Experimental in vivo measurements of light emission in plants: a perspective dedicated to David Walker

Abstract

This review is dedicated to David Walker (1928–2012), a pioneer in the field of photosynthesis and chlorophyll fluorescence. We begin this review by presenting the history of light emission studies, from the ancient times. Light emission from plants is of several kinds: prompt fluorescence (PF), delayed fluorescence (DF), thermoluminescence, and phosphorescence. In this article, we focus on PF and DF. Chlorophyll a fluorescence measurements have been used for more than 80 years to study photosynthesis, particularly photosystem II (PSII) since 1961. This technique has become a regular trusted probe in agricultural and biological research. Many measured and calculated parameters are good biomarkers or indicators of plant tolerance to different abiotic and biotic stressors. This would never have been possible without the rapid development of new fluorometers. To date, most of these instruments are based mainly on two different operational principles for measuring variable chlorophyll a fluorescence: (1) a PF signal produced following a pulse-amplitude-modulated excitation and (2) a PF signal emitted during a strong continuous actinic excitation. In addition to fluorometers, other instruments have been developed to measure additional signals, such as DF, originating from PSII, and light-induced absorbance changes due to the photooxidation of P700, from PSI, measured as the absorption decrease (photobleaching) at about 705 nm, or increase at 820 nm. In this review, the technical and theoretical basis of newly developed instruments, allowing for simultaneous measurement of the PF and the DF as well as other parameters is discussed. Special emphasis has been given to a description of comparative measurements on PF and DF. However, DF has been discussed in greater details, since it is much less used and less known than PF, but has a great potential to provide useful qualitative new information on the back reactions of PSII electron transfer. A review concerning the history of fluorometers is also presented.

Appendix 1 (by V. Goltsev)

DF measurement

In view of the not-so-common use of DF, we describe here in some details this method and analysis of data. Two experimental approaches are used for the measurement and the analysis of the DF signal: (a) monitoring of relaxation of DF intensity in the dark (the so-called “dark DF decay”); and (b) recording of “DF induction curve” (IC) during the transition of dark-adapted samples to light-adapted state. In the first approach, the samples are pre-illuminated by short (single turnover) light pulse (flash) or by continuous light to form redox states of PSII that lead to emission of DF, for example S ₃Z⁺P680Q ^–_A Q _B. Here, S ₃ is one of the oxidized S-states of the oxygen evolving complex of PSII (Joliot and Kok 1975). After turning off the actinic light, the rate of DF quanta emission and the kinetics of DF decay are analyzed. Such an approach is usually applied for measuring the DF kinetic components decaying in ns (Christen et al. 2000), in μs (Grabolle and Dau 2005; Buchta et al. 2007), in ms (Goltsev et al. 1980) or in s (Rutherford et al. 1984; Rutherford and Inoue 1984). To evaluate long-lived light emission (>seconds or minutes), the samples are excited by continuous light (Hideg et al. 1991; Katsumata et al. 2008; Berden-Zrimec et al. 2008).

When a DF emission is monitored during illumination of a dark-adapted sample by continuous light, the DF induction curve is measured, as is done for recording the OJIP transients of PF. Both the measurements reflect changes in the photosynthesis machinery during dark-to-light adaptation. The same population of Chl-proteins of the PSII antennae complexes that emit PF emits DF quanta. The main difference is that the quantum yield of PF is 3-10 orders of magnitude higher than that of DF, and the DF quanta cannot be distinguished from those of PF during illumination. One effective experimental approach that allows one to distinguish between the two types of light emission is the separation of the two processes as follows: PF is recorded simultaneously with illumination and DF—after turning off the actinic light. For measurement of DF induction, it is necessary to use alternate light/dark cycles. During the light period, PF can be measured, a short time interval after the light is turned off (to avoid measuring PF), DF dark decay is measured (see Fig. 10).

Fig. 10

Reconstruction of kinetics of prompt (chlorophyll) fluorescence (PF) and delayed (chlorophyll) fluorescence (DF) signals measured simultaneously by a M-PEA instrument (Hansatech, UK; see Fig. 8) device during dark-to-light adaptation in bean leaves. The data acquisition for the two signals, PF in the light and DF in the dark, was every 0.01 ms in the 0–0.3 ms range, every 0.1 ms in the 0.3–3 ms range, every 1 ms in the 3–30 ms, and up to 30 s; after this time, data was acquired every 10 s. The black dots show values of PF signal, at specific points, during the O–J–I–P transition—i.e., on the PF induction curve (see the top vertical panel). Points marked as DD (“dark drops”) show the first PF values recorded after short dark periods during which DF was measured. Sheets perpendicular to the back plane show DF dark decays at different times of the O–J–I–P transition. Arrows show DF induction curves, recorded at different decay times. This figure was drawn by two of the coauthors (VG and RS) from their own original data

During each dark interval, the DF signal shows a polyphasic decrease. In most analog phosphoroscope-based DF-measuring devices, the quanta emitted during each dark interval are collected, integrated and presented as a value proportional to DF intensity at definite times. The time course of the measured signal at intermittent illumination of dark-adapted samples is presented as a DF induction curve. The digitalization of the measuring signal (Gaevsky and Morgun 1993; Zaharieva and Goltsev 2003) and the use of fast analog–digital converter devices (<50 μs) allows analysis of the kinetics of DF relaxation at each dark interval during induction (Fig. 11). To construct DF induction curves, a distinct dark time period is chosen within which the values of DF intensities are averaged and used as a single point of DF induction curve. Selecting points from different decay intervals, one can construct induction curves that show DF kinetic components with different lifetimes. In Fig. 11, five DF induction curves are shown that used the following time points in DF decay curves, i.e., after 20 μs, 90 μs, 0.9, 2.3 and 23 ms of the start of dark interval. Thus, the time course of different components (measured at different delay intervals) of the DF decay can be monitored during the dark-to-light transitions.

Fig. 11

Induction curves of delayed fluorescence (DF) recorded in 20–90 µs (left panel) and in 100–900 µs dark decay window (right panel) as a function of actinic light intensity. Primary leaves of decapitated bean plants were dark adapted for 1 h and then illuminated by red (625 nm) actinic light of different intensities from 500 to 4,000 µmol photons m⁻² s⁻¹. DF intensities are normalized to maximal values for each curve. The “I’s” (inflection peaks) refer to the induction maxima, and “D’s” (dips) to the minima. This figure was drawn by one of the authors (VG), using his own original data

Origin of DF induction phases

The DF induction curve reflects processes that occur in the photosynthetic machinery of plants during illumination after a period of dark adaptation. Usually, induction maxima are well pronounced after 5 to 15 min of dark adaptation. A stationary level of DF is reached after the 2-3 min of actinic light (Veselovskii and Veselova 1990; Radenovic et al. 1994).

The DF induction curve is extremely complex: it is multiphasic. Even 61 years after its discovery (Strehler and Arnold 1951), the reasons for the changes in the intensity of delayed light quanta emission during the induction transients, are not clear. DF intensity passes through several maxima and minima before reaching a stationary level. The main factors affecting the DF induction shape are: (1) The photosynthetic sample: plant species; (2) structural status of the sample (whole plant, isolated chloroplast suspension, membrane particles); (3) physiological state of the sample (chemical and physical treatments); (4) measurement details: e.g., dark adaptation duration; actinic light intensity; recording period (duration of time interval when DF is measured; dark interval before DF recording). Thus, measuring conditions determine which kinetic components of DF are being measured in an experiment (Zaharieva and Goltsev 2003).

There is no consensus nomenclature of the maxima that are observed in the DF induction curve, and there is no consensus about the number and interpretation of these maxima. We use here the nomenclature proposed by V. Goltsev and coworkers (Goltsev and Yordanov 1997; Goltsev et al. 1998, 2005, 2009; Zaharieva and Goltsev 2003) where the maxima (denoted by I) and minima (denoted by D) are numbered in a sequence according to their position in the DF induction curve (I₁, D₁, I₂, D₂).

The DF induction curve is easily divided into two main phases, a fast phase and a slow phase (Itoh et al. 1971; Itoh and Murata 1973; Malkin and Barber 1978) (Fig. 11). The fast phase that lasts for about 300 ms coincides with the OJIP transient of PF, and then there is the slow phase that occurs in the minute range, reaching a stationary level at the end. Using a mechanical phosphoroscope with fast signal digitalization (~50 μs) and electromechanical light “cutter” (opening time <1 ms), it is possible to resolve details in the structure of the fast phase. Thus, when DF is measured starting with 5.5 ms of illumination (the working cycle being 11–5.5 ms light and 5.5 ms dark and induction, see Goltsev et al. 2003), two maxima I₁ and I₂ (sometime with a minimum D₁ in between) are observed in the fast phase; after this DF drops to a minimum labeled as D₂ (Goltsev and Yordanov 1997; Goltsev et al. 1998, 2003). After a small step, labeled as I₃, the slow phase begins. During this phase, DF rises to a maximum I₄ and then, through several transient maxima (I₅ and I₆), DF intensity decreases to a stationary level S (Itoh and Murata 1973; Goltsev et al. 2003).

For DF that decays in 100-μs time interval, the time position of the first induction maximum I₁ as well as a ratio of I₁/I₂ are highly dependent on light intensity. At 4,000 μmol photons m⁻² s⁻¹ the I₁ maximum appears at about 3 ms of illumination and it is shifted up to 15 ms at lower light intensity (500 μmol photons m⁻² s⁻¹). A similar effect is observed in the induction curve of sub-millisecond DF component (Fig. 11, right panel).

When DF is compared with PF transient and with the kinetics of the signal of “reflection” of modulated light at 820 nm (called MR820—this photoinduced signal is caused by the appearance and disappearance of the oxidized form of P700 and of plastocyanin, see Schansker et al. 2003), the maximum I₁ coincides with the PF increase from the J-level (F _j) to the I level (F _I) and with decrease of MR820 reflecting P700 and plastocyanin oxidation (Schansker et al. 2003). The growth of DF intensity up to I₁ probably reflects the accumulation of S₃ZP680⁺ Q ^–_A and S₃Z⁺P680Q ^–_A states that have a relatively high yield of DF emission. The kinetics of DF decreases after the maximum I₁ to the minimum D₂ is similar to that of the PF rise from J to I and P phases, and it, possibly, is caused by the formation of “closed” PSII states S_iZP680 Q ^–_A Q ²⁻_B that are not able to do charge recombination in sub-ms and ms time interval and, thus, DF formation. Another process that probably has a part in the kinetics of the fast phase of the DF induction is photooxidation of P700 and of plastocyanin (PC) as a result of the activity of PSI due to the lack of electrons in the plastoquinone pool (Schansker et al. 2003). The accumulation of positive charges in the inner part of thylakoid membrane in the form of P700⁺ and PC⁺ may lead to the formation of a transmembrane potential (Satoh and Katoh 1983). Thus, the appearance of I₁, like the transition from F _o to F _j and F _I, can be related to two phenomena: (1) photochemical—accumulation of certain light-emitting states of the PSII RC, and (2) non-photochemical—increase in the DF due to the electrical gradient formed by PSI when P700 is oxidized (Pospisil and Dau 2002; Vredenberg et al. 2006).

The I₂ maximum (usually, at high light intensities; visible only as a shoulder) is probably related to the prolonged reopening of PSII RCs by the accelerated electron transfer from the reduced Q _B when the PQ pool is actively reoxidized by PSI before the full reduction of the PQ pool (I₂–D₂ transition in the DF induction curve that coincides with the I–P phase in PF transient and with the slow increase phase in the MR820 (Modulated light Reflectance at 820 nm) signal. The relative size of this maximum depends on the ratio between the flow of excitation trapping in the RCs of PSII and of the intersystem electron transfer. The share of I₂ increases under several conditions: at lower actinic light, with the decrease in the size of the PSII antenna; and with increase in temperatures (Zaharieva et al. 2001).

After about 300–500 ms of illumination, the plastoquinone pool is reduced and most of the Q _A is in its reduced state, Q ^–,_A Chl fluorescence is maximal (P step) and MR 820 signal reaches its maximal level. At this moment of induction (phase D₂), DF is emitted from RCs in “closed” state Z⁺P680Q ^–_A Q ²⁻_B (Gaevsky and Morgun 1993; Zaharieva and Goltsev 2003; Goltsev et al. 2005). This is “deactivation” type of light emission (see the main text) and is a result of charge recombination in Z⁺P680Q ^–_A Q ²⁻_B or S_iZP680Q ^–_A Q ²⁻_B PSII states. During this induction phase, the amplitude of the sub-ms DF components decreases, and the lifetime of the ms component increases (Zaharieva and Goltsev 2003). In the presence of an artificial electron acceptor (potassium ferricyanide) and uncouplers of phosphorylation, this increase in the lifetime of DF is insignificant and no I₂–D₂ is observed. This indicates that the I₁–I₂–D₂ phase correlates with the processes of reduction of the PQ pool, and the JIP phase of PF transient (Schansker et al. 2003).

The peak I₃ was first discovered with a phosphoroscope-based DF instrument with low actinic light (~1,200 μmol photons m⁻² s⁻¹ (Goltsev et al. 2003) but it is not visible if DF is recorded at high actinic light (4,000 μmol photons m⁻² s⁻¹); it is visible as a small shoulder after the D₂ phase in the DF induction curve with exposure to 1,000 μmol photons m⁻² s⁻¹ light intensity. The source of DF emission of this phase is weakly luminescent “closed” PSII states.

The increase of DF to the next maximum, labeled as I₄, is well pronounced at relatively low excitation light. The DF growth during the D₂–I₄ phase coincides with a slight decrease in the PF intensity and reduction of MR signals caused by oxidation of P700 (Goltsev et al. 2005). The accumulation of P700⁺ suggests that at this time the light-induced activation of the ferredoxin: NADP⁺-oxidoreductase takes place (Harbinson and Hedley 1993; Schansker et al. 2003), i.e., the linear electron transport is activated, and the transmembrane proton gradient starts to accumulate. The increase of the DF intensity in the slow phase (toward the I₄ maximum) is associated with the formation of a proton gradient (Wraight and Crofts 1971; Evans and Crofts 1973) that increases the rate constant of radiative recombination in the PSII RCs.