1 This paper describes a study of 23 plant species native to the British Isles, to investigate the relative importance of growth conditions and ecological origin in determining the content and composition of the photosynthetic pigments (chlorophylls and carotenoids) within leaves.
2 The species studied reflected a range of ecological types, from deep shade to full sun.
3 Plants were grown in two light environments: high light (HL), filtered through a clear filter; and low light (LL), filtered through a neutral density filter which reduced total irradiance with enhanced far-red.
4 Plants grown at high irradiance contained more carotenoid per unit chlorophyll and showed a marked increase in xanthophyll cycle pigments relative to other carotenoids.
5 Deep-shade plants were slightly less responsive to changes in growth light conditions than plants native to less shaded locations. Intermediate sun/shade plants were slightly more responsive; however these differences were small compared with the extent of the response of all species to HL vs LL.
6 The main conclusion is that that the light conditions to which the plants are exposed are more important than the genetic predisposition of those plants in determining the pigment content and composition of leaves.
The chloroplasts of higher plants normally contain two major groups of pigments: the chlorophylls (a and b) and the carotenoids. Although many different carotenoids occur in plants, only six are commonly found in the photosynthetic apparatus and are normally present in leaves: carotene (mostly in its β form), and five xanthophylls (neoxanthin, lutein, violaxanthin, antheraxanthin and zeaxanthin) (Young 1993). The latter three xanthophylls can be reversibly interconverted in an enzymatic process known as the xanthophyll cycle. In this, violaxanthin is de-epoxidated, via antheraxanthin, to zeaxanthin at high photon flux densities (PFDs), a process which is reversed when the plant is returned to low PFD (Sapozhnikov, Krasovskaya & Mayevskaya 1958; Yamamoto, Nakayama & Chichester 1962). The chloroplast pigments are thought to be associated mostly with proteins in the thylakoid membrane.
Although the complement of different pigments is conserved between species, the relative composition of these varies significantly among and within species in response to developmental conditions. Despite this variation being well documented, the reasons for it remain obscure. Clues to the function of different pigments may come from studying the relationship between pigment composition and the ecological history of plants. One common approach has been to look at the effects of irradiance on pigment composition (Demmig-Adams et al. 1989; Demmig-Adams & Adams 1992; Logan et al. 1996; Thayer & Björkman 1990). Although there is some variation, it is generally agreed that plants growing at higher irradiance have an increased chlorophyll a : b, more of each carotenoid (except neoxanthin) per chlorophyll, and substantially increased xanthophyll cycle pigments as a percentage of total carotenoid.
Given the above acclimative responses to irradiance, pigment composition may also be of adaptive importance, that is, a particular pigment composition confers a selective advantage in terms of sun or shade tolerance. Thayer & Björkman (1990) compared the pigment compositions of 10 sun species grown in full sun with those of nine shade species grown in shade. Shade species had a much smaller xanthophyll cycle pool and a smaller chlorophyll a : b than sun plants. In the studies of Johnson et al. 1993a; Johnson et al. (1993b) a strong correlation was found between the ‘shade tolerance’ (a scale based on species’ tendency to grow in woodland) of 22 plant species, and both xanthophyll cycle pool (which decreased with shade tolerance) and lutein (which increased with shade tolerance). Such studies support the idea that pigment composition is of adaptive importance, but are problematic in that plant material was taken from plants growing in their native (or similar to their native) light conditions. Hence it is not clear whether the patterns observed were due to adaptation, or were merely the result of acclimation to the conditions in which the plants were growing. The aim of the present study was to examine separately, for a range of British plant species, the relative contributions of adaptation and acclimation to differences in pigmentation.
Materials and methods
Plants of 23 species native to the British Isles were grown from either seed or rhizomes, or were collected as young plants from field sites in the UK (Table 1). Seeds were germinated in covered pots filled with Levington M3 compost. Once large enough to handle, the seedlings (or, where appropriate, rhizomes or collected plants) were transferred to pots containing Levington M2 compost. Plants were then grown on until the second set of mature leaves developed, at which time seven to ten plants of each species were placed under the two light treatments. Plants continued growing under these conditions until they were harvested.
Table 1. Native British plant species used in this study
Source of plants: (a) seeds provided by NERC Unit of Comparative Plant Ecology, Sheffield University; (b) plants collected from sites in Oxfordshire; (c) seeds provided by Dr Robin Walters, University of Sheffield; (d) seeds from John Chambers, Kettering, Northants NN15 5AJ; (e) collected locally; (f) seeds collected locally; (g) rhizomes/bulbs from John Chambers, Kettering; (h) rhizomes collected from sites in Oxfordshire.
Winter or summer annual therophyte, overwinters as rosette
Mercurialis perennis L. (Euphorbiaceae)
Perennial protohemicryptophyte, some shoots may overwinter
Senecio vulgaris L. (Compositae)
Summer or winter annual therophyte, leaves may overwinter
Silene dioica (L.) Clairv. (Caryophyllacae)
Perennial protohemicryptophyte or chamaephyte, wintergreen
Sinapis arvensis L. (Cruciferae)
Summer (or winter) annual semi-rosette therophyte, may overwinter.
Taraxacum agg. (Compositae)
Perennial rosette hemicryptophyte, wintergreen
Urtica dioica L. (Urticaceae)
Perennial chamaephyte, some shoots may overwinter
Urtica urens L. (Urticaceae)
Summer annual therophyte
The two different light treatments were produced by growing the plants in a glasshouse under lighting filters (Lee Filters, Andover, UK; Fig. 1) Supplementary light was provided by 400 W metal halide bulbs with a 16 h photoperiod. The treatments were: high light (HL), clear filters (no. 130) used to expose the plant to full sunlight; and low light (LL), neutral density filters (no. 211) used to reduce all wavelengths equally over the visible range. These filters are largely transparent to far-red illumination, and so produce a far-red enhanced light environment similar to that seen under a leaf canopy.
The irradiance of the LL treatment was 10% of that achieved in the HL treatment (as measured using a Skye quantum sensor, spectral range 400–700 nm; Skye Instruments, Llandrindod Wells, UK). All plant material was grown between March and May. Although there was considerable day-to-day variation in irradiance during the growth period, a maximum mid-day PFD of 1200 µmol m−2 s−1 was recorded in the HL treatment during the growth period (120 µmol m−2 s−1 in the LL treatment).
Before harvesting, plants were dark-adapted for 17–18 h to allow complete epoxidation of zeaxanthin into violaxanthin. Two leaf disks (0·4–2·7 cm diameter, depending on leaf size), or similar-sized leaf segments (where disks could not be cut) were harvested from the youngest fully expanded leaves of each plant. One disk was immediately frozen in liquid nitrogen. The second was light-treated before freezing. Light treatment involved floating leaves on water while illuminating the disk with 2500 µmol m−2 s−1 white light for 20 min, to induce maximal physiological de-epoxidation of violaxanthin. All leaf material was stored at −80 °C before pigment analysis.
Leaf disks from four to ten plants per species were analysed for pigment content and composition. Analysis of pigment content for each leaf disk was carried out using reversed-phase high-performance liquid chromatography (HPLC) (Barry, Young & Britton 1990; Rosevear, Johnson & Young 1998).
Pigments were extracted from frozen leaf disks by rapidly grinding with 2 ml re-distilled acetone in a pestle and mortar. If carotenoid breakdown was otherwise observed, leaves were ground in the presence of a small amount of sodium bicarbonate. The pigment extract was then filtered, firstly through sintered glass (porosity = 3), then using a GHP Acrodisc 13, 0·45 µm syringe filter (Gelman Sciences, Ann Arbor, MI, USA) and dried under a gentle stream of oxygen-free nitrogen (BOC, Guildford, UK). Pigments were then re-suspended in 100–1000 µl re-distilled acetone, depending on sample size. Samples were injected into a Spherisorb ODS2 (250 × 4·6 mm, 5 µm) column via either a model 7725i rheodyne valve (Rheodyne, Rohnert Park, CA, USA) or a Hewlett Packard model 1050 autoinjection system. A linear gradient between 100% of a 9 : 1 (v/v) acetonitrile/water solution and 100% ethyl acetate was generated using a Gynkotech Model 450 HPLC pump with a Jour Research X-ACT four-channel degasser over 30 min at a flow rate of 1 cm3 min−1. Pigments were detected using a Hewlett Packard 1040 diode-array detector, and integrals were calculated at wavelengths approximating to the λmax of each pigment (431, 437, 441, 447, 455 nm). Calibration of the column for the pigments was carried out using standards which, with the exception of zeaxanthin (obtained from Roche, Basel, Switzerland), were purified by thin-layer chromatography (TLC) from market spinach. These standards were quantified in ethanol on a Cecil CE5501 dual beam UV/Vis spectrophotometer, using the extinction coefficients of Britton, Liaaen-Jensen & Pfander (1995). Because it was not possible to distinguish α and β carotene in some samples, in the subsequent discussion these are considered together.
Estimations of the ratio of chlorophyll a : b were made by extracting pigments in 80% acetone v/v. Chlorophyll was then estimated using the method of Porra, Thompson & Kriedemann (1989). The total chlorophyll concentration (per unit area or fresh weight) estimated in this way did not differ significantly from that estimated by HPLC; however, the ratio of chlorophyll a : b did differ in some cases due to partial breakdown of chlorophyll. Where carotenoid content is expressed relative to total chlorophyll, this is based on the HPLC estimation.
Where pairs of data were compared, significance was tested using Student’s t-test. Where more than two sets of data were compared, one-way anova was used to test for significant differences between groups, and Scheffé’s post hoc multiple comparison test was used to identify groups that differed from the rest. All tests were carried out using the SPSS software package (SPSS Inc., Chicago, IL, USA). Linear regression was carried out using Microsoft Excel (Microsoft, Seattle, WA, USA). Correspondence analysis was carried out using mvsp (Kovach Computing Services, Pentraeth, UK).
To examine possible relationships between leaf pigment composition and content and plant habitat preference, it was necessary to adopt a rating scheme for the species’ tendency to grow in sun or shade. Most previous studies into the effect of native light on a range of factors in plants have tended to class species simply as either ‘sun’ or ‘shade’. To probe for more subtle variations with habitat preference, some studies (Johnson et al. 1993a, 1993b; Murchie & Horton 1997) have used a scaled system in an attempt to quantify the light environment to which plants are adapted. A number of such scales are available (Ellenberg 1979; Johnson et al. 1993a, 1993b; Murchie & Horton 1997). Here the scale of Murchie & Horton (1997) was used. This scale exploits the large amount of data on habitat preferences that is available from an extensive study conducted by Grime, Hodgson & Hunt (1988). This study provides quantitative data on the occurrence of plant species in different habitat types found in the Sheffield area. The shade association parameter, used previously by Murchie & Horton (1997), is itself a modification of a parameter adopted by Johnson et al. (1993a, 1993b). Both these parameters express the frequency of species presence in woodland quadrats (FW) as a proportion of their frequency in all quadrats surveyed (FA). In the shade association scale, habitat preference is expressed as (FA/FW)/[(FA/FW) + 1], providing a linear scale that allows the use of parametric tests to examine any correlations. A plant that is never found growing in woodland has a shade association of 0, and a plant that is found with equal frequency in woodland and all quadrats will have a shade association of 0·5. Previous analyses using this parameter have shown strong relationships with the capacity to dissipate absorbed energy as heat (high-energy-state quenching; Johnson et al. 1993b) and the ratio of lutein to xanthophyll cycle pigments (Johnson et al. 1993a), as well as with the ability of plants to acclimate to growth irradiance (Murchie & Horton 1997).
Acclimation to growth light
The mean ratios of total carotenoid to total chlorophyll and chlorophyll a : b were each significantly greater in HL-grown than in LL-grown plants for both ratios (Table 2). This was also true for most of the individual species (data not shown).
Table 2. Means over all plant species of leaf pigment concentrations, with standard errors
High light (HL)
Low light (LL)
HL values are all significantly different from corresponding LL values (P < 0·002; n = 70–151).
Total carotenoid : chlorophyll
0·34 ± 0·01
0·26 ± 0·01
Chlorophyll a : b
3·10 ± 0·04
2·76 ± 0·06
Xanthophyll cycle pigments
Percentage total carotenoid
20·3 ± 0·4
14·0 ± 0·4
mol mol−1 chlorophyll
0·070 ± 0·002
0·035 ± 0·001
16·5 ± 0·6
7·4 ± 0·3
Maximum de-epoxidation state (%)
69·5 ± 1·6
55·4 ± 1·8
Percentage total carotenoid
38·8 ± 0·5
45·8 ± 0·5
mol mol−1 chlorophyll
0·129 ± 0·002
0·115 ± 0·003
29·8 ± 1·0
24·6 ± 1·0
Percentage total carotenoid
26·8 ± 0·4
23·8 ± 0·4
mol mol−1 chlorophyll
0·093 ± 0·002
0·063 ± 0·002
22·4 ± 0·9
14·0 ± 0·7
Percentage total carotenoid
14·1 ± 0·2
16·4 ± 0·3
mol mol−1 chlorophyll
0·048 ± 0·001
0·042 ± 0·001
11·4 ± 0·4
9·13 ± 0·5
Relative to chlorophyll, and on a leaf-area basis, all four of the carotenoid classes (carotene, lutein, the xanthophyll cycle pigments and neoxanthin) occurred in significantly greater quantities in HL-grown than in LL-grown plants (Table 2) when averaged over all samples. Relative to each other, cross-species mean concentrations of xanthophyll cycle pigments increased significantly between LL and HL treatments. Total carotene (α + β) as a percentage of total carotenoid exhibited a smaller but still significant increase, while neoxanthin and lutein concentrations decreased significantly. Data for the individual species broadly agree with the cross-species data, with a general increase in all carotenoids per chlorophyll, and in the xanthophyll cycle pool size as a percentage of total leaf carotenoid (data not shown). This occurred largely at the expense of lutein. The only major exception to the above was in Arum maculatum, in which carotenoid concentration decreased significantly when grown in HL. As a percentage of carotenoid (α + β), carotene and neoxanthin in individual species usually responded similarly to the cross-species means.
An important factor in the response of plants to excess light is the capacity to de-epoxidize xanthophyll cycle pigments. The extent of de-epoxidation is typically expressed as de-epoxidation state, defined as 100 × (Z + 0·5A)/(Z + A + V) where Z = zeaxanthin, A = antheraxanthin and V = violaxanthin). There was a large and significant difference in maximum de-epoxidation state between the LL and HL treatments (P < 0·001; Table 2). This response was also seen in most individual species. De-epoxidation state ranged from a maximum of 83·5% in HL-grown Digitalis purpurea to 26·9% in LL-grown Arabidopsis thaliana. The lowest de-epoxidation state in HL-grown plants was 39·9% in Cardamine hirsuta.
Adaptation to light
There was considerable interspecies variation in the mean values for each pigment class, over the full range of shade association, for both treatments (Fig. 2a–d). Despite this, there were no significant correlations between pigment concentration and shade association. In contrast, there was a significant increase in the concentration of individual pigments in most species, when plants were grown under HL, compared with LL, conditions.
In a similar approach to that of Murchie & Horton (1997), the 23 species were split into three groups of seven or nine species, based on their shade association: A (shade association = 0–0·3); B (0·3–0·6); C (0·6–1). When presented in this way, there was usually a significant increase in each carotenoid class relative to chlorophyll, in HL- compared with LL-grown plants (Fig. 2). There were generally no significant differences between the different groups when grown under HL or LL conditions. Exceptions to this can, however, be seen for neoxanthin and lutein. For both, the pigment concentration of group C (deep shade) did not differ significantly between HL- and LL-grown plants. The LL-grown plants contained, instead, a significantly higher concentration of each of these pigments than did group A or B plants.
Photosynthetic acclimation is greatest in plants of intermediate shade association (Murchie & Horton 1997). When data for all individual species were considered (Fig. 2a–d), there was no obvious relationship between shade association and the extent of acclimation for individual pigments. However, from the grouped data (Fig. 2e–h) there was some evidence that the extent of acclimation varies according to shade tolerance, with a large response in xanthophyll cycle pigments for group B and a small response in lutein, carotene and neoxanthin for group C.
A parameter previously observed to correlate strongly with species’ tendency for growth in shade is the ratio of lutein : xanthophyll cycle pigments (Johnson et al. 1993a), but there was no relationship between shade association for plants grown in HL or LL (Fig. 3).
Although no simple relationships were observed between shade association and pigment content, it is still possible that other ecological factors determine pigment content. To examine whether more complex relationships exist, correspondence analysis (CA; Kent, & Corker 1992) was applied (Fig. 4). Using this technique, patterns can be identified based on any ecological or environmental factor, rather than simply looking for relationships with a single parameter such as shade association. As CA is an indirect ordination method, that is, no assumptions are made about environmental factors during initial analysis of the data, the two axes do not necessarily correspond to any particular parameter. Rather, environmental trends can be assigned to gradients and groups at a later date. In Fig. 4, each point represents either HL- or LL-grown plants from each species, with the distance between two points being related to the degree of difference in their pigment content and composition. Species grown in the LL treatment generally cluster to the left of the diagram, indicating again that the plants grown under LL have similar pigment content and composition, regardless of their habitat preference. In contrast, HL-grown plants are mostly clustered tightly to the right of the diagram. However, some species fall well outside this cluster and overlap with the LL-grown plants. The species falling outside the main cluster of HL-grown plants do not belong to a particular shade-association group. However, species with intermediate shade tolerances (group B) are found exclusively in the main (right-hand) cluster in this diagram, while the species that are furthest removed from this group tend to be extreme shade species (group C).
Data presented in this paper indicate that the main determinant of leaf pigment content and composition is the environmental factors experienced during development, rather than any genetic predisposition to a given pigment complement. Thus pigment content is not a selectively important feature determining the tolerance to sun and shade in British plant species. Given the large number of studies that have highlighted the relationship between sun–shade and pigment content, this result is very surprising. However, as in a previous study (Murchie & Horton 1997), there does appear to be a slightly higher degree of plasticity in sun–shade intermediate species than in either full-sun or deep-shade species, while deep-shade species have a tendency towards lower plasticity. This presumably reflects the wider range of light intensities to which the former group may be exposed, and the relatively small range of intensities in which the latter would be expected to grow.
Plants in this study tended to respond to increased irradiance by: (i) increasing the ratio of chlorophyll a : b; (ii) increasing the amount of every carotenoid per chlorophyll and per leaf area; (iii) increasing xanthophyll cycle pigments and, to a lesser extent, the total (α + β) carotene pool size, as a percentage of total carotenoid, at the expense of lutein and neoxanthin; and (iv) increasing the extent to which the xanthophyll cycle pool can be de-epoxidized.
These findings are also generally true for individual species, and are consistent with the results of acclimation studies on other species (Demmig-Adams et al. 1989; Logan et al. 1996; Thayer & Björkman 1990). Our results differ from the general consensus in that neoxanthin increased in response to irradiance. Previous studies indicated that neoxanthin concentrations are largely unchanged by acclimation to different irradiances.
The difference in pigment content and composition in response to different light environments may reflect the effects of irradiance, spectral quality, or both. The shaded environment used was intended to imitate natural shade. The far-red enrichment could alter the state of phytochromes or the relative excitation of the two photosystems. We did not attempt to distinguish these effects. Studies with mutants deficient in phytochrome-signalling pathways suggest that these play a relatively small role in photosynthetic acclimation (Walters et al. 1999).
In contrast to a previous study (Johnson et al. 1993a), no linear correlations were seen between carotenoid concentration and shade association. There was little difference in the pigment contents of sun and intermediate species. Deep-shade species showed a small variation from the other groups by having slightly more lutein and neoxanthin per chlorophyll than plants from less-shaded habitats, when grown in low light conditions. They did not significantly increase these pigments on exposure to HL conditions. The relatively high concentration of these pigments in LL-grown deep-shade plants may reflect a particular importance of these two pigments in light harvesting, increasing the efficiency of light harvesting at very low irradiances.
The lack of plasticity in carotenoid acclimation is consistent with a general lack of acclimative plasticity of deep-shade plants in a range of photosynthetic factors. Murchie & Horton (1997) observed that deep-shade species exhibited very little acclimation to growth light in both chlorophyll content and composition, and in the maximum rate of photosynthesis. The only class of carotenoids in which deep-shade species do show marked acclimation are the xanthophyll cycle pigments.
The lack of any relationship between carotenoid content and ecological background is surprising, especially as a large amount of published data shows correlations between pigment content and species origin. This observation may reflect a lack of range in the light conditions found in deciduous British woodlands, where most plants will be exposed to direct sunlight at some point during their life cycle and so must tolerate transiently high irradiance. The deep-shade species that we studied do, however, have a poor long-term tolerance of high light, some showing signs of stress (e.g. reduced growth rate; purple coloration due to accumulation of anthocyanins) under the conditions used. At any level of shade tolerance a similar range of carotenoid concentrations occur in all species, indicating that the ability to produce carotenoids in general does not play a major role in determining tolerance of high light or shade.
The authors would like to thank Professor Phil Grime (University of Sheffield) and Dr Alan Rosevear (The Environment Agency) for help in obtaining plant material, Mr Thurston Heaton and Mr David Newton for their assistance in growing plants, and Mr Stephen Evans for technical assistance. M.J.R. was in receipt of a UK Natural Environment Research Council studentship.