I'm thinking of crosslinking two proteins. The crystallographic nature of the interaction and the binding motif for each molecule is know. The paper says the intermolecular contacts (over 50; van der Waals radius 5 angstrom) are made between the motifs in these two proteins. I also know that one of the motifs on one protein has a sulfhyryl (Methionine) and for the other motif on the other molecule I (think) can use a primary amide reactive moieties. Now this is where my question is.

Would I be correct in thinking that the best cross linker to use is a heterobifunctional cross linker, with a spacer of 10 (or 5?) angstrom in order to pick this interaction up specifically and stabilise it?

I desperately need a guidance on this. I have used the life technologies crosslinker selection tool and the crosslinking technologies hand book but I can't quite figure out based on the information I have, what cross linker length I need to use. I'm not worried about the cleavability, solubility and membrane permeability parameters.


I know that the binding motif on the first protein it has Tyrosine, Proline and Methionine.

Looking at the groove contacts, the second protein is interacting with this motif, using Tyrosine and Glutamine BUT the second protein also has Tryptophan, Lysine, Glutamic acid and Methionine.


I found specially close Glutamic acid (E) and leucine (L) on my first protein, specially close to Lysine (L) on my second protein although its not in the consensus sequence for the second protein. They are less than 10 angstrom apart. is that any better? So presumably I can use the homobifunctional linker?

Many thanks in advance

  • $\begingroup$ Is there a cysteine available, or just methionine? Methionine won't be very useful. What about lysines? $\endgroup$
    – user137
    Jul 15, 2015 at 21:59
  • $\begingroup$ Are site specific mutations a possibility? $\endgroup$
    – user137
    Jul 16, 2015 at 15:47
  • $\begingroup$ Also, does the linkage need to be at the binding site? Because bioconjugate chemistry usually won't get that kind of specificity if your reactive amino acid appears elsewhere on the protein. $\endgroup$
    – user137
    Jul 16, 2015 at 15:58
  • $\begingroup$ I can't do point mutations but I was thinking of doing linkage at the binding site since that's one of the closest contact points that is known and hence using a 10 angstrom linker would reduce non specific linkages. I'm completely new to cross linking hence any guidance would be appreciated. Currently I am planning and amide amide homobifunctional linker at 10 angstrom. Since I calculated the molecular distance from the binding residues on two proteins. I used the X-ray data and SPDBV for calculations. Is that advisable? $\endgroup$ Jul 16, 2015 at 20:09
  • 3
    $\begingroup$ Met does not contain a sulfhydryl, Cys does. $\endgroup$
    – MattDMo
    Jul 16, 2015 at 22:18

2 Answers 2


There are general guidelines for cross-linking protein-protein interactions, but specific parameters (i.e. cross-linker sub-types, spacer length, linker concentration, reaction pH and temperature) are often determined experimentally. Common protein cross-linkers target reactive functional groups that are more likely found on the surface of the protein than in its "active site" or "interface" that is often buried.

Given your relatively hydrophobic, non-reactive binging motif (Tyr/Pro/Met), if you do not need a large amount of cross-linked protein sample, then you could label it with photoactivatable amino acid such as L-photo-Met. Since Met is not often found on the surface, internal cross-linking would stabilize your protein and its binding partner.

If you need a large amount of cross-linked protein (e.g. in mg) for further structural studies, labeling can be costly. You could possibly cross-link the protein (complex) crystal itself (since there were crystals as you mentioned), often just by glutaraldehyde. If the complex is there in the crystal, it would now be as solid as rock.

  • $\begingroup$ Glutaraldehyde would probably knock out any activity though, so the best method depends on what the proteins will be used for. I glanced through that photolabeling paper, did they say how protein yield was affected by the substitution? I've seen non-natural amino acids added through mutant tRNAs and aaRSs, but yields are typically lower. $\endgroup$
    – user137
    Jul 17, 2015 at 0:06
  • $\begingroup$ Thank you for the response. I can't use Gluteraldehyde as I need the cells alive. Yes I have given the different parameters a though since I need the cross-linking under very specific conditions. Labelling would be costly and is not guaranteed to get integrated into my protein of interest on the cell but that is an interesting option! many thanks :) $\endgroup$ Jul 17, 2015 at 8:07

If protein 1 contains a glutamate near the binding site with protein 2 and is close enough to the lysine on protein 2, it may be possible to link the glutamate to the lysine through EDC coupling. enter image description here

EDC is 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide, which reacts with carboxyl groups to form an "active ester". This ester is highly reactive with primary amines, but can also react with water or hydroxyl ions. The general reactions are outlined below: enter image description here

The active ester is compound 2, which you probably want to react with the lysine on protein 2 to produce the amide in compound 3, with the isourea 4 as a byproduct. This shows that other products are possible, depending on what hits the active ester first. This reaction is most effective at pH of 4 - 5, and produces a 0 length linkage, so if you have to reach 10 angstroms across, this might not work.

EDC coupling is often followed by NHS ester formation, which is more stable than the EDC active ester, but is usually done when you intend on storing the active ester as a dry product for some time.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.