Tips & Tricks: Dealing with tandems

Tandems, tandems, tandems. We can't seem to avoid them in high parameter flow experiments--they're just too useful. How do we deal with the main problem of this class of fluorophore, which is that they break down into the donor and acceptor molecules, creating confusion in our data? Here are my tips for best practices with tandem fluorophores.

Example of breakdown of PE-Cy5.5 on CD4+ T cells, as evidenced by the FMO for Nur77-PE:

First, what do we mean by a tandem dye? A tandem dye is two fluorophores stuck together to create one with new, distinct properties. This works through process called fluorescence resonance energy transfer (FRET) where the photon energy moves from one fluorophore to the other due to them being really close together. The donor fluorophore absorbs high energy (narrow wavelength) light, and rather than directly emitting this again, passes the energy to the acceptor molecule, which emits the light at a much longer wavelength. In this way we get long wavelength light in response to a narrow excitation, which we wouldn't get by exciting the acceptor molecule alone.

Colloquially, what we mean in flow cytometry by tandems dyes are labile tandems, which are almost always a small molecule attached to a protein-based fluorophore. In actuality, the majority of the fluorophores we use are tandems: for instance, most of the Brilliant dyes are technically tandems. What's unique about the labile tandems is that the chemical linkage between the two molecules is weaker because the protein fluorophore can't withstand harsher conjugation reactions. So, for the purposes of this post, we're talking about tandems based on PE, PerCP and APC.

Why is it such a problem if the tandem breaks down? The fluorophores PE, PerCP and APC are actually huge in relation to the Cyanine dyes that are usually employed as acceptors. There actually many Cy7 molecules per PE in PE-Cy7. So, when the tandem bond start to degrade, not all the Cy7 moieties will fall off at once, leaving us with a fluorophore that has intermediate properties. This causes spillover-like artefacts in our data--false positives.

What we think PE-Cy7 looks like:

A slightly better representation of PE-Cy7:

I’m going to show you a mistake I made when I was first starting to run high parameter panels to suggest how to design around tandem breakdown. We were studying Tregs in a wide array of tissues. I’d cobbled together a panel using CD45 in a relatively dim APC-Cy7-type tandem to identify leukocytes and Foxp3 in bright APC to identify Tregs. While I knew there might be tandem breakdown, I’d tested the panel using lymphoid tissues and it seemed to work fine.

What I hadn’t realized is that tandem breakdown is catalyzed by cellular enzymes. So, in the non-lymphoid tissues with a more complex cellular composition, tandem breakdown was higher and variable from tissue to tissue. The data were useless.

Here, then, are my suggestions for working with tandems:

• Because breakdown signal will be proportional to tandem signal, you want to avoid cases like this where the tandem signal is brighter than the parent signal.

• Avoid tandems on widely expressed markers. You'll get widespread problems.

• Avoid using tandems on the same cell type as the parent fluorophore.

• Use newer breakdown-resistant tandems or tandem preservatives. For example, BioLegend's Fire dyes (with the exception of Fire 810 conjugates) are more stable than the traditional versions.

• As always, protect from light and heat. Don’t make up master mixes with tandems ahead of time.

• Use tandems after fixation if possible.

• Swap to a new, non-tandem dye. For instance, BioRad's StarBright dyes or BD's RealBlue dyes are very stable.

Swapping to a new dye can really help in high parameter panels. The new dyes coming out are not only stable, but usually also have cleaner spectra, meaning less spread and better resolution. I did this for a panel I showed at the top where I'd initially been using CD25 in PE-Cy5.5 and it was causing breakdown-dependent false positives in PE. This prevented me from accurately measuring PE, which I was using for the critical, low expression marker in my data.

Swapping CD25 from PE-Cy5.5 to StarBright Violet 515 (also moving Foxp3 from APC to Alexa Fluor 660 for unrelated reasons):

Example of results with the new panel: