This is a nice chemical biology paper that hits on a hot topic of the day: the uptake and function of mRNA when administered to cells. You can always look for downstream effects to show that you achieved both those goals, but it would be very useful to get images of this process in real time. Here’s a heavily cited paper from 2013 that was able to do this for siRNA species, but the colloidal gold labeling technique used is not a straightforward one.
To that end, there have been many attempts at generating other fluorescent RNA and DNA species, but as usual, you always have to look out for side effects of the fluorescent labeling. These things take up space and alter physical properties, and there’s not much way around that, so you just have to see if you can live with the results. Cyanine dyes (attached by cycloaddition reactions) have been used on both RNA and DNA, but it’s already known that the fluorescence on the latter can vary according to the sequence surrounding them, and both the type and positioning of such labels can impede the function of mRNA in general (such as choking up the ribosomal machinery). In general, you can go hard and label RNA with things that will be easy to pick up with good signal/noise, at the cost of degrading its ability to function, or you can label it with a lighter hand at the cost of being able to see it in living cells.
This new paper references these and several other methods that have been tried, and proposes a new variety of fluorescent nucleobase. Those as the name implies, are things that look like the existing bases used to form RNA and DNA species, but the purine/pyrimidine part has been modified to be fluorescent. The same constraints apply – can you get one that’s bright enough and long-lasting enough that’s still acceptable to the (rather highly evolved) enzymes that handle such building blocks? This paper has data on a fluorescent cytosine analog incorporating a 1,3,-diaza-2-oxophenoxazine instead of cytidine, and it seems to make the cut. In fact, the paper shows data indicating that they can exchange all the C residues in a 1200-base stretch of mRNA and still see it translated in live human cells, which is pretty impressive. It also features a new solid-support method to synthesize nucleoside triphosphates, which sounds pretty useful (since that can be a painful process) which looks to be the subject of a forthcoming paper itself.
That big stretch of mRNA is, in fact, the transcript for everyone’s favorite fluorescent species, green fluorescent protein (GFP). The team did do some codon-optimizing to cut down on the chances of having two modified-C residues right next to each other, which is probably going to gum up the works no matter what, but they report that a nonoptimized transcript for calmodulin went through just fine, on the other hand. The RNA polymerase enzymes seem to take it up with nearly the same alacrity that they do regular cytosine. As the amount of incorporated fluorescent base goes up, the emission gets a bit red-shifted, and the quantum yields and fluorescent lifetimes drop a bit, all of which are to be expected – some of that is due to self-quenching by residues that are close to each other. They seem to have maxed out at about 75% incorporation; higher than that is self-defeating. But 25% incorporation is bright enough for cellular imaging, so you can stop things there and have less chance of downstream troubles.
As for the results, let me send you here so you can watch a time-lapse (from the supporting data, free to view/download). That’s part of a frame from the movie at right. Red mRNA comes into the cell in little round endosomal packets (as shown by separate orange fluorescent labeling of an early endosome biomarker), followed by the development of what is obviously correctly expressed and folded GFP over several hours. Watching red mRNA get processed into green protein is quite satisfying; I can only imagine how happy the authors must have been to see it work. This is, I believe, a first for the fluorescent-base technique in general. This sort of thing will surely be very useful for RNA applications in general, as you can (in theory) watch your proposed vaccine or drug sequence getting taken up by live cells, trafficked to the right locations in the cytosol, and subsequently degraded in real time. I get a lot of questions along the lines of “How do we know that these mRNAs are going to the right places?” Well, from what I can see, this could be the most straightforward way to answer them!