Fluorescent fish

Divers, reef and fish at Little Brother, Red Sea, Egypt
Photo by Flickr user Derek Keats

A paper out in PLoS ONE last month announces the discovery that a lot more fish fluoresce than we thought. 180 species of fish biofluoresce, with emission colors ranging from greens to reds: eels and rays are in the greens (λem around 530 nm) while scorpionfish and gobys are fairly red (λem about 620 nm) and others are in between (see their Figure 1 for some photos).1

The authors consider several evolutionary advantages that biofluorescence might bestow on these species. Some fish may use it to blend in with fluorescent corals, others may use it to communicate. Certain deep sea fishes are thought to use fluorescence to lure prey.2

What really gets me excited about this paper isn't the fish, though. It's the fluorophores: the molecules that do the fluorescing. These 180 species of fish have the potential to harbor a large variety of fluorescent proteins. Perhaps they are similar to the famous GFP, which comes from the jellyfish Aequorea victoria, or the proteins from coral that are the basis for mCherry, the fluorescent protein I use in lab. But perhaps they're different. They might inform efforts in engineering existing proteins to do new and different things. Current fluorescent proteins have a wide range of properties: sensitivity/resilience to pH, likelihood of staying separate or forming pairs or clusters,3 ability to be "switched on" from a non-fluorescent state,4 and ability to change color,5 for example.

In addition to the potential variety of structures, I'm excited about the possibilities for the photophysics of these molecules, specifically the Stokes shift. The Stokes shift is basically the difference between the color of the light you shine on a fluorophore, and the color of light it emits. We typically calculate the Stokes shift as the difference in wavelength for those two colors.6 The authors say they shined blue light7 on the fishes to get the fluorescence they saw. For the red fluorescent fish like the false stonefish in Figure 1G, that could mean a really big Stokes shift. Blue light is around 500 nm, and red light is around 620 nm. That's potentially a Stokes shift of about 120 nm. For comparison, mCherry's Stokes shift is about 30 nm.

Before I get too excited about huge Stokes shifts, I should mention a few things. The blue light used may not have a wavelength of 500 nm, that's just a ballpark value. The wavelength of the blue light may not be the excitation maximum of the red fluorophore; you can excite fluorophores with other wavelengths if they're close enough. For example, the excitation maximum for mCherry is 587 nm, but the laser I use to excite those molecules has a wavelength of 561 nm. I could even use a 488 nm laser to excite it, but it would do a pretty terrible job because mCherry doesn't absorb that color of light very well. Lastly, the authors may not have used blue light to get the red fluorescence at all! The methods section mentions that the excitation filters used were for 450-500 nm and 500-550 nm. 550 nm light is fairly green. That could really take a bite out of my hypothetical Stokes shift.

I'm still excited by the news in this article. Even if it turns out that these fish all use proteins we're already familiar with, it's some pretty cool biology.

1: I'm estimating emission wavelengths from Figure 4 of the paper.

2: How's this for cool science? Some kinds of fish apparently bioluminsce in one color, then use Förster resonance energy transfer (FRET) to excite fluorophores which emit in another color. It takes teams of PhDs months and years to get those kinds of systems working, and it turns out some fish just grew up knowing how it's done. Of course, they do have eons of evolution on their side. I don't have eons to work on a PhD.

3: Molecules that tend not to clump together are called monomers. The "m" in "mCherry" means "monomeric." (The "cherry" means it's red, like the fruit. There are a bunch of other fluorescent proteins named for foods, including my favorites mBanana, mHoneydew, and dTomato.) Molecules that pair up are called dimers (that's the "d" in "dTomato"), and pairs of these pairs are tetramers.

4: When you shine the right color of light (let's call it λ1) on a typical fluorophore, it will emit light of a different color. Some fluorophores start in "dark" state, though. When you shine another color of light (λ2) on them, they "switch on" and can now fluoresce when lit by the first color (i.e. λ1). This is called photoactivation. It's the "PA" in the PAmCherry that I also use in lab.

5: Changing colors, called photoshifting, is similar to photoactivation, except instead of starting in a "dark" state, the fluorophore starts in another (usually higher-energy) state. An example of this is the fluorescent protein Dendra2 (which I have also worked with). When you shine blue light on Dendra2, it emits green light. After you expose it to violet light, though, that no longer works. Instead, if you shine green light on it, it will emit red light.

6: Actually, it's not quite that simple because there's a range of wavelengths that a fluorophore will get excited by or give off. We use the difference between the wavelength that is absorbed best (the excitation maximum) and the wavelength that is given off most (emission maximum). Also, it's only called the Stokes shift if the light given off is lower in energy (i.e. closer to the red end of the spectrum, longer wavelength) than the light that excited the molecule. If it's the other way around (yes, it can happen), it's called an anti-Stokes shift.

7: Why blue? Woods Hole Oceanographic Institution tells us that blue light penetrates seawater better than other colors, so it makes sense that these fluorescent proteins would use the most abundant color of light.