Spark UV 387 and RB670 interact

We know that certain dye families are prone to interactions, creating non-specific staining. We can work around these, though, often using specific buffers designed to combat these effects. For example, the SIRIGEN polymers (Brilliant and Super Bright dyes) can bind non-specifically to each other, but this can be mitigated through the use Brilliant Stain Buffer. Borate buffer is used to reduce Qdot aggregation. The CellBlox buffer can often reduce interactions between NovaFluor dyes. Today let’s look at an interaction between two seemingly unrelated dyes, Spark UV 387 and Real Blue 670, and what can be done to stop it.

This effect was found (to the best of my knowledge) by researchers in Leuven, Belgium. Vera Dermesrobian (KUL) and Panos Barlampas (Amsterdam UMC) were trying to optimize a panel, and they were seeing diagonal staining between two markers. I’ve replicated this with some unrelated markers, so it seems to be a general phenomenon for these two dyes.

Human PBMC on the Sony ID7000

Below is an example using mouse splenocytes. Across the top row, there's no Real Blue 670 conjugate. The cells are CD45.2 origin, so we don’t get any substantial staining with the anti-CD45.1. On the bottom row, we add in the anti-TIM-3 in Real Blue 670. When combined with either the anti-CD4 or the anti-CD11b, we get a diagonal staining pattern, which is not what we expect biologically. The odd part, though, is that this is unlike the effects that you get with NovaFluors or Brilliant dyes. It is not symmetrical. There is a population of RB670+ cells that remains vertical, whereas all the Spark UV387+ cells become positive for RB670 in direct proportion to how much Spark UV387 they have. This is most striking with the CD45.1 in Spark UV387, where the RB670 staining pattern is essentially unchanged. So, cells covered in Spark UV387 seem to bind RB670, but not the other way around.

Mouse splenocytes on the Cytek Aurora

The advantage of this asymmetrical non-specific binding is that it means you can avoid the effect through sequential staining. As long as you stain for the RB670 first, and thoroughly wash it away, staining afterwards with the Spark UV387 conjugate seems to give the expected pattern for both markers. We are currently using both dyes in a couple of panels, using the RB670 for low expression surface markers, and the Spark UV387 for high expression markers stained after fixation.

Sequential staining helps

Here’s another example of the effect, using human PBMCs:

Human PBMCs on the Cytek Aurora

Hint: CD8 T cells do not generally express high levels of CD14.

It’s good that the sequential staining approach works, because the blocking reagents do not appear to do anything:

Human PBMCs with various blocking reagents. We expect the CD8+ cells to be CD14-.

Another solution is to simply change the combination of dyes you’re using to avoid the problem all together. And if you’re working with fewer than 40 colours or so, that’s pretty easy to do nowadays. BUV395, Spark PLUS UV395 and StarBright UV400 are all good alternatives to Spark UV387, although on several spectral instruments you can use Spark UV387 and BUV395 simultaneously. The other dyes closest to RB670 are PerCP, BB660, NovaFluor Blue 660 or StarBright Blue 675, but some of those have more spillover into other parts of the spectrum.

Examples of the narrowest UV dyes on the DiscoverA8/S8. Figure from BD Spectrum Viewer.

Examples of dyes (somewhat) similar to RB670. Figure from BD Spectrum Viewer.

I find this asymmetric interaction quite odd, but presumably it would make sense if you knew the chemistry, which we don’t. Spark dyes are small molecules, likely related to the Alexa Fluors. Several of them spectrally copy dyes from Biotium’s CF range, but in the case of Spark UV387, it looks almost exactly like the much dimmer mFluor UV375 from AAT Bioquest. Information on the structure of the Real dyes is not available. Understanding the structure of the dyes can be important to know whether we should expect this interaction elsewhere, between other pairs of dyes with corresponding structures. It’s not clear that the Real dyes actually represent a coherent family of dyes with similar chemical properties, and I haven’t seen this interaction between Spark UV387 and any other Real dyes. I also haven’t seen interactions between RB670 and other Spark dyes. Both of these dyes are relatively new, though, and we should probably expect more off-target effects as we build larger panels and shove ever increasing varieties of chemicals into small tubes.

Cinnamon Hummingbird, Costa Rica