Light and temperature, but not UV radiation, affect chlorophylls and carotenoids in Norway spruce needles (Picea abies (L.) Karst.)



    1. Lehrstuhl für Botanik II, Universität Würzburg, Julius-von-Sachs-Platz 3, D-97082 Würzburg, Germany and
    Search for more papers by this author

    1. Lehrstuhl für Botanik II, Universität Würzburg, Julius-von-Sachs-Platz 3, D-97082 Würzburg, Germany and
    Search for more papers by this author
  • K. HAUFF,

    1. Institut für Meteorologie und Klimaforschung, Bereich Atmosphärische Umweltforschung, (IMK-IFU), Forschungszentrum Karlsruhe GmbH, Kreuzeckbahnstr. 19, D-82467 Garmisch-Partenkirchen, Germany
    Search for more papers by this author

    1. Institut für Meteorologie und Klimaforschung, Bereich Atmosphärische Umweltforschung, (IMK-IFU), Forschungszentrum Karlsruhe GmbH, Kreuzeckbahnstr. 19, D-82467 Garmisch-Partenkirchen, Germany
    Search for more papers by this author

    1. Institut für Meteorologie und Klimaforschung, Bereich Atmosphärische Umweltforschung, (IMK-IFU), Forschungszentrum Karlsruhe GmbH, Kreuzeckbahnstr. 19, D-82467 Garmisch-Partenkirchen, Germany
    Search for more papers by this author

    Corresponding author
    1. Lehrstuhl für Botanik II, Universität Würzburg, Julius-von-Sachs-Platz 3, D-97082 Würzburg, Germany and
    Search for more papers by this author

Correspondence: Erhard Pfündel. Fax: +49 931 888 6235; e-mail,


Concentrations of chlorophyll a/freshweight (Chl a FW) and photosynthetic pigments/chlorophyll a were studied during one growing season in the current year's (CYN) and last year's needles (LYN) from Norway spruce (Picea abies (L.) Karst.) grown under natural or close-to-natural climate. Climate regimes differed in photosynthetic active radiation (PAR), temperature (T) and UV-B radiation. Pigments were not affected by UV-B but most of the differences between climate regimes, and also seasonal variations within climate regimes, could be related to PAR and T. Generally, two types of response to climate were observed: firstly, pigments reacted primarily to PAR without marked sensitivity to T and exhibited slow response times (> 30 d), and, secondly, pigments were affected by the combined action of PAR and T and responded faster than 20 d. The Chl a FW and chlorophyll b/chloprophyll a ratio exhibited slow-type response in CYN and fast-type response in LYN. Higher amplitudes in CYN than in LYN were observed for the latter two parameters, which are known to be associated with levels of pigment–protein complexes. It is suggested that slow response in CYN ensures that the high investments in proteins in these needles occur only in response to longer-lasting climate episodes.







Chl a

chlorophyll a

Chl a FW

chlorophyll a/freshweight

Chl b

chlorophyll b


current year's needles


experimental field side




last year's needles




photosynthetically active radiation

SD1 and SD2

solar dome 1 and 2






Plants in nature are exposed to a wide range of visible (400–700 nm) and UV (280–400 nm) radiation intensities. As energy of visible radiation fuels photosynthesis, sufficient availability of light is vital for autotrophic plants. Light intensities exceeding the demands of photosynthesis, however, can severely damage the photosynthetic apparatus (Melis 1999). Under low light therefore plants optimize light-harvesting and, under high light, they restrict access of light to chloroplasts, safely dissipate excess absorbed light energy and scavenge radicals formed under these excessive conditions (Björkman & Demmig-Adams 1994). Furthermore, natural UV-radiation, especially UV-B, can inhibit photosynthesis with the reaction centre of photosystem II (PSII) being one of the most-studied targets (Jordan 1996; Vass 1997); hence, plants also require processes to control UV damage.

In response to excess light, leaves can change orientation, optical properties and anatomy; high light is also known to reduce chloroplast size and diminish grana stacking in chloroplasts and, in many evergreen sclerophytes, decrease chlorophyll levels (Björkman & Demmig-Adams 1994; Smith et al. 1997; Adams et al. 2001). The changes in chloroplast ultrastructure have been linked to reduction of the light-harvesting complex II (LHCII) which forms the major part of PSII (Anderson, Chow & Park 1995). Because LHCII binds most of the chlorophyll b (Chl b) in thylakoid membranes (Siefermann-Harms 1985; Melis 1991), photosynthetic acclimation to high light is consistently paralleled by decreased ratios of Chl b/Chl a (Anderson et al. 1995).

In addition, carotenoids are important for light-harvesting but protect the photosynthetic apparatus from the deleterious effects of strong light as well (Siefermann-Harms 1985; Demmig-Adams 1990; Young & Britton 1990). In response to excess light, a rapid increase in dissipation of absorbed light energy by carotenoids can occur by the formation of zeaxanthin (zea) and antheraxanthin (ant) within the xanthophyll cycle (Demmig-Adams 1990; Gilmore & Yamamoto 1993; Pfündel & Bilger 1994). Extended periods of strong light are often accompanied by increased ratios of carotenoid/chlorophyll probably reflecting the permanently increased needs for photoprotection by carotenoids (Demmig-Adams, Gilmore & Adams 1996; Demmig-Adams 1998). Unlike strong light acclimation, UV increases Chl b/Chl a (Strid, Chow & Anderson 1990; Pfündel, Pan & Dilley 1992; Jordan et al. 1994; Deckmyn & Impens 1995); however, a decrease in Chl b/Chl a under elevated UV has also been reported (Deckmyn, Martens & Impens 1994). Response of carotenoids to UV is variable: decreased carotenoid levels were observed under UV (Strid et al. 1990; Musil, Midgley & Wand 1999) but they were also stimulated by UV (Steel & Keller 2000; Xiong & Day 2001). Currently, there is no clear concept explaining why UV-induced pigment responses vary so markedly.

The exact way by which radiation intensities control long-term pigment changes has not been finally established. Studies exploring effects of visible radiation at different temperatures suggest that it is not light per se but the balance between absorbed light energy and the energy used by photosynthesis which triggers pigment acclimation (Maxwell, Falk & Huner 1995; Huner, Öquist & Sarhan 1998; Pfannschmidt, Allen & Oelmüller 2001). This energy balance may depend on ambient temperature because light absorption and charge separation by reaction centres are temperature-independent under most natural conditions but subsequent photosynthetic events are markedly slowed down with decreasing temperature (Raven & Geider 1988). The effects of UV radiation on pigments have been discussed mostly in terms of direct destruction after absorption of the high-energy photons; however, increased carotenoid levels in the presence of UV (see citations above) might reflect acclimation to climate by improving carotenoid-mediated quenching of UV-induced radicals.

So far, the action of UV on photosynthetic pigments has been demonstrated in flowering plants but not in conifers. Conifer needles exhibit considerably higher UV-screening of epidermal cell layers than most flowering plants (Day, Vogelmann & DeLucia 1992; Day 1993; Sullivan et al. 2000). Therefore, the question arises if UV intensities inside the photosynthetic tissue of conifers might be too low to affect pigment concentrations. Further, evidence for combined effects of visible radiation and temperature on photosynthetic pigments was abstracted from experiments under controlled environment or by comparing data from summer and winter material: the continuous pigment acclimation to natural climate fluctuations is incompletely understood. Although light and temperature can be predicted to be important determinants for pigment concentrations, they are also affected by leaf history (Yamashita, Koike & Ishida 2002); however, data on the relation between leaf history and pigment–climate interaction is still sparse. To obtain information on these issues, we studied the effects of natural and close-to-natural climate on the pigment concentrations in both the current and previous year's needles of Norway spruce. In addition, we present an explanation of the quantitative relationships existing between photosynthetic pigments, and light and temperature.


Plants and growth conditions

Four-year-old-spruce trees (Picea abies (L.) Karst.) representing plants of a vegetatively propagated clone were obtained from a local nursery (Staatliche Samenklenge, Laufen, Germany). Plant growth condition were as described in detail by Fischbach et al. (1999). Approximately 8 months before the experiment, the trees were planted in aluminium-wrapped pots, 24 cm in diameter, containing 7 L Molasse sand soil (horizone II Bg, from a depth of > 85 cm; originating from a spruce stand near Freising, Germany) covered with white quartz sand. The plants were transferred to experimental sites before bud break. The trees were irrigated regularly to keep the soil water potential between −0.5 and −0.9 MPa (corresponding to a relative soil humidity of 23 ± 5%). Under these conditions, the typical biomass ratio of shot/root was 2.7. Plants were sprayed with Apollo TM (Schering, Berlin, Germany) in April to prevent mite infestation, and fertilized in April and June by applying 50 mL per pot of a solution containing 60 mm NH4NO3, 40 mm KH2PO4, 10 mm MnSO4, 30 mm MgS04, and 10 g per pot of Osmocote, a commercial slow-release fertilizer (Spiess & Sohn, Kleinkarlbach, Germany).

From May to October, 1998, the trees were cultivated at three different experimental sites situated in southern Germany (47.3°N, 11.1° E) providing different climate regimes: a field site (FLD) located 800 m above sea level represented natural climate conditions, and two dome-shaped greenhouses provided close-to-natural conditions at an altitude of 1780 m (‘solar domes’; see also Zimmer et al. 2000). The solar domes were covered with UV-B transparent Plexiglas (4 mm, GS 2458; Röhm, Darmstadt, Germany). Radiation conditions differed between domes because one (SD1) was covered inside with a UV-B-transparent 25 µm Teflon film inlay (ET 6235; Nowofol, Siegsdorf, Germany) and the other (SD2) with a 50 µm polyester film inlay (Folanorm; Folex, Dreieich, Germany) screening out the UV-B portion of natural radiation.

From six trees per experimental site, two exposed twigs per site were randomly selected, harvested between 0900 and 1100 h on the days indicated in Fig. 1, and kept for 24 h in a moist and dark container at room temperature. Subsequently, CYN and LYN were collected separately under dim light, pooled for each experimental condition and divided into portions of about 100 mg fresh weight. Samples were frozen in liquid nitrogen and stored at −80 °C until high-performance liquid chromatography (HPLC) analysis.

Figure 1.

Climate parameters. Panel (a) shows spectral irradiances of the three study sites of the 1998 experiment: FLD, field; SD1 and SD2, solar dome 1 and 2. Panels (b) and (c) are irradiances in the UV range (280–400 nm) and biologically weighted irradiances calculated according to Caldwell et al. (1983), respectively, derived from spectra in (a). Panels (d) and (e) record PAR and T, respectively, and correspond to the running average of the preceding 10 d. Running averages were derived from daily means. A daily mean corresponds to the average of half-hourly means measured during the 8 h interval centred at solar noon. Dashed lines, FLD; solid black, SD1; solid grey lines SD2. Open and closed triangles indicate the time of harvest from solar domes only and from all three experimental sites, respectively.

Means of air temperatures were measured half-hourly at the field and solar dome site using HP-100-A (Imko, Ettlingen, Germany) and HMP 143 sensors (Vaisala, Helsinki, Finland), respectively. Spectral irradiance was recorded at solar noon ± 1 h under cloudless skies on 18 and 19 October 1997 using a spectroradiometer with double monochromator (Bentham DTM 300; Bentham, Reading, UK). Both, half-hourly means of global radiation and photosynthetically active radiation (PAR) were monitored in the SD1 and SD2 using solar radiation sensors (pyranometer CM 11; Kipp & Zonen, Hamburg, Germany) and quantum sensors (Li-190SA; LiCor Inc., Lincoln, NE, USA), respectively. At the FLD site, only solar radiation was measured and PAR was estimated from the corresponding relations between PAR and solar radiation in SD1 and SD2.


Approximately 70 mg of liquid nitrogen-cooled needles were ground to a fine powder with about 70 mg of a quartz sand–CaCO3 mixture (1 : 1, w/w) in a 1.5 mL reaction tube using a glass pestle which had both been immersed in liquid nitrogen to prevent sample thawing. Pigments were extracted from the powdered needles at 4 °C using 400 µL pure acetone (HPLC-grade; Applichem, Darmstadt, Germany) and the suspension was centrifuged at 0 °C. The pellet was extracted once more with 200 µL acetone at room temperature followed by centrifugation at 0 °C. Prior to chromatography, the pooled supernatants were filtered at 0 °C using 0.45 µm nylon filters (Micro-Spin centrifuge filter; Roth, Karlsruhe, Germany).

Pigments were separated on a 5 µm particle Allsphere ODS-1 non-endcapped RP18 column of 250 × 4.6 mm inner diameter, thermostated at 20 °C (Alltech, Deerfield, IL, USA) using an injection volume of 20 µL and a flow rate of 2 mL min−1 (Gilmore & Yamamoto 1991). Chromatography started with 10 min isocratic elution using pure solvent A (acetonitrile : methanol : H2O : 0.1 m Tris buffer pH 8.0 (72 : 28 : 6 : 3, v/v)), followed by a 4 min linear gradient from 100% solvent A to 100% solvent B (methanol : n-hexane (9 : 2, v/v)). After another 8 min isocratic elution, the starting conditions were restored during a 2 min gradient followed by column equilibration for 4 min. HPLC grade solvents (AppliChem or Eurolab, Darmstadt, Germany) were used.

The HPLC system consisted of a solvent degasser (Model ERC-3312; Erma Inc, Tokyo, Japan), two solvent pumps (Model 510; Waters, Eschborn, Germany) controlled by a Waters automated gradient controller, a Peltier-cooled column oven (Gynkotek, Germering, Germany), and a diode array detector (Model 1000S; Applied Biosystems, Weiterstadt, Germany). The HPLC eluants were monitored at 440 nm.

System calibration for chlorophylls used solutions of chloroplast pigments in buffered 80% acetone (v/v) with known Chl a and Chl b concentrations determined according to Porra, Thompson & Kriedemann (1989); other calibrations utilized lutein (lut), violaxanthin (vio) and neoxanthin (neo) isolated by thin-layer chromatography (Egger 1962), and commercially available standards, α-carotene (α-car; The International Agency for 14C Determination, Hørsholm, Denmark) and β-carotene (β-car; Fluka, Deisenhofen, Germany). The dissolved carotenoids were quantified using published extinction coefficients (Jeffrey, Mantoura & Bjørnland 1997). Antheraxanthin (ant) and zeaxanthin (zea) were identified by their retention times, spectral properties and increased concentrations observed under high-light conditions in various plant leaves (Gilmore & Yamamoto 1991; Jeffrey et al. 1997); the latter xanthophylls were quantified using the data obtained with lutein as standard. The cis- and trans-isomers of β-car were differentiated by spectral properties and retention times (Siefermann-Harms 1994; Bialek-Bylka et al. 1995).


For statistical analyses, we utilized Sigma Stat for Windows Version 2.03 statistical software (SPSS, München, Germany). To test whether the different treatments produced statistically significant differences, one way analysis of variance (anova) was used: significance was defined as P < 0.05. If anova yielded P < 0.05 then the Student–Newman–Keuls method was used to determine which data groups were different; significant difference was defined as P-values < 0.05.


Spruce needles from three experimental sites were investigated: the light climates of these sites are characterized in Fig. 1. Natural radiation conditions were present at the field site (FLD, Fig. 1a). In solar dome 1 (SD1), natural radiation was also present but at reduced intensities. In solar dome 2 (SD2), radiation was similar to SD1 but the UV-B spectral range was screened out. Irradiance in the UV range in SD1 and SD2 was 70 and 50%, respectively, of that under FLD conditions (Fig. 1b). Biologically effective radiation (calculated according to Caldwell et al. 1983) was 60 and 0% of the FLD value under SD1 and SD2 conditions, respectively (Fig. 1c). In SD1 and SD2, the values of PAR and temperature (T) were almost similar but lower and higher, respectively, than under FLD conditions (Fig. 1d & e).

Figure 2 shows the behaviour of Chl a expressed as nmol/mg fresh weight (Chl a FW). In both solar domes, Chl a FW in current year's needles (CYN) increased markedly during June and July but exhibited a comparably moderate rise in last year's needles (LYN) during this period (Fig. 2). The Chl a FW between SD1 and SD2 conditions differed frequently but these differences were unsystematic and the SD1 and SD2 treatments did not result in statistically significant differences (see Material and methods for information on statistics used). In comparison with the solar domes, moderate changes in Chl a FW occurred under FLD conditions. In both needle-age classes, the FLD treatment resulted in significantly lower Chl a FW than either of the solar dome conditions. With the exception of the minor xanthophylls ant and zea, concentrations/fresh weight of all other pigments were linearly related to Chl a FW: coefficients of determination of linear regressions (R2) ranged between 0.68 and 0.96; P-values were always < 0.001 (data not shown).

Figure 2.

Chlorophyll a content. Chlorophyll a per fresh weight (Chl a FW) during the 1998 experiment is given for current year's needles (CYN) and last year's needles (LYN) in (a) and (b), respectively. Bars indicate standard errors; 3 ≤n≤ 8. Circles, FLD; inverted triangles, SD1; upright triangles, SD2.

Pigment stoichiometry was studied by expressing pigment concentrations relative to Chl a (Fig. 3). In general, differences between corresponding data sets from SD1 and SD2 were statistically insignificant. With the exception of neoxanthin/Chl a (neo/Chl a; Fig. 3o & p), significant differences existed between FLD conditions and solar domes.

Figure 3.

Pigment concentrations normalized to chlorophyll a. Molar ratios of different pigments per chlorophyll a of the 1998 experiment are shown. Panels below ‘CYN’ depict data from current year's needles, other panels shown data from last year's needles (LYN). See Fig. 2 for symbols and error bars. α-car, α-carotene; β-car, β-carotene; ant, antheraxanthin; Chl a, chlorophyll a; Chl b, chlorophyll b; lut, lutein; neo, neoxanthin; vio, violaxanthin; zea, zeaxanthin.

In the solar domes, Chl b/Chl a in CYN decreased steeply during June and July and exhibited a tendency to increase thereafter but Chl b/Chl a in LYN changed only moderately (Fig. 3a & b). Both needle-age classes showed smaller Chl b/Chl a ratios under FLD conditions than under solar dome conditions. Similar to Chl b/Chl a, in solar domes lutein/Chl a (lut/Chl a) steeply decreased during June and July in CYN but exhibited moderate variations in LYN (Fig. 3c & d). Under FLD conditions, lut/Chl a was higher than in solar domes. In solar domes, both needle-age classes exhibited minimum values for β-carotene/Chl a (β-car/Chl a) around August; β-car/Chl a was increased under the FLD treatment relative to solar domes (Fig. 3e & f). The patterns of α-carotene/Chl a (α-car/Chl a) resembled the inverted pattern observed for β-car/Chl a (compare Fig. 3e& f with Fig. 3g & h).

Xanthophyll cycle pigments per Chl a are depicted in Fig. 3(i–n). Note, that here we are dealing with completely dark-adapted needles in which the presence of ant and zea reflects the inaccessability of these pigments to the epoxidase enzyme rather than the light-driven de-epoxidation by the violaxanthin de-epoxidase (Adams et al. 2001). Generally, needles grown at the FLD site showed higher contents of xanthophyll cycle pigments per Chl a than needles from solar domes. In CYN under FLD conditions, these pigments varied markedly but the time courses of variations differed: violaxanthin/Chl a (vio/Chl a) decreased from July till November, antheraxanthin/Chl a (ant/Chl a) increased during July, and zeaxanthin/Chl a (zea/Chl a) increased between end of July and middle of August (Fig. 3i, k & m). At all three experimental sites, neo/Chl a in CYN decreased from June to the middle of August and recovered slightly thereafter; neo/Chl a in LYN remained rather constant (Fig. 3o & p).

In a separate experiment, we investigated CYN of plants exposed for 2.5 months in SD1 with individual branches wrapped in filter foils which either transmitted all natural radiation, or screened out UV-B, or UV-B and short-wavelength UV-A, or all UV radiation (data not shown). By comparing all data after filter treatment with those prior to exposure, we observed statistically significant increases for lut/Chl a, β-car/Chl a, ant/Chl a and zea/Chl a, and a statistically significant decrease for α-car/Chl a. Differences between filter treatments, however, were not statistically significant.

Subsequently, we concentrated on the effects of PAR and T on pigment parameters of Figs 2 and 3. An essential consideration is that needles have to be exposed to these climate factors for an unknown interval (Δt) to acclimate. We sought for the value of Δt which yields the best correlation between pigment and climate within a range of 2–50 d with 2 d increments. For each Δt, means of PAR and T prior to needle harvest (PARΔt and TΔt, respectively) were calculated for each date of harvest for solar dome and FLD conditions. The PARΔt and TΔt were calculated with daily means which were derived from measurements during 8 h intervals centred at solar noon. Pigment data from CYN and LYN, respectively, were plotted as a function of each set of PARΔt or TΔt (plots not shown). The different relations between pigment parameters (Pi) and PARΔt were analysed by linear regression according to:


and between Pi and TΔt according to


where A0 and A1 denote ordinate intercept and slope, respectively [values for A0 and A1 in Eqns 1 and 2 are different].

At early dates of harvest, long Δt extended into the range for which climate data were not recorded and, hence, climate means could not be calculated (see Fig. 1). This results in varying numbers of data points in pigment versus climate plots depending on the time interval considered. Therefore, P-values, which take into account the actual number of observations, rather than R2, were always used to compare linear regressions. The lowest P-values, obtained using the equations above, together with R2 and sign of slope of linear regression are compiled in Table 1. The Δt which resulted in lowest P can be read from Fig. 4 as will be explained below.

Table 1.  Parameters of linear regressions representing best correlation between pigment and climate
Age class
Dependence on PARDependence on TDependence on PAR/(TTOFF)Minimum improvement
  1. Column 1 identifies pigment parameters and column 2 the needle-age class analysed. The table compiles the lowest P-values and corresponding R2 from linear regression analyses of three types of pigment versus climate plots as defined in Eqns 1–3 (the type of correlation is also given: Pos, positive; Neg, negative; NS, not statistically significant, that is P > 0.05). For all types of plots, pigment data were taken from Figs 2 and 3. Climate data were means of time intervals prior to needle harvest (Δt) of PAR, T or SQ[=PAR/(TTOFF)]; corresponding regression data are listed in columns 3–5, 6–8, and 9–11, respectively. Best correlation was obtained by varying the Δt in the case of PAR and T, and the Δt and TOFF in the case of SQ. Column 12 compiles quotients of smallest P-value of pigment versus PAR or versus T plots divided by the smallest P-value when pigment is plotted versus SQ: numbers > 1 indicate better correlation between pigment and SQ compared to PAR or T; if numbers were < 5 and > 10, parameters were classified as SQ-independent and SQ-dependent, respectively. In the case of SQ-independent parameters, the ranges of Δt at which pigment versus PAR plots yielded lowest P-values correspond to the Δt of lowest P-values in contour plots (Fig. 4); for all parameters, the Δt and TOFF of areas which produced lowest P-values in pigment versus SQ plots are shown in Fig. 4.

Chl a FWCYN6.52e−040.527Neg2.60e−030.488Pos1.23e−030.537Neg0.5
Chl b/Chl aCYN3.58e−040.668Neg2.03e−030.562Pos1.91e−040.700Neg1.9
lut/Chl aCYN6.42e−050.749Pos1.23e−040.721Neg2.80e−070.857Pos229.3
β-car/Chl aCYN2.11e−030.559Pos1.34e−040.717Neg6.70e−070.795Pos200.0
α-car/Chl aCYN1.07e−030.498NS3.69e−040.608Pos1.92e−050.691Neg19.2
neo/Chl aCYN1.15e−010.148NS1.70e−020.309Neg1.28e−010.139NSn.s.
vio/Chl aCYN4.99e−040.650Pos4.08e−030.511Neg1.16e−040.666Pos4.3
ant/Chl aCYN1.72e−040.705Pos2.90e−030.536Neg1.76e−060.769Pos97.7
zea/Chl aCYN2.80e−020.266Pos1.80e−020.383Neg1.01e−040.622Pos178.2
Figure 4.

Dependence of P-values on Δt and TOFF. Contour plots of P-values are shown. P-values results from linear regression analyses of the pigment parameters depicted in Figs 2 and 3 plotted against SQ[= PAR/(TTOFF)]. The SQ was calculated for a range of temperature offset (TOFF; ordinates). In addition, means of SQ for different time intervals (Δt) before needle harvest were considered (abscissas; see Eqn 3 and text). Panels below ‘CYN’ depict analyses of current year's needles, the remaining data are derived from last year's needles (CYN). Pigment parameters are identified in Figs 2 and 3. P-values of bold contour lines are indicated as exponents of 10 (negative integers). In regions of minimum P-values, thin lines show profiles of P-values at higher resolution (corresponding to nine equal intervals between bold lines). In panels (a), (l) and (n), the P-values given fully refer to thin lines. For all panels, minimum P-values, coefficients of determination and type of relation (positive or negative) are listed in Table 1.

The R2 obtained with Eqns 1 and 2 show that the degree to which variations in PAR or in T explain pigment behaviour varied largely (Table 1). Despite these variations, two general features were observed: (1) if for a given pigment parameter P < 0.05 in CYN and LYN, then both age classes showed the same type of dependence on PARΔt or TΔt (positive or negative); and (2) if P < 0.05 for both, the correlation between pigment and PAR and between pigment and T, then a positive dependence on PARΔt was always paralleled by a negative dependence on TΔt. and vice versa.

The latter observation indicates opposite action of PAR and T on pigments and caused us to analyse their combined action by considering the ratio between PAR and T. The zero point of the temperature scale, however, which is important for pigment response is unknown. Hence, we define a ‘stress quotient’:


and searched for the temperature scale that yields best correlation between pigment and SQ by varying the temperature offset (TOFF) from 240 to 280 K in 1 K increments: values > 280 K produced negative temperature differences (TTOFF) and were not used. For each TOFF, a range of Δt was considered as described above. Pigment parameters were plotted (not shown) against the various SQ followed by linear regression analyses according to:


where A0 and A1 denote ordinate intercept and slope, respectively, which assume different values than the A0 and A1 in Eqns 1 and 2.

The latter procedure yields P-values as function of Δt and TOFF which are depicted as contour plots in Fig. 4; the lowest P-values in these plots together with R2 and sign of slope of linear regression are listed in Table 1. Each contour plot provides two principal types of information: (1) the TOFF at which minimum P-values occur: high values for TOFF indicate high temperature sensitivity of the parameter as the contribution of variations in T to variations in SQ increase with TOFF (see definition of SQ); and (2) the Δt of a minimum P-values indicates the time interval that an average SQ has to exist to produce the maximum response of a pigment parameter.

To classify pigment–climate interaction, the ratio of the smallest P-value obtained when only PAR or T was considered divided by the P obtained with SQ was used. These ratios were < 5 in the case of Chl a FW and Chl b/Chl a in CYN, and of neo/Chl a and vio/Chl a in both needle-age classes (Table 1), indicating that the SQ was not convincingly better suited to describe pigment behaviour than PAR or T alone. The latter pigment parameters are denoted as SQ-independent. Still, these SQ-independent parameters exhibited P-values < 0.05 in contour plots except neo/Chl a in CYN (Fig. 4). Consistent with low sensitivity to T, contour plots of SQ-independent parameters failed to show defined areas of lowest P-values located at high TOFF. Because the ranges of Δt of minimum P-values in contour plots agreed reasonably well with the Δt which gave best correlation between pigment and PAR (data not shown), Fig. 4 demonstrates that response times of SQ-independent pigment parameters extended well into the range above 30 d.

For all other pigment parameters, SQ describes pigment behaviour with at least 10-fold improved P-values than the best P obtained with PAR or T alone, and the R2 was increased by 10–100% (Table 1). These ‘SQ-related’ pigment parameters exhibited reasonably defined areas of minimum P which are often curbed by the upper limit of our range of TOFF and located at Δt < 20 d except the lut/Chl a which responded maximally to climate at around 25 d (Fig. 4).


Our work suggests that UV-B radiation has little influence on pigments in spruce needles: UV-B radiation was present under SD1 and screened out under SD2 conditions (Fig. 1) but, in both needle-age classes, neither absolute pigment concentrations as represented by Chl a FW nor pigment concentrations normalized to Chl a differed between SD1 and SD2 to a statistically significant degree (Figs 2 & 3). In various flowering plants, UV-B decreased Chl a FW or chlorophyll a+b per fresh weight (Jordan et al. 1994; Correia et al. 1999; Lingakumar, Amudha & Kulandaivelu 1999; Yuan et al. 2000); also, UV-B has been reported to increase Chl b/Chla (Strid et al. 1990; Pfündel et al. 1992; Jordan et al. 1994; Deckmyn & Impens 1995). These studies were carried out under a wide range of conditions which also include UV-exclusion in field experiments. An obvious explanation for the lack of UV-B effects on pigments in spruce needles is the considerably higher UV screening capacity of epidermal layers in conifer needles in comparison with leaves of most flowering plants (Day et al. 1992; Day 1993; Sullivan et al. 2000); thus, UV-B intensities inside needles are most likely too low to induce any significant alterations in absolute and relative pigment contents. Certainly, this conclusion is limited to UV-exclusion experiments in solar domes in which moderate UV-B levels (Fig. 1) could have obscured any existing effect of ambient UV-B radiation.

PAR and T have been demonstrated to induce acclimation of photosynthetic pigments in conifers (Adams & Demmig-Adams 1994; Siefermann-Harms 1994; Ottander, Compbell & Öquist 1995; Savitch et al. 2002). Here, we introduced a novel approach to quantify the relationship under natural and close-to-natural conditions between pigments and the two climate parameters. The new aspects of our analysis are (1) consideration of varying time intervals preceding needle harvest of natural or close-to-natural climate; and (2) introduction of the ‘stress quotient’SQ (= PAR/(TTOFF); see Results) to take into account the combined action of PAR and T on pigments.

Of our 18 data sets, 12 are better predicted by SQ than by PAR or T alone (Table 1). The SQ-related pigment data include Chl a FW and Chl b/Chl a in LYN, and all carotenoids per Chl a in both needle-age classes except the most oxidized xanthophylls, namely, violaxanthin and neoxanthin. Dependence on SQ is consistent with the involvement of chloroplast redox poise in the control of expression of photosynthetic pigment genes (Huner et al. 1998; Pfannschmidt et al. 2001): those pigment parameters not clearly associated with SQ are probably controlled by a different mechanism.

The SQ-independent parameters include Chl a FW and Chl b/Chl a in CYN, neo/Chl a in LYN, and vio/Chl a in both needle-age classes. Contour plots of SQ-independent data differed characteristically from plots of SQ-related data: only the latter exhibited defined areas of minimum P-values mostly located Δt < 20 d but the former showed smallest P-values well above 30 d indicating a more sluggish response to climate of SQ-independent pigment parameters (Fig. 4).

The Chl a FW and Chl b/Chl a were SQ-independent in CYN and SQ-related in LYN (Fig. 4, Table 1) and exhibited much higher amplitudes in CYN than in LYN (Figs 2, 4a & b). Increases in Chl a FW and Chl b/Chl a imply increases in the abundance of entire photosystems and in the size of LHCII, respectively, as virtually all Chl a is organized in pigment-protein complexes and LHCII binds most of Chl b (Siefermann-Harms 1985; Anderson et al. 1995); synthesis of photosynthetic pigment-proteins therefore appear higher in CYN than in LYN. Delayed response to climate of Chl a FW and Chl b/Chl a in CYN might indicate that these needles only invest the potentially high energy costs in synthesis of photosynthetic pigment-proteins after stable climate changes have occurred; that these two parameters react in a faster SQ-related way in LYN can be explained by their lower potential for acclimation which implies moderate costs for protein synthesis.

Because all major carotenoids per fresh weight changed in parallel to Chl a FW, the switch to a SQ-type response from CYN to LYN basically affects all pigment concentrations. In addition to these changes, most carotenoids/Chl a vary in a SQ-related way in both needle-age classes. According to our idea presented above, the short response times observed for the carotenoids/Chl a would suggest that enrichment or exchange of carotenoids in pigment-protein complexes involves no major costs for protein synthesis.

Among all carotenoids per Chl a, only α-car/Chl a was negatively related to SQ (Table 1) indicating that α-car is enriched in photosystems when PAR limits photosynthesis during most of the daily light period. Shade conditions have been reported to stimulate synthesis of α-car (Grill & Pfeifhofer 1985; Thayer & Björkman 1990; Königer et al. 1995; Demmig-Adams & Adams 1996; Logan et al. 1996; Demmig-Adams 1998; Krause et al. 1999). Increased α-car, however, was observed under summer conditions (Grill & Pfeifhofer 1985; Cardini 1983; Adams & Demmig-Adams 1994; Siefermann-Harms 1994) suggesting that high-light favours synthesis of this carotenoid. These apparently different effects of PAR on α-car are not necessarily contradictory because SQ, but not PAR alone, is important for α-car: for instance, low SQ can result from relatively high PAR at elevated T as well as from low PAR at any temperature permitting energy consumption by photosynthesis at reasonable rates.

The parameter, β-car/Chl a, was positively related to SQ, and the sum of α-car/Chl a plus β-car/Chl a varied to a much lesser degree than α-car/Chl a or β-car/Chl a (Fig. 3e–h). This is consistent with replacement of α-car by β-car in some plant species as has been suggested by Demmig-Adams (1998). The significance of this putative carotene exchange is unclear. As carotenes are mostly located in the core complexes of both photosystems (Siefermann 1985; Young & Britton 1989; Green & Durnford 1996) the effect of carotene exchange, if any, should influence the properties of these complexes. Quenching by β-car of singlet oxygen formed in isolated PSII reaction centres has been demonstrated (Telfer et al. 1994); further, β-car exhibits markedly higher antioxidant activity than α-car (Stahl, Sies & Sundquist 1994; Miller et al. 1996). Hence, exchange of α-car for β-car under high SQ could increase scavenging of free radicals produced in core complexes under conditions of high SQ. In turn, enrichment of α-car at the expense of β-car under low SQ could imply a so-far unidentified function of α-car in improving light-harvesting.

The positive relations between lut/Chl a, ant/Chl a or zea/Chl a and SQ are consistent with other data from conifers showing that these xanthophylls are positively correlated to PAR or negatively correlated to T (Adams & Demmig-Adams 1994; Adams et al. 2001; Savitch et al. 2002). Among all parameters investigated, the ant/Chl a and zea/Chl a in both needle-age classes, and the lut/Chl a in LYN responded fastest to SQ with Δt < 10 d (Fig. 4f and Fig. 4o–r). In our dark-adapted needles, these three xanthophylls are presumably bound to the light-harvesting complexes of photosystems where they probably provide sustained protection against photo-inhibition by dissipating excess absorbed light energy (Niyogi, Björkman & Grossman 1997; Adams et al. 2001). The particularly short response times observed for these xanthophylls suggest that variation of sustained energy dissipation is one of the first steps in acclimation of photosystems to natural climate changes.

In CYN, the lut/Chl a exhibited the slowest response to climate among SQ-dependent carotenoids (Fig. 4e). Furthermore, in CYN, we observed slow response of Chl b/Chl a indicating slow LHCII acclimation (Fig. 4c). Kühlbrandt, Wang & Fujiyoshi (1994) have shown that lut forms a structural component in LHCII; therefore, the structural role of lut might link lut/Chl a to the levels of LHCII and delay response to climate change in CYN.

The reason for SQ-independence of vio/Chl a in CYN (Fig. 4m) is uncertain: it is probably different from the slow response of lut/Chl a because a structural role of vio in LHCII appears less likely as a significant part of vio can be de-epoxized in conifers (Adams & Demmig-Adams 1994; Savitch et al. 2002) and it seems therefore that most vio is not tightly associated with LHCII. The neo/Chl a in both needle classes and the vio/Chl a in LYN were only weakly related to our climate parameters (Fig. 4 and Table 1). It is difficult therefore to categorize their relation to climate. Possible reasons for weak associations between the latter xanthophyll data and climate could be their markedly non-linear relation to climate or their sensitivity to environmental parameters not investigated here.


Much of the variations of photosynthetic pigments in spruce needles can be attributed to variations in PAR and T but they were independent of UV radiation: pigments either responded to PAR without exhibiting marked sensitivity to T, or to the combined action of PAR and T; which type of response prevailed was partly dependent on the developmental state of the needle. Two types of response of pigments to climate were also observed in developing barley leaves by Montanéet al. (1998). These authors, however, reported that the sum of xanthophyll cycle pigments relative to chlorophyll responded to the redox state of the chloroplast which is consistent with dependence on PAR and T. Considering that vio generally constitutes more than 80% of the xanthophyll cycle pigments in spruce needles (Fig. 3), the above finding of Montanéet al. (1998) contradicts the mostly PAR-dependent acclimation of vio/Chl a observed in spruce (Fig. 4m). Such difference could well reflect different acclimation strategies used in various plants (Savitch et al. 2002). Further studies are needed to better understand the full spectrum of plant acclimation strategies in nature.


This work was supported by a grant from the state of Bavaria (BayForUV program). We are grateful to Bob Porra for help in preparing the manuscript.