Excitation energy flow at 77 K in the photosynthetic apparatus of overwintering evergreens


Correspondence: Dr A. M. Gilmore. Fax: +612 6125 5095; e-mail: mball@rsbs.anu.edu.au


The flow of excitation energy from the antennae to photosynthetic reaction centre complexes at 77 K was studied in leaves of two evergreen species, namely, snow gum (Eucalyptus pauciflora Sieb. ex Spreng.) and a hemiparasitic mistletoe (Amyema miquelii, Lehm. ex Miq.). The leaves that were naturally acclimated to winter conditions of freezing temperatures and high irradiance displayed the recently discovered cold-hard-band or CHB feature of the chlorophyll a fluorescence spectra (Gilmore & Ball, Proc. Nat. Acad. Sci. USA 97:11098–11101, 2000). A streak-camera-spectrograph was used and the double convolution integral method for global analysis was applied to simultaneously acquire and simulate, respectively, the time- and wavelength-dependence of all major chlorophyll a components (Gilmore et al. Phil. Trans. Roy. Soc. B-London 355:1371–1384, 2000). The CHB coincided with changed amplitudes and decreased excited state lifetimes for the main F685 nm and F695 nm emission bands from the photosystem II (PSII) core-inner-antenna. The CHB dissipates energy as heat separate from PSII while also reducing the PSII quantum yield by competing for both photon absorption and antenna excitation. The CHB did not correlate with changes in the decay kinetics of the PSI antenna F740 nm band. The spectral-kinetic features of the altered energy flow were similar in the unrelated evergreen species. These results are consistent with a functional association between the CHB, PSII energy dissipation and protective storage of chlorophyll in overwintering evergreens.


During winter, the leaves of temperate evergreen species undergo changes in photosynthetic characteristics including light-dependent reductions in both the concentration of chlorophyll (Chl) and the photochemical efficiency of photosystem II (PSII). The sustained reduction in PSII efficiency is associated with increased rate constants of photoprotective thermal energy dissipation (Groom, Baker & Long 1991; Verhoeven, Adams & Demmig-Adams 1999; Gilmore & Ball 2000). The photoprotective energy dissipation correlates with increased levels of the xanthophyll cycle (Yamamoto 1979) components: violaxanthin (V), antheraxanthin (A) and zeaxanthin (Z). However, in contrast to the well-characterized energy dissipation mechanism involving the xanthophyll-cycle, PsbS-protein and trans-thylakoid membrane ΔpH (Demmig-Adams, Gilmore & Adams 1996; Gilmore 1997; Li et al. 2000; Müller, Li & Niyogi 2001) a substantial portion of energy dissipation during winter is maintained without a ΔpH (Verhoeven et al. 1999; Gilmore & Ball 2000; Gilmore, Itoh & Govindjee 2000; Ebbert et al. 2001). The latter fact indicates that semi-permanent, stable changes in the pigment-proteins of PSII are involved, possibly including amino acid or pigment oxidation, protein phosphorylation, protein conformational changes and/or macro-reorganization (Ottander, Campbell & Öquist 1995; Xu et al. 1999; Ebbert et al. 2001).

Gilmore & Ball (2000) first showed that winter-acclimated leaves of snow gum, Eucalyptus pauciflora Sieb. ex Spreng, display large changes in the steady-state spectral and kinetic patterns of the Chl a fluorescence that correlate with the PSII efficiency changes. The changes influence the emission wavelength regions associated with both PSII and photosystem I (PSI) when measured at either 77 K or room temperature (Matsubara, Gilmore & Osmond 2001; Matsubara et al. 2002). Importantly, the changes are most clearly resolved at 77 K where they are a common feature of over-wintering temperate evergreens, as shown by their occurrence in the hemiparasitic mistletoe species, Amyema miquelii (Lehm. Ex Miq.), that is evolutionarily divergent from its host eucalyptus trees (Matsubara et al. 2002).

The altered leaf spectra at 77 K are dominated by the cold-hard-band (CHB), a broad spectral feature centred around 715 nm (Gilmore & Ball 2000). Under conditions of maximal expression the CHB may exhibit amplitudes larger than the main F740 band reportedly emitted from the PSI antennae complexes Lhca1 and Lhca4 (Mullet, Burke & Arntzen 1980; Melkozernov et al. 1998). The CHB is believed to be an important indicator of winter stress acclimation because it exhibits steady-state amplitudes that reciprocate, and overlap with bands reportedly emitted from the PSII core-inner-antennae complexes, namely, F685 from CP43 and F695 from CP47 (Nakatani et al. 1984). It is undetermined how or if the CHB is associated with the emission from peripheral light-harvesting complexes that have been reported to be represented by a relatively low-amplitude 77 K component centred around 680 nm known as F680 (Rijgersberg et al. 1979). The weak F680 is difficult to resolve at steady state but more readily seen in the early portions of time-resolved emission spectra at temperatures of 77 K and below (Mimuro et al. 1987b; Lin & Knox 1988; Govindjee 1995).

In this study the samples were quick-frozen to 77 K, which arrests all biochemical reactions including photosynthetic electron transport from PSII, degradative enzymatic reactions as well as changes in pigment composition associated with both light-harvesting and energy dissipation. The 77 K conditions also alter the vibrational behaviour of electronic excited states of the Chl-complexes such that the separate antenna and reaction centre fluorescence band widths and decay kinetics also become accentuated (Mimuro et al. 1987b; Lin & Knox 1988). The thermal properties facilitate a structural interpretation of the observable energy transfer pathways and rate constants. Most important to this study however, is that the conditions cryo-stabilize the thermal dissipation processes of interest allowing resolution and interpretation, respectively, of their structural localization and functional influence.

In addition to the above considerations, there are several advantages of the simultaneous time- and wavelength-resolution and data simulation, respectively, afforded by the streak camera-spectrograph and the double-convolution-integral (DCI) simulation method (Gilmore et al. 2000). Primarily, the methods facilitate resolution of energy transfer from one Chl complex to another by revealing the reciprocity in the band amplitudes as a function of time following the exciting laser pulse (Mimuro et al. 1987b; Lin & Knox 1988, 1991). In addition, thermal dissipation can  be resolved as changes in the rate constant(s) of the band com-ponents (Govindjee 1995). Aside from the signal-related aspects of the methodology, there are signal-distortion factors associated with the leaf's optical properties. For example fluorescence re-absorption could be distinguished by comparing the kinetic profile of a complex undergoing re-absorption with the emission kinetic profile from the complex emitting the exciting light. This is because an excitation pulse intensity is convolved as a function of time with the rise–decay profile of the absorbing (and fluorescing) component according to the convolution decay law (Lackowicz 1999). Therefore, re-absorption of a fluoresced photon, like a photon from the exciting laser pulse, will not alter the decay rate constants of the emitting or re-absorbing component but will only alter the re-absorber's rise kinetics by convolution. Additionally, light-scattering is proportional to the energy of the photon and therefore scattering always reduces the relative intensity of emission from higher (bluer), versus lower (redder), energy components that reach the detector (Lackowicz 1999). Light scattering, as commonly observed in suspensions such as subchloroplast membrane particles is proportional to the Chl concentration by virtue of the coincidence of the concentration of Chl-containing particles that act as the scattering agents. Scattering is the most uncertain parameter mainly because all leaf tissue samples exhibit unique structural features and scattering properties. Nevertheless, although scattering and re-absorption alter the amplitudes, these processes do not prevent the use of leaf tissue for quantitative modelling of changes in the PSII efficiency or kinetic rate constants at a given wavelength of light energy or qualitative resolution of major spectral features such as those associated with well-characterized antenna mutants (Gilmore & Ball 2000; Gilmore et al. 2000; Matsubara et al. 2002).

Given the above methodological considerations and novel advantages provided thereby, this study was aimed primarily at first observing and then understanding the unique spectral-kinetic behaviour of the CHB at 77 K in relation to the prevailing physiological light and temperature conditions. We investigated the influence of the CHB on the pathways for energy flow and thermal de-excitation in relation to all the other major PSII and PSI emission components. This included asking how or if the new CHB-dependent components relate to emission coming directly from PSII; that is, does the CHB accept energy from PSII then dissipate the energy as heat, does it compete for antenna excitation with PSII or does it represent a completely detached complex? The major hypothesis tested was whether and how the CHB is functionally associated with the storage and protection of chlorophyll in overwintering evergreens.


Plant material and 77 K sample preparation

Snow gum (Eucalyptus pauciflora Sieb. ex Spreng) seedlings were grown under sunny field conditions as previously described (Ball et al. 1997). Sun- and shade-acclimated mistletoe (Amyema miquelii Lehm. ex Miq.) leaves were collected from the field in winter (June 2001). Leaf samples were collected based on having either low or high PSII efficiencies as measured with a Hansatech plant efficiency analyser (Hansatech Instruments Ltd, Kings Lynn, UK). Leaf pieces were cut to fit the adaxial surface flush into the polished surface of a 1 cm by 1 cm cuvette. Snow gum samples were then rapidly frozen in liquid nitrogen and stored at −80 °C. The mistletoe samples were vacuum-infiltrated for 30 min with 0.3 m glucose, 50 mm HEPES (pH 7.6), and a saturating concentration of 3-(3,4-dichlorophenyl)-1,1′-dimethylurea (DCMU). The saturating concentration for PSII trap closure for mistletoe leaves was used because they were relatively impervious to concentrations of DCMU < 100 µm within the 30 min time frame. Leaf samples were quick-frozen then scanned to determine the steady-state 77 K fluorescence emission spectra (Gilmore & Ball 2000) and samples with distinct and representative 77 K spectral features were packed in dry ice and shipped to Nagoya. Upon arrival, samples were stored at −80 °C until transfer under liquid N2 to the streak-camera spectrograph dewar for analysis as described below.

The 77 K time-resolved chlorophyll fluorescence emission spectral data acquisition

Time-resolved fluorescence emission spectra were collected with a streak-camera spectrograph (Hamamatsu 4334; Hamamatsu Photonics Inc., Hamamatsu City, Japan). The sample compartment consisted of an evacuated, electronically thermostat-controlled liquid nitrogen dewar (Oxford Instruments, Inc., Abingdon, Oxfordshire, UK) fitted to accept the leaf samples in a 1 cm by 1 cm cuvette in a front surface configuration. Excitation was provided by a pulsed laser diode (635 nm, 50 ps pulse width) with the excitation filtered before the sample by an interference filter (635 nm, 10 nm band width, 90% transmittance). The total spectrum of the sample fluorescence emission was collected with the spectrograph (Chromex 2501-S; Hamamatsu Photonics Inc.) centre wavelength at 710 nm (grating = 50 grooves mm−1, blaze = 600 nm and slit = 40 µm) to exclude wavelengths below 650 nm including the laser excitation. Sample data were collected in photon counting mode with a 5 ns time window until a total of 500 000 excitation shots were recorded at a rate of approximately 32.4 Hz. The image data was saved to hard disk in intervals of 100 000 shots (approximately once per hour). There was no indication of sample or signal instability, excessive trigger-jitter (i.e. beyond manufacturer's specification of 20 ps) or diode laser intensity changes during data acquisition. The stability of the system was confirmed by the goodness of fit of the simulation obtained by convolution with the scattered excitation laser pulse profile which was rapidly collected (10 000 shots, ∼ 5 min exposure) with the spectrograph centre wavelength at 650 nm after placing a 1% neutral density filter between the sample compartment and the spectrograph. The 640 (wavelength axis) by 480 (kinetic axis) pixel charge-coupled device image was exported from the Hamamatsu software in ASCII format and imported into Excel 97. The pixels were binned and integrated at a rate of 5 pixels per time channel and 3 pixels per wavelength channel and subjected to a further 3-channel adjacent averaging of the spectral profiles. The chosen spectral region of interest was restricted to 100 time channels (λ = 1.25 nm ch−1) and 160 wavelength channels (t = 31.25 ps ch−1) for global analysis using the double convolution integral method as described below.

Global analysis of the time-resolved emission spectra using the double convolution integral method

The double convolution integral (DCI) method was employed as described earlier (Gilmore et al. 2000). The intensity profiles of each Gaussian spectral band i . . . n were first determined as a function of wavelength to yield I(λ)i. The decay function defining the intensity of band i . . . n as a function of time, F(t)I = ∫α(τ)i exp(–t/τ) dτ, was calculated by summing the integral contributions of up to four Gaussian distribution modes for the pre-exponential amplitude factor as a function of the fluorescence lifetime represented as α(τ)i. F(t)i was then convolved with the laser pulse excitation profile as a function of time, L(t)i, to yield I(t)i = F(t)I ⊗ L(t)i. The time- and wavelength-dependent functions were then used to compute the double convolution integral I(t,λ)i = I(t)I ⊗ I(λ)i. Images were typically simulated by summing 6–8 such Gaussian spectral-kinetic band components represented as I(t,λ)i. The free model parameters included the widths and modes of spectral distribution functions, the width, mode and amplitudes of the kinetic pre-exponential amplitude functions, a constant offset applied to all time-wavelength channel co-ordinates and a time-axis shift between the laser pulse profile and the data. The spectral-kinetic model was robustly fit (Draper & Smith 1998) using the L1 method by minimizing the sum of absolute deviations; the absolute deviation for a given time-wavelength channel coordinate was Li = |Di − Mi|, where Di is the datum and Mi is the model prediction of the datum. Standard errors on fluorescence lifetime mode centre and width parameters longer than the instrument response function (50 ps) were estimated from repeated simulations to be less than 10% whereas lifetime modes and widths below this value were estimated to have ambiguity factors approaching 2.

The global analysis program, was written for use in Microsoft Windows NT4.0 (32 bit) and Excel 97, with Visual Basic for Applications 97 and utilizes a large scale general reduced gradient (LSGRG) minimization engine developed by Frontline Systems Inc. (Incline Village, NV, USA). The LSGRG engine is capable of handling 4000 parameters and 4000 constraints and utilizes a sparse matrix representation. The LSGRG engine is capable of solving, in Excel 97 with a Pentium III computer with 500 Mb RAM and an 800 MHz processor, all 26 test problems prescribed by the National Institute of Standards and Technology (NIST) website for non-linear regression (http:www.itl.nist.govdiv898strdnlsnlsinfo.shtml) with 10 digit precision for the sum of squares parameter.

Pigment analyses

Pigments were extracted and assayed using high-performance liquid chromatography as described (Gilmore & Yamamoto 1991). Absorbance at 440 nm was measured with a Waters 490 (Waters Associates Inc., Milford, MA, USA) variable wavelength detector. Standard errors were less than 10% for both xanthophylls and chlorophylls per sample injection.


Time-resolved emission spectral comparison of winter-acclimated leaves with normal and inhibited PSII efficiency

During winter, sunlit leaves of snow gum exhibit a range of PSII photochemical efficiencies, dependent on the combined effects of minimum temperature (Ball et al. 1997) and exposure to direct sunlight (Blennow et al. 1998). Exemplary winter-acclimated snow gum leaves with contrasting PSII photochemical efficiencies, here measured as Fv/Fm, were selected. The leaves were grown under similar fully exposed field locations where the variation in Fv/Fm in leaves intercepting similar high irradiance strongly associated with small-scale variation in minimum freezing temperatures (Ball et al. 1997). The high- and low-efficiency leaves had Fv/Fm values of 0.710 (Fig. 1a & b) and 0.390 (Fig. 1c & d), respectively. The high-efficiency leaf had more chlorophyll (342 versus 137 µmol m−2 Chl a + b) and a lower xanthophyll cycle de-epoxidation (DPS = [Z + A]/[V + Z + A]) ratio (0.16 versus 0.34) than the low-efficiency leaf.

Figure 1.

Spectral-kinetic contour plots and representative time-dependent wavelength profiles of liquid nitrogen temperature (77 K) chlorophyll a fluorescence emission of winter-acclimated evergreen leaves. Data represent both Eucalyptus pauciflora Sieb. ex Spreng (a–d) and Amyema miquelii, Lehm. ex Miq. (e–h) leaves with respective normal (a, b and e, f) and low PSII efficiency values (c, d and g, h). The coloured contours (left panels) represent the DCI model simulations of the four respective images; the symbols and solid lines (behind the symbols in the right panels) represent corresponding data and DCI simulations, respectively. The colour coded and numbered horizontal lines in the left panels correspond to the time-dependent wavelength profiles in the left panels. The contour line intervals in the right panels represent 5% of the maximum amplitude value that was normalized to unity. The wavelength profiles in the right panels represent the integral of four time channels (125 ps) centred at the times noted in the corresponding left panels. The inset in panel (b) represents the wavelength horizontal axis from 660 to 705 nm. The sums of the least absolute deviation statistic were L1 = 82.53 (a, b); L1 = 113.83 (c, d); L1 = 92.48 (e, f); and L1 = 108.04 (g, h).

In the high-efficiency snow gum leaf (Fig. 1a & b), energy flow in the photosynthetic apparatus at 77 K proceeded in a thermodynamically predictable pattern consistent with previous studies on higher plants (Mimuro et al. 1987a, b). The first spectral features to arise following the laser pulse were bands below 700 nm. In addition to the rapidly rising and decaying F680 band associated with the LHCII (Mimuro et al. 1987a, b; Lin & Knox 1988, 1991), snow gums exhibited the F685 and F695 bands from the respective CP43 and CP47 core-inner-antenna complexes of PSII (Nakatani et al. 1984). Mimuro et al. (1987b) showed that PSI associated bands around 690–695 nm exhibit rapid antenna rise–decay kinetic functions similar to the F680 band of LHCII. As indicated later we observed evidence for overlapping PSII and PSI antennae components in the time-resolved emission spectral region < 700 nm especially for the band resolved and referred to as F695 nm. The kinetic development of the PSII spectral region from 675 to 700 nm is shown in greater detail in the inset (Fig. 1b). Comparison of the transition from the (1) blue to (2) light green to (3) yellow and (4) red colour profiles shows that the 680 nm region decayed most rapidly and revealed, in the approximate 1 ns time span, that excitation flows to the 685 and 695 nm bands of PSII. Finally, the 740 nm region from the PSI antenna complexes (Mullet et al. 1980; Mimuro et al. 1987a, b; Melkozernov et al. 1998) exhibited both the latest rise and longest decay time.

The CHB complex markedly altered the spectral kinetic features of chlorophyll fluorescence at 77 K as shown by comparison of the data for the high- (Fig. 1a & b) and low- (Fig. 1c & d) efficiency snow gum leaves. In general, the high-efficiency leaf showed significantly slower decay times in the PSII-associated spectral regions, enhanced cleavage between the PSII and PSI regions and a greater predominating amplitude of the slowly decaying F740 band region. It is possible that a small amplitude of the CHB or its precursor structural elements is present in the high-efficiency leaf.

Comparative studies were undertaken with winter-acclimated mistletoe leaves that, prior to freezing, exhibited contrasting PSII photochemical efficiencies with Fv/Fm values of 0.717 (Fig. 1e & f) and 0.365 (Fig. 1g & h). As in the snow gum, the high-efficiency mistletoe leaf had more Chl (530 versus 291 µmol m−2 Chl a + b) and a lower xanthophyll cycle de-epoxidation ratio (0.29 versus 0.56) than the low-efficiency leaf. However, unlike the snow gum leaves, the high- and low-efficiency mistletoe leaves had grown, respectively, in shaded under-storey and sun-exposed light environments.

A similar pattern of energy flow was observed in the mistletoes as in the snow gums. The normal mistletoe leaf (Fig. 1e & f) exhibited the characteristic PSII energy flow pattern from the 675 to the 685 and 695 nm regions and a large amplitude in the 740 nm band region. In the low-efficiency leaf, the CHB spectral feature emerged most clearly after the main PSII bands had decayed, that is proceeding from the blue (1) to the red (4) traces (Fig. 1g & h). In comparison with the low-efficiency snow gum (Fig. 1c & d), the low-efficiency mistletoe leaf (Fig. 1g & h) exhibited a larger relative amplitude of the CHB.

Normalized time-resolved emission spectral comparison of winter-acclimated mistletoe leaves with normal and inhibited PSII efficiency

The altered energy flow in the high versus low PSII efficiency mistletoe leaves is illustrated in a different format in Fig. 2 which shows the spectra normalized to the peak intensity wavelength for each 31.25 ps time channel in a descending format. The time-series for the high-efficiency leaf in Fig. 2a began at time channel 1 (black silhouette) when the peak F680 band was initially resolved from the background noise. The thick white line (676 nm) demarcates the peak of F680 and helps to illustrate the rapid shift of the peak within 120–150 ps to 685 nm. The subsequent shifts in the spectral region below 700 nm are much slower. Nevertheless, it is clear that at the last shown (white profile) time channel 40 (1220 ps) the peak emission had further shifted to 688 nm. The spectral region> 700 nm was predominated by the PSI F740 nm band that rises and decays more slowly than the PSII core and antenna bands. In the low-efficiency leaf the time series (Fig. 2b) again the F680 band peaked first and the changes in the spectral region < 700 nm changed much more rapidly than in Fig. 2a. It is crucial to point out that the peak emission was> 690 nm within 60–90 ps. There was then a peak resolved at 695 nm that slowly diminishes to yield a broad shoulder in the decay region between 680 and 720 nm. The relative intensity of the 715 nm region compared to the 740 nm region gradually decreased indicating a more rapid decay for all bands centred < 740 nm. At the final time frame shown (white profile) the broad spectral shape of the CHB was clearly revealed along with the 740 nm band. Comparison between the final shown time frames of the high- and low-efficiency leaves indicates the CHB amplitude replaced and strongly overlapped the main PSII core bands in the energy transfer sequence.

Figure 2.

Time-resolved spectral profiles from winter-acclimated Amyema miquelii, Lehm. ex Miq. leaves with normal (left) and low PSII efficiency (right). The time series is presented in 31.25 ps increments in a descending manner starting with the initial signal resolution (black silhouette) and ending with the white time profile. The incremental scaling on the left axis refers to the incremental (− 0.1 × time channel number) subtracted from the normalized spectra in the descending series. The white line demarcates the first time-resolved peak of the antenna emission components.

Simulation of the time-resolved emission spectra in winter-acclimated mistletoe leaves with the double-convolution integral method

Figure 3 profiles the data and DCI simulation (Gilmore et al. 2000) representing the integral spectral (left) and kinetic (right) components of the emission bands, the excitation profile and the model residuals (subplots) for the high- (upper) and low- (lower) efficiency mistletoe images from Figs 1 and 2. The DCI analysis illustrates only the five main spectral band components common to both the low (upper) and high (lower) efficiency mistletoe leaves, plus the additional CHB component (yellow) in the leaf with low PSII efficiency. The low-efficiency leaf exhibited more rapid decay kinetics for the F685 (red) and F695 (green) bands than the normal leaf. In the low-efficiency leaf spectrum, the CHB (yellow) spectral-kinetic component exhibited intermediate decay kinetics compared to the F740 (white) and F685 bands, respectively. In the low-efficiency mistletoe leaf the F680 band (blue) exhibited both a very rapid kinetic phase and a minor fraction of a much slower (approximately 4000 ps) phase. The F680 band in the low-efficiency mistletoe also exhibited elimination of the long> 100 ps component of the antenna. These major changes in F680 were consistent with the large loss of Chl and high ratio of the CHB : F740 often observed in the low-efficiency mistletoe leaves (Matsubara et al. 2002).

Figure 3.

Time-integrated wavelength profiles (left) and wavelength-integrated time profiles (right) of 77 K chlorophyll fluorescence emission from winter-acclimated Amyema miquelii, Lehm. ex Miq. leaves with normal (upper) and low PSII efficiency (lower). The coloured Gaussian spectral band components in panels (a) and (c) coordinate with their kinetic profiles in panels (b) and (d). The raw fluorescence data in both the left and right main panels are represented by diamond symbols and overlay the model denoted by a thick black line. Each coloured line in the left and right main panels corresponds to the total time and or wavelength integral of the acquired image. Hence the kinetic profiles represent the integrals of the spectral bands accounting for both their widths and amplitudes not simply their centres of gravity. The time-integrated wavelength profiles represent a definite integral (0–5000 ps) and it was clear the decays for the bands with longer lifetimes were not completed in the observed time frame. The thin dotted lines in the main panels (b) and (d) represent the laser excitation profile. The subplots represent the definite time- or wavelength-dependent integrals, respectively, of the residual errors weighted by the square root of the model prediction assuming Poissonian statistical noise (Bevington 1969; Gilmore et al. 2000).

Comparing the low- (left) and high- (right) efficiency mistletoe leaves in Figs 4a and b, respectively, shows that the F685 (red) and F695 (green) bands decay more rapidly in the low-efficiency mistletoe leaf and that the total spectrum is dominated by the broad CHB (yellow). When comparing the mistletoe (upper) and snow gum leaves (lower) the F685 is narrower and red shifted (dashed line) about 3 nm in the mistletoes whereas the F695 is broader but centred similarly. The mistletoe leaf also exhibited a larger decrease in the F695 band lifetime than F685 and the changes in the antenna F680 component were also visibly different between the two species. The F680 component did not change nearly as much, if at all, in the snow gum samples. The F740 bands were also slightly red shifted by 2–3 nm in the mistletoes (thin dotted line). Spectral differences in PSI and PSII components in both species were corroborated by steady state spectral analyses at 77 K (data not shown, Gilmore & Ball 2000; Matsubara et al. 2002).

Figure 4.

Spectral-kinetic contour plots of the major DCI model spectral-kinetic components of the 77 K chlorophyll fluorescence emission from Amyema miquelii, Lehm. ex Miq. (upper) and Eucalyptus pauciflora Sieb. ex Spreng (lower) with low (left) and normal (right) PSII photochemical efficiencies. The colour coding is the same as that used in Fig. 3. The streak-camera spectrograph image data correspond to the same samples as in Fig. 1. The amplitude axis scale for the contours for each band-decay ranged from 0 to 1 with an interval of 0.025.

Pathways of energy flow in the two species were explored further by comparing kinetic profiles for the major PSII, PSI and CHB emission bands and the laser excitation pulse after normalizing each kinetic profile to its maximum value. The normalized view in Fig. 5 facilitates visualization of the reciprocity of the band amplitudes and the sequence of peak excitations associated with energy transfer among the bands. Both species exhibited strong amplitudes with the longest lifetimes and rise-times for PSI (F740) indicating it is the terminal acceptor which is consistent with it possessing the lowest absorbance energy of all the major bands studied here. The decay of the F680 band coincided with the delayed rise, after the laser excitation, in the F685, CHB and F740 indicating energy transfer initiated from the F680 band. In both species the CHB rise time coincided closely with the F685 band whereas the CHB exhibited faster initial decay kinetics in the mistletoe than in the eucalypt. The F695 band exhibited a rapid rise in all samples from both species. As mentioned above it is clear in the sun mistletoe (Fig. 4a) that a small amplitude (∼ 5%) of a 4 ns lifetime component exists because the fluorescence of F680 did not decay completely to the ground state in the 1 ns time window shown here or the 4 ns window in Fig. 3.

Figure 5.

Normalized kinetic-profiles of the excitation pulse and Gaussian spectral-kinetic band amplitude components corresponding to the model fit to the data for the Amyema miquelii, Lehm. ex Miq. leaves (left) and the Eucalyptus pauciflora Sieb. ex Spreng leaves (right) with normal (upper) and low (lower) Fv/Fm values. The colour coding of band-profiles correspond to those in Fig. 3. The data for each profile were normalized to unity based on the maximum amplitude value of the profile.

Comparison of the kinetic parameters describing the time-resolved emission spectral components in winter-acclimated leaves with normal and inhibited PSII efficiency

In Fig. 6 we compare the distributions of pre-exponential amplitude factors of the fluorescence lifetimes (α(τ)) for each major spectral band to illustrate three things: (1) the changes in the fluorescence lifetime distribution modes and amplitudes as an indicator of changes in energy dissipation; (2) the correspondence between components with rapid positive amplitude modes (possible energy donor modes) and components with corresponding negative amplitudes (possible energy acceptor modes); and (3) evidence for fluorescence self- or re-absorption. Figure 6a and b showed the F680 band exhibited two positive amplitudes both < 500 ps. The major positive mode, associated with the antenna donation function of F680 was less than 10 ps in all samples. The longer mode of F680 was strongly diminished in the low-efficiency mistletoe (Fig. 6a) and the longer mode in the snow gum (Fig. 6b) was centred at a shorter lifetime than in the mistletoe (Fig. 6a). The relative amplitudes of both F680 modes increased in the low-efficiency snow gum, although, the lifetime centres and modes did not; the average lifetimes of F680 were virtually identical in the snow gum samples at 125 ps each. The relative increase in both F680 modes of the snow gums is explained by the relative decrease in the other spectral bands noted earlier.

Figure 6.

Profiles of the distributions of the pre-exponential amplitude factor as a function of the fluorescence lifetime of the major photosystem II and I emission components from Eucalyptus pauciflora Sieb. ex Spreng (right) and Amyema miquelii, Lehm. ex Miq. (left). The kinetic parameters correspond to the model simulations presented in Figs 1, 4 and 5 with normal (grey solid line) or low (dotted line) PSII efficiencies before freezing. The colour coding convention described in the legends below the locants in the left panels corresponds to the band components described in Figs 3–5 for both the corresponding left and right panels.

The F685 distributions in Fig. 6c and d exhibited bimodal decay kinetics with corresponding negative acceptor modes corresponding closely with the F680 band's rapid decay mode. Both the F685 mode centres and amplitudes exhibited changes in the low-efficiency mistletoe and snow gum to result in a significant lowering of the average fluorescence lifetimes. As mentioned above the F695 kinetic distributions (Fig. 6e & f) exhibited two or three positive decay modes in the high-efficiency leaves. The shortest mode's lifetime centre did not change when comparing the high- and low-efficiency leaves in either species which was consistent with the suspected antenna functionality of these modes (Lin & Knox 1988). The low-efficiency leaves of both species exhibited changes in the lifetime modes and amplitudes of F695, like F685, to yield shorter fluorescence lifetime averages.

Figure 6g and h show data only for the low-efficiency leaves exhibiting resolvable amplitudes of the CHB. The data indicate that the CHB exhibits a negative acceptor mode (<10 ps) in both species consistent with it receiving energy from the aforementioned antenna components < 700 nm. It is also clear, however, from the predominating positive decay amplitudes, compared to the rapid acceptor modes, that the CHB complex also directly absorbs a considerable fraction of light in proportion to its physical absorbance cross-section. Importantly, Fig. 6g & h shows the CHB exhibited two positive decay modes in both species. The longest CHB mode was similar and between 1500 and 1800 ps. The shorter mode was much faster in the mistletoe (<100 ps) than in the snow gum (∼ 590 ps) and coincided with the larger CHB : F740 and CHB : F685 amplitude ratios noted earlier for the mistletoe.

Figures 6i and j show that the mistletoe (left) and snow gum (right) samples exhibited major positive F740 decay modes that were not significantly attenuated, remaining close to 2800 ps in the low (solid grey) compared to the high Fv/Fm samples (dotted). The F740 bands all exhibited significant negative amplitude acceptor modes with rapid lifetimes (<10 ps) that closely corresponded with the positive decay or donor modes in the lower panels (Fig. 6a & b) for the F680 band (and the rapid decay mode of the F695 band). The only remarkable change in the F740 bands that correlated with the PSII efficiency was observed for the snow gums in which the low-efficiency leaf exhibited an increased negative amplitude in the hundreds of picoseconds range (≈ 350 ps). The low-efficiency mistletoe also exhibited significant negative F740 amplitudes> 100 ps. The stronger negative or rising amplitudes in the F740 band correlate with the larger relative F740 : F685 ratio in both snow gum samples compared with the mistletoes.

Figure 7 shows that the average lifetime of the PSII F685 and F695 bands averaged together were around 1500–1800 ps in the high-efficiency leaves and were attenuated to around 700–800 ps in the low-efficiency leaves. The CHB average lifetimes were clearly resolved in the low-efficiency leaves at around 1500–1600 ps. Thus both PSII and CHB were considerably faster in average decay time than the PSI F740 band which remained unaltered around 2800–2900 ps.

Figure 7.

The average fluorescence lifetimes for all the major emission bands of the PSII core (F685 and F695), the cold-hard-band (CHB) and photosystem I antennae (F740). Data represent both the Eucalyptus pauciflora Sieb. ex Spreng (SG) and Amyema miquelii, Lehm. ex Miq. (MT) leaves with respective high (High) and low (Low) PSII efficiencies presented in Fig. 6. The fluorescence lifetime averages were calculated according to (Gilmore & Ball 2000; Gilmore et al. 2000) by integrating the lifetime-weighted fractional intensity distribution. The PSII F685 and F695 bands were averaged together in this representation for simplicity.


PSII excited state behaviour and energy dissipation in winter-acclimated leaves at 77 K

The 77 K PSII excitation energy flow in leaves, with uninhibited PSII photochemical efficiency, initiates with the rapid rise and decay of antenna bands centred <700 nm. This data is consistent with studies of isolated PSII particles and chloroplast thylakoid membranes (Mimuro et al. 1987b; Lin & Knox 1988). Notably, the normal mistletoe sample and both snow gums also exhibited F680 band modes>100 ps possibly consistent with loosely detached antenna components. In the normal mistletoe the longer F680 mode and enhanced amplitude was interpreted to possibly indicate antenna size over-development during shade acclimation (Anderson 1986). Our previous instrument-corrected excitation spectra confirmed the Chl b antenna contribution (>470 nm) was larger in shade compared to sun-acclimated leaves (Matsubara et al. 2002). However, the rapid (∼ 10 ps) decay components in all samples with uninhibited PSII efficiency indicated the primary antenna energy transfer function was still comparable and kinetically independent of the longer lifetime F680 mode.

The rapid antennae F680 modes of both species exhibited clear kinetic reciprocity with the F685 component of PSII when its efficiency was uninhibited. We propose, consistent with Mimuro et al. (1987a), that the F695 band resolved here, in addition to CP47, overlaps irresolvably with antennae rise components possibly donating to PSI and/or the CHB; the rapid rise–decay may obscure the expected delayed rise in F695 emission from CP47. It was clear, however, that in all uninhibited PSII samples the F695 band exhibits the longest PSII lifetime component. Both the core inner antennae components, F685 plus F695, average together to be around 1.5–1.7 ns in the leaves exhibiting Fv/Fm ratios around 0.70. As indicated in our previous room-temperature studies these uninhibited conditions correlate clearly with PSII lifetime components in the 2–2.5 ns range observed when PSII photochemistry is blocked by herbicides (Gilmore & Ball 2000; Matsubara et al. 2002).

The picture of PSII energy flow changes markedly when photochemical efficiency is compromised by thermal dissipation induced during extended periods of excess light, frost and low temperatures. The major changes in the PSII bands include separately increased decay rate constants for both F685 and F695 as indicated by the advent of new, and/or more rapid fluorescence lifetime distribution modes (Fig. 6). According to Lin & Knox (1988), antenna band modes (like F680 and F695) should exhibit rapid lifetimes that are independent of energy dissipation in the reaction centre complexes. We thus propose that the low amplitude of the long-lifetime (∼ 4 ns) F680 component observed in the low PSII efficiency mistletoe leaf, with the strongest PSII inhibition, could be associated with detached chlorophylls or damaged Chl-containing complexes not excitonically coupled to a reaction centre (Govindjee 1995). Most importantly, the rapid decay modes of the antenna bands and reciprocal rise modes of the PSII core-inner-antenna bands in this study do not themselves exhibit significant kinetic changes. The data indicate that the major changes in thermal dissipation are caused by structural-chemical changes in the PSII core-inner-antennae proteins.

Consistent with our earlier room temperature fluorescence lifetime studies (Gilmore et al. 1996; Gilmore & Ball 2000; Matsubara et al. 2002), we propose that the changes in the core-inner-antennae bands of PSII may be complex, involving both (1) photo-inhibitory changes possibly due to amino-acid, electron transport component and or Chl oxidation in addition to (2) sustained xanthophyll cycle- and PsbS-dependent energy dissipation. As outlined by Gilmore et al. (1996) the photo-inhibition process is believed to be associated primarily with attenuation of the main PSII lifetime mode(s) whereas the xanthophyll energy dissipation is more clearly correlated with increased fractional intensities for modes with shorter lifetimes (<1 ns). Structurally, this implies that the PSII core-inner-antennae complexes may be influenced as an average population, by random, cumulative oxidative damage to various amino acids, chlorophylls or electron transport components, but still also be influenced separately by the PsbS-dependent conformational changes required for the xanthophyll-dependent energy dissipation. The combined photo-inhibition + xanthophyll energy dissipation hypothesis is supported by our recent room temperature comparison of sun and shade-acclimated mistletoes in which all major lifetime modes were shorter in the sun leaves with lower PSII photochemical efficiencies than the shade leaves (Matsubara et al. 2002). We speculate that any putative sustained PsbS-dependent heat dissipation that is maintained in the absence of thylakoid lumen acidification may involve oxidation and or phosphorylation of key amino acid residues of the PsbS itself or neighbouring proteins.

The different F685 and F695 spectral widths, centre, lifetimes and amplitudes between the mistletoe and snow gum may be attributed to several items including primarily: (1) factors that may influence vibrational behaviour of the electronic excited states and energy transfer pathways of the PSII Chl-proteins; and (2) different structural leaf anatomy/cell morphology that may influence optical properties such as light-scattering and re-absorption. With respect to the former, we postulate there are probably species-specific amino acid sequence changes or other factors that could alter the organization and architecture of the pigment-proteins including the PSII core-inner-antenna. It has recently been resolved by X-ray crystal structure that the PSII holocomplexes are organized such that CP43 and CP47 are diametrically opposed to each other relative to the D1/D2 heterodimer and other antenna components (Zouni et al. 2001). Hence different rise and decay times for F685 and F695 are not unexpected, given the two components must exhibit different relative distance and orientation parameters from the peripheral antennae complexes. Furthermore, our observations of interspecific differences in leaf anatomy, cell morphology and thylakoid ultrastructure (S. Matsubara unpublished results) must directly influence the leaf optics (Schmuck & Moya 1994).

The influence of the CHB on energy flow in PSII and PSI in winter-acclimated leaves

Figure 8 outlines our current view comparing the flow of energy for PSII and PSI in normal leaves at 77 K (Fig. 8a) to how this flow is influenced by the CHB when PSII efficiency is inhibited (Fig. 8b). The clearest initial events in both the normal and inhibited leaves are the rapid (∼ 10 ps) decays of the LHC compartment with the highest absorbance energy. The LHC compartment may comprise components from both PSII (LHCII) and PSI (LHCI), both with main emission bands centred at < 700 nm. Importantly, the LHC compartment is identified as the antenna by virtue of the fact that its most rapid decay mode exhibits kinetic reciprocity (see Fig. 6) with a rapid rise mode of all the other major bands including the PSII core-inner-antennae, PSI (F740) and the CHB. As explained above PSII inhibition correlates with a decreased average decay time for PSII (Fig. 8b) and the presence of the CHB including its rapid (< 600 ps) decay mode. The PSII inhibition is as indicated before largely independent of the rapid antenna transfer event.

Figure 8.

Scheme of excitation energy flow in leaves of winter-acclimated evergreens with either normal (a) or inhibited (b) photosystem II photochemical efficiency. Solid arrows indicate coupled energy transfer from the antennae complexes associated with PSII (LHCII) or PSI (LHCI). The average lifetimes are listed below each complex, which is drawn to an approximately proportional length that serves as a relative indicator for the rate constants of energy dissipation. The asterisk labels the average lifetime mode of the most rapid and variable CHB decay mode. The gradient region for PSI denotes a delayed rise time separate from the primary (≈ 10 ps) antennae excitation transfer event; the dashed arrow represents the negative amplitude lifetime component used to simulate the delay and the uncertainty of the nature of the process.

In addition to the rapid antennae components < 700 nm which indicate the CHB is accepting energy from the peripheral antennae (Fig. 8b), we also concluded from the strong positive amplitudes that the direct absorbance cross-section of the CHB is also quite considerable. By acting as both a strong new sink for photon absorbance and excitation energy transfer from the peripheral antennae the CHB effectively competes with and lowers the relative quantum yields for both the PSII core-inner-antennae bands and PSI antennae components. The broad CHB band simply overlaps quite strongly, if not completely, with PSII which explains the previous observations of relatively diminished amplitudes for F685 and F695 in the steady-state spectra (Gilmore & Ball 2000; Matsubara et al. 2002).

The CHB exhibited rise kinetics more closely paralleling those of the PSII bands, being significantly earlier than the PSI F740 band. Thus unlike F740, simulation of the CHB kinetics did not require the use of negative amplitude components> 100 ps. Figures 8a and b depict the negative amplitudes (dashed arrow) associated with the delayed rise of F740 (gradient pattern) that are interpreted to be probably associated with energy transfer from components uphill in energy and possibly from PSII. We however, cannot unequivocally distinguish within the framework of this empirical simulation whether the PSI delay is purely transfer, also commonly referred to as ‘spillover’ (Govindjee 1995; Franck, Juneau & Popovic 2002) from PSII or PSI components lying uphill in absorbance energy or partly due to re-absorption. Despite the uncertainty concerning the interpretation of the F740 delayed rise we firmly conclude, from the unaltered decay times of the major (∼ 2800 ps) mode, that the long-wavelength absorbing Chl complexes of PSI which emit F740 are not as strongly influenced by the effects of excess irradiance as either the PSII core-inner-antenna or CHB.

Overall, we interpret the data to indicate that the PSII core-inner-antennae and CHB components respond in parallel to the winter excess light and low-temperature stress by forming modes with increased rate constants of heat dissipation. It remains to be determined in future, more in-depth experiments, precisely how the decay distributions of the CHB, PSI and PSII components relate during the acclimation–de-acclimation processes. For example, we are most interested in studying how the CHB components influence the PSII lifetime distributions in combination with the ΔpH and PsbS-dependent non-photochemical quenching as well as in combination with changes in PSII photochemical quenching. Although we can speculate, based on the similar kinetic and overlapping spectral behaviour that the structural source of the CHB is from damaged, photo-oxidized PSII holocomplexes we emphasize that the pigment-protein composition of the CHB remains undetermined and that it could possibly involve PSI components as well. The latter speculations are corroborated by the data of Ottander et al. (1995) which show that winter-acclimation in pines is associated with novel protein aggregates of PSII components including the PsbS and the LHCs in addition to PSI components.

Key physical and physiological considerations of the 77 K time-resolved emission spectra of winter-acclimated leaves

As outlined in the Introduction the streak camera-spectrograph and DCI methods employed herein theoretically consider the spectral and kinetic influence of: (1) energy transfer events; (2) changes in energy dissipation rate constants; (3) self- and re-absorption events; and (4) the influence of energy-dependent light scattering. The acquisition and analysis techniques employed herein thus provide a functional alternative to the conventional ‘solid dilution’ (Weis 1985) technique that is intended to balance the re-absorption properties of different leaf samples with inert ingredients to compensate for changes in the Chl concentrations of the leaf tissue. The solid-dilution method, which functions at the level of cellular and organellar debris, cannot correct for changes in Chl concentration, Chl a to b ratio, energy transfer and thermal dissipation, that occur primarily at the level of the thylakoid membrane.

With respect to absorption, here it is of interest to note that none of the PSII-associated components exhibited significant re-absorption. The rapid decaying antenna components with the highest absorbance energy, emitting < 700 nm, most clearly are not exhibiting significant re-absorption of themselves or any of the fluorescence components downhill in energy. This is obvious from the fact that their simulated rise-decay kinetics indicate they only receive, and are convolved with, the brief excitation from the laser pulse. Likewise the more rapid F685 and F695 components show no evidence for re-absorption of either themselves or the slower, more intense CHB or F740 band. The above lack of significant re-absorption signals from PSII is not surprising because we note that re-absorption could at most, assuming the impossibility that 100% of all photons emitted as fluorescence are re-absorbed, be only 3–10% of the original absorbed intensity of the excitation pulse; this is based on a coarse estimate of a 3–10% quantum yield of Chl fluorescence (Govindjee 1995). We thus emphasize that re-absorption as a phenomenon is not the major artefact influencing PSII kinetics in the streak camera-spectrograph technique for leaves frozen at 77 K.

We also consider the possibility that re- and self-absorption may be more strongly influencing the F740 band since it has the lowest absorbance energy and highest quantum yield in the system. Evidence supporting this speculation is mainly the slow (> 100 ps) rise components needed to simulate the F740 band, which may suggest some PSII emission components or other uphill components are selectively re-absorbed by F740. However, as mentioned earlier, the negative components may also relate to direct or possibly complicated spillover pathways for energy transfer from PSII or other uphill components. Further one must consider that the delayed rise time is an intrinsic excited state property of this uniquely intense band that has in fact been the subject of several recent investigations involving mutants of either the Lhca1 or Lhca4 proteins (Knoetzel, Bossmann & Grimme 1998; Melkozernov et al. 1998; Morosinotto, et al. 2002).

As indicated earlier it appears likely that ubiquitous light scattering is a larger influence on the blue to red ratios of the Chl fluorescence band intensities than re-absorption per se. Fortunately, light scattering did not seem to be a major artefact preventing resolution of kinetic changes in PSII in this study as evidenced by: (1) the ability to simulate all the major spectral bands with symmetric Gaussian shapes; and (2) the ability to correlate the kinetics with the PSII efficiencies and lifetime distributions measured at room temperature (Gilmore & Ball 2000; Matsubara et al. 2002). It is clear that the quantitative differentiation between scattering and re-absorption of fluorescence may have a very significant impact on methods designed to interpret the ratios of PSI and PSII contributions to fluorescence from leaves at either room temperature and or at 77 K (Schmuck & Moya 1994; Pfündel 1998; Agati, Cerovic & Moya 2000; Gilmore et al. 2000; Franck et al. 2002). A case in point is the counterintuitive observation of lower F685 : F730 ratios in Chl-b-less barley (Gilmore et al. 2000; Peng & Gilmore 2002) and Chl-b-less Arabidopsis (unpublished results) leaves compared to Chl b-replete strains. Because the former have less than half the Chl per leaf area or volume of the latter one would offhand predict more re-absorption and hence lower F685 : F735 ratios in the b-replete strains. Thus, it is clear that the blue to red intensity ratios in leaves are not simply a factor of Chl concentration and re-absorption as again exemplified in the case of the CHB, which is usually most strongly observed in leaves with less Chl.


The CHB competes for photon absorption and antenna excitation to effectively decrease the relative quantum yields of both PSII and PSI. The kinetic features of PSII in this 77 K study and our previous room temperature studies (Gilmore & Ball 2000; Matsubara et al. 2002) also indicate that PSII quantum yield is further down-regulated independent of the competition by the CHB. This may occur by several other processes including primarily both the ΔpH, PsbS and xanthophyll cycle-dependent energy dissipation processes and oxidative photo-inactivation of the PSII reaction centre (Gilmore et al. 1996; Gilmore 1997; Li et al. 2000). We conclude the spectral-kinetic behaviour of CHB is consistent with the formation of protective winter storage chlorophyll-protein complexes.


We thank Jack Egerton and Wayne Pippen for expert field assistance and Tim Hobbs for permission to grow eucalypts on his farm.