Section 2 highlights the importance of a reporter protein-based real time redox assay. The advantages of using GFPs and its color variants (e.g., yellow fluorescent proteins) for a variety of applications and how the chromophore is formed have been reviewed previously (see ref. 16, and Chap. 4).
In the context of the use of GFP for redox-sensing in vivo, the opportunity to conduct noninvasive experiments with a protein which is regarded as neutral in its impact on biological systems, contains no disulfide bonds, can be expressed in a tissue-specific manner and be targeted to specific subcellular compartments makes this an attractive starting point (17). However, in plant cells transformed with GFP-based genes, high-level expression (i.e., the strongest fluorescers) is sometimes not recovered, suggesting that deleterious effects on plant cells can occur (18). Thus, at lower levels of expression of GFP, more subtle perturbations may operate, which the researcher ought to keep in mind. Nevertheless, the potential advantages of a noninvasive system have spurred the independent development of two types of redox GFPs; redox-sensitive yellow fluorescent proteins (rxYFPs; ref. 17) and redox GFPs (roGFPs; ref. 2).
3.1. A Comparative Both systems, although developed independently, have several Description of the features in common, but also some differences, which will influ-
rxYFP and roGFP ence the choice of which system to deploy for leaves of higher
Redox Reporter plants. Both types of GFP have been made redox-sensitive by the
Systems introduction of solvent-exposed cysteine residues by site-directed mutagenesis into adjacent strands of the P-barrel that surrounds the GFP chromophore. Upon oxidation of the redox reporters, the principle behind the creation of a disulfide bridge in the P-barrel is to bring about a conformational change that perturbs the structure of the chromophore and favors its protonated (neutral) form over its anionic form, bringing about a change in the fluorescence properties of the GFPs. This feature is important for considerations on the choice of rxYFP or roGFP for leaves.
In both initial reports describing rxYFP and roGFP, the same P strands of the protein were targeted. For the rxYFP variant protein harboring the mutations N149C and S202C, the difference between maximal fluorescence of the reduced form and diminished fluorescence of the oxidized form was the greatest (17) whereas in the roGFP variants produced (roGFP1 and roGFP2), the mutations were S147C and Q204C (2).
Despite the introduction of similarly positioned disulfide bridge, these proteins do have different midpoint redox potentials and therefore may well be in equilibrium with different cellular redox couples. RxYFP has a midpoint redox potential (E 0) of -261 mV, determined by titration against a range of GSH to GSSG ratios (17) and was suggested to accurately report the redox status of the glutathione pool in vivo. This was subsequently confirmed in yeast with the use of mutants with altered GSH levels and GSH-to-GSSG ratios (19). In wild-type Escherichia coli and yeast cells, it was estimated that steady-state oxidized protein levels were present at approx 50% and 90%, respectively, of the total amount of rxYFP present (17, 19). One problem with the rxYFP system is that the best fluorescence difference that could be obtained between oxidised and reduced form was approx. 2.2fold, which possibly is a limitation for measuring subtle changes in cellular redox potential (17).
In contrast, the roGFPl and roGFP2 variant proteins have consensus midpoint redox potentials of -291 and -287 mV, respectively, using redox titrations with dithiothreitol, lipoic acid, and bis(2-mercaptoethyl)sulfone redox couples (2, 20). The highly reducing redox potential of roGFPs makes it unclear as to the cellular redox couple(s) the protein would be in equilibrium with. In addition, the rates of changes in redox potential are in the order of tens of minutes (21) making them less than ideal to respond to transitory changes in redox states associated with the production of short bursts of ROS.
The original roGFPs and rx YFPs have been modified further to address some of the deficiencies in the original system. In both cases, changing the electrostatic environment by introducing positively charged amino acids around the disulfide bridge improved the properties of redox reporters. In the case of roGFPl, the changes N149K, F223R, and S202K (roGFP1-R12 variant) were particularly effective at lowering the midpoint redox potential of the reporter to -265 mV and increasing the rates of reduction and oxidation of fivefold and sixfold, respectively (21). In the case of rxYFP, the combined changes of Y200R, Q204R, and A227R resulted in a 13-fold difference in fluorescence between its maximal oxidised and reduced forms (22).
3.3. Ratiometric Versus Absolute Determinations of Fluorescence
One of the crucial differences between the two types of redox reporters is the original choice of starting fluorescent protein and the consequences this has for the use of these proteins as redox reporters in organs such as the leaf. Wild-type GFP from Aequoria victoria exhibits two excitation maxima at wavelengths of 400 and 475-490 nm. These maxima correspond to absorption of the anionic and neutral forms of the chromophore (23). In wild-type GFP, both of these excitation maxima elicit fluorescence (maximum at 508 nm). However, in the production of "color" forms of GFP, including YFP, the dual excitation maxima have been replaced by a single "red-shifted" excitation and emission maximum (23). In roGFPl, derived from wild-type GFP (2), the dual excitation maxima are retained and, as the protein becomes more oxidized, there is a decrease in the fluorescence from the 400-nm excitation and an increase in the fluorescence from the 470-nm excitation (2, 20). This means that a ratio-metric approach to measuring changes in fluorescence can be adopted in which the fluorescence values can be expressed as a ratio of the excitation at 400/470 nm, thus resulting in an increase in the ratio as roGFPl becomes more oxidized, the maximum difference in ratios between fully oxidized and reduced forms of roGFPl being sixfold in vitro (2, 20, 21). Ratiometric determinations of fluorescence eliminate or reduce possible artefactual influences on the data caused by photobleaching or inadvertent modification of the protein, protein concentration, illumination stability, excitation path length and non-uniform distribution of the protein in groups of cells (2, 20).
Determination of redox potential using rxYFP relies upon direct measurement of fluorescence quenching as the protein becomes more oxidised and vice versa when being reduced (17). The ratiometric approach to collection of fluorescence data is not an option for this system, but nevertheless with appropriate estimations of protein concentrations, this system has proved very effective in the experimental systems where it has been reported so far (17, 19).
A further problem that might also dictate choice of redox probe is that YFP fluorescence and its variants can be quenched by anions, in particular halide ions, and is more susceptible to pH change. This need not be a problem in some experimental systems but it should be noted that roGFPl is relatively insensitive to these problems (2, 20).
Both version of redox GFPs have been used successfully in living cells. RxYFP has been used in E. coli and yeast (17, 19), whereas roGFPs have been targeted to the cytosol and mitochondria in mammalian cells and in transgenic Arabidopsis and measured in roots (2,12, 20).
A clear demonstration of the potential of redox reporters to provide important information about the glutathione redox couple in cells was first shown for the cytosol of E. coli harboring rxYFP and the impact of the loss of the thioredoxin system for maintaining reduced proteins. Mutants deficient in thioredoxin (trxB) showed 30% more oxidized rxYFP than wild-type cultures. Thioredoxin does not react with rxYFP and the cytoplasmic GSH/GSSG redox potential and concentration was close to the value estimated in vitro for when the rxYFP and glutathione redox couples were in equilibrium. Therefore, these data suggested that loss of the thioredoxin system impacts on the redox state of the glutathione pool, which partly compensates for the loss of the thioredoxin system.
It should be noted that corrections for the amount of reduced and oxidized rxYFP have to be performed in both the bacteria and yeast experiments (17, 19) to ensure that fluorescence yield was not a consequence of variation in the amount of protein. This was achieved by extracting cultures directly into strong acid to prevent further oxidation and reduction of the rxYFP protein,
3.4. Monitoring Cellular Redox Status In Vivo Using rxYFP androGFPl followed by fractionation on nondenaturing polyacrylamide gels, which were used to resolve oxidized and reduced forms of the reporter and semiquantification by an immunoblotting technique. The rxYFP was also used to determine the in vivo redox potential of the cytosolic glutathione pool in yeast, which was estimated to be highly reducing at -289 mV, indicating a very low concentration of GSSG.
3.5. Choice of System Considering the differential properties of rxYFP and roGFP dis-
for LeaVes cussed previously, the development of a redox reporter system in a layered multicellular organ, such as intact green leaves, favors roGFP over rxYFP for a number of reasons. Despite the advantage of having a midpoint redox potential in equilibrium with the GSH redox couple, rxYFP fluorescence emission could result from changes in halide concentrations and pH (24), as well as reflecting the redox state of the targeted environment. With the use of rxYFP, it would be necessary to use a variety of controls to deconvolute the emission spectra to separate cellular redox signals from those reflected changes in halide ion and pH. In addition, rxYFP, with a single excitation maximum, prevents the use of ratiometric methods, a desirable approach in complicated tissues that eliminates the need to separately determine protein concentrations.
These limitations of rxYFPs point towards the use of a roGFP-based reporter system for application in leaves in vivo. Some of the initial problems associated with roGFPs have been overcome with the development of a pH-insensitive variant of roGFP1, roGFP1-R12, which has a less negative midpoint potential of -265 mV (approaching the midpoint potential of glutathione) and an increased response time (21).
The roGFP1 variant has been successfully reported in Arabi-dopsis roots to determine the average resting redox potentials of root cytoplasm and mitochondria. Monitoring real time dynamic changes in redox in vivo, the authors demonstrated that the mitochondria are better at buffering redox changes compared with the cytoplasm (12). However, in Arabidopsis thaliana, Haseloff and co-workers (18) showed that expression of GFP cDNA was curtailed by aberrant mRNA splicing. These researchers altered the codon usage of GFP to avoid recognition of a cryptic intron, which resulted in restored expression of the fluorescent protein. To develop a redox state reporter system targeting specific cellular compartments within Arabidopsis leaves, a similar modification would be necessary to prevent any loss of fluorescence due to splicing of the transcript by a cryptic intron.
The main advantage of using a roGFP-based system over rxYFP in leaves is that the two excitation maxima allow for the use of a ratiometric approach (2, 12, 20), eliminating the need to determine protein concentrations (see above). A possible pitfall of using roGFPs in leaves is the fact that blue excitation wavelengths overlap considerably with the absorption spectrum of leaf chlorophylls (25). Using a single excitation energy of 470 nm led to interference of chlorophyll concentration with GFP fluorescence, disrupting the proportional relationship between GFP content and fluorescence that is intrinsic to its use as a quantitative reporter (25). This complication is more prevalent in plant species that show substantial changes in chlorophyll concentrations with leaf age and was less important in Arabidopsis where changes in chlorophyll with leaf maturation were less pronounced (25).
3.6. Problems of Light The relationship between chlorophyll content and roGFP signal
Absorption reiterates the advantages of using a ratiometric approach to assess
GFP status. However, in multilayered structures such as whole leaves, complications using duel excitation wavelengths can still be envisaged. The redox status of roGFPl is determined with the use of a ratio of emission at 510 nm after excitation at 410 and 474 nm. However, these different excitation wavelengths are likely to be differentially absorbed dependent upon pigment concentrations, leaf thickness, and anatomy, resulting in different path-lengths and, therefore, differential penetration depths in the leaf.
It is, however, possible to establish wavelengths absorption profiles (26), which would allow determination of pathlengths for excitation wavelengths and possible correction factors. It is worth noting that Arabidopsis leaves are significantly thinner (236 pm) compared with many other species, thus possibly shortening the pathlength in these plants (27). However, it should be recognised that numerous environmental cues and stresses can cause changes in leaf pigment concentrations and composition. Modification of pigments with stress could result in different pathlengths and absorption profiles between control and treated plants. A decrease in the chlorophyll content in conjunction with an increase in xanthophyll pigments has been observed in Pinus halepensis needles treated with ozone, whereas in the same study, drought was shown to also increase the levels of glutath-ione reductase (28). Pigment adjustment is also found in many mutants, for example temperature-sensitive Arabidopsis mutants (29) and UV-B tolerant Arabidopsis (30).
Another possible solution to overcome differences in excitation pathlengths is to simultaneously excite and measure fluorescence emission from both the abaxial and adaxial surfaces of leaves, with appropriate excitation intensities and band-pass filters for detection. Appropriate control plants would be necessary to determine the influence of excitation pathlength on emission signals. The emission of roGFP fluorescence at 509 nm carries fewer complications and represents an advantage of roGFP over other colored variants. Green light has been shown to have one of the greatest pathlengths through the leaf (26) and is therefore less likely to be reabsorbed by leaf pigments as it is emitted. Emission from coloured variants of GFP, such as the blue (emission 440 nm) and cyan (emission 477 nm) will be heavily re-absorbed by leaf pigments. The red variant of GFP with an emission wavelength at 683 nm overlaps significantly with that of chlorophyll a fluorescence, the intensity of which would greatly exceed that emitted from any GFP.
3.7. GFP Fluorescence Although historically GFP fluorescence has been detected with
Monitors confocal microscopy, which requires some sort of tissue prepara tion, commercial systems are on the market with the capability of determining GFP signals from intact leaves. Hand-held fiberoptic fluorometers that clip onto the leaf are available (Opti-Sciences, Hudson, NH), which support a variety of source and detector combinations permitting the detection and measurement of several fluorescence markers, including GFP. Imaging systems are also available that can detect whole plant or leaf GFP signals (Qubit systems, Canada) permitting heterogeneity to be ascertained. Blue LEDs are used to provide the excitation at 490 nm whereas 510-nm emission filters in front of a CCD camera allow a GFP image to be achieved. A variety of lighting systems and optical emission filters are available for use with the other variants of GFP.
Although the commercial systems described in this chapter are for single GFP emission detection, it is possible to envisage the development of either a fibre-optic or imaging system encompassing dual excitation with detection of two emission wavelengths, which would allow roGFP detection within intact leaves. However, such a system would necessitate addressing the problems regarding differential absorption and differing path-lengths first.
Was this article helpful?