Work / Technology & tools ILLUSTRATION BY THE PROJECT TWINS THE PROJECT BY ILLUSTRATION THE HUNT FOR RED FLUORESCENT PROTEINS By pushing fluorescent proteins further into the red, bioengineers are expanding the palette and penetration depth of biological imaging. By Amber Dance reen fluorescent protein is one requires. No oxygen, no fluorescence. being the same, redder is better,” says Robert of the most popular items in the So, she turned to a label that can do with- Campbell, a protein engineer who spends microscopist’s toolbox. It is a out oxygen. A relatively new addition to the half his time at the University of Tokyo and Nobel-prizewinning innovation that fluorescent-protein palette, IFP2.0 fluoresces the other half at the University of Alberta in brilliantly lights up molecules of mainly in the near-infrared — a portion of the Edmonton, Canada. It also provides a way Ginterest across a diverse range of biological electromagnetic spectrum that is barely visi- to add another hue, or two, to experiments. fields, laboratories and techniques. But it does ble to the human eye but readily apparent to “The more channels we can pack into an exper- not work for physical chemist Julie Biteen. microscope cameras1. “We’re really excited,” iment, without significant bleed-through, the Biteen studies gut bacterial communities at Biteen says. “We could see single cells and more interactions we can study,” says Talley the University of Michigan in Ann Arbor, and identify them.” Lambert, a microscopist at Harvard Medical was eager to use fluorescent proteins to iden- Imaging at the red end of the spectrum School in Boston, Massachusetts. tify individual species in complex mixtures. offers other advantages, too: lower back- Reddish fluorescent proteins have existed But gut bacteria don’t like oxygen — something ground fluorescence, reduced toxicity and for decades, but they are still generally no green fluorescent protein (GFP) absolutely deeper tissue penetration. “All other factors match for GFP in terms of both brightness and 152 | Nature | Vol 596 | 5 August 2021 ©2021 Spri nger Nature Li mited. All rights reserved. ©2021 Spri nger Nature Li mited. All rights reserved. hue. Even the ‘red’ fluorescent protein RFP is tags if they are to stand out from the crowd. directly. Timothy Wannier, a synthetic biologist closer to orange. Scientists are making head- A simpler approach is to use a colour that at Harvard Medical School, used both molec- way in developing fluorescent proteins that are doesn’t overlap with any of the others. That’s ular evolution and computer-based protein truly red — often called ‘far red’ to distinguish where far-red and near-infrared labels are analysis and design on GFP relatives during them from earlier attempts. Infrared offers handy; they make it easy to get at least four his PhD studies at the California Institute of similar advantages. Development is still in its non-overlapping signals from the same cells. Technology in Pasadena. His goal was to turn infancy, but advances in bioprospecting, pro- In his lab at Westlake University in Hang- dimeric far-red fluorescent proteins into mon- tein engineering and synthetic chemistry are zhou, China, tool developer Kiryl Piatkevich omers, which would help to prevent undesir- helping to improve the labels. Most are avail- routinely records five signals from the same able interactions. But he also had to engineer able, in gene form, on the plasmid repository microscope slide by using three visible colours mutations to stabilize the lone monomers4. Addgene. and two in the near-infrared range. Such exper- One of the resulting tags, mKelly1, caught There’s clearly a need. Scientists are des- iments can often be performed without major the eye of Yi Shen, a protein engineer at the perate for tags and sensors that they can use equipment upgrades. University of Alberta, who used it to build far- alongside standard tools, such as GFP, blue red calcium sensors named FR-GECOs (ref. 5). DNA stains and channelrhodopsins, which Where the fluorophore grows are activated by green and blue light. Brian Whereas many fluorescent proteins are found Redder all the time Almond, senior manager for product manage- in sea creatures, far-red and near-infrared Despite these advances, far-red and ment at Thermo Fisher Scientific in Carlsbad, molecules tend to come from bacteria. But near-infrared fluorescent proteins remain dim California, says an alternative hue is often cus- unlike GFP and similar proteins, bacterial bulbs. Whereas some green proteins push the tomers’ first request for new fluorescent tools. light receptors lack a component to absorb limits of brightness, the best near-infrared tags “Everything is green,” they tell him. “Please, light (called a chromophore) of their own; they hover at around 10–20% of the maximum. don’t make it green.” require the addition of a pigment known as One solution is a workaround based on far- biliverdin. The good news is that biliverdin is red and near-infrared chemical dyes, which are Scarlet solution a natural intermediate in the breakdown of also becoming available more widely. Organic A typical fluorescence-microscopy experiment haem, which binds to oxygen to transport it chemist Luke Lavis at the Howard Hughes can use around three colours without overlap. through the blood, so it’s naturally present Medical Institute Janelia Research Campus But picking labels that will work together is in Ashburn, Virginia, developed bright, non- not as simple as yellow-green-blue. There are “Everything is green. toxic dyes that can switch between colourless hundreds of fluorescent proteins to choose and fluorescent forms according to their sur- from, and they vary in factors such as hue, Please, don’t make roundings6. Lavis has since teamed up with brightness and fluorescence longevity. Some it green.” protein-engineer colleague Eric Schreiter to are single-unit proteins, but others have the turn these dyes into ‘chemigenetic’ cellular potential to stick to each other and perhaps sensors that couple a chemical dye with a even glue the protein of interest to others like in mammals. The bad news is that biliverdin protein partner. it, interfering with the results. No one protein is also quickly degraded, so it’s far from The pair use the genetically encoded will be the best for every application. abundant. ‘HaloTag’ as a dock for the synthetic dyes, When choosing a tag, it’s best not to rely One solution to that problem is to add more and hook it up to sensor proteins that change too closely on published data, warns Roberto biliverdin, either from standard chemical sup- shape in the presence of calcium or electrical Chica, a protein engineer at the University of pliers, or by altering organisms to make more voltage7. The shape change alters the local Ottawa. Proteins that work well in a test tube of it. Another is to engineer natural far-red and environment of the dye such that it fluoresces might not shine in a model organism, and data near-infrared proteins so that they work bet- — and is about ten times brighter than previous tables are often incomplete. It’s best to test a ter outside their usual host, for example by red sensors, says Schreiter. Lavis gives the dyes few fluorescent proteins and pick the best for boosting the efficiency of the binding between to other scientists free of charge. He is now your experiments. the protein and the pigment, says Vladislav testing next-generation dyes that he expects Several free online resources can help Verkhusha, a molecular bioengineer at Albert will penetrate deeper into tissue for in vivo scientists to choose candidate fluorescent Einstein College of Medicine in New York City. applications. proteins, including Lambert’s FPbase; Thermo His group induces random mutations in the “The future is bright for this class of sensors,” Fisher’s Fluorescence SpectraViewer; and the relevant genes in vitro, then expresses those says Schreiter. He adds that it should be pos- Fluorescence Spectra Analyzer developed by genes in the bacterium Escherichia coli and sible to replace GFP in any pre-existing sensor BioLegend in San Diego, California. Users can selects for the reddest or brightest products. with the HaloTag to create a new, red, chemi- view the excitation and emission curves for In one example2, the team used 17 rounds genetic sensor. But for the rest of the rainbow, hundreds of fluorescent proteins and dyes, of this molecular-evolution approach to Lavis says, standard fluorescent proteins and by inputting their light sources, filters and obtain a near-infrared protein tag called should suffice “because they’re awesome”. detectors, they can choose accordingly. miRFP670nano. The tag is about 60% of the One strategy for using multiple colours is size of GFP, binds to biliverdin efficiently and Amber Dance is a freelance science journalist to let computers sort out any overlapping fluoresces brightly in mammalian cells. in Los Angeles, California. emission spectra after the data are collected. Piatkevich uses molecular evolution, too, 1. Yu, D. et al. Nature Commun. 5, 3626 (2014). With this technique, called ‘spectral imaging but in mammalian cells, which fold up the 2. Oliinyk, O. S., Shemetov, A. A., Pletnev, S., Shcherbakova, and linear unmixing’ or ‘fluorescent unmixing’, proteins such that they match those of the D. M. & Verkhusha, V. V. Nature Commun. 10, 279 (2019). researchers can probe many more colours — up target cells more closely than those evolved 3. Piatkevich, K. D. et al. Nature Chem. Biol. 14, 352–360 (2018). to 40 tags in the same flow-cytometry experi- in bacteria. His team has used this approach 4. Wannier, T. M. et al. Proc. Natl Acad. Sci. USA 115, E11294– ment, says BioLegend product manager Kenta to brighten a near-infrared fluorescent voltage E11301 (2018).