Hi David,

usually the biotinilation level should be kept below 10biotins/mab. Linking of more biotin tends to partially denature the antibodies, makes them sticky and therefore unspecific. Many companies choose a reaction stochiometry that is aimed at a mean of 2,5 bio/mab to prevent this effect.

If you're not sure about the incorporation level of your product - why not request the info from the manufacturer or biotinilate yourself. It's quick and easy using one of the commercial biotinilation kits and you can adjust the incorporation rate as you need it.

Good luck,

Ingo

I'm assuming that they've used a standard biotinylation process which targets amino acid side chains such as Lysines. There will be a large number of such potential biotinylation sites in an IgG in both the constant regions of the heavy and light chain as well as in the variable regions. The exact number will depend upon the subclass and also the sequence of the V-regions. In general the aim is to get an average of just a few biotin molecules per antibody, since there is a risk if you over conjugate the antibody that you might inactivate the antigen binding site. The exact postion of the biotins will tend to vary from one molecule to another.

Using 150,000 as the Mr (relative molecular mass) for an IgG is a good approximate estimate but the exact Mr will vary with subclass, V-region sequence and also the glycosylatiion of the C-region and possibly V-region. Glycosylation varies from one molecule to another.

Thus given that there is already an assumed error in the estimate of the Mr of the antibody I don't think that it will make much difference if you simply were to make a guestimate of ~1 biotin per antibody and still use 150,000 as your figure. That is unless you have good reason from the data sheets to believe otherwise.

As labeling is via the lysine it is difficult to say exactly how many - would probably vary for each antibody in the batch, but I believe it is in the range of 3 to 6 biotins per antibody. Taking 4 biotins that would be a difference of 0.6%, so potentially negligible.

Hi David, if it's just about the molecular mass, you may neglect it. I doubt that one will see a difference between biotinylated and non-biotinylated AB on a SDS PAGE, as the smear produced by the glycosylation side chains might obscure it.

There are just a few positions, where biotin can attach: Biotin is introduced by adding a Biotin-NHS-ester to the antibody, this will react with lysines and the terminal amino groups of the individual chains. Get a model antibody sequence and count. You'll get the theoretical maximum.

There is a very useful kit to quantify the number of biotin molecules attached to a protein: HABA assay from pierce http://www.piercenet.com/product/pierce-biotin-quantitation-kit.

An interesting paper about the "biotin counting" is this one:

Cholesterol Loading and Ultrastable Protein Interactions Determine the Level of Tumor Marker Required for Optimal Isolation of Cancer Cells

Jayati Jain, Gianluca Veggiani, and Mark Howarth

check out figure 2C, supplementary methods and supplementary figure S1

basically very simple, boil the protein in sample buffer, add streptavidin and run

An optimal molar coupling rate for IgG is normally at 20 - 40 Biotins per IgG. So you would find the difference in an SDS-Page. This coupling itself is a random process so you can have IgG's without any Biotin an may some with more than 80. For the recognition of the biotin-label by avidin, the number of biotins per IgG plays also no role, one is sufficient for the process and if you are using enzyme labeled avidins, there will be a stoichiometrical hindrance anyhow.

If are coupling by yourself, residual biotin can be measured simply after precipitating the protein by TCA. Free Biotin will be in the supernatant.

I agree the suggestion from Fortunato Ferrara. Please order the HABA assay from pierce and then you knew how much biotin could be in the biotinylated antibody. (http://www.piercenet.com)

Dear David,

is it the molecular mass that affects your assay or the diffusion coefficient? In fluidics systems its usually the diffusion that affects the assay, not really the molecular mass. Alsdo, 150 kDa is only the average estimate, the exact mass can differ, some antibodies have long CDR3. Also isotypes have an influence, e.g. human IgG3 with its long hinge has a different diffusion coefficient. Fc-glycans do exhibit microheterogeneity and some antibodies even have additional N-glycans in the variable domains. Antibodies can also have the tendency to aggregate, which has a BIG impact on molecular mass and diffusion coefficient. Aggregation / cross-linking may also occur during biotin conjugation, so you may also want to know if those aggregates are present or were removed (they probably werent cause it adds to the costs). You may also want to consicer degradation products, Fab or Fab'2 fragments would behave quite differently.

On average, 3-10 biotin molecules can be conjugated per IgG molecule.

The best seems to me to bring 30:1 M:M Biotin: IgG, to get 2-3 biot/Ig in bad conditions, 10 biot/Ig generally , and up to 20 biot/Ig providing you avoid the various pitfalls of biotin coupling: free AA, too low concentrations of Ig....

When you're trying to get an heavier coupling ratio, a part of conjugates is non longer functional as too strongly conjugated at ratio up to 100 biot/.Ig.

Hi David,

usually the biotinilation level should be kept below 10biotins/mab. Linking of more biotin tends to partially denature the antibodies, makes them sticky and therefore unspecific. Many companies choose a reaction stochiometry that is aimed at a mean of 2,5 bio/mab to prevent this effect.

If you're not sure about the incorporation level of your product - why not request the info from the manufacturer or biotinilate yourself. It's quick and easy using one of the commercial biotinilation kits and you can adjust the incorporation rate as you need it.

Good luck,

Ingo

Biotinylating antibodies yourself is very informative. It is easy and kits come with details of how to work out how many biotin molecules have been added. The conditions of biotinylation make an enormous difference and you can control this. If you are unlucky and your antibody has lots of lysines in the variable region this can be crucially important as over-biotinylation can have a marked impact on antibody affinity.

A note of caution. Determining the average number of biotins per antibody molecule in your preparation may not be the same as the average number of biotins per active antibody! As Richard Smith points out some antibodies can have one or more lysines in their antigen binding site. I remember an occasion when I labelled my humanised antibody Campath-1H (alemtuzumab) only to find it worked very well as a labelled human IgG antigen, but no longer worked in binding to CD52 on human lymphocytes. Following that experience I reduced the average number of biotins for following batches and the binding activity for CD52 antigen was retained. In any given preparation of such antibodies there are likely to be individual molecules that are thus labelled, but because of the position of the label, may not be biologically active.

I would also stress on the point that to get the best biochemical yield of coupling should be put behind the need of keeping the Ab functional!!

The problem of the relationship of the numbers of labels per molecule and the retention of biological activity can be understood from a simple thought experiment. Imagine a molecule that has 10 potential label sites and two of them are crucial to biological activity. Also assume that labelling is random. If you put one label on each molecule in a batch then you expect that 8/10 molecules within your preparation will still be biologically active. If however you put two labels on it is (8/10)x(7/9) = 0.62. For 3 it is (8/10)x(7/9)x(6/8) = 0.467

Thus with this hypothetical molecule where 80% of the labelling sites are not involved in biological activity, adding 3 labels to each molecule is likely to destroy more than half of the biological activity of the preparation. The inactive molecules will be those in which either one, the other or both of the crucial residues have ended up being labelled.

This same principle holds true for an antibody, but the exact numbers of potential labelling sites in the binding site compared to those outside will vary dependent on the sequence of the particular antibody.

You've been very helpfull with your answers and explanations.

I am going to neglect the biotin mass. The manufacturer claims there are very few biotin per antibody and that the functionality of antibodies has been confirmed.

Nevertheless, I fell more secure after your answers.

The quantitation kit is a very good idea. But we are working at a physics laboratory and there is no spectrophotometer avaliable.

We are conjugating the antibodies to magnetic nanoparticles, and we want to estimate the magnetic dragging f force

Thank you all

i got now real values:

BD answered in one hour: they have approx. 6-10 molecules per mab. so more than 2.5.

I read in most publications (check older books too) that 4-40 are normal. but of course, depends... In heidelberg they had the same values.

miltenyi i got a weird answer: confidential.... They did not even offer a CDA.

(i have a guess: strep tag and recombinant antibody... mh. so they don't even need to biotinylate them anymore i guess...)

But we will know in 2 days because we tested it on a gel with an ideal standard: a pure peptide of 42 AAs and 4 kD and ONE biotin by synthesis. So we have a molar attenuation coefficient to check the amount of the biotin and can run this peptide on the gel.

we also test thermo/invitrogen. but always only one or few.

Couldnt reach biolegend, but will continue to try again.