Root cooling strongly affects diel leaf growth dynamics, water and carbohydrate relations in Ricinus communis


Achim Walter, ICG-3 (Phytosphere), Forschungszentrum Jülich, 52425 Jülich, Germany. Fax: +49 2461 61 2492; e-mail:


In laboratory and greenhouse experiments with potted plants, shoots and roots are exposed to temperature regimes throughout a 24 h (diel) cycle that can differ strongly from the regime under which these plants have evolved. In the field, roots are often exposed to lower temperatures than shoots. When the root-zone temperature in Ricinus communis was decreased below a threshold value, leaf growth occurred preferentially at night and was strongly inhibited during the day. Overall, leaf expansion, shoot biomass growth, root elongation and ramification decreased rapidly, carbon fluxes from shoot to root were diminished and carbohydrate contents of both root and shoot increased. Further, transpiration rate was not affected, yet hydrostatic tensions in shoot xylem increased. When root temperature was increased again, xylem tension reduced, leaf growth recovered rapidly, carbon fluxes from shoot to root increased, and carbohydrate pools were depleted. We hypothesize that the decreased uptake of water in cool roots diminishes the growth potential of the entire plant – especially diurnally, when the growing leaf loses water via transpiration. As a consequence, leaf growth and metabolite concentrations can vary enormously, depending on root-zone temperature and its heterogeneity inside pots.


The immediate environment of growing roots and leaves differs strongly in temperature, light, humidity and other environmental factors to which the growing organs are exposed. For example, soil temperature, like air temperature, has sinusoidal oscillations on a diel (24 h) scale (Hillel 1998). However, depending on soil depth and moisture, changes in soil temperature are delayed and lower in amplitude than the atmospheric temperature variation. In a typical diel pattern (shown in Walter, Silk & Schurr 2009), atmospheric temperature varies by 16 K and reaches a maximum around 1400 h, whereas soil temperature at 10 cm depth varies by merely 3 K and reaches a maximum around 1600 h. Soil temperature variation is even more damped at greater depth and the diel average temperature to which shoot and root are exposed can differ strongly, depending on weather and season. The importance for the plant of different temperatures at root and shoot has been analysed in relatively few studies throughout the last decade. Yet, in studies that did expose root and shoot to different temperatures, a clear influence of root temperature on leaf growth and on partitioning of biomass between roots and leaves has been reported (Watts 1972; Davies & Van Volkenburgh 1983; DeLucia, Heckathorn & Day 1992; Engels 1994).

Most greenhouse and growth chamber experiments use plants cultivated in relatively small, black containers although a small rooting volume can strongly hamper shoot growth (Thomas & Strain 1991). Moreover, it has to be expected that roots experience uncommon temperature regimes there and that this may affect shoot growth. In a study with soybean it has been shown for example that watering the plants with cold water (15 °C instead of 25 °C for control plants) leads to a 10% growth decrease (Passioura 2006). Yet, more often, root temperature in pot experiments will be higher than in the field because of radiative heating of the black pots. Increased root temperature can also decelerate leaf growth because of strongly increased root respiration and other factors, which was demonstrated in a study on the C4-grass Andropogon gerardii (DeLucia et al. 1992) that had optimal leaf growth at a root temperature of 25 °C. At low root temperature, it is well known that root water uptake is reduced because of decreased transmembrane transport (Markhart et al. 1979), which can lead to wilting (Bloom et al. 2004) or other symptoms resembling salt stress that are induced by the decreased shoot water availability (Davies & Van Volkenburgh 1983).

From experiments where the leaf growth zone temperature has been controlled along with root temperature, it seems that the close proximity of monocot leaf meristems to the soil leads to a strong influence of soil temperature on monocot leaf processes (DeLucia et al. 1992; Engels 1994). In fact, for maize (Watts 1972), leaf extension rate was proportional to the root temperature when the meristem temperature was not controlled, but when meristem temperature was actively kept constant, leaf extension rate was constant too, regardless of the soil temperature. In dicot plants, it is much less clear, how root temperature affects the dynamics of leaf growth, which are governed to a much higher extent by intrinsic temporal patterns probably regulated by the circadian clock (Walter et al. 2009). Experiments exposing the root system of Phaseolus to a constantly low temperature throughout 24 h indicate that leaf growth is inhibited during the day but not at night (Davies & Van Volkenburgh 1983). Such a daytime leaf growth inhibition seemed to prevail even when root temperature was increased for 8 h within a 12 h day (Ainsworth, Walter & Schurr 2005).

As carbohydrates regulate leaf growth in a diel pattern (Smith & Stitt 2007; Wiese et al. 2007; Sulpice et al. 2009), it is important to learn how carbohydrate concentrations react to altered root temperature. Presumably, shoot photosynthesis will not be directly affected by cooling the root. Yet, shoot carbohydrate would be likely to increase if the root's demand was reduced, leading to end-product inhibition. As the diel leaf growth cycle in dicot plants is strongly affected by the interaction of water relations and carbohydrate metabolism (Walter et al. 2009), it is important to perform physiological experiments taking all of these aspects into account.

Hence, it was the aim of this study to investigate how cooling the root system throughout the diel cycle affects leaf growth dynamics in the dicot plant Ricinus communis. Leaf growth analyses were complemented by investigations of water relations, carbon translocation and carbohydrate concentrations that are a vital part of the physiological framework facilitating the diel leaf growth cycle. Ricinus communis was selected as an experimental system as, in contrast to rosette plants, leaves and roots are separated widely enough to independently control their temperature, and as leaf growth dynamics, water relations and carbohydrate metabolism have been analysed intensively in this plant species.


Plant material

Ricinus communis L. (ecotype Carmencita rot) plants were germinated and grown in a greenhouse for 10 d. The plants were then translocated to the growth chamber some days before the experiment started (usually 14 d after germination). Plants were grown in 40 pots (1 L) filled with a pre-fertilized soil (ED73, Balster Einheitserdewerk, Fröndenberg, Germany). During germination and for the duration of the experiment soil water content was kept at optimal level.

When these pots were exposed to a temperature change in a preliminary experiment, temperature inside the pot acclimated to a new steady-state value within 1 to 4 h, depending on the exact location within the pot (Fig. 1). In this experiment, a soil-filled pot was cooled down overnight to 5 °C and then placed in a climate chamber (air temperature = 20 °C, relative humidity = 60%, light intensity = 600 µmol m−2 s−1 photosynthetically active radiation). At each position within the pot, a temperature higher than air temperature was reached within 15 to 100 min. At the bottom of the pot, steady-state temperature was reached fastest, exceeding air temperature by 3 to 5 °C. At the top, temperature still increased markedly after 4 h of incubation in the new temperature regime and it exceed air temperature already by 10 K at that time.

Figure 1.

Temperature at nine different locations of a soil-filled pot. The pot was translocated from a cold room (5 °C) to a climate chamber. Exact locations of temperature analysis are indicated in the scheme.

Experimental schedule and environmental parameters

The entire study consists of a suite of four subsequent experiments with different root temperature treatments. In each of these experiments, different parameters were monitored. In each of these experiments, air temperature was kept at 22 °C in the growth chamber and the light intensity measured with a quantum light sensor (LI-190SB, Li-Cor, Lincoln, NE, USA) was 200 µmol m−2 s−1, 12 h a day. Photosynthetic active radiation, ultraviolet (UV)-A and UV-B were measured using an optometer (X12, Gigahertz-Optik GmbH, Puchheim, Germany). The light was provided by an array of fluorescent lamps. The root cooling system consisted of an insulated chassis containing a heat exchanger. Coolant was pumped through the heat exchanger and thus lowered temperature inside the box compared with the outside. Several apertures were made in the upper side of this box, allowing to place the pots inside. The cooling pump allowed controlling of coolant temperature and it took about 30 min to reach the desired temperature within the entire soil volume of the pot (Füllner 2007). Up to four cooling boxes (each containing maximally 10 plants) were used for these experiments.

In the first experiment, root temperature was kept constant for 19 d with three different treatments (10, 15 and 20 °C). Leaf growth and shoot biomass was monitored. In experiment two, root temperature was modified in each cooling box every 4 d. In box A, roots were first exposed to 20, then to 15, then to 10 °C for 4 d, respectively. In box B, roots were first exposed to 15, then to 10 and then to 20 °C. In box C, roots were first exposed to 10, then to 20 and then to 15 °C. In this experiment, leaf growth and diel leaf growth cycles were analysed. In experiment three and four, only the two extreme root temperature treatments were applied, as experiments one and two showed intermediate results for the intermediate temperature treatment. In experiment three, roots were exposed to 20 °C for 7 d, then to 10 °C from day 8 to day 12 and again to 20 °C thereafter. Here, leaf growth, root growth, carbohydrate concentrations, transpiration rate and xylem pressure were analysed. In experiment four, roots were exposed to 20 °C except for a 24 h period, during which they were exposed to 10 °C. Here, carbon partitioning was analysed.

Ten thermocouples (type T, cooper/constantan and 0.5 mm in diameter) were used (three on the meristems, six in the soil, and one measuring air temperature at canopy level) to record temperature. All temperature data were recorded using a laptop PC and signal interfaces (cFP-1804; cFP-TC-120; cFP-CB-3, National Instrument, Austin, TX, USA). Data were pooled every second and an averaged value recorded every 10 s using a Labview-based (National Instrument) program. Air temperature and air relative humidity were also recorded every 10 min using an independent data logger (Testo 175-H2, Testo AG, Lenzkirch, Germany).

Leaf growth

The length and width of initiated leaves were manually measured every other day during experiment one. Rank specific leaf form factors (0.78 for cotyledon, 0.72 for primary leaves and 0.79 for the following leaves) were used to calculate leaf area from leaf dimensions (length × width × form factor = area). For diel leaf growth assessment, primary leaves were mechanically fixed in a stationary position (compare Walter, Feil & Schurr 2002a; Fig. 1 therein) to keep growing leaves in the focal plane of the camera. Each investigated leaf was fixed at the petiole using a stripe of Parafilm (Parafilm M®, Pechiney Plastic Packaging Company, Chicago, IL, USA) and kept flat with five nylon threads, clamped to the edge of the leaf using shortened hair clips and fabric tape to protect the leaf. Each of the threads was pulled with a weight of 12 g and spun over a metal ring surrounding the leaf, a procedure that does not affect temporal or spatial growth patterns (Walter et al. 2002a). Leaf images were acquired with high resolution, monochrome CCD cameras (Scorpion SCOR-20SO, Point Grey Research, Vancouver, BC, Canada), positioned above the plants and equipped with a standard objective lens (25 mm; Cosmicar/Pentax; The Imaging Source, Bremen, Germany) and an infrared interference filter (880 nm; Edmund Optics, Karlsruhe, Germany). Constant illumination throughout day and night was provided by six infrared diode clusters (880 nm; Conrad Electronics, Hirschau, Germany). Grey value images were taken every 180 s and were saved in a multi-tiff format. Image sequences were evaluated with algorithms based on a structure-tensor approach (optical flow via the brightness constancy constraint equation (BCCE) (Bigün & Granlund 1987; Schmundt et al. 1998) that calculates velocities from all moving visible structures at the leaf surface within the image sequence of a growing leaf. Area relative growth rates (RGRs) were calculated as the divergence of the estimated velocity field by selecting an area of interest (AOI) within the image and tracking the structure within this AOI with time. For more details, see Schmundt et al. (1998), Walter et al. (2002a) and Matsubara et al. (2006).

Root growth and water losses

During experiment three, 10 plants were grown with the root system and the surrounding soil (1 L) in a transparent plastic bag sealed at the base of the stem. This allowed carefully taking the bag out of the pot every second day for direct access to the roots growing at the surface of the bag. The position of each root tip was marked every other day. From this data, the increase in root length and root number was calculated. Furthermore, the pots were weighed at the beginning and end of each day and the lost water was replaced to give a direct estimate of transpired water as the increase in plant biomass was negligible (less than 1% of the weight difference caused by water loss via transpiration).


In experiment three, tissue sections of roots and leaves were sampled for carbohydrate analysis in the afternoon. Two discs per leaf (diameter: 8 mm) were punched out of full-grown cotyledon and growing primary leaves, were weighed, pooled, frozen in liquid nitrogen and stored at −80 °C for further extraction. Two pieces of root tissue were sampled from the main root (length: 1 cm; diameter: 1–2 mm), were rinsed with water, gently dried with tissue, pooled, frozen and stored at −80 °C. Soluble carbohydrates were extracted from frozen leaf material and glucose, fructose, sucrose and starch concentrations were analysed in a coupled enzyme assay (Jones, Outlaw & Lowry 1977) using a multiplate spectrophotometer (ht II; Anthos Mikrosysteme GmbH, Krefeld, Germany) as described in Walter et al. (2002a) and Wiese et al. (2007).

Carbon partitioning

The short-lived carbon isotope 11C decays with a half-life of 20.4 min, resulting in the emission of photons that can be detected non-invasively, because of their high energy. For plants in experiment four, 11CO2 was applied to one cotyledon, and we monitored all of the radiolabelled photoassimilate exported from the leaf, using scintillation detectors to measure radiotracer in both shoot and root, with radiation shielding arranged to limit the regions of each plant ‘seen’ by the detectors. Pairs of plants were set up in standard conditions, and after 24 h of acclimation to the shielding and detection setup, with root and shoot at 20 °C, pulses of 11CO2 were applied three times on two sequential days at 0900, 1130 and 1400 h. Soil temperature of one plant of the pair remained at 20 °C, and in the other plant it was reduced to 10 °C at 1600 h on day 1, and returned to 20 °C at 1000 h on day 2. From the time-series of radiotracer in shoot and root we used transfer-function analysis with allowance for tracer decay (Minchin & Thorpe 2003), to calculate the fraction of exported tracer being transported to root and shoot, both before and after changing the soil temperature. To show the treatment effects relative to controls, the shoot/root partitioning ratio of the treated plant calculated for each tracer pulse was divided by the partitioning ratio for the control plant measured at the same time, to allow for strong developmental effects in the young plants. The shoot/root ratio for a cotyledon's photoassimilate changes with plant development in favour of the roots because the true leaves become carbon sources for shoot growth.

Xylem pressure

The xylem pressure probe consisted of a water-wettable pressure transducer mounted in a 50 µL Plexiglas chamber and sealed to a glass microcapillary that was inserted into the xylem vessel as described in more detail in Zimmermann et al. (2004). The plant stem was attached to a fixed metal rod, with tape above and below the insertion zone, and the microcapillary was inserted slowly into the stem using a micromanipulator. When the tension dropped suddenly, indicating that the tip had reached a xylem vessel, the microcapillary was held at that position. Pressure equilibrium between the vessel and the probe was established within a few seconds, and was usually stable for several hours, allowing a minimally invasive direct reading of the pressure of the xylem sap.

Statistical analysis

Comparisons between treatments were performed using two-tailed Student's t-tests (software: SPSS Sigmastat).


Shoot growth

In experiment one, low root temperatures strongly reduced leaf development (Fig. 2a). At day 34, plants with 20 °C root-zone temperature reached 881 cm2 of total leaf area, whereas plants with root temperatures of 15 and 10 °C gave leaf areas of only 471and 288 cm2, respectively. Lower root-zone temperatures led to decreased leaf growth rates throughout the entire experiment and not only during an initial phase (Fig. 2b). As leaf area and biomass might differ in their response, we also monitored shoot biomass (Fig. 2c). The biomass difference between populations with a root-zone temperature of 10 and 20 °C is in the same range as the difference in leaf area: The final fresh weight of plant shoots grown at 10 °C root-zone temperature was a factor of three lower compared with the fresh weight of plants grown at 20 °C. Plants grown at a root-zone temperature of 15 °C showed intermediate fresh weights.

Figure 2.

(a) Total leaf area of three populations of Ricinus communis plants exposed to three constant root-zone temperatures (10 °C white symbols, 15 °C grey symbols and 20 °C black symbols; n = 10; mean value and standard error). (b) Leaf relative growth rate calculated from the increase in total leaf area depicted in (a). (c) Shoot fresh weight determined at the end of the experiment.

When plants were exposed to controlled root-zone temperatures in experiment two for 12 d and root-zone temperature was altered two times during this period, leaf growth acclimated reversibly within some days to altered temperature regimes (Fig. 3). In this experiment, the effect of root-zone temperature on leaf RGR is most clearly shown in the population exposed to 15 °C in the beginning of the experiment (Fig. 3b). Leaf RGR was initially low (7% d−1) but increased to values around 17% d−1, which were comparable with the values in experiment one at that developmental stage (Fig. 2b). Transfer to 10 °C root-zone temperature resulted in a rapid decrease of leaf RGR to values around 6% d−1, which were comparable with growth rates reached by the 10 °C population in experiment one. When root-zone temperature was finally set to 20 °C, leaf RGR increased again to a similar level as that of the respective population at this developmental stage in experiment one (around 11% d−1, Fig. 3b). Populations exposed to other permutations of root-zone temperature (Fig. 3a,c) showed similar effects, clearly demonstrating that the inhibition of leaf growth induced by lowered root-zone temperature is reversible and can be relieved within some days. The reversibility of this effect is demonstrated by the fact that the final area of the three populations shown in Fig. 3 were almost the same at the end of the experiment (a: 276 ± 31 cm2; b: 275 ± 37 cm2; c: 248 ± 26 cm2), regardless of the root-zone temperature sequence they experienced.

Figure 3.

Relative growth rate (RGR) of the total leaf area of three populations of Ricinus communis plants exposed to three root-zone temperature sequences: a = 20, 15 and 10 °C; b = 15, 10 and 20 °C; c = 10, 20 and 15 °C. n = 10, mean value and standard error.

Diel leaf growth variation

Because leaf growth varies strongly throughout the diel cycle, the growth of single leaves was investigated at higher temporal resolution. The general decrease of leaf growth with decreasing root-zone temperature was also obvious here (Fig. 4a–i). Compared with the diel leaf growth cycle without any control of root-zone temperature (Fig. 4k), three main effects of low root-zone temperature on diel leaf growth activity can be pointed out: firstly, diurnal (daytime) growth activity almost vanished, independent of the exact temperature setting (Fig. 4j,k). Secondly, average nocturnal growth intensity was correlated with the selected root-zone temperature (Fig. 4k). Nocturnal average RGR decreased in the first population shown in Fig. 4a–c from the first investigated stage shown in a (root-zone temperature at 20 °C) to b (15 °C) to c (10 °C). In the second population, it decreased from d (15 °C) to e (10 °C) and increased again at the latest investigated stage to f (20 °C). Finally, in the third population, average nocturnal RGR increased from g (10 °C) to h (20 °C) and decreased to iI (15 °C). Also, when average nocturnal RGR was compared between populations with developmental stage fixed, growth rates correlated with root-zone temperature: a > d > g; h > b > e, f > i > c. These results correspond very well to the results for the growth of the total leaf canopy (Fig. 2b). Thirdly, timing, amplitude and shape of the nocturnal RGR peak were not strongly affected by root-zone temperature. A nocturnal maximum RGR of approximately 1.5% h−1 was reached a few hours before the onset of light, whether or not root temperatures were regulated.

Figure 4.

(a–i) Average diel cycle of leaf relative growth rate (RGR) of three populations of Ricinus communis plants exposed to different sequences of root-zone temperature. Population one (a–c) was exposed to 20 °C for 4 d, then to 15 °C for 4 d and then to 10 °C for another 4 d (populations two and three: root-zone temperatures as specified in panels d–i, respectively). (j) Average diel cycle of leaf RGR with unregulated root temperature (‘control’ plants of this experiment). (k) Normalized average diurnal and nocturnal leaf RGR from a–i. Average diurnal (or nocturnal) RGR was related to average diurnal (or nocturnal) RGR of control plants, respectively. n = 4–12, mean value and standard error.

Root growth

In experiment three, decreasing root-zone temperature from 20 to 10 °C showed that both the velocity of root tips and the induction of new lateral roots was strongly inhibited (Fig. 5). With 3 d cold, velocity of roots was a factor of five lower than in control plants: the main root grew by only 5 mm whereas control roots grew by 25 mm. The reduction in root growth was much stronger than that in leaf growth. Nevertheless, velocities of root tips and lateral root initiation rate never decreased to zero and they resumed when the root-zone temperature was reset to 20 °C.

Figure 5.

Root growth of Ricinus communis plants exposed to different root-zone temperatures. (a) Number of root tips counted in control plants (T = 20 °C) and cooled plants (T = 10 °C between day 18 and 23 and T = 20 °C else). (b) Length of the main root relative to initial value at day 17. (c) Increase in number of root tips per day. (d) Velocity of the tip of the main root. n = 3–7, mean value and standard error.

Carbohydrates and water relations

In experiments three and four, several physiological parameters defining the framework of leaf growth control were monitored. Carbohydrates in roots, cotyledons (full grown) and primary (growing) leaves (Fig. 6) of plants with low root-zone temperature were all increased compared with the control plants. Strongest differences between cooled and control plants were observed for the glucose and fructose content of growing leaves, with values increased by a factor of four, whereas sucrose and starch concentrations were increased by a factor of two to three. Roots contained negligible amounts of starch either during or after root cooling; for sucrose, fructose and glucose, lower concentrations compared with leaf samples were found, with much higher concentrations in treated plants.

Figure 6.

Starch and soluble sugar content of tissues from two populations of Ricinus communis plants exposed either to constant root-zone temperature throughout the experiment (control, T = 20 °C) or to 5 d of cold root-zone temperature (cooled, T = 10 °C) followed by a return to control temperature (recovery, T = 20 °C). n = 6–9, mean value and standard error. The inset shows relative growth rate (RGR) of the total leaf area of the two populations during cold root-zone temperature (n = 10, mean value and standard error).

When cooled plants were allowed to recover in the following week, carbohydrate concentrations were comparable in control plants and formerly cooled plants in all tissues – with the exception of sucrose and starch values in growing leaves. Sucrose and starch was significantly lower in recovered cooled leaves compared with control leaves (P ≤ 0.001 and P = 0.002). During the cooled phase, leaf RGR of the sampled leaves was a factor of two lower compared with control leaves in the same developmental stage (inset of Fig. 6).

Transpiration rates of plants from cooled and control groups were comparable both during night and day, respectively (Fig. 7a), with even somewhat higher values in cooled plants compared with control plants. In contrast to this, xylem pressure of cooled plants was significantly lower (P ≤ 0.001) compared with control plants: In the late afternoon, control plants reached values around −2.6 bar, whereas cooled plants had a xylem pressure of −3.8 bar (Fig. 7b).

Figure 7.

(a) Nocturnal and diurnal transpiration rates of two populations of Ricinus communis plants exposed to different root-zone temperatures (control: T = 20 °C; cooled: T = 10 °C). n = 3–7, mean value and standard error. (b) Diurnal xylem pressure values of both control and cooled Ricinus communis plants. n = 3, mean value and standard error.

We used 11C tracer in experiment four to show the partitioning of carbohydrate between shoot and root, after export from a cotyledon. There was a dramatic increase in shoot/root partitioning, relative to controls, in plants with root-zones cooled from 20 to 10 °C (Fig. 8). Two cases are shown, and the shoot/root partitioning increased threefold in one case, and fivefold in the other. The effect of cooling was reversible – when roots were returned to the warmer control temperature, partitioning immediately decreased, the shoots then receiving a smaller fraction of the carbon exported from the labelled cotyledons.

Figure 8.

11C gain data from two pairs of Ricinus communis plants exposed either to a constant root-zone temperature (T = 20 °C) throughout the experiment or to a transient root-zone cooling event (10 °C during 24 h). Depicted values show shoot/root ratio of a cooled plant divided by shoot/root (S/R) ratio of the control plant (open symbols: pair one, closed symbols: pair two).


Growth reaction from day to day

Cold root temperatures resulted in decreased leaf growth of Ricinus communis (Fig. 2). The decrease was reversible and the acclimation to an altered root temperature occurred within days (Fig. 3). These results correspond to findings from monocot plants (Watts 1972; DeLucia et al. 1992) and they are consistent with the view that dicot leaf growth responds linearly to a range of atmospheric temperatures when analysed from day to day and not at a higher temporal resolution (Granier & Tardieu 1998). For the root itself, a similarly rapid and very intense reaction was found that was also reversible (Fig. 5). This finding is a logical consequence of the fact that root growth follows alterations in temperature almost linearly and immediately as long as temperature is within a physiological range (Pahlavanian & Silk 1988; Walter et al. 2002b; Hummel et al. 2007). The severe decrease of root elongation and ramification when temperature was decreased from 20 to 10 °C indicates a strong loss of functionality that could then indirectly affect shoot growth, for example, via diminished transport of water or nutrients to the shoot, via diminished uptake of metabolites provided by the shoot or via altered exchange of signalling substances between shoot and root.

Leaf growth reaction during the diel cycle

In contrast to the almost linear response of day-to-day leaf growth and biomass accumulation to root-zone temperature (Fig. 2), diel leaf growth patterns show a highly non-linear response (Fig. 4). Although nocturnal average growth rates scale well with the temperature, diurnal average growth rates are severely decreased to values close to zero and nocturnal maximal growth rates are more or less independent of root-zone temperature. This result is a cornerstone of our study. It demonstrates that leaf growth reaction to root-zone or root temperature alteration cannot simply be scaled between long-term and short-term responses, but that diel and probably circadian variations of regulating mechanisms have to be taken into account. As the molecular pathway, with which plants can perceive temperature, is just beginning to be unravelled, this is a valuable insight that can help to identify process regulation in the plant temperature signalling network. For example, low temperature has been shown to stall the rhythmic expression of LHY and TOC1, two central elements of the circadian clock (Ramos et al. 2005). Because the circadian clock plays a prominent role to evoke physiological responses to alterations of the temperature (Penfield 2008) and because growth processes are also controlled by the circadian clock (Nozue et al. 2007; Harmer 2009), it will be important to clarify to which extent growth responses to temperature are affected by the circadian clock. Whereas LHY is an active component of the circadian clock in shoot and root, TOC1 is only involved in the circadian clock of the shoot (James et al. 2008), showing that differential regulation of temperature-induced processes in shoots and roots are likely. If, for example, the shoot–root temperature difference at a decisive ‘Zeitgeber’ time point of the circadian cycle is completely unnatural, growth might be disturbed for quite some time until physiological processes get back to a normal level again. Because the onset of light is such a Zeitgeber in a lot of situations (Harmer 2009), prominent morning differences between treatments with controlled, lowered root-zone temperature and uncontrolled root-zone temperature might be evoked as an output of the circadian clock.

The possible role of water relations

The reduction of leaf growth at reduced root-zone temperatures may also be caused by a reduction of water availability. Because the resistance to water uptake into roots increases at low temperature because of altered membrane fluidity and xylem conductivities (Holbrook et al. 2002; van Ieperen 2007), leaves suffer from a restricted availability of water (Turner 1994). Nevertheless the same transpiration rate as in control plants was observed both night and day (Fig. 7). This means that water had to be withdrawn or withheld from other process such as growth. Our observation of more negative xylem tension at lower root-zone temperature supports this suggestion, as it confirms that there was a high demand for water going into the transpiration stream in plants with cooled roots. Moreover, reduced xylem tension can correlate with reduced turgor, which decreases growth processes according to the Lockhart equation (Lockhart 1965). Because diurnal transpiration was much higher than nocturnal transpiration, root cooling would have caused a severe decline of the water available for growth processes by day and a less severe decline at night, which would fit with our observations.

The importance of the transpiration stream for coordinated dicot leaf growth was recently shown in the facultative crassulacean acid metabolism (CAM) plant Clusia minor (Walter et al. 2008): When Clusia is in the C3 mode (stomata open at day), leaf growth increases during the night and is maximal at dawn. When the plant shifts to CAM (stomata open at night), nocturnal growth is strongly reduced and diurnal growth is enhanced. These effects might also be coupled to the altered carbohydrate metabolism when the plant shifts its mode of photosynthesis.

The possible role of carbohydrates

The significance of carbohydrate metabolism for the phasing of diel leaf growth cycles has recently been shown in a study on Arabidopsis thaliana (Wiese et al. 2007). There, starch-free mutants showed strongly reduced nocturnal but enhanced diurnal growth activity, demonstrating the importance of proper access to carbohydrate stores to facilitate nocturnal growth. Increased pools of carbohydrates, in turn, can support higher nocturnal leaf growth, which was shown in a study with tall fescue exposed to differing amounts of daily irradiance (Schnyder & Nelson 1988).

In our study, carbon accumulated during the day as transitory starch and plants with low root temperature even showed higher starch concentration than plants with high root temperature (Fig. 6). Hence, this transitory starch was used by the Ricinus communis plants of our study to utilize the full growth potential of the plants at night. The increased pools of carbohydrates stored in all investigated organs when roots were cooled were mobilized rapidly when plants recovered from cold root-zone temperature and provided essential substrates for leaf growth to recover again.

Also, our 11C analyses support the notion that root cooling increased carbohydrate supplies to the shoot, to the detriment of the root (Fig. 8). The reduction of carbon import after roots are cooled is likely caused by an increase in root carbohydrates after sink metabolism slows (Minchin, Farrar & Thorpe 1994). Cooling the roots also affects the phloem pathway itself, and so may have also caused an effect like that of cold girdling, reducing phloem solution flow into the roots, and thus increasing shoot carbohydrate concentrations (Peuke, Windt & Van As 2006).

Our results demonstrate that leaf and root carbohydrate content depends strongly on root temperature. Moreover, root temperature will not be uniform within a pot (Fig. 1) and will adjust to a steady-state temperature faster or slower, depending on the exact distribution of the root system within the soil of a pot and depending on the exposure of the pot to the irradiation within the growth chamber. Hence, root system architecture, exact location of the pot in an experimental setup, or simply the extent of the root system, can have a considerable effect not only on leaf growth but also on carbohydrate content of leaves.


In conclusion, this study of a dicot shows that leaf growth, plant water relations and carbohydrate transport and concentration depend strongly on root-zone temperature. Growing cells of the leaf have restricted access to water, especially during the day, when the transpiration is high, but the consequent growth decrease can be alleviated at night when water availability is improved and stored carbohydrate can be provided. Moreover, the distribution of the root system within pots is usually inhomogeneous and rarely known, and yet soil temperature can show strong gradients within a pot, and change dramatically in time. Therefore, the strong influence of root temperature on shoot physiology is likely to be the source of considerable variability in many laboratory and greenhouse experiments, and should be considered seriously as the variation will not be random.


We thank Beate Uhlig for her support during the breeding of the plants.

Richard Poiré thanks the International Helmholtz Research School of Biophysics and Soft Matter for the stimulating discussions emerging from this interdisciplinary study, and acknowledges the support of his PhD thesis at the Heinrich-Heine-Universität Düsseldorf.