Correspondence: Rainer Hofmann, AgResearch, Grasslands Research Centre, Private Bag, Palmerston North, New Zealand. Fax: + 64 (0)6351 8042; e-mail: email@example.com
This study used comparisons across nine populations of Trifolium repens (white clover) in conjunction with drought to examine physiological responses to ultraviolet-B radiation (UV-B). Plants were exposed for 12 weeks to supplementation with 13.3 kJ m−2 d−1 UV-B, accompanied by 4 weeks of drought under controlled environmental conditions. UV-B increased the levels of UV-B-absorbing compounds and of flavonol glycosides and this effect was synergistically enhanced by water stress. These changes were more pronounced for the ortho-dihydroxylated quercetin, rather than the monohydroxylated kaempferol glycosides. UV-B increased leaf water potential (ψL) by 16% under drought and proline levels by 23% under well-watered conditions. The intraspecific comparisons showed that higher UV-B-induced levels of UV-B-absorbing compounds, of quercetin glycosides and of ψL were linked to lower plant productivity and to higher UV-B tolerance under well-watered conditions. These findings suggest that: (1) slow-growing T. repens ecotypes adapted to other stresses have higher capacity for physiological acclimation to UV-B; and (2) that these attributes also contribute to decreased UV-B sensitivity under drought.
Furthermore, it has yet to be tested whether natural adaptation of plants to stress, such as drought, can provide UV-B tolerance (Gwynn-Jones et al. 1999b). Comparisons of numerous plant species in other investigations have demonstrated that tolerance to one form of stress can be related to adaptation against other stress forms and to lower plant productivity (e.g. Bungener et al. 1999; Grime 2001). In the absence of large-scale species comparisons in UV-B studies, another approach to address such questions is the examination of a variety of populations within a species, thus avoiding confounding species-specific differences. Information on intraspecific differences in UV-B effectiveness depending on water availability is limited, and usually based on single comparisons between two cultivars or genotypes (Teramura et al. 1990; Schmidt et al. 2000).
Several studies suggest UV-B-protective functions for UV-B-absorbing compounds and phenylpropanoid derivatives such as phenolic acids and flavonoids (e.g. Markham et al. 1998; Meijkamp et al. 1999). Some findings attribute the primary role for flavonoid induction to UV-B (Balakumar et al. 1993), whereas other results suggest an interaction with water stress (Nogués et al. 1998). Photosynthetic parameters often show little sensitivity to UV-B (Allen, Nogués & Baker 1998), but this may also depend on water availability, for example for photochemical efficiency of photosystem II or chlorophyll levels (Nikolopoulos et al. 1995). Investigations of UV-B effects in combination with drought also warrant examination of physiological aspects associated with the plant water status. In particular, there is a need to study relationships between plant water potential and UV-B-induced accumulation of stress-related solutes (Schmidt et al. 2000). Studies in drought-stressed pea plants have shown improvements of the leaf water potential (ψL) under UV-B (Nogués et al. 1998). Accumulation of the osmoregulator proline has been observed in response to a number of environmental factors, including drought and UV-B (Shetty, Atallah & Shetty 2002).
Recent studies have demonstrated decreased UV-B sensitivity when T. repens plants are drought stressed (Hofmann, Campbell & Fountain 2003). Our objective was therefore to examine the interactive physiological effects of UV-B and drought for this species. Comparisons of nine T. repens populations also showed that under well-watered conditions, UV-B sensitivity is associated with the productivity of the populations and of their habitat of origin (Hofmann et al. 2003). A further aim was thus to test whether functional responses which could confer UV-B tolerance across populations under drought conditions would be confined to slow-growing, stress-tolerant T. repens populations under well-watered conditions.
MATERIALS AND METHODS
The experimental design included two levels of UV-B (with and without UV-B supplementation), two watering regimes (well watered and drought stressed) and nine populations of Trifolium repens L. Plants were grown in pots that were placed on trolleys in two climate-controlled growth rooms. One room contained a rig with UV-B lamps whereas the plants in the second room were grown under a rig without UV-B supplementation. In each room, the plants on three trolleys remained well watered, whereas the plants on the other three trolleys were drought stressed. In each treatment, there were five replicated pots per population, arranged randomly on the trolleys. Each pot contained 10 genotypes of a population. To minimize the effects from differential shading under the UV-B rigs, the trolleys were rotated daily using a standardized pattern. This allowed different positions for 24 consecutive days before the starting position was resumed.
Plant cultivation and UV-B irradiation
Nine T. repens populations were chosen to provide a variety of different genetic and geographical backgrounds (Hofmann et al. 2000). This included three slow-growing ecotypes collected in the wild from habitats exposed to multiple stress conditions (Häggås, Sarikamis and Tienshan), four cultivars selected in breeding programmes for agricultural yield (Haifa, Huia, Kopu and Prestige) and two breeding lines (Octoploid and Syrian).
Trifolium repens plants were grown from 3- to 5-day-old seedlings in square gallon pots containing sterilized sand. The plants were supplied with half-strength Hoagland's nutrient solution at regular intervals using an automatic drip irrigation system and were flushed with water weekly. The experiment was conducted for 12 weeks in growth rooms of identical design. In other studies these chambers have demonstrated excellent environmental control with insignificant differences between rooms (Warrington et al. 1999). The average photosynthetic photon flux (PPF) during the 12 h light period was 425 µmol m−2 s−1. Temperatures were 22 °C during the day and 18 °C at night and relative humidity was 70–80%. From one hour after onset of the light period until one hour before darkness, UV-B radiation was supplied by Philips TL 40 W/12 RS UV fluorescent tubes (Philips Lighting NZ Ltd, Auckland, New Zealand), enclosed in cellulose diacetate filters. The irradiance levels were 13.3 kJ m−2 d−1 biologically effective UV-B, normalized to 300 nm (Caldwell 1971), representing 25% summer ozone depletion above Palmerston North, New Zealand. Further details on plant cultivation conditions and the UV-B supplementation system can be found in Hofmann et al. (2000).
Harvesting of plant material and drought application
Plants were clipped four times near pot height during the 12-week experimental period. This minimized the effects of shading and prevented the plants from exceeding the pot size. Drought was applied during the last 4 weeks of the experiment and was monitored gravimetrically every day (Barbour et al. 1996). Whereas the moisture level of the sand ranged around 11% in the well-watered pots, the drought-stressed plants received reduced amounts of nutrient solution to maintain the moisture content of the sand at 3–4%. All plant measurements were conducted at the conclusion of the 12 week UV-B exposure period.
The following treatments were applied during the 4 week drought period: UV– WW, no UV-B supplementation and well-watered conditions; UV+ WW, UV-B supplementation and well-watered conditions; UV– DR, no UV-B supplementation and drought conditions; UV+ DR, UV-B supplementation and drought conditions.
UV-B-absorbing compounds and flavonoid analysis
Five replicates were analysed for each T. repens population under each treatment condition, using fully open trifoliate laminae. UV-B-absorbing compound levels were estimated spectrophotometrically at 300 nm after extraction of 15 mg oven-dried leaf powder in 1.2 mL MeOH : H2O : HCl (79 : 20 : 1) (Hofmann et al. 2001). For analysis of flavonoids, 50 mg freeze-dried leaf powder were extracted in 3 mL MeOH : H2O : HOAc (89 : 10 : 1), followed by centrifugation and de-fatting of the supernatant with 1 mL petroleum ether (Hofmann et al. 2000). After solvent evaporation under nitrogen and vacuum drying, the residue was taken up into 700 µL CH3CN : HOAc : H2O (32 : 3 : 65). The solution was then centrifuged and a 25 µL sample was used for high-performance liquid chromatography (HPLC) analysis.
The instrumentation used for analytical HPLC has been described in Hofmann et al. (2000). The gradient solvent system comprised solvent A [1.5% H3PO4] and solvent B [HOAc : CH3CN : H3PO4 : H2O (20 : 24 : 1.5 : 54.5)], mixed using a linear gradient starting with 80% A, decreasing to 33% A at 30 min, 10% A at 33 min and 0% A at 39.3 min. Peaks of quercetin and kaempferol derivatives were identified on the basis of the on-line spectra. The integrated areas of all flavonol peaks (measured at 352 nm) were added to calculate flavonol glycoside levels from a standard curve prepared using rutin (quercetin 3-rutinoside).
Extraction of chlorophyll from fully open T. repens laminae was performed in N,N-dimethylformamide (DMF). Three samples for each population and treatment condition were ground, freeze-dried and weighed (10 mg) into centrifuge tubes. Samples were extracted with 1.2 mL DMF in darkness at 4 °C for 24 h under occasional vortex shaking, followed by centrifugation at 20 000 × g and 4 °C for 3 min. Absorbance of the supernatant was read at 664.5 and 647 nm and chlorophyll a and b content was calculated using the equations by Inskeep & Bloom (1985).
Chlorophyll fluorescence measurements were performed as described in Hofmann et al. (2001). Fully open distal leaflets of four to five replicate samples were used for each population under each treatment condition. Dark-adapted samples had been in darkness for at least 10 h. For light-adapted fluorescence measurements, samples were collected in the illuminated growth room and allowed to equilibrate for 4 min at the PPF level equivalent to that in the growth room. The measurements were conducted at room temperature with a pulse-amplitude-modulation (PAM) fluorometer (Pam 101; Walz, Effeltrich, Germany), using a flash intensity of 10 000 µmol m−2 s−1. Fv/Fm and ΔF/Fm′ (ΔF = Fm′ − Ft) were calculated (Genty et al. 1990) to determine the maximum and effective quantum efficiency of photosystem II photochemistry, respectively.
Leaf water potential and proline
The pressure chamber technique was used to determine ψL (Scholander et al. 1965). Fully open distal leaves were enclosed in a pressure chamber and the cut surface of the protruding petiole was monitored while pressure in the chamber was gradually increased. The pressure required to bring xylem water to the surface (equivalent to the water tension inside the xylem) was recorded. Measurements were conducted on five replicate samples per population and treatment condition.
Levels of free proline in fully open T. repens laminae were determined using the method by Magné & Larher (1992). In contrast to the commonly used approaches of proline determination, this method precludes interference by carbohydrates. For each population under each treatment condition, three samples were ground in liquid nitrogen, freeze-dried and 10 mg weighed into centrifuge tubes. Precipitation of protein was carried out in 1.2 mL of 3% (w/v) sulphosalicylic acid under vortex shaking, followed by centrifugation at 12 000 × g for 7 min. The supernatant was removed and after renewed centrifugation, 500 µL were collected and supplemented to 1 mL with water. This was followed by the addition of 2.0 mL ninhydrin reagent [1% (w/v) ninhydrin in 60% (v/v) glacial acetic acid], vortex shaking and reaction of the solution for 1 h at 98 °C. The reaction was stopped in an ice-water bath, followed by addition of 2 mL toluene and vortex shaking for 20 s. Phases were allowed to separate for at least 5 min and the products extracted in toluene were examined spectrophotometrically in a 1 mL glass cuvette at 520 nm. Free proline content in the T. repens samples was calculated from a standard curve of known proline concentrations (0–25 µg mL−1) prepared in an identical manner alongside each batch of samples.
The General Analysis of Variance procedure in GENSTAT (Genstat 1993) was used for analysis of main effects (UV-B, drought, population) and their interactions. The GENSTAT Regression Analysis and Correlation procedures were used to examine relationships of plant attributes and UV-B sensitivity at final harvest. Previously we observed significant differences between the nine T. repens populations for constitutive productivity (i.e. biomass production without UV-B supplementation) as well as for UV-B sensitivity (biomass reduction) under well-watered conditions (Hofmann et al. 2003). However, no population differences in UV-B sensitivity were present in plants exposed to drought. Correlation analyses to aspects of constitutive biomass production and UV-B sensitivity were thus performed on well-watered plants.
UV-B-absorbing compounds and flavonoids
UV-B-absorbing compound levels displayed an overall increase in response to UV-B (P < 0.001). This effect was mainly due to changes in water-stressed plants (Table 1), with a significantly (P = 0.016 for the UV–B × drought interaction) greater average change (+12%) than in well-watered conditions (+3%). Changes in UV-B-absorbing compound levels under UV-B displayed differences among the T. repens populations (P = 0.038), with particular increases for Tienshan and Sarikamis across water treatments (Fig. 1).
Table 1. Biochemical and physiological attributes averaged across nine populations of T. repens (mean ± 1 SE), grown with (UV+) and without (UV–) supplementation of 13.3 kJ m−2 d−1 UV-B, under well-watered (WW) or droughted (DR) conditions
Probabilities reflect significance of the UV-B effect under the two water regimes. ***P < 0.001; **P < 0.01; *P < 0.05; +P < 0.10; NS, P≥ 0.10.
Flavonoids in the T. repens leaf samples were identified by HPLC analysis as glycosides of the flavonols quercetin and kaempferol. There were significant differences between water treatments for UV-B-induced flavonol glycoside increases (P < 0.001 for the UV–B × drought interaction). Under well-watered conditions, a more than double increase of flavonol glycoside levels was observed in response to UV-B, and the increase was more than triple in the UV-B × drought combination across populations (Table 1). These responses were also observed for individual flavonols, especially quercetin glycosides, which under UV-B showed higher absolute levels and larger increases than kaempferol compounds (Table 1). Comparison with the results for UV-B-absorbing compound levels showed similar direction, but much more pronounced quantitative changes for flavonol glycoside levels in response to UV-B (Table 1).
Furthermore, there were significant differences in the UV-B-induced flavonol glycoside responses among T. repens populations (P < 0.001). Particularly high flavonol glycoside accumulation was observed for the T. repens ecotypes Tienshan, Sarikamis and Häggås (Fig. 2). Flavonol glycoside accumulation in response to UV-B was consistently higher in all populations than under unstressed conditions or under drought alone (Fig. 2). The UV-B-generated changes in flavonol glycoside accumulation were directly related to those for UV-B-absorbing compound levels (r = 0.698, P = 0.037).
Photosynthetic pigmentation and photochemistry
Photosynthetic pigmentation experienced little change in response to UV-B (Table 1). UV-B led to a marginal increase in chlorophyll levels by 4% under well-watered conditions and to no change in the chlorophyll a : b ratio (data not shown). Drought stress resulted in an average 16% decrease of chlorophyll levels (P < 0.001). No marked UV-B effects could be observed for photochemical efficiency of photosystem II, with a small 3% UV-B-generated increase in ΔF/Fm′ under both watering regimes being the only measurable change (Table 1). No significant interactions between UV-B and drought or T. repens populations were found for chlorophyll levels or photochemical efficiency.
Leaf water potential and proline
On average, ψL was increased by UV-B across populations and water treatments (P < 0.001). This was mainly due to a UV-B-induced increase of ψL under drought (16%, Table 1), which was significantly (P < 0.001) different from an increase of 3% under well-watered conditions. In comparison with the control plants, there was a pronounced average decrease of about 50% for ψL under drought (P < 0.001). Significant differences were observed among well-watered T. repens populations in the constitutive (i.e. UV-B-independent) ψL (P = 0.005), and in the degree of the UV-B-induced changes for ψL (P = 0.029). In particular, increases of ψL under well-watered conditions could be observed for Tienshan and Sarikamis and decreases for Haifa and Octoploid (Fig. 3). Under drought conditions, the UV-B-induced increase in ψL occurred consistently for all populations (Fig. 3). A direct relationship between constitutive ψL and UV-B-generated ψL changes under well-watered conditions (r = 0.729, P = 0.026) showed that populations with low intrinsic ψL were able to increase their ψL under UV-B, while less negative constitutive values for ψL were linked to UV-B-elicited decreases in ψL (Fig. 3).
Free proline levels were consistently enhanced by UV-B under well-watered conditions, leading to an overall increase by 23% (Table 1, Fig. 4a). In comparison with the well-watered plants, proline levels under drought were an order of magnitude higher. Averaged across T. repens populations, there was no significant overall difference for free proline accumulation in response to UV-B under drought (Table 1). Nevertheless there were differences in this response among the T. repens populations (P = 0.047), in particular with decreases in proline accumulation for Syrian and Sarikamis (Fig. 4b). Significant inverse relationships were found between intrinsic proline concentration (i.e. without UV-B supplementation) and increases in these levels under UV-B (r = −0.867, P = 0.003 for well-watered and r = −0.708, P = 0.033 for droughted plants). Thus in both water regimes, free proline concentrations were more strongly increased by UV-B in T. repens populations that had low intrinsic proline levels (Fig. 4a & b). Furthermore, populations with low ψL contained high proline levels under UV-B in drought conditions (r = −0.694, P = 0.038) and in well-watered plants (r = −0.610, P = 0.081). Examination of the proline absorption spectrum showed no absorbance in the UV region above 240 nm (data not shown).
Intraspecific relationships to biomass production and UV-B sensitivity
Physiological factors that were associated with the productivity and UV-B responsiveness of the nine T. repens populations included accumulation of quercetin glycosides, UV-B-absorbing compounds and ψL. Productivity of the T. repens populations under well-watered conditions displayed an inverse relationship to constitutive (r = −0.826, P = 0.006) and to UV-B-induced (Fig. 5a) accumulation of quercetin glycosides, and this accumulation in turn was linked to UV-B tolerance (r = 0.629, P = 0.069). No relationships to productivity or UV-B tolerance could be found for kaempferol glycosides.
Similarly under well-watered conditions, productive populations had higher (less negative) constitutive ψL (Fig. 5b), and this in turn tended to be inversely related to UV-B sensitivity (r = −0.605, P = 0.085). Productive populations decreased their ψL most under UV-B (Fig. 5c, indicated by changes exceeding 100% in Haifa and Octoploid). This decrease again showed an inverse trend with UV-B sensitivity (r = −0.637, P = 0.065). Furthermore, UV-B tolerance in the T. repens populations correlated with the capacity to increase accumulation of UV-B-absorbing compounds under UV-B (Fig. 6). A similar relationship was found for increases in flavonol glycoside levels (r = 0.588, P = 0.096). In drought conditions there was a significant inverse correlation of proline concentration with plant productivity (r = −0.728, P = 0.026), showing that the less productive stress-tolerant ecotypes contained the highest proline levels (Fig. 4b).
Increased levels of UV-B-absorbing compounds (Fig. 1) and higher accumulation of UV-B-generated flavonol glycosides (Fig. 2) in less productive, UV-B-tolerant T. repens ecotypes are in agreement with ecological models, proposing higher biochemical acclimation for slower-growing, stress-tolerant plants (Grime 2001). Pronounced UV-B-induced increases of UV-B-absorbing compounds and flavonol glycosides in combination with drought are supported by findings in pea (Nogués et al. 1998) and Cistus (Stephanou & Manetas 1997). The higher UV-B-generated increase for flavonol glycosides than for UV-B-absorbing compounds shows underestimation of the flavonoid effect by the crude UV-B-absorbing compound extraction, which also contains compounds that are not involved in UV-B responses (Meijkamp et al. 1999).
We observed an inverse relationship of quercetin – but not kaempferol – glycoside accumulation to plant productivity (Fig. 5a) and UV-B sensitivity. Higher UV-B-induced increases of the B-ring ortho-dihydroxylated quercetin, rather than the monohydroxylated kaempferol glycosides (Table 1) suggest that other processes apart from a mere alteration of carbon budgets are involved in secondary metabolite accumulation of relevance for UV-B protection. This response pattern is well conserved in the plant kingdom (e.g. Markham et al. 1998; Ryan et al. 2002) and could be a reflection of higher antioxidant activity or energy dissipation in dihydroxylated flavonoids (Smith & Markham 1998; Kostina, Wulff & Julkunen-Tiitto 2001).
Photosynthetic pigmentation and photosystem II photochemistry were not negatively affected by UV-B and the studies in T. repens show that this is independent of water availability (Table 1). This may reflect adequate photoprotective responses for these parameters, including the observed increases of UV-B-screening secondary metabolites. However, photosynthesis could still be inhibited, for example due to reductions in carbon assimilation or smaller leaf areas. Such UV-B-induced changes of leaf morphology and other growth reductions have been observed in T. repens (Hofmann et al. 2001, 2003) and could help explain the increases of ψL when UV-B-treated plants are subsequently exposed to drought (Table 1; Fig. 3). Increases of ψL for less productive T. repens ecotypes under well-watered conditions (Fig. 5c) may be related to the previously observed UV-B-generated increases in the root : shoot ratio for such populations (Hofmann et al. 2001).
We found consistent UV-B-induced enhancements in proline accumulation across well-watered T. repens populations (Fig. 4a). UV-B-generated increases in proline levels have recently been observed in another legume and have been related to the synthesis of phenolic secondary metabolites (Shetty et al. 2002). UV-B-induced accumulation of proline was not linked to UV-B tolerance or plant productivity, contrary to the relationships with quercetin glycosides (Fig. 5a) and UV-B-absorbing compounds (Fig. 6). Furthermore, under drought conditions, the slower-growing populations (already high in proline accumulation without UV-B) generally did not increase, or even decreased their proline levels in response to UV-B (Fig. 4b), which is opposite to what was found for flavonol glycosides (Fig. 2). This could indicate a fundamental difference at the level of amino acid biosynthesis between production of the heterocyclic proline and that of the aromatic flavonoid precursors tyrosine and phenylalanine. The slower-growing populations appear more efficient in directing their aminocarbon flow towards the latter process under UV-B.
The limitations of controlled environmental (CE) studies are recognized, for example due to disproportionate spectral conditions such as high ratios of UV-B to PPF. CE studies can be used: (1) to investigate UV-B responses in combination with drought without other confounding environmental influences; and (2) to screen a number of populations within a species for responsiveness to UV-B ( Corlett et al. 1997). To our knowledge, this is the first study using multiple population comparisons in conjunction with drought to examine functional plant responses to UV-B. It has frequently been stated that the metabolic costs of UV-B protection are unknown (e.g. Rozema et al. 1999). Our results suggest that one cost of a plant strategy towards UV-B-protective functions may be lower constitutive carbon allocation towards growth. The pattern of accumulation for flavonols and UV-B-absorbing compounds suggests that similar means of UV-B protection operate in slow-growing, stress-tolerant ecotypes under well-watered conditions, and in the general decrease of T. repens UV-B sensitivity by drought. Using modulated UV-B supplementation in interaction with drought (Campbell, Hofmann & Hunt 1999), we have started to test these effects on T. repens plants growing under field conditions.
We thank Yvonne Gray, Margaret Greig, Chris Hunt, Derryn Hunt, Christine van Meer and Linda Robinson for technical assistance. We thank Dennis Greer for advice on fluorescence measurements. We are grateful to Dave Barker, Brian Jordan and Tessa Mills for helpful advice. R.W.H. was supported by an AGMARDT Fellowship. The research was funded by the New Zealand Foundation for Research, Science and Technology, Contract Number C10632.
Received 2 July 2002; received in revised form 9 October 2002; accepted for publication 14 October 2002