Deriving information from chlorophyll a fluorescence
Estimating PSII photochemical efficiency
It is primarily because the closure of PSII centres increases the yield of fluorescence from PSII in the manner described above (by reversing so-called ‘photochemical quenching’ of fluorescence) that it is possible to use Chl a fluorometry to investigate the functioning of PSII. This is achieved by measuring the fluorescence signal when the system being investigated is in one or more known states. For example, by measuring the fluorescence signal from dark-adapted material under very low photon irradiance, when virtually all PSII centres are open in the dark-adapted state (Fo), and during a pulse of super-saturating photon irradiance, when virtually all PSII centres are closed in the dark-adapted state (Fm), it is possible to estimate the dark-adapted photon efficiency of PSII photochemistry (the PSII maximum efficiency), as (Fm–Fo)/Fm. In a similar fashion, measurement of the light-adapted fluorescence signal (F') and the fluorescence signal when all PSII centres are closed in the light-adapted state (F'm) allows for estimation of the ‘operating’ photon efficiency of PSII photochemistry (the PSII operating efficiency), as (F'm–F')/F'm (Genty et al., 1989). Figure 2 shows an illustrative fluorescence curve, from which PSII maximum efficiency and PSII operating efficiency can be calculated. Within this figure, the terms Fv and F'q are used to describe Fm–Fo and F'm–F', which allows (Fm–Fo)/Fm and (F'm–F')/F'm to be rewritten as Fv/Fm and F'q/F'm, respectively.
Equations 1 and 2 describe the maximum and operating quantum efficiencies of PSII photochemistry in terms of rate constants for the de-excitation pathways shown in Fig. 1A and B. It is worth noting that the PSII operating efficiency can be lowered from the PSII maximum efficiency by a decrease in the effective rate constant for photochemistry ([QA]xkP) and/or an increase in the rate-constant for down-regulation (kSV). Equation 2 assumes perfect connectivity among PSII centres (equivalent to the entire population of PSII reaction centres being connected to a single pigment matrix).
where: kP, rate constant for PSII photochemistry; kD, dark-adapted rate constant for non-radiative decay within the pigment matrix associated with PSII; kF, rate constant for Chl a fluorescence within the pigment matrix associated with PSII; kSV, rate constant for the light-dependent increase in non-radiative decay within the pigment matrix associated with PSII; [QA], the fraction of PSII centres in the open state or the probability of finding a particular PSII centre in the open state.
The actual level of connectivity among PSII centres is probably somewhere between this extreme and zero connectivity (where each PSII reaction centre is embedded within a pigment matrix that is not connected to the pigment matrix of another PSII reaction centre). The effect of intermediate levels of connectivity on the yield of Chl a fluorescence was first described mathematically by Joliot and Joliot (1964) and is discussed further by Lavergne and Trissl (1995). In terms of Chl a fluorometry, the only situation where connectivity is an issue is when the parameter qP is used as a proxy for the fraction of PSII centres in the open state (see ‘Other fluorescence parameters in common usage’, below).
Quantifying down-regulation at PSII
At PSII, the loss of excitons through non-radiative decay at PSII is a regulated process, termed down-regulation. Down-regulation is a complex process, which is linked to lumen acidification (see Horton et al., 1996, for a review). Typically, an increase in down-regulation occurs when the incident photon irradiance is increased or the supply of CO2 is decreased. This brings about a ‘non-photochemical’ quenching of the fluorescence signal, which is clearly evident within the fluorescence trace shown in Fig. 2. This quenching can be quantified through changes in Fm/F'm – 1 (Bilger and Björkman, 1990), which treats down-regulation as a Stern–Volmer quenching mechanism. Equation 3 shows Fm/F'm – 1 expressed in terms of rate constants, derived from the model presented in Fig. 1A and B. It is important to note that the value of Fm/F'm – 1 is dependent upon both down-regulation (kSV in Fig. 1B and equation 3) and the initial rate constant for non-radiative decay (kD in Fig. 1 and equation 3). It follows that two samples with identical levels of down-regulation (same values of kSV) but different initial levels of non-radiative decay (different values for kD) will have different values for Fm/F'm – 1. Differences in kD will normally show up in the values of Fv/Fm; a higher level of kD resulting in a lower value for Fv/Fm. However, since Fv/Fm includes the term for photochemistry (kP), which is not represented in Fm/F'm – 1, it is possible (though admittedly unlikely) that two samples with the same values for kD and kSV, but different values for kP in the dark-adapted state, will have different values for Fv/Fm, but the same values for Fm/F'm – 1. The important point here is that although Fm/F'm – 1 can reliably be used to quantify changes in down-regulation within a single sample (since these are independent of changes in kP), this parameter should only be used across multiple samples with reference to Fv/Fm measurements and, even then, with a certain degree of caution.
Quantifying photoinactivation of PSII reaction centres
During the period of constant photon irradiance in Fig. 3, there is an increase in F' over the first few seconds, followed by a slower decrease to a constant level. F'm also decreases during the period of constant photon irradiance, reaching a minimum level at 9 min. After the light is switched off, Fo and Fm approach, but never attain, the initial dark-adapted levels. The higher level of Fo observed can be attributed to an increase in the fraction of PSII reaction centres that are in a photoinactivated state (reviewed by Barber, 1998), which results in a decrease in the PSII photochemical capacity (the maximum electron flux through the entire population of PSII reaction centres). The lower level of Fm is indicative of an increase in down-regulation, which may be wholly dependent, partly dependent, or completely independent of an increase in the fraction of PSII centres that are photoinactivated.
Because Fv/Fm contains terms for both photochemistry and non-radiative decay, any difference between Fv/Fm measured before and after light treatment could be due to photoinactivation, down-regulation, or both. It is possible to isolate changes due to photoinactivation, using Fv/(Fm.Fo) (Dominy and Baker, 1980), which is equivalent (both mathematically and conceptually) to 1/Fm–1/Fo (introduced by Havaux et al., 1991). This parameter must be normalized to gain meaningful information. It then becomes (FvRFmFo)/(FvFmRFoR), where the subscripted R represents the ‘reference’ values (usually the dark-adapted measurements made at the start of an experiment).
Quantifying photochemical and non-photochemical limitations to the PSII operating efficiency
The PSII operating efficiency (calculated as F'q/F'm) is the product of the PSII maximum efficiency in the light (calculated as F'v/F'm) and a factor relating the two efficiencies, the PSII photochemical factor (calculated as F'q/F'v). Put simply, F'q/F'v provides the fraction of the PSII maximum efficiency (what it would be if all PSII centres were in the open state) that is actually realised. The fraction of the PSII maximum efficiency that is not realised (1 – F'q/F'v) can therefore be attributed to the closure of PSII centres in the light. It is important to note, however, that the relationship between 1 – F'q/F'v and the fraction of PSII centres in the closed state is non-linear (see ‘Other fluorescence parameters in common usage’, below). A question arises as to how PSII centres that become inactivated during an experiment should be treated. The calculation of F'v/F'm and F'q/F'v requires determination of F'o (since F'v = F'm – F'o). F'o can be calculated from Fo, Fm, and F'm, using equation 4 (Oxborough and Baker, 1997b). If the pre-illumination values of Fo and Fm are used in this calculation, any centres that have become inactivated by subsequent light treatment will be treated in the same way as PSII centres that are closed by the super-saturating pulse used to measure Fm or F'm, which decreases the calculated value of the PSII photochemical factor. If post-illumination values of Fo and Fm are used, the inactivated PSII centres will decrease the value of the PSII maximum efficiency.
Other fluorescence parameters in common usage
The efficiency of non-radiative decay processes at PSII has been estimated using 1 – F'v/F'm (Demmig-Adams et al., 1996; Verhoeven et al., 1997; Barker et al., 1998). What this parameter actually provides is an estimate of what the combined efficiency of non-radiative decay processes and chlorophyll fluorescence would be if all PSII centres were open at the point of measurement. As noted earlier, the fraction of PSII centres in the open state tends to decrease with increasing photon irradiance. Consequently, 1 – F'v/F'm becomes an increasingly inaccurate parameter for estimating the efficiency of non-radiative decay at PSII as the incident photon irradiance is increased.
A set of parameters termed ‘quenching coefficients’ (qP, qE, qT, qI, and qN) have been used in a wide range of studies over the past 20 years or so (Horton et al., 1996). qP is the so-called ‘coefficient of photochemical quenching’, which has frequently been used as a proxy for the fraction of PSII centres in the ‘open’ state (Maxwell and Johnson, 2000). Unfortunately, this application of qP takes no account of the curvilinearity introduced to the relationship between fluorescence yield and the fraction of PSII centres in the open state, which was first described by Joliot and Joliot (1964). It is the case that qP is calculated in exactly the same way as the PSII photochemical factor (calculated as F'q/F'v), which is described in ‘Estimating PSII photochemical efficiency’, above. Interpreting F'q/F'v as a factor relating the maximum and operating efficiencies of PSII photochemistry makes no assumptions about the level of connectivity.
All of the remaining quenching coefficients relate to non-photochemical quenching processes: qE is the coefficient of ‘energy-dependent’ quenching, qT is the coefficient of non-photochemical quenching associated with state-transitional changes, qI is the coefficient of non-photochemical quenching associated with photoinhibition, and qN is the coefficient of total non-photochemical quenching. All of the non-photochemical quenching coefficients are usually calculated as the normalized variable fluorescence (Fv) at the point of measurement. A significant problem with this method is that any change in the value of the rate constant for PSII photochemistry (kP) will decrease the value of the supposedly non-photochemical quenching coefficient that is being calculated. By contrast, Fm/F'm – 1 (see ‘Quantifying down-regulation at PSII’ above) is unaffected by changes in kP and could, therefore, be considered a better parameter for quantifying non-photochemical quenching processes.
Imaging specific considerations
It is generally the case that images of Fo are far more difficult to generate than images of Fm, F', or F'm (see ‘Imaging of Fo’, below). An obvious consequence of this is that parameterized images requiring an Fo image (Fv/Fm, Fv/[FmFo], F'v/F'm, and F'q/F'v) are generally much more difficult to generate than ones that don’t have this requirement (F'q/F'm and Fm/F'm – 1). Of the two parameterized image types that do not require an Fo image, F'q/F'm can be used in isolation (see ‘Estimating photochemical efficiency’, above), while Fm/F'm – 1 images frequently require reference to an associated image of Fv/Fm (see ‘Quantifying down-regulation at PSII’, above).
Although generating images of Fo often presents the greatest technical challenge, movement of samples between images that were taken a long time apart (for example the Fm and F'm images required for construction of an image of Fm/F'm – 1), can also present problems. Where movement of samples between images does occur, it is often possible to use image processing tools to ‘nudge’ one image against another (Oxborough et al., 2000). Where this is not possible, the last resort is to derive a single value from each object (e.g. individual chloroplasts within a cell, or individual plantlets within a population) within each raw image and derive single parameter values for each of these object (Lawson et al., 2002).
Accurate determination of Fm and F'm requires the application of a super-saturating pulse of light that is typically several hundred ms in length at a photon irradiance of several thousand µmol m–2 s–1. When imaging large areas, this requirement for a high photon irradiance presents a significant technical challenge. With some types of application, such as the screening of large numbers of plants for certain physiological characteristics, it may seem feasible to use a subsaturating pulse to induce a ‘comparative’ Fm or F'm. However, it should be noted that this comparative value could be affected by a number of factors, including chlorophyll content and down-regulation at PSII. Consequently, observed differences for a particular fluorescence parameter, among samples or within different areas of a sample, may be more artefact than real.
Sources of error
The two best documented (and generally the most important) sources of error, when using Chl a fluorescence in a quantitative analysis of PSII function, are Chl a fluorescence from PSI (Genty et al., 1990; Pfündel, 1998) and the quenching of Chl a fluorescence by plastoquinone (Vernotte et al., 1979; Kramer et al., 1995).
It has been estimated that PSI fluorescence can represent as much as 30% of the fluorescence signal at Fo in C3 species and 40% in C4 species (Pfündel, 1998). Since the yield of PSI fluorescence is generally thought to be insensitive to changes in incident photon irradiance, the impact of PSI fluorescence on fluorescence parameters, although non-proportional, is at least progressive in nature. For example, with the decrease in F'q/F'm values that generally accompanies an increase in incident photon irradiance, the error due to PSI fluorescence will also increase.
It has long been appreciated that plastoquinone can quench a significant fraction of the fluorescence from PSII (Vernotte et al., 1979). Current evidence suggests that taking the plastoquinone pool from a fully oxidized state (100% in the form of plastoquinone) to a fully reduced state (100% in the form of plastoquinol) increases the yield of Chl a fluorescence by approximately 20% (Vernotte et al., 1979; Kramer et al., 1995). This has significant implications for the measurement of fluorescence parameters, since the light-addition method that is routinely used with Chl a fluorescence imaging systems and integrating fluorometers actually relies upon reduction of the plastoquinone pool by the application of a multiple-turnover, saturating pulse for measurement of Fm and F'm. The largest errors to arise through changes in the redox state of the plastoquinone pool are likely to occur during measurement of Fv/Fm, since this normally involves the largest change in the redox state of the plastoquinone pool, between measurement of Fo (when the plastoquinone pool is normally at its most oxidized) and Fm (when the plastoquinone pool is very highly reduced).