Fluorescent proteins

From Scholarpedia
Robert E. Campbell (2008), Scholarpedia, 3(7):5410. doi:10.4249/scholarpedia.5410 revision #91274 [link to/cite this article]
Jump to: navigation, search
Post-publication activity

Curator: Robert E. Campbell

Figure 1: The molecular structure of the archetypical Aequorea victoria green fluorescent protein (often abbreviated as GFP) . The polypeptide chain is shown in cartoon format with the chromophore in stick representation. PDB ID 1EMA.

Fluorescent proteins are members of a structurally homologous class of proteins that share the unique property of being self-sufficient to form a visible wavelength chromophore from a sequence of 3 amino acids within their own polypeptide sequence. It is common research practice for biologists to introduce a gene (or a gene chimera) encoding an engineered fluorescent protein into living cells and subsequently visualize the location and dynamics of the gene product using fluorescence microscopy.

Contents

Aequorea victoria green fluorescent protein

Discovery and characterization

Figure 2: Aequorea victoria jellyfish. This picture does not show luminescence or fluorescence. Image credit: Steven Haddock.

The presence of a fluorescent component in the bioluminescent organs of Aequorea victoria jellyfish (phylum Cnidaria, class Hydrozoa) (Figure 2) was noted by Davenport and Nicol in 1955 (Davenport and Nicol 1955), but it was Osamu Shimomura of Princeton University who was the first to realize that this fluorophore was actually a protein. In a footnote to his paper describing the isolation of the bioluminescent protein aequorin from Aequorea, Shimomura wrote, "A protein giving solutions that look slightly greenish in sunlight though only yellowish under tungsten lights, and exhibiting a very bright, greenish fluorescence in the ultraviolet of a Mineralite [a handheld ultraviolet lamp], has also been isolated from squeezates" (Shimomura et al. 1962). "Squeezate" refers to the solution that resulted from the squeezing of the excised bioluminescent tissues of the jellyfish through a cotton bag. Soon after, Shimomura reported the fluorescence emission spectrum of this protein and suggested that energy transfer from aequorin to this green fluorescent protein could explain why the in vivo luminescence of Aequorea is greenish and not blue like the luminescence of purified aequorin (Johnson et al. 1962). In 1971 Morin and Hastings described very similar green fluorescent proteins isolated from Obelia (phylum Cnidaria, class Hydrozoa) and the sea pansy Renilla reniformis (phylum Cnidaria, class Anthozoa). The nature of the chromophore itself remained a mystery until 1979 when Shimomura correctly determined the chromophore to be a 4-(p-hydroxybenzylidene)-5-imidazolidinone moiety covalently linked within the polypeptide chain (Shimomura 1979). Achieving this breakthrough necessitated not only the laborious harvest of 100 mg of the naturally occurring green fluorescent protein from Aequorea, but also remarkable chemical intuition on the part of Shimomura.

Initial cloning and recombinant expression

The complete primary sequence of the 238 amino acids of Aequorea green fluorescent protein (accession P42212) was not revealed until the cloning and sequencing of its cDNA by Prasher in 1992 (Prasher et al. 1992). Just two years later came the first dramatic demonstrations that the gene was self-sufficient to undergo the post-translational modifications necessary for chromophore formation. Specifically, Chalfie reported the gene encoding Aequorea green fluorescent protein could be functionally expressed in the sensory neurons of the worm Caenorhabditis elegans (Chalfie et al. 1994) and Inouye and Tsuji showed that expression of the gene in Escherichia coli resulted in green fluorescent bacteria (Inouye and Tsuji 1994).

The biological research community, armed with the arsenal of powerful molecular biology techniques developed during the preceding two decades, was quick to recognize the unique utility of a genetically-encoded fluorophore as a marker of gene expression and protein localization. Accordingly, by November 1995 there were at least 36 additional examples of applications of recombinant Aequorea green fluorescent protein (Cubitt et al. 1995) and in the decade that followed the number of publications per year with "GFP" or "fluorescent protein" as a keyword rose at a rate of 500-600 publications per year per year (Wouters 2006). In hindsight, the cloning of the gene and the first demonstrations of recombinant expression in non-jellyfish organisms marks a clearly discernible turning point in the history of fluorescent protein research.

Molecular structure

Shimomura's assignment of the chromophore structure was unambiguously confirmed by a detailed reinvestigation of a chromophore-containing peptide fragment in 1993 (Cody et al. 1993) and two independent reports of the x-ray crystal structure of Aequorea green fluorescent protein in 1996 (PDB IDs 1EMA (Ormo et al. 1996) and 1GFL (Yang et al. 1996)). These x-ray crystal structures revealed that the protein has a unique overall fold comprised of an 11-stranded \(\beta\)-sheet wrapped into a cylindrical \(\beta\)-barrel protein that is 42 \(\AA\) in height and 24 \(\AA\) in diameter (Figure 1). This type of protein fold has been given the blithe and befitting name of a \(\beta\)-can (Yang et al. 1996) due to its resemblance to a soup can in terms of shape and proportions. The chromophore is located near the center of the protein, attached to a helical segment of the protein that threads through the center of the \(\beta\)-can along its long axis (Figure 3A). Numerous subsequent structures of fluorescent protein homologues from other organisms have revealed that all fluorescent proteins share the \(\beta\)-can fold and thus belong to a protein superfamily defined by similarity to the Aequorea green fluorescent protein archetype (SCOP Superfamily: GFP-like).

Figure 3: A. The molecular structure of Aequorea green fluorescent protein as viewed from the top in Figure 1. The chromophore is shown in stick representation where gray represents carbon, red represents oxygen, and blue represents nitrogen. Helices closest to the viewer are rendered transparent. PDB ID 1EMA. B. A proposed mechanism for the series of post-translational modifications that converts the serine 65, tyrosine 66, glycine 67 tripeptide sequence into the fluorescent chromophore (Heim et al. 1994). The conjugated system of the mature chromophore is represented as in A. Dashed lines represent connections to the polypeptide chain.

Chromophore formation and excited state dynamics

The spontaneous formation of the Aequorea green fluorescent protein chromophore within the folded \(\beta\)-can protein structure must necessarily involve at least three key steps: cyclization of the main chain, loss of a molecule of water (dehydration), and oxidation with molecular oxygen. The exact order and mechanism of these steps is a matter of ongoing investigation (Zhang et al. 2006; Barondeau et al. 2006). An early, and still generally accepted, proposed mechanism is shown in Figure 3B (Heim et al. 1994). In this mechanism, chromophore formation starts with the nucleophilic glycine 67 amide nitrogen attacking the electrophilic serine 65 carbonyl carbon to form a 5-membered ring in the main chain of the protein. The resulting tetrahedral hemiaminal intermediate undergoes an elimination of water to form a second intermediate. In the final step, the C\(\alpha\)-C\(\beta\) bond of tyrosine 66 is oxidized to a double bond with consumption of molecular oxygen and generation of hydrogen peroxide (Zhang et al. 2006). The installation of this double bond simultaneously converts the 5-membered ring into an aromatic system and puts it into conjugation with the aromatic phenol ring of the tyrosine side chain. Chromophore formation is spontaneous only within the context of the fluorescent protein \(\beta\)-can structure where steric constraints force the peptide into a tight turn conformation (Branchini et al. 1998) and the side chains of highly conserved residues, such as glutamate 222 and arginine 96, are positioned to facilitate the reaction.

In the wild-type Aequorea green fluorescent protein the chromophore exists as an equilibrating mixture of the neutral phenol (absorbance \(\lambda_ {max}\) = 397 nm, extinction coefficient = 25,000 M\(^ {-1}\)cm\(^ {-1}\)) and anionic phenolate (absorbance \(\lambda_ {max}\) = 475 nm, extinction coefficient = 9,500 M\(^ {-1}\)cm\(^ {-1}\)) (Morise et al. 1974; Heim et al. 1994; Patterson et al. 1997). Regardless of whether excitation is at 397 nm or 475 nm, the fluorescence emission occurs from the anionic phenolate species (fluorescence \(\lambda_ {max}\) = 504 nm) with a quantum yield of 0.79 (Patterson et al. 1997). Excitation of the neutral phenol species results in the fast (tens of picoseconds) excited state proton transfer (ESPT) (Chattoraj et al. 1996) of the phenol proton to an internal hydrogen bond network (Brejc et al. 1997). Variants of Aequorea green fluorescent protein with the ground state equilibrium shifted to either the phenol (Tsien 1998; Zapata-Hommer et al. 2003) or phenolate (Heim et al. 1995) species are particularly useful for fluorescence imaging applications. A blue fluorescent variant (fluorescence \(\lambda_ {max}\) = 456 nm) with both the ground and excited state equilibriums shifted toward the phenol has been reported (Ai et al. 2007).

Fluorescent proteins in other organisms

Anthozoan fluorescent proteins

Figure 4: A close-up of the adult red-cyan morph of the stony coral Acropora millepora. The red fluorescence is attributable to amilRFP and the cyan fluorescence is attributable to amilCFP (Alieva et al. 2008).Image credit: Mikhail V. Matz and Jörg Wiedenmann.
Figure 5: The molecular structure of Discosoma red fluorescent protein. One \(\beta\)-can of the homotetramer is colored magenta. The polypeptide chain is shown in cartoon format and the conjugated system of the chromophore (refer to Figure 7B) is shown in stick representation. PDB ID 1G7K.

The growing popularity of Aequorea green fluorescent protein, and the demand for additional variants that fluoresced at wavelengths other than green, prompted researchers to begin the search for homologues in other marine organisms. This effort came to fruition in late 1999 when at team of researchers from the Russian Academy of Science reported that reef Anthozoa contain fluorescent proteins with hues ranging from cyan to red (Figure 4) (Matz et al. 1999). The inspiration behind this breakthrough discovery is credited to the evolutionary biologist Yulii A. Labas who had prompted Mikhail V. Matz, then a graduate student in the lab of Sergey A. Lukyanov, to attempt to clone Aequorea green fluorescent protein homologues from the brightly colored tentacle tips of a sea anemone and several other Anthozoan organisms. An amusing recountal of this episode can be found here. A red fluorescent protein (commonly known as DsRed) from Discosoma sp. mushroom anemone served as the focal point of much of the initial excitement and just one year after the initial report, papers describing its chromophore structure (Gross et al. 2000), obligate tetrameric structure (Baird et al. 2000), and x-ray crystal structure (Figure 5) (Wall et al. 2000; Yarbrough et al. 2001) began to appear in the literature. These reports revealed that Discosoma red fluorescent protein was very similar to Aequorea green fluorescent protein in terms of its overall fold (a \(\beta\)-can) and chromophore-formation chemistry. The key difference between the red and green proteins is that Discosoma red fluorescent protein undergoes an additional fourth step in the chromophore maturation pathway, oxidizing the adjacent C\(\alpha\)-N bond to form an acylimine moiety that extends the conjugated system by two double bonds (Figure 6 and Figure 7B).

It is now generally understood that many of the bright and varied colors of reef coral are due to fluorescent proteins and their nonfluorescent homologues (Dove et al. 2001). Indeed, 4-years prior to the groundbreaking paper from Lukyanov, the pink pigment of the Anthozoan coral Pocillopora damicornis had been isolated and shown to be a proteinaceous pigment, as opposed to a non-protein pigment, which the authors dubbed pocilloporin (Dove et al. 1995). The authors did not, at that time, recognize that the protein was indeed a member of the GFP-like superfamily. This revelation was reported sometime after the Lukyanov paper, when the gene encoding pocilloporin was sequenced and shown to encode a protein with 19.6% identity with Aequorea green fluorescent protein (Dove et al. 2001). It is now apparent that, in terms of extinction coefficient and quantum yield, naturally occurring fluorescent proteins lie on a broad continuum. Pocilloporin (often referred to as a nonfluorescent chromoprotein) can be conceptualized as a fluorescent protein that has a quantum yield of effectively zero and thus lies at one extreme of the continuum.

Figure 6: The molecular structure of one \(\beta\)-can of Discosoma red fluorescent protein as viewed from the top in Figure 5. The conjugated system of the chromophore (refer to Figure 7B) is shown in stick representation where gray represents carbon, red represents oxygen, and blue represents nitrogen. Helices closest to the viewer are rendered transparent. PDB ID 1G7K.

Fluorescent proteins in non-Cnidarian animals

In recent years fluorescent proteins have been identified in several non-Cnidarian animals. Specifically, fluorescent proteins have been identified in 3 species of lancelet (phylum Chordata, subphylum Cephalochordata) (Deheyn et al. 2007) and a copepod crustacean (phylum Arthropoda, subphylum Crustacea) (Shagin et al. 2004; Masuda et al. 2006). Although a peer-reviewed report has not appeared, Axxam (Milan, Italy) has filed a patent on a fluorescent protein gene derived from a comb jelly, possibly adding the phylum Ctenophora to this list (WIPO Pub. No.: 2008/041107).

Variations in chromophore structure

The archetypical Aequorea green fluorescent protein chromophore is remarkably tolerant of chemical modifications that change its spectral properties. The limits of this tolerance have been explored by both natural protein evolution (undoubtedly) and unnatural laboratory engineering (see below). In terms of the range of natural variations of fluorescent protein chromophores, researchers have now discovered at least 5 chromophores with chemically distinct conjugated systems (Figure 7). Each of these chemical structures is associated with a range of fluorescence hues, following the general trend that more extended conjugation produces longer wavelength fluorescence. For example, the structures shown in A, C, and B of Figure 7, are associated with greenish, yellowish, and reddish hues, respectively. The exact excitation and emission wavelengths for a particular chromophore structure also depends strongly on the protein microenvironment that surrounds the chromophore. For example, naturally occurring cyan fluorescent proteins from Anemonia majano (Henderson and Remington 2005), Discosoma striata (Malo et al. 2008), and Clavularia sp (Ai et al. 2006) have Aequorea green fluorescent protein-type chromophores (Figure 7A), but substantially blue-shifted emission due to electrostatic interactions with the surrounding protein microenvironment.

Another type of variation in chromophore structure is the stereochemistry about the exocyclic double bond of the imidazolidinone moiety (Figure 7F). The two possible stereoisomers are often referred to as cis and trans, but the Z and E designation is the more appropriate nomenclature. In principle, every type of fluorescent protein chromophore could exist as either the Z or E stereoisomer depending on the steric constraints of the protein microenvironment and/or prior illumination. For example, the Discosoma red fluorescent protein chromophore is the Z stereoisomer shown in Figure 7B, while a nonfluorescent chromoprotein from Montipora efflorescens (Prescott et al. 2003) and a far-red fluorescent protein from Entacmaea quadricolor (Petersen et al. 2003) exist as the E stereoisomer of this same chromophore structure. Photoinduced isomerizations between Z and E isomers in the same protein have also been shown to occur in some cases (Henderson et al. 2007), and are likely responsible for the phenomenon of on/off blinking in Aequorea green fluorescent protein (Dickson et al. 1997). A photoinduced isomerization of Z and E stereoisomers of the chromophore shown in Figure 7D is also the likely mechanism by which the Anemonia sulcata "kindling" nonfluorescent chromoprotein converts into a red fluorescent protein upon illumination (Chudakov et al. 2003; Quillin et al. 2005; Andresen et al. 2005), though it remains unclear which stereoisomer is the fluorescent species (Henderson and Remington 2006).

Figure 7: Fluorescent protein chromophore structures known to occur in nature. All structures are shown as the Z stereoisomer. A. The archetypical Aequorea green fluorescent protein chromophore (Shimomura 1979). B The Discosoma red fluorescent protein chromophore (Gross et al. 2000). C The Zoanthus yellow fluorescent protein chromophore (Remington et al. 2005). D The Anemonia sulcata "kindling" fluorescent protein chromophore (Chudakov et al. 2003; Quillin et al. 2005; Tretyakova et al. 2007). E Trachyphyllia geoffroyi "Kaede" red fluorescent protein chromophore (Ando et al. 2002; Mizuno et al. 2003). F Representation of the more common Z and less common E stereoisomers of fluorescent protein chromophores.

In addition to inducing interconversion of stereoisomers, illumination can also result in photochemical reactions that change the covalent structure of the chromophore. For example Trachyphyllia geoffroyi "Kaede" red fluorescent protein (Ando et al. 2002; Mizuno et al. 2003) initially forms a typical Aequorea green fluorescent chromophore. Subsequent illumination with ultraviolet light results in an elimination reaction across the C\(\alpha\)-C\(\beta\) bond of the adjacent histidine side chain (Hayashi et al. 2007), thus cleaving the main chain of the polypeptide and producing the red fluorescent chromophore shown in Figure 7E. Proteins that undergo effectively identical reactions have also been cloned from Lobophyllia hemprichii (Wiedenmann et al. 2004; Nienhaus et al. 2005) and Dendronephthya sp. (Gurskaya et al. 2006).

Biology of fluorescent proteins

Function

In bioluminescent organisms such as Aequorea and Renilla, fluorescent proteins are found to be partnered with the bioluminescent proteins aequorin (Johnson et al. 1962; Morin et al. 1971) and luciferase (Ward and Cormier 1976), respectively. The function of the fluorescent protein is to act as a bioluminescence resonance energy transfer (BRET) acceptor that converts the otherwise blue emission of the bioluminescent protein into a longer wavelength green emission. A possible role of bioluminescence in Aequorea may be to attract secondary predators upon being attacked; the so-called "Burglar-alarm" hypothesis (Burkenroad 1943; Abrahams et al. 1993). This is consistent with the observation that Aequorea rarely, if ever, exhibits bioluminescence in the absence of stimulation (Mills, C.E. Bioluminescence of Aequorea, a hydromedusa).

Fluorescent proteins (and nonfluorescent chromoproteins) in Anthozoans seem to be capable of functioning in one or more species-dependent capacities. One key biological function may be to provide photoprotection to symbiotic photosynthetic algae in high light environments (Kawaguti 1944; Salih et al. 2000). However, this function alone can not explain the variety of different colors that are observed in reef Anthozoans. The coloring provided by fluorescent proteins may also be important in species identification by reef fish (Matz et al. 2006). These are but two of the numerous speculative hypotheses regarding the biological roles of fluorescent proteins (Alieva et al. 2008; Schnitzler et al. 2008). It is remarkable that, despite the critical role of coral reefs in supporting a wide diversity of ocean life and providing food for humans, so little research has been devoted to understanding the role of color and fluorescence in reef biology and ecology.

Evolution

The first member of the fluorescent protein superfamily likely arose early in the metazoan lineage, sometime between 500 and 1000 million years ago. Evidence for this early date comes from the discovery of fluorescent proteins in multiple phyla including Cnidaria, Arthropoda, and Chordata, indicating that the gene was present in the common ancestor. A study of reconstructed ancestral fluorescent proteins has demonstrated that the earliest fluorescent protein in coral was probably green fluorescent (Ugalde et al. 2004). Red fluorescent proteins, as well as other colors that require additional chromophore modifications, originated later and in more than one species. This feat of evolution has recently been accomplished in a laboratory setting (Mishin et al. 2008).

A fluorescent protein ortholog (Hopf et al. 2001), known as nidogen globular fragment 2, has been discovered in humans and a number of other animals (Kohfeldt et al. 1998). Although they share only 10% sequence identity, the structural similarity between Aequorea green fluorescent protein and mouse nidogen globular fragment 2 is remarkable. The two structures can be superimposed such that root mean square deviation over 195 \(\alpha\)-carbon atoms is 2.5 \(\AA\ .\) In nidogen globular fragment 2 the ability to form the intrinsic chromophore has been lost due to non-productive substitutions at numerous residue positions that are known to be critical for chromophore formation. The biological function of the nidogen globular fragment 2 domain is to provide structural stability to the basement membranes that serve as a scaffold for epithelial cell growth and differentiation. Mice in which both nidogen-1 and -2 have been genetically deleted have defects in limb development and die soon after birth (Bose et al. 2006).

Fluorescent proteins in biological research

Engineering improved variants

As noted above, the scientific community was quick to embrace Aequorea green fluorescent protein in the mid-1990s. However, researchers quickly discovered that, in some regards, the wild-type protein is sub-optimal with respect to the demands of typical biological fluorescence imaging applications. For example, the wild-type protein had evolved to fold and undergo the chromophore-forming reaction most efficiently at the cool temperatures of the ocean. Engineering of Aequorea green fluorescent protein to fold more efficiently at the physiological temperature of 37 degrees celsius was an early example of how protein engineering could provide substantial improvements in the protein's usefulness with respect to imaging applications (Crameri et al. 1996; Cormack et al. 1996). Another example of engineering Aequorea green fluorescent protein to be more useful for certain imaging applications involved the introduction of mutations that abolished the tendency of the protein to dimerize at high concentrations (Zacharias et al. 2002).

Figure 8: Engineered fluorescent protein chromophores not known to occur in nature. A. The histidine variant of the Aequorea green fluorescent protein chromophore (Figure 7A) (Heim et al. 1994). This is the chromophore of blue fluorescent protein and its descendants (Heim and Tsien 1996). B The tryptophan variant of Aequorea green fluorescent protein (Heim et al. 1994). This is the chromophore of cyan fluorescent protein and its descendants (Tsien 1998). C The phenylalanine variant of Aequorea green fluorescent protein chromophore (Cubitt et al. 1995). This variant requires ultraviolet excitation and has not yet been used in a research application (Tsien 1998). D The tryptophan variant of Discosoma red fluorescent protein. This is the chromophore of the yellow fluorescent mHoneydew variant (Shaner et al. 2004). E The phenylalanine variant of Discosoma red fluorescent protein. This is the chromophore of the blue fluorescent mBlueberry variant (Ai et al. 2007). F The chromophore of the orange fluorescent mOrange variant (Shaner et al. 2004). This chromophore has a 3rd ring that provides a conjugated system effectively identical to that of Zoanthus yellow fluorescent protein chromophore (Figure 7C) (Shu et al. 2006).

Another perceived limitation of Aequorea green fluorescent protein is that it was originally available in just the one fluorescent color. However, this limitation was also soon overcome through the efforts of protein engineers who created a series of variants with distinct excitation and emission maxima during the mid-to-late 1990s (Tsien 1998). Examples of each of the main color classes include: a blue fluorescent protein known as BFP (Heim et al. 1994, 1996); a cyan fluorescent variant known as CFP (Heim et al. 1996; Tsien 1998); a yellow fluorescent variant known as YFP (Ormo et al. 1996; Wachter et al. 1998); a violet-excitable green fluorescent variant known as Sapphire (Tsien 1998; Zapata-Hommer et al. 2003); and a cyan-excitable green fluorescing variant known as enhanced green fluorescent protein or EGFP (Yang et al. 1996). As shown in Figure 8AB, the blue and cyan fluorescing variants were created through modification of the chromophore structure. In contrast, the yellow variant retains the Aequorea green fluorescent protein chromophore structure but the fluorescence is red-shifted due to an amino acid substitution that creates a \(\pi\)-\(\pi\) stacking interaction with the chromophore. In the Sapphire and EGFP variants, the chromophore environment has been modified such that the ground state equilibrium is shifted either towards the neutral phenol or anionic phenolate form, respectively.

As discussed above, a breakthrough in the area of fluorescent protein research came with the discovery and cloning of cyan, yellow, orange, and red fluorescent proteins from reef Anthozoan species (Matz et al. 1999). These fluorescent proteins presented protein engineers with new opportunities but also new challenges. For example, some of the initial excitement regarding the Discosoma red fluorescent protein was tempered by the realization that the protein was limited to only a subset of live cell imaging applications due to its tetrameric structure (Figure 5) and slow and incomplete chromophore formation (Baird et al. 2000; Lauf et al. 2001; Bevis et al. 2002; Soling et al. 2002; Bulina et al. 2003). It was only by combining rational protein engineering with an extensive process of directed evolution that researchers were ultimately successful in creating a more useful monomeric variant of Discosoma red fluorescent protein (Campbell et al. 2002). The resulting monomeric Discosoma red fluorescent protein served as the progenitor for a series of variants, known as the mFruits (m for monomeric), with hues that span the yellow to far-red region of the spectrum (Shaner et al. 2004; Wang et al. 2004; Shaner et al. 2008). While some of the mFruit variants have modified chromophore structures (Figure 8D-F), most retain the standard Discosoma red fluorescent protein chromophore (Figure 7B). The growing number of naturally tetrameric fluorescent proteins that have been "monomerized" by protein engineering includes variants from Galaxea sp. (Karasawa et al. 2003), Fungia concinna (Karasawa et al. 2004), Lobophyllia hemprichii (Wiedenmann et al. 2004), Pectiniidae (Ando et al. 2004), Dendronephthya sp. (Gurskaya et al. 2006), Montipora sp. (Kogure et al. 2006), and Clavularia sp. (Ai et al. 2006).

Applications

The most popular applications of fluorescent proteins involve exploiting them for imaging of the localization and dynamics of specific organelles (Rizzuto et al. 1995) or recombinant proteins (Ballestrem et al. 1998) in live cells. For imaging of a specific organelle, standard molecular biology techniques are used to fuse the gene encoding the fluorescent protein to a cDNA encoding a protein or peptide known to localize to that specific organelle. This fusion is done such that the chimeric gene will be expressed as a single polypeptide, creating a covalent link between the targeting motif and the fluorescent protein. A plasmid containing the chimeric gene under control of a suitable promoter is used to transfect mammalian cells that then express the gene to produce the corresponding chimeric protein. The chimera localizes to the target organelle and thus renders it fluorescent. Through the use of fluorescence microscopy, the morphology, dynamics, and distribution of the organelle can be imaged as a function of time. The procedure for imaging of a fusion between a fluorescent protein and a specific protein-of-interest (in order to gain insight into its localization and dynamics) is identical (Figure 9). The availability of a broad selection of colors of fluorescent protein (Shaner et al. 2005) has provided researchers with the means to image the localization of multiple organelles and/or proteins-of-interest, simultaneously (Rizzuto et al. 1996; Kogure et al. 2006). Numerous examples are available to view at the Hamamatsu digital video gallery.

Figure 9: A schematic representation of how a monomeric fluorescent protein is employed for imaging of \(\beta\)-actin. A mammalian cell is transfected (a) with a cDNA chimera composed of a fusion of the genes encoding the fluorescent protein and \(\beta\)-actin. The gene is transcribed (b) to produce mRNA that is then translated (c) to form the chimeric protein. The trafficking and localization (d) of the protein is dictated by the protein-of-interest and the fluorescent protein, ideally, does not interfere. In the case of \(\beta\)-actin, the chimera is incorporated into actin filaments (e) along with the endogenous protein. Shown in the inset is a fluorescence image of a gray fox lung fibroblast (FoLu) cell that has been transfected with mTFP1-\(\beta\)-actin (Ai et al. 2008). Scale bar represents 10 microns.

The availability of a broad selection of colors of fluorescent protein has also enabled researchers to develop methods to probe whether two proteins are within a distance of less than 10 nm of each other. The observation of colocalization for two different proteins-of-interest fused to different colors of fluorescent protein is insufficient to address this question, since the theoretical resolution limit of conventional optical imaging is several hundred nanometers. To obtain information on the proximity of two proteins at better than 10 nm resolution, investigators exploit the phenomenon of Förster (or fluorescence) resonance energy transfer (FRET) (Förster 1948). FRET is the distance- and orientation-dependent radiationless transfer of excitation energy from a donor fluorophore to an acceptor chromophore. The greater the extent of the spectral overlap between the donor emission and the acceptor excitation, the more efficient the energy transfer is for a particular FRET pair of fluorescent proteins. Since spectral overlap is constant, and orientations are assumed to be random (dos Remedios and Moens 1995), the efficiency of FRET is generally a good indicator of distance between the donor and acceptor fluorescent protein. Accordingly, by expressing the donor fluorescent protein as a fusion with one protein-of-interest and the acceptor fluorescent protein as a fusion with a second protein-of-interest, the distance between the two proteins-of-interest can be inferred from the FRET efficiency measured using live cell fluorescence microsopy (Bastiaens et al. 2000; Selvin 2000). In contrast to intermolecular FRET for the investigation of protein-protein interactions, intramolecular FRET between two fluorescent proteins fused in the same polypeptide chain can be used to investigate small molecule dynamics and enzyme activities in a live cell. These intramolecular FRET constructs are often referred to as "reporters" or "biosensors" of biochemical activities (Zhang et al. 2002, 2007). One of the most well known examples of such reporters are the "Cameleons" which enable imaging of intracellular calcium ion concentrations (Miyawaki et al. 1997). Although the BFP-EGFP pair was the first set of engineered fluorescent proteins that had the appropriate spectral overlap to allow their use as a FRET pair, it was soon superseded by the still popular CFP-YFP pair that has been used in vast majority of FRET experiments to date (Miyawaki et al. 1997). However, with the advent of the numerous Anthozoa-derived fluorescent proteins, a wide variety of new FRET pair combinations has been explored. One application that has been made possible with these new combinations is the simultaneous imaging of two different FRET pairs in a single cell (Piljic et al. 2008; Ai et al. 2008).

Variants derived from Aequorea green fluorescent protein have proven to be fairly tolerant of a variety of dramatic structural manipulations including the genetic insertion of a second protein (Baird et al. 1999; Doi et al. 1999), circular permutation (Baird et al. 1999; Topell et al. 1999), and even splitting into two polypeptide chains that are competent to fold into a functional fluorescent protein when brought into close proximity (Ghosh et al. 2000; Hu et al. 2002). In certain cases, fluorescent protein chimeras that incorporate a genetically inserted second protein, or circularly permuted fluorescent proteins with interacting proteins or peptides fused to the new N- and C-termini, can be used as single fluorescent protein-based (as opposed FRET-based) biosensors. These biosensors are designed such that the binding of the second protein to its ligand, or the ligand-dependent interaction of the attached proteins and/or peptides, results in a change in the protein environment (and thus the fluorescence properties) of the chromophore. Single fluorescent protein-based biosensors have been successfully used for the imaging of localized calcium ion (Baird et al. 1999; Nagai et al. 2001; Nakai et al. 2001) and hydrogen peroxide (Belousov et al. 2006) concentrations. Yet another strategy for the creation of single fluorescent protein-based biosensors is to modify the \(\beta\)-can itself, such that the fluorescence properties of the chromophore are dependent on external factors. Some representative examples of biosensors of this type include ones for halide ions (Jayaraman et al. 2000) and cellular redox potential (Ostergaard et al. 2001; Hanson et al. 2004).

As mentioned in the Variations in chromophore structure section, illumination can induce conversions between chromophore stereoisomers or photochemical reactions of the chromophore in certain fluorescent proteins. Researchers have exploited these illumination-dependent changes in spectral properties to create "optical highlighters" that can be turned "on", and sometimes "off", by illumination at specific wavelengths (Lukyanov et al. 2005). These tools are often referred to as photoactivatable, photoconvertable, or photoswitchable fluorescent proteins, depending on the particular mechanism by which the illumination-dependent change occurs (Remington 2006) (Shaner et al. 2007). These optical highlighter fluorescent proteins have enabled numerous new applications including imaging of sub-populations of cells during organism development (Chudakov et al. 2003), imaging of fast protein dynamics (Ando et al. 2004), and imaging with sub-diffraction limit resolution (Betzig et al. 2006; Willig et al. 2006).

While the creation of fluorescent cells and tissues has now become a relatively mundane practice for many practicing scientists, the creation of a transgenic animal expressing visible amounts of a fluorescent protein is still considered exotic and worthy of media attention. Some of the most highly publicized examples from recent years include: fluorescent zebrafish commercially available from GloFish, a green fluorescent pig (Lai et al. 2002), and red fluorescent cats (Yin et al. 2008). Perhaps the most dramatic and technically sophisticated example of creating a fluorescent transgenic animal is the recently reported "Brainbow" mouse in which each cell of a living mouse's brain was colored one of about 90 different colors (Livet et al. 2007). For more examples please see Marc Zimmer's website. Not all examples of transgenic organisms receive so much attention, and numerous researchers are using fluorescent proteins in organisms ranging from plants (Stewart 2006) to mice (Hadjantonakis et al. 2003) for the study of the basic biological processes associated with normal and diseased cells. Fluorescent mice have been particularly useful for the study of cancer cells in living animals (Hoffman 2005).

Concluding remarks

The field of fluorescent protein research has progressed from squeezing jellyfish through cloth bags to the creation of transgenic mice with rainbow brains. During the course of this progression, the fluorescent proteins have been studied, modified, and applied to an extent that is enjoyed by only a handful of other classes of proteins. This attention is clearly warranted due to the incalculable value of fluorescent proteins for the study of cells, tissues, and even whole animals at level of detail and subtlety that would otherwise be experimentally inaccessible. Unfortunately, Aequorea victoria has not enjoyed a similar level of interest, and in recent years there has been scant research on this animal that gave science so much.

Notes on terminology

  • The term fluorescent protein is used in this article to refer specifically to proteins that are structural homologs of Aequorea green fluorescent protein and that are able to form an internal visible wavelength fluorophore from their own polypeptide sequence. However, it is important to note that any protein that has a tryptophan, tyrosine, or phenylalanine residue within its sequence is, strictly speaking, a "fluorescent protein" due to the inherent but non-visible fluorescence of these amino acids. A number of other proteins, such as the phycobiliproteins, are visibly "fluorescent proteins" due to the presence of non-proteinaceous chromophores that are associated with the protein.
  • Fluorescence is the emission of a lower energy (more red-shifted) photon from a molecule (a fluorophore) that is in a singlet excited state due to prior absorption of a higher energy photon (excitation). The singlet excited state is short-lived (less than 100 ns) so fluorescence is observed, effectively, only during excitation. Emission from the long-lived (greater than milliseconds) triplet excited state is known as phosphorescence and can be observed long after the illumination has ceased. For bioluminescence (i.e. enzymatically catalyzed chemiluminescence), the excited state is first produced by a chemical reaction and not by absorption of a photon.

Resources

Web resources

Fluorescent proteins section of Molecular Expressions
Hamamatsu fluorescent protein digital video gallery
Roger Y. Tsien's website
Sergey A. Lukyanov's website
Mikhail V. Matz's website
Claudia E. Mills's website
An interview with Martin Chalfie
Marc Zimmer's website
Discovery of Renilla green fluorescent protein by John E. Wampler
A discussion of coral fluorescence for aquarium hobbyists by Anthony Calfo
A multipart series on coral color, fluorescence, and fluorescent proteins by Dana Riddle: Part 1, Part 2, Part 3, Part 4, Part 5, Part 6, Part 7

Recommended reading

Internal references

  • Eugene M. Izhikevich (2007) Equilibrium. Scholarpedia, 2(10):2014.


References

External Links

Robert E. Campbell's website

Personal tools
Namespaces

Variants
Actions
Navigation
Focal areas
Activity
Tools