Spatio-temporal leaf growth patterns of Arabidopsis thaliana and evidence for sugar control of the diel leaf growth cycle


Author for correspondence: A. Wiese Tel: +49 2461618688 Fax: +49 2461612492 Email:


  • • Leaf growth dynamics are driven by diel rhythms. The analysis of spatio-temporal leaf growth patterns in Arabidopsis thaliana wild type and mutants of interest is a promising approach to elucidate molecular mechanisms controlling growth. The diel availability of carbohydrates is thought to affect diel growth.
  • • A digital image sequence processing (DISP)-based noninvasive technique for visualizing and quantifying highly resolved spatio-temporal leaf growth was adapted for the model plant A. thaliana. Diel growth patterns were analysed for the wild type and for a mutant with altered diel carbohydrate metabolism.
  • • A. thaliana leaves showed highest relative growth rates (RGRs) at dawn and lowest RGRs at the beginning of the night. Along the lamina, a clear basipetal gradient of growth rate distribution was found, similar to that in many other dicotyledonous species. The starch-free 1 (stf1) mutant revealed changed temporal growth patterns with reduced nocturnal, and increased afternoon, growth activity.
  •  The established DISP technique is presented as a valuable tool to detect altered temporal growth patterns in A. thaliana mutants. Endogenous changes in the diel carbohydrate availability of the starch-free mutant clearly affected its diel growth rhythms.


Post-emergent leaf growth is a complex process characterized by strong spatial and temporal variations which are controlled by numerous endogenous and exogenous factors (Walter & Schurr, 2005). Strong and species-specific diel variations of leaf growth have been reported for a number of species (Walter & Schurr, 2005). Carbohydrate relations seem to play a very important role in determining the shape and the amplitude of the diel leaf growth cycle. It has been shown that diel changes in metabolite concentrations (Walter et al., 2002, 2005) and diel carbon allocation patterns (Kehr et al., 1998) can significantly alter the dynamics of leaf growth. Rapid alterations of growth as responses to altered carbohydrate relations have also been shown for roots (Aguirrezabal et al., 1994; Muller et al., 1998; Nagel et al., 2006).

During the day, carbohydrate metabolism is mainly driven by photosynthesis, while at night, carbohydrate availability for growth processes is retained by degradation products of transitory starch (Geiger & Servaites, 1994). Sugars support growth via their role as energy carriers and as important building blocks of, for example, cell wall compounds (Crosgrove, 2005). Because of the important role of sugars in many plant processes, elaborate sensing mechanisms have evolved that can distinguish between different forms of sugars and that interact with almost all physiological processes within the plant (Smeekens, 2000; Rolland et al., 2006).

Carbohydrates act as controlling substances of gene expression for many diurnally controlled genes (Bläsing et al., 2005) and thus they may also be responsible for controlling diel changes in growth rate. Indeed, mutants with altered starch metabolism show greatly reduced growth (Caspar et al., 1991; Schulze et al., 1991; Zeeman et al., 1998; Sun et al., 2002) and the influence of sugars and sugar sensing mechanisms on growth has been shown for Arabidopsis thaliana wild-type seedlings (Jang et al., 1997) as well as for plants altered in their sugar sensing mechanisms (Jang et al., 1997; Mita et al., 1997; Moore et al., 2003). One well-known sensor for glucose is the A. thaliana hexokinase AtHXK1 (Moore et al., 2003).

The exact correlation between the timing of carbohydrate metabolism and leaf growth has not been unravelled, as highly time-resolving leaf growth measurements for the dynamics of A. thaliana leaf growth are not yet available. The spatio-temporal growth processes of single leaves and roots can be quantified noninvasively by digital imaging sequence processing (DISP) analysis, without affecting the plant significantly (Schmundt et al., 1998). In this method, the organ of interest is oriented and fixed in the focal plane of a charge-coupled device (CCD) camera, and illuminated by infrared diodes. Changes in surface area observed in the time-lapse videos obtained are quantified and used as a measure of growth intensity. Here we present the DISP technique for the analysis of A. thaliana leaf growth dynamics and investigate diel leaf growth of the starch-free 1 (stf1) mutant. A mutation in the gene encoding plastidic phosphoglucomutase impedes starch biosynthesis in stf1 (Kofler et al., 2000) and thus alters the diel carbohydrate metabolism of the mutant. We investigated whether stf1 shows a different diel leaf growth cycle from wild-type plants and, if so, whether sugar sensing via the A. thaliana hexokinase AtHXK1 plays an important role in this process.

Materials and Methods

Plant material

Arabidopsis thaliana (L.) Heynh. plants were grown in soil in a growth room (illumination: photosynthetic photon fluence rate (PPFR) 50–60 µmol m−2 h−1 with 12 h:12 h light (08:00–20:00 h):dark period; temperature: 23°C during the day; 20°C at night). The starch-free mutant stf1 (Kofler et al., 2000) carries a 55-bp deletion in the plastidic phosphoglucomutase gene (pgm) in the background of the Landsberg erecta (Ler) ecotype. Seeds for stf1 were obtained from Dr H. Kofler. The sugar-sensing mutant glucose insensitive 2-1 (gin2-1) in the background of the ecotype Ler carries a mutation in the gene encoding A. thaliana hexokinase 1 (Moore et al., 2003).

Leaf growth analysis

Plants were grown to an age of at least 1 month. Growing leaves were chosen on the basis of their size and position on the rosette, always being the 2nd to 4th youngest visible leaf and being less than 50% of the size of fully grown leaves (c. 5–10 mm in length). Manual, nondestructive fixation of leaves smaller than 5 mm for DISP analysis was not possible. Final leaf length was comparable for leaves at different positions as the selected leaves came from later stages than leaf 8 (Donnelly et al., 1999). Leaves chosen for the experiments reported here were in later stages than the youngest stages analysed by Donnelly et al. (1999), which showed cell division all over the leaf blade. We consider the chosen leaf stage to still show low cell division rates at the basal leaf blade, as shown by Donnelly et al. (1999). According to the terminology of Beemster et al. (2005), they were at an intermediate stage of ‘leaf expansion’ (characterized by cell division rates of c. 0.01 cells cell−1 h−1) and had definitely left the ‘proliferative’ stage. Comparison of growth and the expression of marker genes for cell division and cell elongation in the future should provide insight into the cellular nature of growth.

To keep growing leaves in the focal plane of the camera and to prevent measurement artefacts, investigated leaves had to be mechanically fixed in a stationary position (Fig. 1a). The leaves were fixed with five threads, glued (Pattex; Henkel, Düsseldorf, Germany) to the edge of the leaf, one at the very tip of the leaf and two each along the leaf sides. Each of the threads was stretched with a weight of 1.8 g and spun over a metal ring surrounding the whole plant (Fig. 1b). Additionally, a sixth weight was fixed to the tip of a fully grown leaf opposing the fixed young leaf to stabilize the plant in the soil.

Figure 1.

Experimental set-up for digital image sequence processing (DISP) leaf growth analysis in Arabidopsis thaliana. (a) Fixation of strings on an A. thaliana leaf (ecotype Columbia (Col-0)). (b) Fixation of a leaf with weights in a metal ring. (c) Complete set-up with camera, infrared diodes and a plant with a fixed leaf. (d) Effect of fixation with different weights on relative growth rate (RGR; average of 24 h).

Leaf fixation with appropriate weights does not affect temporal and spatial growth patterns, as shown previously in Ricinus communis (Walter et al., 2002). To establish the leaf growth analysis procedure for A. thaliana, we analysed the influence of different weights on diel growth and on growth intensities. Growth of fixed leaves was determined crudely by photographing leaves and calculating the leaf size of fixed and unfixed leaves (n = 3). The results showed that fixation with five weights of 1.8 g each per leaf was sufficient to mount the leaf in the optical plane of the camera and did not affect leaf growth intensity (Fig. 1d).

Images of the fixed leaves were acquired with CCD cameras (Sony XC55 or XC75; Sony, Köln, Germany) or CMOS cameras (Flea BW by Point Grey Research Inc., Vancouver, Canada), positioned on top of the plants and equipped with a standard objective lens (12 or 25 mm; Cosmicar/Pentax; The Imaging Source, Bremen, Germany) and an infrared interference filter (880 or 940 nm; Edmund Optics, Karlsruhe, Germany). Constant illumination throughout the day and night without affecting plant growth behaviour was provided by infrared diodes (880 nm or 940 nm; Conrad Electronics, Hirschau, Germany). Grey value images were taken every 120 s and were saved in a multi-tiff format.

Imaged plants were illuminated with a PPFR of 50–60 µmol m−2 h−1 and a photoperiod of 12 h:12 h light:dark. Light was reduced in the DISP set-up as a result of shading by the camera and infrared diodes to a minimum PPFR of 30 µmol m−2 h−1.

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. Areal relative growth rates (RGRs) in percentage h−1 were calculated as the divergence of the estimated velocity field by either selecting an area of interest (AOI) within the image and tracking the structure within this AOI with time (GrowFlow) or determining the mean RGR over 4 h, resulting in colour-coded RGR maps (RGRFlow). For more details, see Schmundt et al. (1998), Walter et al. (2002) and Matsubara et al. (2006).

Because of the variance of the RGR between leaves from different plants of the same line, a normalization procedure had to be performed for each investigated diel cycle of leaf growth (Walter et al., 2005). Normalized data then allowed comparison of diel leaf growth cycles between the two different plant lines. For each leaf analysed, 1-h average values of RGR were calculated and normalized on the average RGR of a 24-h growth period (% of diel mean; see Fig. 3 below), giving a value of 100% at that time of day, when the average RGR for a 24-h cycle was reached. Fractions of integral growth intensity were calculated accordingly by taking average values for periods longer than 1 h.

Figure 3.

Normalized diel leaf growth cycle of Arabidopsis thaliana ecotypes. (a) Landsberg erecta; (b) Columbia 0. 1-h average values of relative growth rate (RGR) are shown for six leaves of each ecotype (error bars: standard error). Shaded areas indicate night.

Carbohydrate analysis

For carbohydrate analysis, leaves were selected according to the same criteria as for growth measurements. These leaves were sampled at 18:00, 07:00 and 13:00 h (late afternoon, at the end of the night, and at the middle of the day), frozen in liquid nitrogen after their fresh weight had been determined, and stored at −80°C for further extraction. Soluble carbohydrates were extracted from frozen leaf material and glucose, fructose and sucrose concentrations were analysed in a coupled enzyme assay (Jones et al., 1977) using a multiplate spectrophotometer (ht II; Anthos Mikrosysteme GmbH, Krefeld, Germany) as described in Walter et al. (2005). The remaining leaf material was prepared for starch analysis as described by Walter et al. (2005); the starch content was determined enzymatically as glucose concentration using the same procedure as described above for soluble carbohydrates.


Spatio-temporal growth of A. thaliana leaves

Diel leaf RGRs of A. thaliana ecotypes Ler and Columbia 0 (Col-0) showed maxima in the morning soon after dawn, and a subsequent decrease in growth during the day (Fig. 2). Growth was at its minimum at the beginning of the night, and then increased again until the end of the night. Peak RGR values of > 5% h−1 were observed; average RGR over a 24-h period was often > 1% h−1 (or 24% d−1). RGR declined from day to day. In both ecotypes, 70% of the integral growth activity occurred during the day, while only c. 30% of growth activity during a 24-h period occurred at night (Table 1). Normalized values for replicate measurements demonstrate the reproducibility and significance of the DISP detection of the diel leaf growth pattern for both ecotypes (Fig. 3).

Figure 2.

Diel leaf growth cycle of Arabidopsis thaliana ecotypes. (a) Landsberg erecta; (b) Columbia 0. Data are shown for one representative leaf of each ecotype. Shaded areas indicate night. RGR, relative growth rate.

Table 1.  Fractions of the integral growth intensity (% of total growth during 24 h) of Arabidopsis thaliana
Plant lineIntegral growth intensity (% of total growth during 24 h)
NightSecond half of nightAfternoon
  1. Values are mean ± standard error.

  2. Night, 20:00 to 08:00 h; second half of night, 02:00 to 08:00 h; afternoon, 16:00 to 20:00 h. Col 0, n = 17 d, eight individual plants; Ler, n = 21 d, seven individual plants; stf1, n = 14 d, six individual plants.

  3. Col 0, Columbia; Ler, Landsberg erecta; stf1, starch-free 1 mutant.

Ler36.0 ± 9.120.8 ± 4.511.4 ± 4.1
Col 033.9 ± 6.821.4 ± 6.514.0 ± 4.0
stf124.2 ± 6.4 8.3 ± 2.924.8 ± 4.7

A basipetal gradient of leaf growth was observed for both ecotypes, similar to that reported for most other dicotyledonous species. The RGR of the leaf base (region 1) was higher than the RGR near the tip (region 4) during the day and at night (Fig. 4a). The inclination of this gradient was correlated to overall leaf growth intensity: the gradient was steeper during the day than at night (Fig. 4a) and was steeper in fast-growing leaves compared with slow-growing leaves (data not shown). A basipetal distribution of growth intensity was observed throughout the diel growth cycle, but with some superimposed heterogeneity (Fig. 4b,c).

Figure 4.

Basipetal leaf growth pattern of Arabidopsis thaliana (ecotype Landsberg erecta (Ler)). (a) Average relative growth rate (RGR) of leaf regions during the day (open bars) and at night (closed bars; data from 3 d of growth of a typical leaf). The region ‘base 1’ starts at 20% of total leaf length measured from base to tip. Leaf region ‘tip 4’ starts at c. 65% of total leaf length. The length of each region comprises c. 15% of total leaf length. (b) Time series of 1-h average values of RGR for regions 1–4. Shaded areas indicate night. (c) Colour-coded spatial distribution of RGRs in Ler (upper row) and Columbia (Col-0) (lower row). Spatial growth was averaged over 4 h as indicated in the Materials and Methods.

Starch mutants show a changed temporal growth pattern

Defects in starch metabolism cause alterations of the normal diel carbohydrate cycle (Gibon et al., 2004; Bläsing et al., 2005). A severe defect is present in the mutant stf1, which is unable to synthesize starch as a result of an altered gene for plastidic phosphoglucomutase (Kofler et al., 2000). Hence, stf1 is an ideal candidate to analyse the interplay between sugar metabolism and leaf growth dynamics.

It is well known that stf1 and other mutants with defects in starch metabolism develop a reduced vegetative rosette area when grown in a 12-h photoperiod (Kofler et al., 2000), but the effect of such genetic aberrations on diel leaf growth cycles has not been investigated before (Caspar et al., 1991; Schulze et al., 1991; Zeeman et al., 1998).

The diel growth cycles of Ler wild type and stf1 differed markedly (Fig. 5). stf1 showed lower growth rates than Ler in the second half of the night. The fraction of integral growth intensity of stf1 was significantly lower than that of Ler during the second half of the night (Table 1; unpaired t-test with P < 0.0005). While 20% of the total leaf growth activity of Ler was confined to the second half of the night, only 8% of the growth activity was found in stf1 during this period. In the afternoon, stf1 showed significantly higher RGR than Ler. In wild-type plants, only 11% of the total diel leaf growth activity was confined to the afternoon, while in stf1 this fraction was 25%.

Figure 5.

Diel leaf growth cycle of the starch-free 1 (stf1) mutant of Arabidopsis thaliana. A comparison is shown of the diel growth patterns of Landsberg erecta (Ler) (closed circles) and stf1 (open circles). One-hour average values of relative growth rate (RGR) are shown for six leaves of each line (error bars: standard error). Shaded areas indicate night, *P < 0.05; **P < 0.01.

Diel changes in sugar and starch contents of growing leaves

As expected, almost no starch was detected in stf1 at any time of the day, while in Ler, the starch concentration increased during the day and decreased at night (Table 2). During the day, sucrose concentrations in stf1 and Ler were comparable, while glucose and fructose concentrations were significantly higher in stf1 than in Ler. At 13:00 h, stf1 contained 1.5-fold more glucose than Ler, and at 18:00 h it contained 2-fold more. For fructose, concentrations were 2- and 3-fold higher in stf1 at 13:00 h and at 18:00 h, respectively. At the end of the night, the concentrations of hexoses were comparable in stf1 and Ler, while the sucrose content was significantly lower in stf1 than in Ler. The results show clearly that growing leaves of stf1 have a surplus of carbohydrate growth substrates (hexoses) in the afternoon while they are depleted of carbohydrates in the form of sucrose at the end of the night.

Table 2.  Concentration of soluble sugars and starch in growing leaves of Arabidopsis thaliana
Plant lineConcentration (µmol g−1 FW)
13:00 h18:00 h07:00 h13:00 h18:00 h07:00 h13:00 h18:00 h07:00 h13:00 h18:00 h07:00 h
  1. Starch is represented in glucose units. Values are mean ± standard error (n = 4 or 5). Values that are significantly different from the wild type: *, P < 0.01; **, P < 0.005.

  2. FW, fresh weight; Ler, Landsberg erecta; stf1, starch-free 1 mutant.

± 0.95± 0.60± 0.70± 0.18± 0.11± 0.21± 0.29± 0.42± 0.19± 1.00± 1.45± 0.46
± 0.99± 1.81± 0.66± 0.61± 1.03± 0.51± 0.40± 0.59± 0.12± 0.10± 0.05± 0.09

The glucose sensor AtHXK1 is not required for normal spatio-temporal growth pattern

After showing that the availability of carbohydrate growth substrates affects the diel leaf growth cycle, the effect of sugar sensing on leaf growth dynamics was investigated by analysis of the diel leaf growth pattern of plants lacking the glucose sensor AtHXK1. Plants lacking AtHXK1 show strong growth retardation under high light conditions as a result of reduced cell elongation (Moore et al., 2003). This glucose sensor might thus be responsible for the diel control of growth in a normal day:night cycle. We investigated the diel growth behaviour of the gin2-1 mutant under the selected light conditions. Only a slight decrease in the rosette area size of gin2-1 compared with Ler plants was observed in our growth conditions (data not shown). The diel growth cycle of leaves of gin2-1 did not differ from that of Ler (Fig. 6). This indicates that AtHXK1 is not required for the control of diel leaf growth of A. thaliana in the experimental conditions of our study.

Figure 6.

Normalized diel leaf growth cycle of the Landsberg erecta (Ler) ecotype of Arabidopsis thaliana (closed circles) and the mutant glucose insensitive 2-1 (gin2-1) (open circles). A comparison of the diel growth patterns of Ler and gin2-1 is shown. One-hour average values of relative growth rate (RGR) are shown for six leaves of Ler and five leaves of gin2-1 (error bars: standard error). Shaded areas indicate night.


Spatial and temporal leaf growth patterns in A. thaliana

In our growth conditions, the strongest leaf growth was found in the early morning hours, soon after dawn, while minimal growth activity was observed during the first half of the night (Fig. 3). Similar growth patterns have previously been observed for leaves of Nicotiana tabacum (Walter & Schurr, 2005) and Ricinus communis (Walter et al., 2002), but the variability among species seems to be large; leaves of Populus deltoides, for example, show growth maxima in the evening (Walter et al., 2005; Matsubara et al., 2006).

No clear differences were found between the diel leaf growth cycles of the two investigated ecotypes Ler and Col-0 (Figs 2, 3), although differences in plant growth parameters such as root elongation rates (Beemster et al., 2002), hypocotyl length (Sergeeva et al., 2006) and rosette size (Granier et al., 2006) have previously been reported for a number of A. thaliana ecotypes.

Concerning the spatial distribution of growth across the leaf lamina, the two ecotypes also grew in comparable ways, showing the base–tip gradient of RGR which is present in most, but not all, dicotyledonous species. This gradient is caused by earlier completion of cell division and expansion in the distal part of the lamina than in the basal part (Schmundt et al., 1998; Donnelly et al., 1999; Walter & Schurr, 2005).

In linearly organized growth zones of the leaves of monocotyledonous plants (Ben-Haj-Salah & Tardieu, 1995; Beemster et al., 1996) or roots (Goodwin & Stepka, 1945; Silk & Erickson, 1979), distributions of relative elemental growth rates can be used to quantify the expansion of individual cells. Even cell division rates can be calculated in such organs based on determination of cell lengths at different positions and on determination of root tip/leaf tip velocity via the continuity equation (Silk & Erickson, 1979; Silk, 1984; Beemster & Baskin, 1998). In leaves of dicotyledonous plants, the situation is much more complex as considerable cell division activity is still present in regions of cell expansion (Beemster et al., 2005). Initial attempts have been made to connect cellular processes with local growth rates of tissue regions based on models of relative elemental growth rate distribution (Coen et al., 2004; Prusinkiewicz, 2004). Such approaches will only be successful, for example for a deterministic analysis of the development of leaf shape, if precise quantitative analyses of growth processes as performed in this study and precise analyses of cellular development are carried out in an orchestrated way in the future.

Metabolite control of growth

The results of this study show clearly that diel leaf growth variations are altered by diel variations of carbohydrate metabolism. While the reduction of overall growth activity in plants with defects in starch metabolism is well known from the literature (Caspar et al., 1985; Huber & Hanson, 1992), a relationship between the diel timing of starch metabolism and the phasing of diel leaf growth cycles has only been discussed on the basis of correlative data, comparing carbohydrate concentrations and growth in wild-type plants (Walter et al., 2002; Walter & Schurr, 2005). Here, differences of diel carbohydrate metabolism between mutant and wild-type plants were reflected in diel growth alterations. stf1 showed increased growth in the afternoon compared with the wild type (Fig. 5, Table 1); this can be assumed to be induced by the excess of free hexoses, which are not incorporated into starch storage. Hexose contents are significantly increased, while the sucrose concentration of stf1 mutants remains comparable to that of the wild type during the day (Table 2). In the starch-free mutants stf1 (Ler background) and pgm (Col-0 background), increased concentrations of hexoses but also sucrose have been reported in the literature for mature leaves (Kofler et al., 2000; Gibon et al., 2004; Bläsing et al., 2005; Gonzali et al., 2006), but comparable findings for young growing leaves have not been reported.

Towards the end of the night, stf1 grew more slowly than the wild type, showing that the inability to withdraw carbohydrate metabolites from transitory starch causes a distinct decrease in leaf growth. Nocturnal growth reductions in mutants with defects in starch biosynthesis or remobilization can also be inferred from studies in the literature reporting a positive correlation between the extent of growth reduction and the duration of the night phase (Caspar et al., 1991; Schulze et al., 1991; Zeeman et al., 1998; Kofler et al., 2000). Longer light periods lead to smaller differences in growth between mutants and wild type; continuous light even restores the wild-type phenotype in mutants (Caspar et al., 1985; Huber & Hanson, 1992).

At the end of the night, sucrose concentrations in growing leaves were significantly lower in stf1 mutants compared with the wild type (Table 2). This finding is in agreement with results for whole rosettes of the starch-free mutant pgm (Gibon et al., 2004; Bläsing et al., 2005).

Our results on the diel growth of a starch-free A. thaliana mutant confirm findings from the literature which showed, based on much cruder growth measurements, that a starch-deficient mutant of N. tabacum with a modified plastidic phosphoglucomutase gene (pgm) grew more quickly than wild-type plants during the day and more slowly during the night (Geiger et al., 1995). To check the correlation between sugar availability and growth in more detail, a direct overlay of spatio-temporal growth rate maps with maps for noninvasively visualized spatio-temporal sugar concentrations, for example using specific sugar sensor proteins (Lager et al., 2006), should be performed.

It has to be pointed out, however, that the mutants investigated in our study did not show a profound alteration, such as a reversal or extinction of diel leaf growth activity. The general phasing of the leaf growth cycle was still comparable to that of wild-type plants, indicating that leaf growth is controlled not only by carbohydrate availability but by a complex network of factors (Walter & Schurr, 2005).

Growth and sugar sensing

The pgm mutant (Caspar et al., 1985) has been utilized as a tool for the investigation of in vivo changes in sugar signals, with observed changes corroborated by comparison with glucose or sucrose feeding or starvation data (Thimm et al., 2004; Bläsing et al., 2005; Gonzali et al., 2006). Most of the observed changes in gene expression in pgm were triggered by the low sugar concentrations at the end of the night (Bläsing et al., 2005). Some sucrose-inducible genes were up-regulated in pgm at the end of the light period (Gonzali et al., 2006). It has been shown recently that the diel availability of sugars in pgm controls the diel variation in expression of many genes (Bläsing et al., 2005). At the end of the night, genes involved in cell wall synthesis/breakdown, cell wall modification and protein synthesis are indeed differentially regulated in pgm (Thimm et al., 2004), which supports our finding of reduced growth of starch mutants in the second half of the night.

The integration of sugar signals into processes such as growth or gene expression requires sensors and signalling cascades (Smeekens, 2000; Rolland et al., 2006). One well-known sensor for glucose is the A. thaliana hexokinase AtHXK1. Hexokinase activity is required for triggering high glucose signals (Moore et al., 2003) as well as for the detection of the reversal of sugar starvation (Fujiki et al., 2000). Hence, we tested the hypothesis that AtHXK1 can act as a pacemaker of the diel growth rhythm. This hypothesis can be rejected as no difference in the temporal growth pattern of the AtHXK1 mutant compared with the wild type (Fig. 6) could be detected. Still, AtHXK1 might be required for the detection of endogenous changes in glucose concentrations other than those found in normal cycling.


In the course of this study, we adapted an existing image sequence processing method to analyse spatio-temporal leaf growth patterns of the model plant A. thaliana. We were able to show a direct link between diel sugar metabolism and diel leaf growth cycle. The door is open now for analysis of the effect of genes involved in a wide range of processes that putatively control growth.


The authors wish to thank Heike Kofler for providing seeds of stf1 and for helpful comments on the manuscript. We thank Beate Uhlig for assistance in plant cultivation and Hanno Scharr and Andrés Chavarría-Krauser for support with image sequence processing algorithms. AW was funded by a grant from the Forschungszentrum Juelich.