reactivities start to get patchy and hard to predict.

This new paper details a way to modify methionine side chains, by first oxidizing them selectively. That’s a welcome addition to the toolbox, because Met is much less likely to be involved in something crucial as compared to Cys. (On the down side, Met is a much less common amino acid, but you could spin that as giving you fewer possibilities for multiple products). The method uses an oxaziridine reagent to produce a sulfimine (via the sulfoxide), and that gives you access to a whole variety of things hanging off the new nitrogen atom. For instance, you can label the Met side chain with a “click”-ready alkyne or azide group (as is demonstrated in the paper).

If there’s not a handy methionine, you can always engineer one, and the authors demonstrate this on an antibody fragment to green fluorescent protein. They were able to functionalize the new Met residue (and thus the antibody) with a variety of species and retain antibody potency, which gives you a potential new route into antibody-drug conjugates. Indeed, they went on to show just that on trastuzumab (Herceptin), engineering a methionine into it and attaching the chemotherapy agent monomethyl auristatin E. The resulting species showed enhanced activity in a cell assay, as it should.

The other thing you can do with this technique, as with the Cys labeling methods, is distinguish various types of them in the proteome. Not all cysteines are created equal, and not all methionines are, either, since they can be involved in oxidative reactions in vivo, which can be important for protein function and regulation. The oxaziridine treatment can identify the hyper-reactive Met residues, in the same way that electrophile screens can pick out the hyper-nucleophilic Cys ones.

So this looks like just the sort of thing that chemical biology people have been looking for – a new handle, with new chemistry, on a newly useful amino acid. We’ll see what people make of this now that it’s out there!

13 comments on “Methionine Is Now Your Friend”

  1. anonymous says:

    I wonder if the author tried this bioconjugation technique on small peptides? My experience is that one can get away doing these thing with with antibodies but small peptide could pose a problem.

    1. CuriousPeptide says:

      It’s not immediately obvious why smaller peptides would be more problemnatic. Can you elaborate? I could imagine scalability and different conformations may prove challenging, but not reactivity. Thanks

      1. Roger Moore says:

        I think the problem with peptides is that each amino acid is much more likely to be involved with the function. With an antibody, you can find something away from the CDRs and know you aren’t going to kill binding. With a peptide, you’re much less likely to be able to hang a big group off it, or even modify a single amino acid, without worrying about how it will affect function.

        The flip side, though, is that it’s practical to synthesize peptides chemically. Many peptide drugs already incorporate unnatural amino acids to tweak their biological activity- often by making them more resistant to degradation- so once you’ve found an amino acid that can be replaced without affecting function, you have much more freedom in what to replace it with.

        1. Sebastian Worms says:

          And this also make the authors’ finding less useful for peptides: if you have a permissive site you can already incorporate all kinds of probes directly or through insert a non-canonical aa witha clikc chemistry handle.

          1. anonymous says:

            @ Curious peptide : Well, both Roger Moore and Sebastian Worms could not have been more eloquent expressing my thoughts! Bottom line- AB molecules offer several opportunity unlike the small peptide. Specifically, am thinking of Bombesin receptors natural ligand. There is that methionine subunit and you tinker with it you are screwed!

  2. Next level says:

    Imagine how powerful photosynthesis could have been if nature ever questioned whether thio tyrosine could be tolerated

  3. Kelvin says:

    Seems like one of those very academic, overly engineered solutions to a non-existent problem. If you want to label a protein and there are no cysteines, you just engineer one in on the surface somewhere. Job done.

    1. Isidore says:

      Usually it’s not a problem of cysteine absence, after all most interesting proteins have cysteines. The problem is that you often need to partially reduce disulfides in order to make cysteines available for conjugation and then you have to figure out which cysteines were modified and to what extend, all of which is doable but takes time and resources. And you may also be disrupting protein folding by reducing disulfides. With methionines the problem is simpler. Time will tell how robust this method is and also how scalable.

      1. Kelvin says:

        Most internal disulfides (and even unpaired cysteines) are resistant to reduction and modification within the folded proteins, whereas too few or many on tge surface can be easily engineered in or out, as appropriate, and usually without causing any disruption as long as it’s away from any critical binding sites.

        But seriously, has anyone ever had any real problems with this, which could not be addresed with a bit of simple protein engineering? And I mean in practice, not just in theory?

        1. Ted says:

          Why should you have to hassle with engineering a cysteine into a protein when there is a readily available methionine to label? What about a commercial IgG that gives problems when labelled via cysteine or lysine? Label, clean, screen – one day, maybe two.

          Our lab spends a lot of time thinking about protein presentation/labeling – new methods are always welcome, and this is one we’ll be trying in the near future.


          1. RAM says:

            let us know how it worked out

  4. milkshake says:

    FYI, methionine sulfoximine is seriously neurotoxic – it is a metabolic poison that blocks production of glutamine and glutathione. This was discovered accidentally in humans and animals fed flour bleached with nitrogen trichloride; NCl3 flour bleaching got phased out as a result. L-methionine sulfoximine is now used as a post-transfection tool to block endogenous glutamine synthease, so as to weed out all non-transfected mammalian cells grown in glutamine-starved media.

    One needs to be very cautious about non-proteinogenic aminoacids, at least a dozen of them are vicious toxins

  5. Me too. my experience is that one can get away doing these thing with with antibodies but small peptide could pose a problem.

Comments are closed.