Warm and cold parental reproductive environments affect seed properties, fitness, and cold responsiveness in Arabidopsis thaliana progenies


A. Polle. Fax: +49 551 392705; e-mail: apolle@gwdg.de


Conditions in the parental environment during reproduction can affect the performance of the progenies. The goals of this study were to investigate whether warm or cold temperatures in the parental environment during flowering and seed development affect Arabidopsis thaliana seed properties, growth performance, reproduction and stress tolerance of the progenies, and to find candidate genes for progeny-related differences in stress responsiveness. Parental plants were raised at 20 °C and maintained from bolting to seed maturity at warm (25 °C) or cold (15 °C) temperatures. Analysis of seed properties revealed significant increases in nitrogen in seeds from warm temperature and significant increases in lipids and in the ratio of α-linolenic to oleic acid in seeds from the cold parental environment. Progenies of the warm parental environment showed faster germination rates, faster root elongation growth, higher leaf biomass and increased seed production at various temperatures compared with those from the cold parental environment. This indicates that under stable environmental conditions, progenies from warm parental environments had a clear adaptive advantage over those from cold parental environments. This parental effect was presumably transmitted by the higher nitrogen content of the seeds developed in warm conditions. When offspring from parents grown at different temperatures were exposed to chilling or freezing stress, photosynthetic yield recovered faster in progenies originating from cold parental environments. Cold acclimation involved up-regulation of transcripts of flavanone 3-hydroxylase (F3H) and pseudo response regulator 9 (PRR9) and down-regulation of growth-associated transcription factors (TFs) NAP and AP2domain containing RAP2.3. NAP, a regulator of senescence, and PRR9, a temperature-sensitive modulator of the circadian clock, were probably involved in mediating parent-of-origin effects, because they showed progeny-related expression differences under chilling. Because low temperatures also delay senescence, cold responsiveness of NAP suggests that this factor is linked with the regulatory network that is important for environmental acclimation of plants.


Low temperatures are among the most important abiotic environmental factors affecting geographical distribution, growth and yield indices of plants. Therefore, it is of considerable interest to understand how plants respond and adapt to cold stress. Low temperatures decrease the rate of photosynthesis and affect electron transport rates as evident from decreases in photosynthetic quantum yield Φ (Maxwell & Johnson 2000). In Arabidopsis thaliana, photosynthesis is strongly inhibited after transfer from warm (22 °C) to cold conditions (5 °C), but recovery occurs during long-term exposure to low temperatures (Savitch et al. 2001). During cold stress, the expression of nuclear-encoded photosynthetic genes is suppressed in A. thaliana, and recovery is associated with a strong increase in enzyme activities of the Calvin and sucrose biosynthesis (Strand et al. 1997; Strand et al. 1999; Hurry et al. 2000).

Furthermore, physiological responses of plants to low temperatures involve accumulation of cryoprotectants, changes in lipid membrane composition and detoxification of reactive oxygen species (Thomashow 1999). These physiological changes are brought about by a regulatory network of altered gene expression (Shinozaki, Yamaguchi-Shinozaki & Seki 2003). In A. thaliana, 100–1000 cold-responsive genes have been identified depending on experimental conditions and ecotype (Fowler & Thomashow 2002; Kreps et al. 2002; Seki et al. 2002; Jung, Lee & Lee 2003; Provart et al. 2003). The transcripts of more than 2000 genes showed significant changes after long-term cold acclimation (Hannah, Heyer & Hincha 2005).

In addition to the immediate impact of environmental restrictions, growth and development of plants can also be influenced by their parental life history (Martienssen & Colot 2001; Grant-Downtown & Dickinson 2005). This phenomenon, which has been denominated as ‘imprinting’ or ‘epigenetic memory’, has been studied in some herbaceousspecies, including Arabidopsis, as well as in tree species. In A. thaliana, some parent-of-origin effects (e.g. seed mass development) have been related to gene-specific differences in DNA methylation (Finnegan et al. 1998; Adams et al. 2000; Finnegan, Peacock & Dennis 2000). In Norway, spruce (Picea abies) seed production in cold environments has resulted in twofold higher DNA methylation in progenies than seed production in warm environments (Baumann 2004). Cold reproductive environments of spruce mother trees advance bud set and cold acclimatization of their progenies in fall, and accelerate dehardening and flushing in spring, whereas a warm reproductive environment delays timing of these traits (Skrøppa & Johnsen 2000; Johnsen & Skrøppa 2001). It has been hypothesized that temperature in the reproductive period regulates an epigenetic memory involving differential expression of genes with putative functions in bud phenology, cold acclimatization and embryogenesis. Johnsen et al. (2005) showed that spruce progenies from cold reproductive environments were more freezing tolerant than those from warm reproductive environments. These imprinting effects may last for several years in the filial generation (Baurens et al. 2004; Johnsen et al. 2005). In contrast to spruce, freezing tests on progenies of the C4 weed Echinochloa crus-galli from warm and cold reproduction environments did not reveal any differences in stress tolerance (Charest & Potvin 1993). It is still open whether low or high ambient temperatures during parental reproduction modulate stress responses in the next generation of plant species with short life cycles.

In herbaceous plants, parent-of-origin effects have most frequently been studied in seeds, where they are most pronounced (Gehring, Choi & Fischer 2004). Larger seeds, differences in flowering or germination time may result in competitive advantages and may increase the fitness of a given species (Lacey et al. 2003). For example, the maternal photoperiod affects seed germination percentages and germination speed in A. thaliana (Munir et al. 2001). Two different parental reproductive environments (spring with low radiation and low temperature, and summer with high radiation and high temperature) have pronounced effects on total seed mass production and seed volume in the parental Arabidopsis generation (Andalo et al. 1999). Smaller seeds are formed in summer than in spring. Despite smaller seeds, the first generation derived from ‘summer’-parents produced more biomass and total seed mass than those derived from ‘spring’-parents (Andalo et al. 1999). Whether differences in the quantity and quality of reserve compounds are important factors for the transmission of parental effects has not been investigated. Furthermore, it is important to know whether progenies of A. thaliana with different parental life histories differ in their respective stress tolerance.

In the present study, we asked whether the temperature of the parental environment affected A. thaliana seed properties and the performance of the offspring such as growth, reproduction or stress tolerance. To investigate these questions, parental plants were grown at warm (25 °C) or cold (15 °C) temperatures, and properties of the seeds like storage lipids, carbohydrates and nitrogen content were determined. Growth parameters of the progenies from cold and warm parental environments were tested under different temperatures. We hypothesized that the offspring from parental plants grown at low temperature would have an adaptive advantage at low temperature over offspring from high-temperature-grown parents and vice versa. We further tested whether offspring from parents grown at low temperature would be less cold sensitive than offspring from parents grown at high temperature.


Plant growth and reproduction

Seeds from A. thaliana[ecotype Columbia, Columbia-0 (Col-0) N1093] were sown on Murashige–Skoog (MS) root growth medium (Murashige & Skoog 1962) and were germinated under the following conditions: 8 h day/16 h night [250 ± 10 µmol m–2 s–1 photosynthetically active radiation (PAR), Osram HQL-R 400 W; Osram, Munich, Germany], 20 °C and 60% air humidity. After 21 d, the seedlings were potted into a commercial soil mixture for horticulture (Fruhstorfer Erde Typ N; Industrie-Erdwerk Archut, Lauterbach/Wallenrod, Germany) and randomized regularly during subsequent cultivation. Potted plants (= 28 per temperature) were grown under the same conditions as above in two climate chambers (Weiss-Technik, Reiskirchen, Germany) for 3 weeks. Subsequently, the day length was extended to 16 h, and when the first peduncles started to emerge, the temperature was increased to 25 °C in one chamber and decreased to 15 °C in the other chamber, respectively. Arasystems (Ara System 360 KIT; Beta Tech bvba, Gent, Belgium) were used to isolate each individual to avoid cross fertilization. Plants flowered and seeds ripened under the two different temperature settings. Seeds were harvested 7 and 9 weeks after flowering from plants in the warm (25 °C) and cold (15 °C) environments, respectively. The seeds were stored at 4 °C for further analysis. The experiment was conducted twice, in 2003 and 2004.

Seed analyses

To determine seed weight, 100 seeds were counted and weighed (SuperMicro S4; Sartorius, Göttingen, Germany). For each treatment, six replicates obtained by pooling seeds from different maternal plants were analysed.

For analysis of carbon and nitrogen, the seeds were dried (60 °C, 3 d), ground to a fine powder (MM 2; Retsch GmbH, Haan, Germany), filled in tin capsules (0.8 mg per sample) and analysed in a carbon and nitrogen (C/N) elemental analyser (EA-1108; Carlo Erba Instruments, Rodano, Italy).

Total carbohydrates were analysed in fresh seeds employing the anthrone method (Carroll, Longley & Roe 1956). For detailed analyses, glucose, fructose, sucrose and starch were measured employing enzymatic methods (Beutler 1978; Schopfer 1989). Fatty acid profiles were measured after alkaline transmethylation by gas chromatography-flame ionization detector (GC-FID) as described by Hornung et al. (2005).

Growth experiments and fitness parameters

Arabidopsis thaliana seeds from different environments were sown and germinated on MS root medium at 8 h light (100–150 µmol m–2 s–1 PAR), 20 °C and 60% air humidity in an acclimatized room (MobyLux GroBank, model AR 75-L CLF; Plant Climatics, Emersacker, Germany). To avoid positional effect, the place of the pots or trays in the growth rooms was changed regularly.

Growth performance at 20 °C

After 3 weeks, the seedlings were potted individually in pots (n = 15 per environment) and continued to grow under the same conditions as previously mentioned. Leaves were counted every fourth day to estimate growth rates. The experiment was repeated five times. A representative experiment is shown.

Competition studies and fitness parameters in different environments

After 3 weeks, progenies from cold and warm environments were planted alternately into trays in six rows, each containing five plants (30 plants in total). The trays were placed in two climate chambers and maintained at 15 and 25 °C, respectively. Leaf formation was recorded regularly during the first 7 weeks of growth; subsequently, main and side siliques were counted. The plants were allowed to pollinate freely and at seed maturity, they were harvested; rosette diameter was determined, and dry mass of the leaves was determined. Dry mass of seeds per plants was measured, and seed weight was determined by counting and weighing 100 seeds per individual plant (five replicates per growth environment and seed origin).

Root growth at different temperatures

Petri dishes (foursquared plates; Greiner Bio-one, Solingen, Germany) were three-fourths filled with MS rooting medium. Each plate was equipped with 10 seeds, five from each environment. The plates were exposed at temperatures of 10, 15, 20, 25 or 30 °C, respectively. At each temperature, five plates were analysed per environment. The position of the root tips was marked every second day and was used to calculate root growth rates. The experiment was conducted twice for each temperature using seeds from the independent seed lots of the two reproduction experiments.

Cold-stress experiments and molecular analyses

Seeds from different environments were sown and germinated on MS root medium at 8 h light (100–150 µmol m–2 s–1 PAR), 20 °C and 60% air humidity in an acclimatized cabinet (model AR 75-L CLF, Plant Climatics). After 3 weeks, the seedlings were potted individually in pots (n = 15 per environment) and continued to grow for 3 weeks under the same conditions as mentioned earlier. Subsequently, one set of plants of each environment was maintained at 20 °C as control. The other set was transferred to a temperature of 3 °C to induce cold stress. The plants were maintained for 3 d at low and high temperatures, respectively, and chlorophyll fluorescence was measured with a photosynthesis yield analyser (Mini PAM; Walz, Effeltrich, Germany) during the light phase in the ‘morning’ and ‘afternoon’ and in darkness. The quantum yield of photochemistry was calculated according to


with Fm and Fo measured in light and darkness yielding actual and maximum quantum yield of photochemistry, respectively (Maxwell & Johnson 2000).

After 3 d, leaves of cold-stressed and control plants originating from cold and warm environments were harvested and stored at −80 °C for further analysis.

Leaves were used to identify possible differentially expressed candidate genes on Regulatory Gene Initiative in Arabidopsis (REGIA) filters containing about 1400 transcription factors (TFs) (Paz-Ares & REGIA Consortium 2002). For this purpose, initially, aliquots of the seven individual samples were pooled, ground in liquid nitrogen (MM2, Retsch GmbH) and used for RNA isolation after Chang, Puryear & Cairney (1993). RNA was transcribed into radioactive cDNA (33P), hybridized and scanned with a phosphorimager (BAS-1500 Bioimaging Analyzer; Raytest, Straubenhardt, Germany). Images were scanned with Aida Software version 4.06 (Raytest), saved as a picture data file and analysed with Array Vision TM Version 8.0 (Imaging Research Inc., Munich, Germany). During this preliminary analysis, the filters with samples from the cold environment were used as the reference, and the relative expression was calculated as the ratio of the signal of the sample from 25 °C/signal sample 15 °C. A ratio > 3 was tentatively considered to indicate significant regulation. Out of 19 potentially significantly regulated genes, two up- and two down-regulated genes were chosen for further analysis by quantitative reverse transcriptase (RT)-PCR (NAP, NAC-family TF, At1g69490; flavanone 3-hydroxylase (F3H), At3g51240; pseudo response regulator 9 (PRR9) gene, At2g790; and AP2 containing protein RAP 2.3, At3g16770).

For these genes, primers were designed (Primer 3 input program, http://edu/cgi-bin/primer3/primer3_www.cgi, cf. Table 1) and produced by MWG Biotech (Ebersberg, Germany). The sequences are shown in Table 1. For quantitative real-time PCR (qRT-PCR), isolated RNA was treated with DNase I (EN0521; Fermentas, St. Leon-Roth, Germany) and RNeasy mini columns (clean up protocol; Quiagen, Hilden, Germany) according to the instructions of the companies and transcribed into cDNA (Kit K1622, Fermentas). After testing the primers in a standard PCR, qRT-PCR was carried out in an iCycler + MyiQ single color real-time PCR (Bio-Rad, Munich, Germany) using 0.5 µg cDNA according to the standard protocol of Bio-Rad (http://www.bio-rad.com/iCycler) for the SYBR Green Supermix Kit (170–8880) and an annealing temperature of 57 °C. 18S rRNA served as the reference. The relative expression was calculated with the relative expression software tool (Pfaffl 2001; Pfaffl, Graham & Dempfle 2002).

Table 1.  Primers of selected genes and expected size of product
AGI codeNamePrimer sequenceProduct size (bp)
At3g51240Flavanone 3-hydroxylase (F3H)5′ gtggcggatatgactcgtct228
3′ cgtcactttcacccaacctt
At3g16770AP2 domain containing RAP 2.35′ gagggatacgtaagcgtcca174
3′ gcagatctgggaagttgagc
At2g46790Pseudo response regulator 9 (PRR9)5′ gtggaattgacaagcgtcct77
3′ aagtccaagctcaggaccaa
At1g69490NAC-like, activated by AP3/PI (NAP)5′ cccgagaaaacagagtttgg154

Freezing test

The plants were grown at 20 °C as described earlier for cold stress. Six-week-old plants were either exposed immediately to −5 °C (3 h) or acclimated for 9 h to 2 °C and subsequently exposed for 3 h to −5 °C. After freezing stress, the temperature was increased to 2 °C and maintained for 24 h. Acclimation and freezing tests were done in darkness. Subsequently, the temperature was increased to 20 °C, and the plants were grown under short-day conditions as before. Chlorophyll fluorescence was determined as quantum yield of photochemistry before, immediately after frost stress, repeatedly during the 24 h recovery phase in darkness and after 5 d of recovery under normal growth conditions.

Statistical analysis

If not indicated otherwise, data are means (± SD). The number of replicates is indicated in figures and tables, and may be lower than the number of plants in the experiment, if plants died or if materials were used for different types of analyses. To investigate treatment effects, either mono- or multivariate analysis of variance was performed using Statgraph (Manugistics, Rockville, MD, USA) followed by a multiple range test. For the comparison of two samples, Student's t-test was applied. To determine root growth rates, means of root lengths were calculated for the offspring of each parental temperature environment on each plate and were used to determine growth rates by regression analysis employing a linear model after an initial lag phase of 3–6 d depending on acute growth temperature. The slopes of the fitted curves, which had regression coefficients of R = 0.9902–0.9999 for the different curves, gave the mean growth rate per plate and parental line. P-value was calculated for paired samples of each plate using a signed rank test. Means were considered to differ significantly, if P ≤ 0.05.


Growth temperatures have profound effects on the composition of storage compounds in A. thaliana seeds

When A. thaliana plants started to bolt, they were transferred from 20 to 15 °C or 25 °C, respectively, and maintained at these temperatures until seed maturity. Analysis of seed quality showed significantly higher relative amounts in nitrogen (+48%) in seeds from the warm environment than in those from the cold environment (Table 2). The seeds from the cold environment contained slightly but significantly increased relative amounts of carbon (+4%, Table 2). The C/N ratio of seeds from the warm environment was significantly lower than that of those from the cold environment.

Table 2.  Relative carbon and nitrogen amounts in seeds of Arabidopsis thaliana from two different growth environments
Parameter15 °C25 °CP
  1. During reproduction, the parental plants were grown at 15 and 25 °C, respectively. Data indicate means (= 6, ± SD).

Nitrogen (%)2.43 ± 0.133.42 ± 0.120.001
Carbon (%)58.95 ± 0.6656.60 ± 1.100.001
C/N24.32 ± 1.0716.59 ± 0.590.001

Storage compounds that accounted only for the carbon content of the seed (i.e. carbohydrates and fatty acids) were analysed in detail. This analysis showed that fatty acids were dominant between these two groups of storage compounds (88%) in A. thaliana seeds and both increased in response to seed development under cold conditions (+14%, Table 3). The fraction of unsaturated fatty acids in the seeds was generally high (86–88% of total fatty acids) and not affected by developmental temperatures, whereas the relative proportion of individual fatty acids differed significantly (Table 3). In seeds from the 15 °C environment, α-linolenic acid (+61%) and homo-γ-linoleic acid (+72%) showed the strongest increases compared with their respective concentrations in seeds from the warm environment (Table 3). Other fatty acids were increased by 20–30%. The relative amounts of vaccenic acid and behenic acid, which constitute small fractions of total fatty acids, remained unaffected by environmental temperatures (Table 3). Only oleic acid relative amounts were significantly smaller (–29%) in seeds from the cold environment than in those from the warm environment (Table 3). The comparison of mol percentages showed that the fatty acid profile was barely affected with two exceptions: seeds from 15 °C contained a higher fraction of α-linolenic acid and a lower fraction of oleic acid than those from the 25 °C environment (Table 3). The influence of the temperature during reproduction on seed carbohydrates was negligible (Table 3).

Table 3.  Carbohydrates and fatty acids (µmol g–1 F.wt) in seeds of Arabidopsis thaliana grown at 15 and 25 °C, respectively
CompoundSymbol15 °C25 °CP%15/%25
  1. %15/%25 indicate mol percentage of a given fatty acid in seeds of parents maintained at 15 or 25 °C, respectively. Data indicate means (= 6, ± SD).

  2. nd, not detected; na, not applicable.

Soluble sugar 148.8 ± 10.7137.6 ± 12.00.067na
Glucose 5.7 ± 1.35.5 ± 0.50.691na
Fructose 1.5 ± 0.31.2 ± 0.20.037na
Sucrose ndnd na
Starch 5.3 ± 0.95.6 ± 0.50.508na
Fatty acids (sum) 1194.6 ± 71.61040.4 ± 80.50.001100/100
Palmitic acid16:096.7 ± 5.474.1 ± 2.80.0018.1/7.1
Palmitoleic acid16:1 (9Z)4.1 ± 0.53.2 ± 0.40.0010.3/0.3
Stearic acid18:041.2 ± 2.429.2 ± 1.20.0013.4/2.8
Oleic acid18:1 (9Z)154.5 ± 13.1217.8 ± 36.90.00112.9/20.9
Vaccenic acid18:1 (11Z)19.9 ± 1.220.3 ± 1.80.5071.7/1.9
Linoleic acid18:2 (9Z, 11Z)353.8 ± 23.2319.0 ± 23.90.00229.6/30.6
α-Linolenic acid18:3 (9Z, 11Z, 15Z)236.5 ± 21.7146.6 ± 12.60.00119.7/14.0
Arachidic acid20:022.3 ± 1.318.5 ± 0.90.0011.8/1.8
Gondoic acid20:1 (11Z)218.5 ± 11.0180.1 ± 16.50.00118.2/17.3
Homo-γ-linoleic acid20:2 (11Z, 14Z)26.0 ± 1.515.1 ± 0.80.0012.1/1.4
Behenic acid22:02.9 ± 1.32.7 ± 0.50.6050.2/0.2
Erucic acid22:1 (13Z)18.1 ± 1.413.8 ± 1.30.0011.5/1.3

Growth performance of progenies from warm and cold parental environments

To investigate whether differences in seed quality as the result of different parental environments influenced the performance of the progenies, germination rate, root development, leaf formation rates and biomass production were investigated. In the temperature range from 10 to 30 °C, seeds originating from the warm environment completed germination always about 24 h faster than those from the cold environment (data not shown). Root growth from seeds originating from the cold environment was always significantly slower than that of roots emerging from seeds produced in the warmer environment (Fig. 1). Progenies from both environments showed optimum root growth rates between 20 and 25 °C, but those originating from the cold environment were ca. 20% retarded compared with those from the warm environment (Fig. 1).

Figure 1.

Growth rates of Arabidopsis roots at different temperatures. Seeds were used from parents grown at 15 °C (open symbols) and 25 °C (black symbols), respectively. Root lengths were measured regularly. Means were calculated for the offspring of each parental temperature environment on each plate and were used to determine growth rates by regression analysis. The means of growth rates (± SD) were plotted against the growth temperatures. If errors bars are missing, they were smaller than the symbol. The P-value was calculated for paired samples using a signed rank test.

In the following text, offspring from parents grown at 15 °C will be abbreviated with O15 and that of parents from 25 °C as O25. When the progenies of these two different parental environments were grown in soil for about 7 weeks at optimum temperatures (20 °C), growth differences relating to different seed origin persisted. Origin 15 (O15) plants showed significantly lower leaf formation rates, which resulted in about 35% lower biomass production than in origin 25 (O25) plants (Table 4).

Table 4.  Growth and biomass of progenies from Arabidopsis thaliana plants grown at 15 and 25 °C, respectively
ParameterO25 (°C)O15 (°C)P
  1. Seeds from parental growth temperatures of 15 °C [origin 15 (O15)] and 25 °C [origin 25 (O25)] were used. The seedlings were grown for 7 weeks in soil at 20 °C. Data indicate means (n = 10–15, ± SD).

Leaf formation rate (number day–1)0.318 ± 0.0140.280 ± 0.0130.001
Fresh mass (g plant–1)1.64 ± 0.601.06 ± 0.380.050

Fitness of progenies from warm and cold parental environments

To investigate fitness of plants originating from environments of 15 and 25 °C, rosette diameter, leaf biomass, number of siliques, total seed mass and seed numbers were determined by growing plants of seeds from both origins in competition at both temperatures, 15 and 25 °C, respectively (Table 5). When grown at 15 °C, O25 plants developed higher leaf area, more leaf biomass and a higher number of siliques than O15 plants. Total mass and number of seeds produced per plant were not affected by seed origin (Table 5). This shows that at low growth temperatures, plants originating from seeds formed at low temperature have no competitive advantage over plants originating from seeds formed at higher temperatures. When O25 and O15 plants were grown at 25 °C, O25 plants performed better than O15 plants as indicated by higher biomass production and higher number of siliques. O25 plants grown at 25 °C showed a trend towards increased production of seed but not greater seed numbers (Table 5). Regardless of seed origin, higher growth temperatures stimulated vegetative growth and seed mass (Table 5).

Table 5.  Fitness parameters of Arabidopsis thaliana grown at 15 or 25°C, respectively
  1. Seeds from parental growth temperatures of 15 °C [origin 15 (O15)] and 25 °C [origin 25 (O25)] were used. After germination, offsprings from these two different parental reproductive environmental temperatures were grown in competition at either 15 °C (T15) or 25 °C (T25), respectively. Data are means (= 32–58, ± SE). Different letters in rows indicate significant differences at ≤ 0.05.

Rosette diameter (mm)107.0 ± 2.0a112.0 ± 2.0b113.0 ± 2.0b117.0 ± 2.0b
Dry mass (g)0.78 ± 0.06a1.00 ± 0.06b0.85 ± 0.08ab1.30 ± 0.07c
Number of main siliques2.6 ± 0.4a3.9 ± 0.4b4.4 ± 0.6b6.6 ± 0.5c
Number of side siliques7.2 ± 0.6a7.7 ± 0.6a11.8 ± 1.0b12.4 ± 0.8b
Seed mass (mg plant–1)34.0 ± 2.0a41.0 ± 4.0a48.0 ± 8.0ab67.0 ± 8.0b
Seed number × 103(plant–1)1.9 ± 0.4a2.1 ± 1.2ab3.9 ± 2.9c3.8 ± 2.9bc

Stress tolerance of progenies from warm and cold parental environments

To find out whether progenies from seeds originating from low temperatures differed in stress tolerance compared to those from the warm environment, A. thaliana plants from both origins were grown at optimum temperature (20 °C). Freezing tests at −5 °C were performed in darkness. Exposure of non-acclimated plants to frost resulted in complete damage of leaves of both seed sources (data not shown). Therefore, the plants were briefly acclimated to low temperature (2 °C for 9 h) before frost exposure. Low temperature acclimation in darkness did not result in differences in the maximum quantum yield of photosynthesis (Φ) compared to controls maintained at 20 °C (Fig. 2). Acclimated plants were exposed at −5 °C, which resulted in drastic decreases in Φ in Arabidopsis plants of both seeds origins (Fig. 2). Recovery of O15 plants was significantly faster than that of O25 plants (Fig. 2), but neither O25 nor O15 plants showed long-term negative effects of the frost event (Fig. 2).

Figure 2.

Relative quantum yield of photosynthesis of Arabidopsis thaliana exposed to freezing stress. Plants originated from seeds of parents grown at 15 °C (open symbols) and 25 °C (black symbols), respectively. The plants were grown at 20 °C, acclimated for 9 h to 2 °C, exposed for 3 h to −5 °C (indicated by inverted arrows on x-axis), subsequently maintained for 24 h in darkness at 2 °C and then transferred to 20 °C. Bars indicate measurements after 4 d at 20 °C. Data are means of Φ(cold stressed)/Φ(control) (= stress/control) measured in darkness in plants from two independent reproduction experiments (= 12, ± SD). Significant differences at P ≤ 0.05 between the two progenies are indicated by *.

To study low temperature acclimation in greater detail, A. thaliana plants from both seed origins were exposed for 3 d to chilling at 3 °C. Cold-stressed plants showed decreases in photosynthetic quantum yield (Φ, Fig. 3). Because Φ recovered fully during the night, the observed decreases reflect down-regulation of photosynthetic electron flux and not injury. Initially, down-regulation was stronger in progenies from the cold environment than in those from the warm environment (Fig. 3). After 3 d, progenies of the warm environment showed stronger depression in Φ than those from the cold environment (Fig. 3). These observations indicate that ambient temperatures in the parental reproductive environment have carry-over effects on the physiological behaviour of the progenies under stress.

Figure 3.

Relative quantum yield of photosynthesis of Arabidopsis thaliana exposed to cold stress. Plants originated from seeds of parents grown at 15 °C (open symbols) and 25 °C (black symbols), respectively. The plants were grown at 20 °C for 7 weeks before exposure to cold stress. Data are means of Φ (cold stressed)/Φ (control) (= stress/control) of plants from two independent experiments (= 10, ± SD). Significant differences at ≤ 0.05 between the two progenies are indicated by *. Hatched sections indicate dark periods.

To analyse the molecular basis of differences in stress responsiveness, a preliminary screen of differentially regulated genes was conducted using a filter with ca. 1400 Arabidopsis TFs and some stress-related structural genes. Of 19 putative candidates, the transcript levels of two up- and two down-regulated genes were selected for further analysis. qRT-PCR showed that transcript levels of PRR9 and F3H in cold-stressed plants were significantly up-regulated compared to those of non-stressed plants (PRR9: 2.22 ± 0.77, P = 0.025; F3H: 3.19 ± 1.73, P = 0.020), respectively, whereas those of AP2 and NAP were relatively decreased (AP2: −3.05 ± 0.01, P = 0.001; NAP: −1.64 ± 0.16, P = 0.071). To address the changes of stressed versus non-stressed plants and of O25 plants versus O15 plants in greater detail, various comparisons were conducted (Fig. 4). Non-stressed O25 plants showed a trend (P = 0.07–0.09) towards higher transcript levels of PRR9 and AP2, respectively, compared with non-stressed O15 plants (Fig. 4a). Transcript levels of NAP in stressed O25 plants were maintained at significantly higher levels than those in stressed O15 plants (Fig. 4b). Stressed O25 plants showed a significant increase in F3H and significant decreases in AP2 transcript levels compared with non-stressed O25 plants, whereas transcript levels of NAP and PRR9 were unaffected (Fig. 4c). In stressed O15 plants, AP2 was also significantly decreased, whereas NAP, PRR9 and F3H showed only trends towards down- respective up-regulation with P-values between 0.058 and 0.077 (Fig. 4d).

Figure 4.

Relative transcript levels in leaves of Arabidopsis thaliana exposed to chilling stress and in non-stressed offspring of parents grown at 15 °C [origin 15 (O15)] or 25 °C [origin 25 (O25)], respectively. Data were normalized with respect to 18S rRNA, and means were calculated for plants harvested after 3 d of chilling at 3 °C from three plants per independent reproduction experiment (= stress) or from plants grown at 20 °C (= control), (= 6, ± SE). (a) Relative transcript levels in leaves of controls of O25 plants versus controls of O15 plants. (b) Relative transcript levels in leaves of stressed O25 plants versus stressed O15 plants. (c) Relative transcript levels in leaves of stressed O25 plants versus O25 control plants. (d) Relative transcript levels in leaves of stressed O15 plants versus O15 control plants. P-values for these comparisons are indicated by the numbers above the error bars. PRR9, pseudo response regulator 9; F3H, flavanone 3-hydroxylase.


Growth conditions of parental environments such as shade, day lengths, elevated CO2 and summer or spring climate can affect performance and reproductive fitness of the following generation in A. thaliana (Andalo et al. 1998; Andalo et al. 1999; Munir et al. 2001; Callahan & Pigliucci 2002). In general, parental effects have been associated with seed size or the quantity of reserves in seeds (Charest & Potvin 1993; Johnsen & Ostreng 1994; Wulff 1995; Galloway 2001a,b). In our study, warm temperatures in the parental environment resulted in seeds with significantly higher relative nitrogen amounts than in those produced at low temperatures (Table 2). This effect transmitted by the parental environment may have resulted in faster germination and root elongation growth, increased biomass and seed production of the offspring (Fig. 1, Tables 4 and 5). The benefits of increased nitrogen respective protein amounts in seed reserves are known for a long time (Gray et al. 1988). However, our study is the first to show that a competitive advantage acquired at warm temperatures in the parental environment is also maintained over a range of higher as well as lower growth temperatures of the progenies.

A further important effect of temperature was found on seed oil composition. Low temperatures resulted in about twofold increases in the ratio of α-linolenic (18:3) to oleic acid (18:1) from 0.67 in the warm to 1.53 in seeds from the cold environment (Table 3). Small shifts in the degree of desaturation can also be caused by vernalization (i.e. maintenance of seeds themselves at low temperatures); however, this was not observed in the Columbia ecotype used here (O’Neill et al. 2003). This indicates that the temperature regime during seed development is a major factor influencing seed oil composition in Arabidopsis. Similar low temperature-induced shifts in 18:3/18:1 desaturation ratios were found in canola (Brassica napus) and sunflower seeds, and it was suggested that desaturases necessary for the conversion of saturated to unsaturated fatty acid may be inhibited at higher temperatures (Deng & Scarth 1998; Pritchard et al. 2000; Sánchez-García et al. 2004). Polyunsaturated fatty acids are important for cold tolerance (Kodama et al. 1994). Because the major fraction of 18:3 is present in the embryo and not in the endosperm of Arabidopsis seeds (Penfield et al. 2004), enrichment of 18:3 versus 18:1 may be required for temperature acclimation of the resting embryo. Because plants originating from 18:3-enriched seeds have no developmental benefit, elevated levels of 18:3 obviously do not provide a competitive advantage during germination and seedling establishment.

Our study clearly shows that under optimum growth conditions, transmission of parental effects as a result of low temperatures do not lead to increased reproductive fitness in progenies from low temperature compared to those from high temperatures. However, until plants reach the reproductive stage, they are likely to encounter fluctuating environmental temperatures. Therefore, we were also interested in the question whether offspring originating from low temperature environments would be better protected against sudden events of freezing and chilling stress. When non-acclimated plants were exposed to below zero temperatures, all died. This has also been reported for E. crus-galli progenies from warm and cold parental environments exposed to freezing shock (Charest & Potvin 1993), but contrasts with data for tree species, where the temperature in the maternal reproductive environment caused imprinting effects improving freezing tolerance of the progenies (Johnsen et al. 1995, 1996). Because the overall degree of DNA methylation is greater in tree species than in Arabidopsis and is strongly affected by the maternal environment (Fraga, Rodríguez & Cañal. 2002; Baumann 2004; Baurens et al. 2004), we may speculate that imprinting is perhaps more important for species with long generation times than for species with short life cycles.

Nevertheless, when Arabidopsis was exposed to chilling temperatures, small but significant differences in the stress responses of the progenies from different parental environments occurred at the physiological as well as at the molecular level (Figs 2–4). Because chilling stress may cause accumulation of reductant in the stroma as a result of diminished rates of CO2 assimilation, it may lead to tailback of electrons into photosynthetic electron transport chain under illumination. As a consequence, the quantum yield of photosystem II in light is decreased but recovers in darkness (Savitch et al. 2001). Our data show that photosystem II was not damaged by chilling because the quantum yield of photosynthesis, which was depressed in illuminated chilled plants, recovered in darkness. Our data imply that progenies from parents of low temperature environments were initially more prone to over-reduction because they showed a stronger decrease in quantum yield of photosystem II in light than progenies from parents of higher environmental temperatures (Fig. 3). While the fluorescence yield of the latter plants showed no recovery during chilling, that of progenies from low temperature parents increased by 30% (Fig. 3). Freezing stress after low temperature acclimation resulted in about 25% injury to photosystem II in both progenies, which was evident from the drop in yield of photosystem II in darkness (Fig. 2). But restoration of photosystem II activity was faster in progenies originating from parents of low temperature environments than in those originating from the warm environment. Taken together, these results show that the parental environment affects the capability of the progenies to restore homeostasis during cold stress.

We identified several candidate genes with possible functions in the recovery of homeostasis. When data from all progenies were taken together, AP2 showed significant down-regulation, and PRR9 and F3H showed significant up-regulation after 3 d of cold acclimation. Analysis of cold responses of these genes using GENEVESTIGATOR (https://www.genevestigator.ethz.ch (Zimmermann et al. 2004) showed up- respective down-regulation of F3H (2.35 ×), PRR9 (7.44 ×) and AP2 (0.8 ×) in various array analyses. F3H is co-regulated with chalcone synthase, both forming the entry point for flavonoid and anthocyanin biosynthesis (Pelletier & Shirley 1996), a pathway that is activated under cold stress (Kreps et al. 2002; Provart et al. 2003; Ray et al. 2003) and that may protect photosystem II from excess light (Huner, Öquist & Sarhan 1998). Sugar signalling is involved in the regulation of the genes of the flavonoid biosynthetic pathway (Solfanelli et al. 2006). Because low temperatures lead to sucrose accumulation in Arabidopsis (Savitch et al. 2001), the modulation of sugar levels may form a regulatory device triggering acclimatory responses. As F3H showed similar regulation in both progenies, we conclude that it does not play a role in parental-transmitted differences in cold acclimation.

Likely candidates involved in progeny-related effects in cold responsiveness are PRR9 and NAP, which showed significant differences in offspring from low compared to those from warm-temperature parents (Fig. 4). PRR9 is involved in the regulation of the internal circadian clock (Eriksson et al. 2003) and is temperature responsive (Salome & McClung 2005). PRR9 probably functions in temperature and light input pathways or represents an element of an oscillator necessary for the clock to respond to temperature signals (Salome & McClung 2005). Our finding that this TF was more strongly up-regulated in progenies from cold than in those from warm parental environments under stress suggests that PRR9 is important for recovery of homeostasis during cold stress. In fact, PRR9 has recently been suggested to enhance fitness in local environments because Arabidopsis accessions from different latitudes have shown significant variation in the period and phase of the circadian clock, allowing adaptation to different environmental conditions (Michael et al. 2003).

The TFs NAP, a homologue of NAM (no apical meristem, Sablowski & Meyerowitz 1998) belonging to the NAC-family, and AP2 (AP2 domain containing protein RAP 2.3) are involved in the regulation of growth and development (Souer et al. 1996; Okamuro et al. 1997; Riechmann & Meyerowitz 1998; Ôoka et al. 2003). Both were down-regulated in response to cold acclimation but NAP was significantly less down-regulated in progenies from warm than in those from cold parental environments. This resulted in a significant progeny-related effect on NAP transcript levels (Fig. 4b), which may explain why GENEVESTIGATOR analysis showed increased (2.11) and not decreased expression. NAP over-expression accelerates and knock-out mutation delays senescence (Guo & Gan 2006). Because low temperatures also delay senescence, cold responsiveness of NAP ties this component into the regulatory network important for environmental adaptation of plants.


We are grateful to C. Kettner for maintenance of the plants, to T. Klein and P. Meyer for technical assistance, and to the Deutsche Forschungsgemeinschaft for financial support.