Dear Andreas Nygaard ,

It depends (of course) on what kind of assay you want to perform.

But in general I would say yes the unlabelled protein is not visible (but the unlabelled is obviously also binding or interacting with whatever you are looking at). Sometimes deliberately an unlabelled (wild-type) protein or peptide is added because of (for example) decreasing the degree of self-quenching:

https://www.researchgate.net/publication/20979704_Tryptophan_fluorescence_study_on_the_interaction_of_the_signal_peptide_of_the_Escherichia_coli_outer_membrane_protein_PhoE_with_model_membranes

But be careful:

Again depending on the type of assay you want to perform (and provided that the labelled and unlabelled protein act the same!) the ratio of protein/lipid, protein/(other)protein or protein/ligand is having some kind of optimum (range).

This ratio is (of course) different with 20% labelled protein compared to 100% labelled protein (for example a 1/40 ratio for 20% labelling is found this would mean (ideally) a 5/40 ratio in the case of 100% labelling). So you 'correct' for the 'real' amount of protein that binds or interacts.

Check: In the perfect world you (once you obtained a significant amount of data) should try to purify a (small) amount of your labelled protein (HPLC or FPLC could work) in order to get a 100% labelled fraction and check whether the above assumption/correction is/was correct.

Hope this helps you somewhat and good luck trying.

If you are not getting a high labeling efficiency, you might be able to change the conditions to improve it. Some of the Cys may be oxidized into a bimolecular disulfide bond, which can be reduced before labeling. The pH may have an effect. You can increase the labeling time and label concentration. If the labeling buffer contains any residual sulfhydryl (e.g. dithiothreitol), you can remove it using a gel filtration desalting column.

With any fluorescent assay you measure the properties of your labelled protein, which may be the same as those of the unlabelled protein, or not. In other words, you'll have to make sure that the labelling did not (noticeably) change the properties of the protein.

The lack of labelling in the 80% of your protein will limit the sensitivity of your assay. However, the signal should not be affected, unless either the exciting or emitted wavelength is absorbed by the protein itself.

What do you mean by "sometimes"? When you use the same procedure on the same material, you should get the same result each time. If that is not the case, search for the cause and correct it.

Hi again.

And thank you very much for the answers! It all makes perfectly sense. Ideally, I know that the labeling should be optimized before proceeding. It was just a theoretical consideration if - and the degree to which - the unlabeled ensemble of proteins would affect the fluorescence-based characterization of the labeled proteins.

Again, thank you kindly for your time

Hi Andreas,

here some additional remarks for you. I'm mainly in line with the others, but I was missing some few aspects. I know this very same problem from cysteine-targeted spin-labeling in electron paramagnetic resonance (EPR) spectroscopy:

1) the labeling/conjugation efficiency depends on the sterical accessibility of the introduced cysteine (surface exposed/buried), so it cannot be always 100%.

2) if you didn't check for the labeling effciency, perform an Ellman's test each time you labelled your protein. It differs a lot depending on preparation and so on !

3) in MALDI TOF or another high-res mass spectrometric method, you might also observe multiple peaks/labeled and unlabeled protein, depending on the mass of the label and the mass res of your device.

4) if at all, carefully use DTT or TCEP to gain better conjugation efficiency as you might mess up your beloved protein quickly. It depends...

5) I'm not sure how it is with fluorescent labels, but in EPR a reduced labeling efficiency naturally results in a lower signal intensity. The unlabeled proteins are totally invisible as they do not contain paramagnetic species, so should it be for fluorescent labels in excitation mode.

6) If you have ligands applied, it becomes kinetically very complex in terms of association constants and concentration-related phenomena, but the ensemble should behave uniform. So, you only see the labeled protein interacting with the ligand/amino acid.

It' not an easy problem, but I hope this also helps a bit to rule out a little more troubleshooting,

best,

Joerg