The present study investigated the interaction of growth irradiance (Qint) with leaf capacity for and kinetics of adjustment of the pool size of xanthophyll cycle carotenoids (sum of violaxanthin, antheraxanthin and zeaxanthin; VAZ) and photosynthetic electron transport rate (Jmax) after changes in leaf light environment. Individual leaves of lower-canopy/lower photosynthetic capacity species Tilia cordata Mill. and upper canopy/higher photosynthetic capacity species Populus tremula L. were either illuminated by additional light of 500–800 µmol m−2 s−1 for 12 h photoperiod or enclosed in shade bags. The extra irradiance increased the total amount of light intercepted by two-fold for the upper and 10–15-fold for the lower canopy leaves, whereas the shade bags transmitted 45% of incident irradiance. In control leaves, VAZ/area, VAZ/Chl and Jmax were positively associated with leaf growth irradiance (Qint). After 11 d extra illumination, VAZ/Chl increased in all cases due to a strong reduction in foliar chlorophyll, but VAZ/area increased in the upper canopy leaves of both species, and remained constant or decreased in the lower canopy leaves of T. cordata. The slope for VAZ/area changes with cumulative extra irradiance was positively associated with Qint only in T. cordata, but not in P. tremula. Nevertheless, all leaves of P. tremula increased VAZ/area more than the most responsive leaves of T. cordata. Shading reduced VAZ content only in P. tremula, but not in T. cordata, again demonstrating that P. tremula is a more responsive species. Compatible with the hypothesis of the role of VAZ in photoprotection, the rates of photosynthetic electron transport declined less in P. tremula than in T. cordata after the extra irradiance treatment. However, foliar chlorophyll contents of the exposed leaves declined significantly more in the upper canopy of P. tremula, which is not consistent with the suggestion that the leaves with the highest VAZ content are more resistant to photoinhibition. This study demonstrates that previous leaf light environment may significantly affect the adaptation capacity of foliage to altered light environment, and also that species differences in photosynthetic capacity and acclimation potentials importantly alter this interaction.
Requirements for efficient light interception versus avoidance vary depending on incident quantum flux density (Q) as well as on leaf physiological potentials for light utilization (Pearcy & Valladares 1999; Valladares & Pugnaire 1999). Unless the absorbed excitation energy is not used in photosynthesis or dissipated as heat, it may lead to photo-inhibition or extensive oxidation of photosynthetic membranes. Thus, absorption of light in excess of physiological capacities of its utilization is potentially disadvantageous to photosynthetic carbon gain (Osmond et al. 1997, 1999). Although the efficiency of photosynthetic light use at high light increases with increasing foliage photosynthetic capacities, net assimilation rate versus Q responses generally saturate at a considerably lower Q than the peak natural irradiances observed in mid-day on clear days, indicating inherent constraints in photosynthetic acclimation to irradiance.
In natural conditions, light constantly fluctuates during the day, among the days and during the season, implying that plasticity in quenching and antioxidative capacity may crucially alter whole plant performance. However, few studies have investigated the limits and response kinetics of re-acclimation in VAZ pool size of leaves grown at a certain light environment to modified incident irradiance (Demmig-Adams et al. 1989; Bilger et al. 1995; Eskling & Åkerlund 1998; Logan, Demmig-Adams & Adams 1998), and to our knowledge, only Logan et al. (1998) have experimentally altered the light environment in natural conditions. Although these studies demonstrate that the size of the foliar VAZ pool plastically adjusts to rapid changes in irradiance, there are large study-to-study and species-to-species variations in the degree and kinetics of re-acclimation (Demmig-Adams et al. 1989; Bilger et al. 1995; Eskling & Åkerlund 1998; Logan et al. 1998) that are currently not entirely understood.
Studies reporting higher VAZ contents at a common irradiance in plants with lower photosynthetic capacity (Bilger et al. 1995; Verhoeven, Demmig-Adams & Adams 1997) or in conditions curbing photosynthesis (Adams & Demmig-Adams 1994; Adams et al. 2001) suggest that the VAZ pool size is strongly linked to the level of excess excitation energy, although the control may be indirect. Given that leaf photosynthetic capacity is primarily determined by the light environment during leaf growth and development (Niinemets et al. 1998; Frak et al. 2001), there may be important interactive effects of previous leaf light environment on the acclimation potential and time constants for re-acclimation of VAZ pool size to an altered light regime. The possible control of VAZ pool size by excess irradiance suggests that for a common modification in irradiance, the acclimation response should be largest for low-light-grown leaves, because the change in excess light is largest for these leaves. However, there is also evidence that the rate of chloroplast protein synthesis is higher in leaves adapted to higher irradiance (Geiken et al. 1992; Lee, Hong & Chow 2001; García-Plazaola et al. 2002). Such a modification of the rate of carotenoid biosynthesis at the level of gene expression complicates prediction of the summary effect of the previous and current light environment on VAZ pool size.
To determine the extent of foliar adjustment of xanthophyll cycle carotenoid pool size and the kinetic constants of acclimation, we artificially illuminated and shaded leaves of temperate deciduous trees Tilia cordata Mill. and Populus tremula L. along the light gradient in the mature tree canopy (4–25 m). Shade-intolerant P. tremula is a main species in early successional natural European temperate forests, whereas shade-tolerant T. cordata reaches the canopy and dominates in late-successional forests. According to previous studies, P. tremula has higher photosynthetic capacity and higher non-photochemical quenching at a common growth irradiance than T. cordata (Niinemets et al. 1998; Niinemets & Kull 2001). We hypothesized that the capacity for changes in VAZ pool size is directly dependent on the previous light environment of leaves . In particular, that the lower canopy leaves most strongly respond to changed light environment, because these leaves are generally exposed to the highest excess quantum flux densities during the duration of lightflecks. We also expected that VAZ/area and VAZ/Chl versus time and irradiance relations saturate as the leaves come to a new equilibrium with changed irradiance conditions, and that the stable VAZ pool size is achieved faster in the upper canopy leaves due to more rapid protein synthesis in these leaves.
MATERIALS AND METHODS
The experiment was accomplished in a naturally regenerated temperate mixed deciduous stand in Järvselja, Estonia (58°22′ N, 27°20′ E, elevation 38–40 m) in 2000. The forest is dominated by shade-intolerant species P. tremula and Betula pendula Roth. in the upper canopy layer (15–25 m), and by shade-tolerant species T. cordata in the lower layer (4–17 m). The total leaf area index is about 6 m2 m−2, absorbing more than 95% of the above-canopy irradiance. Niinemets et al. (1998) provide a thorough description of the study area.
The experiment was started on 30 June (T. cordata) or on 1 July (P. tremula), and continued until 10 July (T. cordata) or 11 July (P. tremula). Fully mature leaves developed in different light environments within the canopy were selected to cover the complete range in natural variation in foliar pigment contents, morphology and photosynthesis potentials (Fig. 1a). Because the canopy is very homogeneous, and previous research has demonstrated no significant tree effect on pigment contents and photosynthetic rates (Niinemets et al. 1998), one representative tree from each species was chosen for detailed sampling. In P. tremula, 16 distinct canopy positions, in T. cordata, eight locations were marked. Next to these ‘control’ leaves, neighbouring leaves were each individually illuminated with additional irradiance of approximately 500–800 µmol m−2 s−1 between 0500 and 2100 h (Fig. 1b) by wide-beam (beam angle of 60°), 65 W halogen dichroic lamps (Decostar Titan; Osram GmbH, Munich, Germany). The lamps were directly fastened onto the trees, leading to very stable extra incident irradiance, in spite of tree and branch movements in the wind.
In addition to these leaves, shade bags were installed on 16 leaves of P. tremula, and eight leaves of T. cordata. The shade-bags were made of neutral density shade-cloth, and transmitted 45% of incident irradiance. A neighbouring control leaf was also selected for each shaded leaf. Thus, the total number of ‘control’ leaves was 32 for P. tremula, and 16 for T. cordata.
Because of the canopy gradient in radiation interception, there is also a confounding variation in temperature with canopy height, that is of the order of 2–4 °C according to previous studies at our site (Niinemets, Oja & Kull 1999b). Estimations of leaf temperature during fluorescence measurements around mid-day demonstrated that the extra illumination increased leaf temperature on average (± SE) by 1.63 ± 0.38 °C with no clear sample height effect on the temperature difference (r2 = 0.00). Although the leaves with shade bags intercepted less radiation than the control leaves, leaf temperature was slightly higher (0.33 ± 0.15 °C, P < 0.05 according to a paired samples t-test) for the shaded than for the control leaves, possibly due to restricted convective cooling of the leaves enclosed in the shade bags. These data demonstrate that the experimental treatments had an effect on leaf temperature, but also that this effect did not interact with leaf growth irradiance.
Estimation of growth and extra irradiance
To determine long-term irradiance during leaf development (growth irradiance), hemispherical photographs were taken above each leaf included in the analyses. From these photographs, the fractions of penetrating diffuse (ID) and direct (IB) irradiance for 15 d after the summer solstice were determined as in Niinemets et al. (1998). This allowed us to calculate the average daily integrated photosynthetic quantum flux density (Qint; mol m−2 d−1) for the period 21 May to 11 July as :
Qint = Qint0[IDpD + IB(1 − pD)](1)
where Qint0 (mol m−2 d−1) is the average integrated above-canopy quantum flux density for the corresponding period, and pD the fraction of diffuse light (Niinemets et al. 1998). We derived a value of 36.7 mol m−2 d−1 for Qint0 from the global solar radiation measurements in an adjacent meteorological station (58°16′ N, 26°28′ E) and the above-canopy Qint measurements during the study period (r2 = 0.98). Solar radiation data also provided an estimate of pD of 0.556.
Leaf irradiance for each experimental day was estimated by combining the hemispherical photo analyses with continuous measurements of quantum flux density in 13 distinct canopy positions as reported in full detail in Niinemets et al. (1998). Briefly, hemispherical photographs were taken above each quantum sensor, and values of ID and IB were determined. These values along with the integrated daily irradiances [Qd(day)] that were calculated from the sensor readings, were used to develop day-specific regressions in the form of Qd(day) =aID + bIB, where a and b are the regression coefficients. Leaf integrated irradiance for each day was determined from these regressions using leaf-specific estimates of ID and IB.
Analogously, we estimated the amount of light intercepted in each day until the sampling for leaf pigments. In this case, leaf- and day-specific regressions were developed between the sensor readings integrated until leaf sampling [Qds(day)], and values of ID and IB corresponding to each quantum sensor.
The extra quantum flux density was estimated from the measurements by a quantum sensor (LI-190SA; Li-Cor, Inc., Lincoln, NE, USA) as the difference between the quantum flux densities incident to the leaf and directly next to the leaf. Leaf irradiance was an average of several point measurements in various locations along the leaf surface. Repeated estimations of the extra quantum flux density in overcast and clear days demonstrated that the intensity of extra irradiance was stable. Total daily light interception [Qt(day)] of leaves with artificial illumination was found as Qd(day) + Qe(day), where Qe(day) is the daily integrated extra irradiance. The amount of light intercepted until leaf sampling for pigments [Qts(day)] was the sum of Qds(day) and the extra irradiance until the sampling [Qes(day)]. These values were also used to calculate the cumulative total amount of natural light and extra light intercepted until leaf sampling.
Foliar sampling for pigment analyses
Samples were collected in the afternoon hours of 1500–1600 h. The leaves receiving extra illumination and their control leaves were sampled on each day for the first 6 d, and on the last (11th) day. Samples from the leaves with shade bags and their control leaves were collected on the fifth and 11th day. Overall, more than 400 samples were taken during the entire experiment. Each time, two samples of 0.22 cm2 area were removed by a cork-borer, put in labelled vials, and plunged in liquid nitrogen within 5–10 s from removal. The percentage of total area removed during the experiment varied from 1.1 to 13.2% for different leaves (average ± SE = 5.05 ± 0.30%). According to separate experiments, sampling of leaf discs did not alter foliar pigment contents. The samples were kept at −80 °C until analysed.
Full specifics of pigment extraction and analysis by high-performance liquid chromatography (HPLC) are reported in García-Plazaola & Becerril (1999, 2001) and in García-Plazaola et al. (1999). Briefly, frozen leaf discs were ground in liquid nitrogen with a pestle and mortar, and pigments extracted with 1 mL ice-cold 100% acetone. The extract was centrifuged at 15 000 g, and the supernatant kept on ice. The extraction was repeated until the residue was colourless, and the supernatant fractions were pooled. The final extract was filtered through a 0.2 µm syringe filter, and 15 µL were injected in the Waters HPLC system (Waters, Milford, MA, USA). The pigments were separated in a Waters Spherisorb ODS-1 reversed-phase column (particle size 5 µm, inner diameter 4.6 mm, length 250 mm) that was preceded by a Waters Novapak C-18 guard column (4 µm, inner diameter 3.9 mm, length 20 mm). Pigments were eluted using a gradient of two solvents – acetonitrile/methanol/water (84/9/7, v/v/v) and methanol/ethyl acetate (68/32, v/v) – as in García-Plazaola & Becerril (1999, 2001), and detected with a Waters diode array (Model 996) and a scanning fluorescence detector (Model 474). The system was calibrated using purified pigment standards (García-Plazaola & Becerril 1999). A lutein standard was repeatedly injected to check for the calibration drifts.
Fitting of cumulative irradiance versus foliar pigment content relations
Initially, we used linear regressions to fit the xanthophyll cycle carotenoid (sum of violaxanthin, antheraxanthin and zeaxanthin, VAZ) content per area, chlorophyll content per area and VAZ/Chl ratio versus cumulative extra irradiance and total cumulative irradiance relationships. Linear fits assume that the change in pigment contents is doze-dependent, and that the rate of pigment accumulation or degradation does not change in time. For VAZ/Chl, these assumptions were justified as the explained variations of linear fits were always high (average ± SE r2 = 0.915 ± 0.013), and were not improved by using second- or third-order polynomial regressions. VAZ/area and Chl/area versus cumulative irradiance relations were in some cases curvilinear, but occurrence of saturating responses with cumulative irradiance apparently did not depend on leaf growth irradiance. To compare all data-sets without the interfering presumption of a particular statistical model, we calculated the slopes of VAZ/area and VAZ/Chl versus cumulative irradiance using the measurements for the first and the last day, whereas all of the data were included to derive the slope of VAZ/Chl versus irradiance according to the linear model.
Rate of photosynthetic electron transport
A maximum rate of linear whole-chain electron transport (Jmax) was measured in situ between 1100 and 1400 h using a pulse-amplitude modulated fluorometer (PAM-2000; Heinz Walz GmbH, Effeltrich, Germany) and a leaf clip holder (Model 2030-B) on days 5, 7 and 11 after the start of the experiment (Schreiber, Bilger & Neubauer 1994):
Jmax = 0.5 ΦPSII Θ Q(2)
where Q is the photosynthetically active quantum flux density, Θ is leaf absorptance, and ΦPSII is the effective quantum yield. In these calculations, Θ was derived from empirical correlations with leaf chlorophyll (Niinemets et al. 1999b). As the partitioning of absorbed light between the two photosystems is conservative among C3 plants (Laisk et al. 2001), Eqn 2 provides a realistic estimate of photosynthetic electron transport rate (Edwards & Baker 1993).
For control leaves and leaves with shade bags, the measurements were started at Q-values of 200–300 µmol m−2 s−1, and the actinic irradiance was gradually raised during 5–10 min since the beginning of illumination to 1200–2500 µmol m−2 s−1 in P. tremula and to 600–1200 µmol m−2 s−1 in T. cordata. In leaves with extra light, photosynthetic electron transport was fully induced, and the measurements were started at Q-values of 1000–2000 µmol m−2 s−1. The value of ΦPSII was determined repeatedly, and the measurements were completed when a stable value of ΦPSII had been achieved (Niinemets et al. 1998). Jmax was defined as the average of the three highest measurements. The illumination period of 5–10 min was long enough to fully induce photosynthetic electron transport in control leaves, but the induction was slower in shaded leaves, possibly due to enhanced stomatal limitations. Thus, the value of Jmax in shaded leaves gives a representative estimate of the responses of electron transport chain to rapid changes in light environment, but may not entirely reflect the true biochemical potentials of these leaves.
Leaf dry mass per unit area and nitrogen content
All leaves were harvested at the end of the experiment. Leaf lamina area was determined with a computer digitizer (QD-1212; QTronix, Taiwan) and a self-developed computer program. Lamina dry mass was determined after oven-drying at 70 °C for at least 48 h. Lamina nitrogen contents were measured with an elemental analyser (CHN-O-Rapid; Foss Heraeus GmbH, Hanau, Germany).
Testing for the treatment and growth irradiance effects on leaf variables
Linear regression analyses were carried out to test for the statistical significance of the growth irradiance (Qint) effects on leaf variables. Regressions between leaf characteristics and Qint were calculated for all sampling dates, and the treatment effect on these relations was tested for by covariation analyses (ancova) using corresponding untreated leaves as controls. A separate slope ancova model that includes Qint × treatment interaction term was used first to check for the slope differences. Whenever the interaction terms was insignificant, the analysis was continued according to a common slope ancova model that lacks the interaction term. For some leaf characteristics (e.g. nitrogen concentration or lutein content), the effect of growth irradiance was not always significant. In such cases, the means were compared by standard analyses of variance (anova), and the treatment effects were separated by Bonferroni test. All statistical effects were considered significant at P < 0.05.
Daily integrated quantum flux densities (Qint) during leaf growth varied approximately 30-fold between the canopy top and the bottom (Fig. 1a). Although there were several lightflecks with high irradiance that comprised approximately 40–50% of total irradiance on sunny days, the leaves at the bottom of the canopy were exposed to low-level diffusive irradiance of 5–20 µmol m−2 s−1 during most of the day. Average (± SE) above-canopy daily integrated quantum flux density for the study period, 30 June to 11 July, was 33.6 ± 3.3 mol m−2 d−1, comparing well with the Qint value during leaf development of 36.7 mol m−2 d−1. The Qint range was 45–47 mol m−2 d−1 on four clear experimental days; this range was 26–37 mol m−2 d−1 on six partly overcast days, and Qint was 14.1 and 16.9 mol m−2 d−1 on two overcast days.
Extra illumination increased on average (± SE) daily integrated leaf irradiance by 43.2 ± 1.2 mol m−2 d−1 (Fig. 1b). Compared with the growth irradiance and the natural irradiance during the experiment, daily light interception was increased by approximately two-fold for the upper and 10–15-fold for the lower canopy leaves.
Natural variation in foliar xanthophyll cycle pigment (VAZ) and chlorophyll contents
In control leaves, the content of xanthophyll cycle carotenoids (sum of violaxanthin, antheraxanthin and zeaxanthin; VAZ) per area was positively related to growth irradiance (Qint) in both species (Fig. 2). Analogously, VAZ to chlorophyll ratio (r 2 = 0.71 for P. tremula, r 2 = 0.44 for T. cordata, P < 0.001 for both) was positively associated with Qint in both species. According to separate and common slope analyses of covariance (ancova), VAZ per area and VAZ/Chl versus Qint relations did not significantly differ between various dates in P. tremula (P > 0.3 for slope differences, and P > 0.6 for intercept differences). The slopes were also not significantly different between the various dates in T. cordata (P > 0.9 for VAZ/area and P > 0.2 for VAZ/Chl). However, in this species, the intercept of VAZ/area versus Qint relation was significantly larger (P < 0.01) in the first than in the last day of the experiment (day 11), and the intercept of VAZ/Chl versus Qint was larger in the first and fifth day of the experiment than in the last day (P < 0.001).
Chlorophyll content per unit area was positively related to irradiance in both species (r 2 = 0.09, P < 0.01 for P. tremula and r 2 = 0.65, P < 0.001 for T. cordata). These relationships did not vary significantly during the course of the experiment (P > 0.07 for slope and P > 0.1 for intercept differences in P. tremula; P > 0.3 for slope and P > 0.2 for intercept differences in T. cordata).
Effects of extra irradiance on leaf pigments
Extra irradiance led to significant changes in pigment contents and composition, in particular in P. tremula (Table 1). After 11 d of elevated irradiance, the content of all pigments except VAZ was significantly reduced. Whereas in plants under natural conditions, lutein generally comprises close to or more than half of total carotenoids (e.g. Logan et al. 1996; Demmig-Adams 1998), VAZ contents significantly exceeded that of lutein in the exposed leaves at the end of the experiment in P. tremula (Table 1).
Table 1. Changes in average (± SE) foliar pigment contents due to extra irradiance and artificial leaf shading in Populus tremula and Tilia cordataa
Means with the same letter are not significantly different (P > 0.05) according to analyses of variance (anova) or covariance (ancova). Common slope analyses of covariance with growth irradiance (Qint) as the covariate were employed for variables that were significantly (P < 0.05) correlated with Qint. anova and ancova were followed by Bonferroni test to separate the means.
Neo, neoxanthin; Lut, lutein; VAZ, sum of violaxanthin, antheraxanthin and zeaxanthin; β-Car, β-carotene; α-Car, α-carotene; Chl a, chlorophyll a; Chl b, chlorophyll b.
P. tremula (n = 16 for every treatment, and n = 32 for the controls)
9.96 ± 0.16a
28.2 ± 0.9ab
24.2 ± 1.3a
28.7 ± 1.0a
0.137 ± 0.011a
237.9 ± 4.4a
68.6 ± 1.3a
10.10 ± 0.22a
29.1 ± 0.6a
24.3 ± 1.2a
29.2 ± 1.3a
0.093 ± 0.011ab
247.8 ± 5.4ab
73.5 ± 1.3ab
7.71 ± 0.21b
25.4 ± 0.7bc
34.9 ± 2.1c
19.3 ± 0.7b
0.084 ± 0.008b
171.4 ± 4.2c
59.3 ± 1.7c
10.78 ± 0.17a
30.3 ± 0.6a
22.7 ± 1.4b
32.1 ± 1.0a
0.142 ± 0.047ab
264.3 ± 5.7b
78.6 ± 1.1bc
10.06 ± 0.19a
28.9 ± 0.6a
23.5 ± 1.3a
28.7 ± 0.7a
0.082 ± 0.009ab
242.7 ± 4.8ab
73.4 ± 1.6ab
6.07 ± 0.34c
22.4 ± 0.7c
40.8 ± 2.4d
16.7 ± 0.8b
0.045 ± 0.011b
146.0 ± 6.7d
45.2 ± 2.2d
10.77 ± 0.33a
29.6 ± 0.9a
21.2 ± 1.7b
29.4 ± 1.1a
0.097 ± 0.008a
253.4 ± 8.0ab
80.1 ± 2.4c
T. cordata (n = 8 for every treatment, and n = 16 for controls)
13.0 ± 0.6ab
31.1 ± 1.5a
16.6 ± 1.1a
20.5 ± 1.4a
5.8 ± 0.6a
282 ± 11a
97.4 ± 4.0a
11.79 ± 0.47ab
27.5 ± 1.3a
15.9 ± 0.9ab
19.3 ± 1.4a
5.24 ± 0.30a
261 ± 11a
95.7 ± 3.3a
8.8 ± 1.2c
25.6 ± 3.3a
15.3 ± 3.0ac
12.6 ± 2.6b
2.10 ± 0.20b
195 ± 24b
74.7 ± 8.5b
13.5 ± 0.7a
33.0 ± 2.2a
17.0 ± 1.1a
22.9 ± 1.6a
5.95 ± 0.16a
291 ± 16a
103.8 ± 4.8a
8.3 ± 0.9c
26.1 ± 3.1a
15.5 ± 2.7ac
11.0 ± 1.9b
1.15 ± 0.10b
168 ± 19b
65.6 ± 6.5b
11.5 ± 0.6b
26.3 ± 2.3a
14.0 ± 0.9bc
19.0 ± 1.5a
4.9 ± 0.34a
256 ± 13a
94.3 ± 4.2a
13.0 ± 0.9ab
30.5 ± 2.5a
16.0 ± 1.3ab
20.5 ± 2.2a
5.7 ± 0.5a
273 ± 19a
104 ± 6a
Comparison of VAZ/area versus growth irradiance relations between the leaves with extra irradiance and controls indicated that extra light resulted in enhanced VAZ/area in all leaves of P. tremula, and in the upper canopy leaves (growth irradiance 7–10 mol m−2 d−1) of T. cordata (Fig. 3a & b), but VAZ/area was not affected or decreased with extra light in the lower canopy (growth irradiance 1–5 mol m−2 d−1) leaves of T. cordata (Fig. 3a & b). Positive effects of extra-light on VAZ/Chl were evident for all leaves, except for the lowermost foliage of T. cordata (Fig. 3c & d). VAZ/area increased at most two-fold, but VAZ/Chl three to four-fold (cf. Fig. 3b & d). This difference is explained by decreases in foliar chlorophyll contents in leaves with extra irradiance in all cases (Fig. 4).
Interaction of extra light with leaf growth irradiance
In the middle of the experiment, at day 5, the slopes of VAZ/area (P > 0.7), VAZ/Chl (P > 0.2) and Chl/area (P > 0.8) versus growth irradiance were not significantly different between the exposed and control leaves of P. tremula (Fig. 3a & c). At the end of the experiment, the slopes of VAZ/area versus Qint relationship were also not statistically different (P > 0.2) between the control and exposed leaves, but the slope of VAZ/Chl versus Qint relationship was significantly larger (P < 0.005) in the exposed than in the control leaves of this species (Fig. 3d). More responsive VAZ/Chl resulted from a larger decrease in leaf chlorophyll contents in the upper canopy leaves of P. tremula (P < 0.05, Fig. 4). As the result of enhanced chlorophyll decline in the upper canopy, the weak positive correlation between chlorophyll content and Qint in the control leaves (r 2 = 0.18, P < 0.05) was no longer significant in the exposed leaves of this species (r2 = 0.11, P > 0.2, Fig. 4).
In T. cordata, the slopes of VAZ/area (P < 0.001) and VAZ/Chl (P < 0.005) versus Qint were significantly larger in the leaves with extra light than in the control leaves at both day 5 and 11 (Fig. 3). This was because of a greater change of the xanthophyll cycle pool size in the upper canopy leaves.
The average slope of VAZ/area versus cumulative extra light was positively related to growth irradiance in T. cordata, but not in P. tremula (Fig. 5a). The slope of VAZ/Chl versus extra light was positively associated with Qint in both species (Fig. 5b). In P. tremula, this relationship was because of a stronger decrease in chlorophyll contents in the upper canopy (Fig. 5c). In T. cordata, this relation was explained by larger canopy differences in changes in VAZ than in chlorophyll contents (cf. Fig. 5a & c). Thus, these data collectively demonstrate a strong interactive effect of growth light environment on the response kinetics of VAZ pool size and leaf chlorophylls, especially in T. cordata.
Shading and pigment content
In P. tremula, leaf shading decreased VAZ/area (Fig. 6a & b, Table 1) and VAZ/Chl (Fig. 6c & d; P < 0.001) on both of the sampling days 5 and 11. No evidence of reduced VAZ/area was evident in T. cordata (Table 1, Fig. 6a & b), and VAZ/Chl was even larger in the shaded than in the control leaves of T. cordata at the end of the experiment (P < 0.05, Fig. 6d). For these relations, shading × Qint interactions were insignificant (P > 0.8), demonstrating that the shading effects were independent of growth irradiance.
Chlorophyll content increased in response to shading in P. tremula (Table 1; P < 0.02 for total chlorophyll), but not in T. cordata. The content of other leaf pigments was largely unaffected by shading (Table 1).
Modifications in nitrogen, leaf structure and photosynthetic electron transport
In P. tremula, the leaves with extra light had a lower foliar nitrogen content per unit dry mass (NM ± SD = 1.76 ± 0.12%) than the control leaves (1.87 ± 0.17%, P < 0.001, Fig. 7a), and the shaded leaves had a larger NM (1.98 ± 0.16%) than the control or exposed leaves (P < 0.001, Fig. 7a) at the end of the experiment. In T. cordata, NM was also lower (P < 0.05) in the exposed (2.77 ± 0.12%) than in the control (2.87 ± 0.22%) leaves, and NM was larger in the shaded (3.06 ± 0.10%) than in the control leaves (P < 0.01, Fig. 7a).
The slopes of leaf dry mass per unit area (MA) versus Qint were not affected by shading or extra irradiance in both P. tremula (P > 0.3) and T. cordata (P > 0.2, Fig. 7b). At a common irradiance, MA of P. tremula was significantly larger in the exposed leaves followed by control and shaded leaves (all treatments significantly different at P < 0.001). However, nitrogen content per unit area (NA = MANM) was independent of the treatment effects in P. tremula (P > 0.9, Fig. 7c). Given also that neither MA nor NA differed between the treatments in T. cordata, these data together suggest that shading and extra irradiance did not modify foliar structure and did not result in an import or export of foliar nitrogen. Rather, the treatment differences in N percentage and MA are in agreement with observations of enhanced non-structural carbohydrate storage in high-light-exposed leaves and depletion of foliar carbohydrate pools in shaded foliage (Chatterton, Lee & Hungerford 1972; Hendrix & Huber 1986).
Compatible with the enhanced N investment per unit area in high irradiance, the rates of photosynthetic electron transport were positively associated with growth irradiance for all treatments (Fig. 8a). Both increases and decreases in leaf irradiance led to reduced maximum photosynthetic electron transport rates (Jmax) measured after 5–10 min illumination to high actinic irradiance (Fig. 8a & b). Whereas in the shaded leaves the decreases in apparent Jmax were probably due to reduced induction rate of photosynthetic electron transport (see Material and Methods), the photosynthetic apparatus was fully induced in the exposed leaves. Thus, the decrease in Jmax in response to extra irradiance was probably due to photoinhibition or photo-destruction of photosynthetic apparatus.
Control leaves of P. tremula had a larger value of Jmax at both the common growth irradiance (Fig. 8a) and leaf nitrogen content (Fig. 8b) than the leaves of T. cordata. This difference was further amplified by extra irradiance, which reduced Jmax to 60–80% of the initial value in P. tremula (Fig. 9a), but to 5–50% of the initial value in T. cordata (Fig. 9b). The value of Jmax was stabilized in P. tremula after approximately 7 d of extra irradiance (Fig. 9a), and Jmax partly recovered by the end of the experiment in T. cordata (Fig. 9b). Leaves with larger photosynthetic capacity retained a larger fraction of photosynthetic electron transport at various times during the experiment (Fig. 9).
The apparent Jmax was reduced to a similar extent after shading in P. tremula and T. cordata (to 50–80% of initial value), and the percentage of electron transport remained was independent of growth irradiance in both species (P > 0.2).
Natural and extra irradiance and xanthophyll cycle carotenoid (VAZ) pool size
Exposure of leaves to extra light resulted in dramatic changes in carotenoid content and composition (Fig. 3). At the end of the treatment, VAZ/Chl had increased up to four-fold and VAZ/area up to two-fold in the uppermost leaves of shade-intolerant species P. tremula (Fig. 3b & d). In shade-tolerant T. cordata, VAZ/area was moderately affected and even decreased in the lower canopy leaves (Fig. 3b & d) as has been previously observed after a combined high irradiance and water stress in several Mediterranean species (Kyparissis, Petropoulou & Manetas 1995; Munné-Bosch & Alegre 2000). Overall, the slope for changes in VAZ pool size after increases in irradiance was larger in the upper canopy leaves of T. cordata, and was essentially constant for P. tremula (Fig. 5a).
These data demonstrate that previous leaf light environment significantly interacted with the acclimation response of the VAZ pool size in T. cordata, but not in P. tremula. However, these observations are not consistent with the hypothesis that the leaves adapted to the lowest growth irradiance and with lowest photosynthetic capacities possess the highest plasticity for adjustment of VAZ pool size. In fact, constrained photosynthesis rates and larger excess excitation energy are not always associated with enhanced VAZ pool size. For instance, genetic reduction in photosynthesis led to elevated VAZ in one species, but not in the other (Bilger et al. 1995), and N-limitation increased spinach VAZ pool size in one (Verhoeven et al. 1997), but not in the other study (Logan et al. 1999), and did not modify VAZ pool size in Clematis vitalba (Bungard, McNeil & Morton 1997). These conflicting data along with our observations suggest that VAZ pool size is not directly controlled by the excess excitation energy.
According to many studies, protein synthesis is faster in high-light-acclimated leaves (Geiken et al. 1992; Baroli & Melis 1996; Lee et al. 2001). Provided some de novo protein synthesis is required for acclimation in VAZ pool size, the range in the capacities for VAZ pool adjustment observed in our study may result from canopy variation in protein synthesis rates. Possible control of VAZ pool size at the level of gene expression may also explain the non-uniform VAZ pool size versus excess light relationships across the studies.
Changes in foliar chlorophyll with extra irradiance
Exposure to extra irradiance resulted in significant decreases in leaf chlorophyll contents (Fig. 4). Such decreases in chlorophyll are commonly observed after increases in irradiance and effectively reduce the light-harvesting antenna size (Andersson & Aro 1997). Given that non-pigmented apoproteins of pigment-binding proteins are not stable in vivo and the non-bound chlorophyll pools are very small (Paulsen 2001), extensive decreases in leaf chlorophyll contents indicate dismantling of chloroplastic light-harvesting pigment protein complexes (Andersson & Aro 1997; Yang et al. 1998). We cannot tell from our data whether the decrease in light-harvesting antenna size is initiated by de novo synthesis of specific proteases attacking the light-harvesting pigment binding proteins (Andersson & Aro 1997; Yang et al. 1998) and followed by release and degradation of chlorophyll, or whether the proteolysis is initiated by possible chlorophyll photo-destruction (Barry, Young & Britton 1990).
In our study, the slope for changes in chlorophyll content with extra irradiance was smaller for the upper canopy leaves of T. cordata, but a reverse response was observed in P. tremula (Fig. 5c). In fact, more extensive chlorophyll decline in the upper canopy leaves of P. tremula (Fig. 4) was also the primary determinant of the larger increase of VAZ/Chl in the upper canopy of this species (Figs 3d & 5b). Such a greater susceptibility for chlorophyll destruction of the leaves adapted to higher irradiance contrasts with previous observations (Baroli & Melis 1998; Barth, Krause & Winter 2001). However, it is important that multiple stress factors inevitably interact in plant canopies. In particular, leaf temperatures are higher and potential heat stress more severe at greater irradiances (Niinemets et al. 1999b). In general, high temperature interacting with excess irradiance results in more extensive photoinhibition than the light stress alone (Valladares & Pearcy 1997; Königer, Harris & Pearcy 1998). Thus, we suggest that the larger chlorophyll decline in the uppermost leaves results from a significant heat stress at higher irradiance in the canopy of P. tremula (Niinemets et al. 1999b).
Adaptation kinetics to extra irradiance
In growth chamber experiments, the pool size of VAZ generally acclimates to increases in irradiance within 5 (Demmig-Adams et al. 1989; Eskling & Åkerlund 1998) to 7 d (Bilger et al. 1995) after changes in light regime. Decreases in leaf chlorophyll content are also complete within 4–5 d after transfer to high irradiance (Andersson & Aro 1997; Yang et al. 1998). However, in some cases, a full adjustment of VAZ pool size may take more than 7 d (Demmig-Adams et al. 1989; Logan et al. 1998). We expected the VAZ/area and VAZ/Chl versus cumulative excess irradiance relationships to saturate in time as the leaves reach a new equilibrium with changed incident irradiance. There was evidence of curvilinearity in VAZ/area versus time and excess irradiance relations for some leaves (data not shown), but not for all (inset in Fig. 5a), and VAZ/Chl increased linearly in time for all leaves in the canopy with no evidence of reaching a plateau (inset in Fig. 5b).
These data collectively indicate that leaf acclimation in natural environments at the background of strongly fluctuating natural irradiance is more time-consuming than the time-span of 5–7 d observed previously in growth chamber experiments. In another field study, Logan et al. (1998) observed that the VAZ pool size increased even for more than 17 d after a step increase in irradiance level. This study together with our observations demonstrates that foliar xanthophyll cycle pool size is generally not in a steady state under natural conditions. Although the confounding canopy variation in leaf temperature may also potentially modify the adjustments in VAZ pool size, high temperature alone does not necessarily alter leaf VAZ contents (Hanson & Sharkey 2001).
What is the role of high VAZ/Chl?
The physiological significance of the high and continuously increasing VAZ/Chl ratios is somewhat puzzling. Combined increases in VAZ/area and decreases in chlorophyll contents resulted in very high VAZ/Chl values on the order of 0.2–0.3 mol mol−1 (Fig. 3d). Comparable values have so far been observed only in a few cases, such as in Euonymus kiautschovicus (0.21 mol mol−1; Verhoeven, Adams & Demmig-Adams 1998) and conifers (up to 0.24 mol mol−1; Adams & Demmig-Adams 1994) in winter conditions, in nitrogen-deficient Spinacia oleracea (up to 0.23 mol mol−1; Verhoeven et al. 1997), and in very high-light-tolerant Chlamydamonas mutants (up to 0.4 mol mol−1; Förster, Osmond & Boynton 2001).
According to the current model of non-photochemical quenching of excitation energy (NPQ; Gilmore 1997, 2001), only the VAZ bound to the light-harvesting antenna pigment–protein complexes affects NPQ. However, the number of VAZ-binding sites in antenna proteins is limited, and a certain fraction of VAZ is apparently free in chloroplast membranes (Verhoeven et al. 1999; Anderson et al. 2001). Stoichiometric analyses based on reconstitution experiments with recombinant apoproteins of light-harvesting pigment-protein complexes and purified pigments indicate that all leaf VAZ is associated with pigment-binding complexes for VAZ/Chl ratios of up to 0.05 mol mol−1 (Verhoeven et al. 1999; Bassi & Caffarri 2000). This suggests that in our study a major fraction of VAZ was free in chloroplast membranes and did not directly participate in NPQ. In fact, studies with zeaxanthin-deficient mutants demonstrate that only a minor amount of zeaxanthin is required for development of a full capacity of zeaxanthin-dependent non-photochemical quenching (Niyogi, Grossman & Björkman 1998; Pogson et al. 1998). Furthermore, there is no single relationship between NPQ and zeaxanthin content when plants containing an extensive range of VAZ/Chl are pooled (Härtel et al. 1996; Förster et al. 2001).
However, enhanced VAZ is possibly involved in processes other than NPQ. According to recent studies, zeaxanthin plays a significant role as chloroplastic antioxidant, specifically protecting highly unsaturated chloroplast membranes from photo-oxidation (Havaux & Niyogi 1999; Havaux et al. 2000). Given that most of the light-related increase in the VAZ pool resulted in increases in the ‘free’ VAZ, membrane protection from peroxidation may be the primary role of light-related enhancements of VAZ pool size. Beside zeaxanthin playing a role in NPQ and acting as an antioxidant, the xanthophyll cycle may also be involved in the protection of plants from heat stress (Havaux et al. 1996).
Shading and leaf pigment content
Shading significantly reduced VAZ pool size in P. tremula, but not in T. cordata (Fig. 6). Again, in P. tremula, there was no interaction with previous leaf light environment and the reduction in VAZ pool size due to shading. Previous work indicates that VAZ pool size may decrease between the days of varying incident quantum flux density (Niinemets et al. 1998), as well as during the night (Demmig-Adams et al. 1989), demonstrating highly dynamic nature of the xanthophyll cycle carotenoid pool size. The decrease in VAZ pool size due to shading in P. tremula (Fig. 6) further indicates that the turnover of VAZ carotenoids is high.
Shading often leads to significant increases in the antenna size of photosystems I and II (Walters & Horton 1994). In herb species, such changes may be fairly rapid and be completed within 3–5 d from the transfer to low light (Walters & Horton 1994). In our study, shading had only minor effects on foliar chlorophyll contents (Table 1), suggesting that only small changes occurred in antenna size. Given that most of leaf VAZ was bound to chlorophyll-protein complexes in T. cordata (VAZ/Chl ratios < 0.05 mol mol−1), insignificant modification in antenna protein content and stoichiometry may explain the invariable leaf VAZ pool size in this species. Thus, whenever the content and stoichiometry of antenna proteins is not modified, it is likely that only non-bound VAZ pool size may undergo rapid changes during the day and between the days.
Changes in photosynthetic capacity in response to altered light level
There are only minor changes in foliar photosynthetic capacity of fully mature leaves after radical changes in leaf light environment (Naidu & DeLucia 1997; Frak et al. 2001). This inflexibility is mainly associated with limited modifications in leaf dry mass per unit area (MA) and foliar nitrogen after changes in irradiance (Frak et al. 2001). We observed some alterations in nitrogen content per unit dry mass (Fig. 7a) and in MA (Fig. 7b), but these effects were probably associated with non-structural carbohydrate accumulation/depletion in response to altered irradiance with no evidence of changes in nitrogen content per unit area (Fig. 7c). Thus, our results suggest that the foliar biochemical potentials remained stable after extreme changes in light environment. This contrasts with the hypothesis of rapid adjustment of leaf photosynthetic potentials and nitrogen contents to day-to-day modifications in irradiance in natural canopies (Kull & Kruijt 1999).
Exposure to extra irradiance led to a significant photo-inhibition of foliar photosynthetic electron transport rates (Fig. 8). The decline in electron transport rates was larger in the lower canopy than in the upper canopy leaves (Fig. 9), in particular in T. cordata (Fig. 9b). In general, shade-tolerant species are more sensitive to photoinhibition (Öquist et al. 1992; Barth et al. 2001; Krause et al. 2001), possibly because of a larger light-harvesting antenna rather than due to differences in xanthophyll cycle characteristics (Krause et al. 2001; see Baroli & Melis 1998 for a discussion). In our study, greater sensitivity of photosynthetic electron transport in T. cordata is correlated with the limited capacity for xanthophyll cycle pool size adjustment, however, it may also be related to a larger light-collecting antenna in this species (Fig. 4). In fact, lower canopy leaves completely failed to adjust the VAZ pool size (Fig. 3a & b), and retained less than 10% of initial electron transport rate (Fig. 9b).
Although foliar nitrogen contents were stable (Fig. 7c), down-regulation in leaf photosynthesis rates was also observed after shading (Fig. 8). However, this decline in apparent maximum photosynthetic electron transport rate observed after approximately 10 min of actinic light probably reflected a decreased rate of induction of foliar electron transport, rather than the true decline in foliar biochemical capacities.
There are large diurnal, day-to-day and seasonal variabilities in above-canopy quantum flux densities as well as extensive intrinsic light gradients within plant canopies. Although plant leaves acclimate physiologically to the large natural variation in irradiance by increasing the VAZ pool size, the leaf acclimation potential may strongly depend on the previous leaf light environment. Contrary to the expectations, highest acclimation potential was observed in leaves exposed to the greatest irradiances. Even at a common irradiance, the shade-tolerant species T. cordata exhibited a significantly lower plasticity in VAZ/area and VAZ/Chl, and stronger limitations in photosynthetic electron transport than the shade-tolerant species Populus tremula. These important species differences that are possibly associated with differences in light-harvesting antenna size and protein turnover may provide an explanation for preferential dispersal of shade species in more shaded forest and canopy locations.
Our experiment in a natural plant canopy also demonstrates that acclimation in VAZ pool size takes longer in the background of strongly fluctuating natural irradiance and temperature than in stable growth chamber environments. Thus, the experiments on the rate of VAZ acclimation in controlled conditions do not directly apply to the field.
We are indebted to Heiko Rämma and Mari Tobias for skilful assistance, and to Anne Jõeveer for solar radiation data of Tõravere Actinometric Station. We also thank the Estonian Science Foundation (Grant 4584), the Estonian Ministry of Education and Science (grant 0182468As03), the Spanish Ministry of Science and Technology (Grant AGL 2001–1131) and the Spanish Ministry of Education (Grant AMB 91–1171) for partial funding of this project.