I've had this happen too but with my A647 "secondary only" control when doing a two-step label. I end up walking the secondary only back around 10^1 and using it as my "negative" but it's a log out from the unstained. Following to see what others say...

Hi

There is absolutely non-specific binding of your isotype with your cells. I had such problems before.

First: Fix your cells with ice chilled methanol for 5 min (no perm is required). Then block with 1% BSA (1 Hour, RT).

Second: Change your isotype as this antibody is binding non-specifically to your samples. It is not about your AF tag but just the isotype.

Both worked for me on different type of antibodies.

Mike are you using anti-mouse antibody and mouse serum together??? If yes then your mouse serum may has nonspecific binding site.. You may consider changing the combination, 1% BSA or Ovalbumin you work for blocking. Further use antibody in 1:10000 dilution (v/v). You also consider giving wash with 1% triton after staining.

Further if it will not work then prabably you should change your antibody.

Hi Mike,

I am assuming your antibodies are titrated so that you are using the minimal Ab quantity for the cleanest possible detection of your antigen. This is essential for reliable results.

Your controls and tests seem to point to an AF647 specific problem and not an isotype problem. I have used AF647 on many occasions w/o observing these issues... So here are the evident questions:

1- do you observe the same problem with another batch of the AF647 isotype ctrl antibody?

2- do you observe the same problem with another antibody tagged with AF647? i.e. does your Oct4 antibody reproduce the same background staining as your isotype?

3- Take adult primary cells that should be neg for your target protein and check your staining with both AF647 coupled antibodies, the isotype ctrl and the relevant antibody.

For now, I think you might have a batch problem. This might be resolved by adding one or two cycles of washing.

Keep us posted.

Good luck

G.

Please STOP using isotype controls. This is a deprecate practice in flow cytometry. The issue is that every isotype has a different non-specific binding. Also, isotype controls are not able to give you any information about the specificity of the signal that you get from your antibody. Instead use specific controls such as FMO, positive control cells, transfected cells or silenced cells. Isotype are only going to give you misleading information. By the way I had a terrible experience with an AF647 antibody from R&D giving not specific staining...For a few days I though I was working on a Nature paper! I believe they don't know how to do a proper conjugation/quenching.

Hi All,

Figured I'd post to wrap all this up. Just a quick note if its not clear. I'm not using an isotype to gate positive or negative. I'm using an isotype to show that I don't get non-specific binding when blocking the cells. I use either unstained or, more practically, FMOs to set boundary gates.

Anyways, a few findings:

(1) I ran isoclonic controls with saturating concentrations of unlabelled target antibody and dilutions of my AF647-conjugated target. This should negate my target signal completely since all specific sites are take up by the unlabelled, but I still see a shift from the unstained by about half a log. So AF647 is not sticking because of the antibody.

(2) I ran isotypes of the same clonetype (MOPC-21) with AF647 conjugation from two manufacturers (BD and BioLegend). They both show signal when cells are blocked but the signal is 2X difference between them (i.e. proof positive that the F:P ratio isn't the same). I did this thinking that maybe my AF647 isotype had free dye or something but that turns out not to be the case.

(3) I confirmed that if I switch the AF647-conjugate target Ab to a AF488-conjugate target Ab this whole phenomena goes away.

So grand conclusion is that the AF647 tag itself is quite adherent in some way on my ES cells but AF488 is not. As to why I don't know but I'm going to steer clear of AF647 unless my panels have a clearly positive population where this adsorption won't change any gating metrics.

Cheers,

Mike

Mike,

Tip of the hat for coming back with results/conclusion and the details of the workup. This will definitely have me picking another fluor for my two-step labeling if I have a choice. I'm curious about the chemistry that makes the A647 molecule "stickier".

best,

Rich

What happen if you reduce the amount of antibody?

Rich,

I've noticed this before when I used to do a lot of microfluidic work with PDMS channels. Flow dyes are commonly used to trace flow patterns especially in valved devices. I found that AF647 would always non specifically adsorb unless I coated the channels significantly with pluronic F68 solution prior to dyes. Chemically speaking the logD values based on structure of AF647 and AF488 indicates they have quite different hydrophobicity - AF647 being more hydrophobic. So pluronic "may" help to alleviate the dye-based binding.

Best,

Mike

Very cool explanation! That's great experience to draw from. Thanks, Mike.

I had very similar experience with secondary AF-647 antibodies. What helped was use of anti-F(ab')2 -AF647 antibodies, so in my case the problem was likely in blocking solution that contained mouse serum. Also if you used "optimal" voltage for AF647 channel (633nm laser) - you will get wider distribution of unstained as well as low fluorescence intensity=isotype stained cells between the lower bins of fluorescent scale (more diversity in low values= broader background for "weak laser"- red dye combination) than for "optimal" voltage of 488nm laser for AF488 and even PECy7 dyes ("strong" 488nm laser- green or even infrared dye combination) as stronger laser but lower or similar to AF647 stain index (brightness) of this dyes compare to AF647 will give you more homogenious negative population, while with weak laser- red dye you will see more "negative value" cells and difference between unstained vs weakly stained-isotype control cells can be more obvious.