SEARCH

SEARCH BY CITATION

Keywords:

  • Arabidopsis thaliana;
  • image processing;
  • leaf growth;
  • light;
  • Nicotiana tabacum;
  • nutrients;
  • phenotypic analysis;
  • root growth

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    Using a novel setup, we assessed how fast growth of Nicotiana tabacum seedlings responds to alterations in the light regime and investigated whether starch-free mutants of Arabidopsis thaliana show decreased growth potential at an early developmental stage.
  • • 
    Leaf area and relative growth rate were measured based on pictures from a camera automatically placed above an array of 120 seedlings. Detection of total seedling leaf area was performed via global segmentation of colour images for preset thresholds of the parameters hue, saturation and value.
  • • 
    Dynamic acclimation of relative growth rate towards altered light conditions occurred within 1 d in N. tabacum exposed to high nutrient availability, but not in plants exposed to low nutrient availability. Increased leaf area was correlated with an increase in shoot fresh and dry weight as well as root growth in N. tabacum. Relative growth rate was shown to be a more appropriate parameter than leaf area for detection of dynamic growth acclimation. Clear differences in leaf growth activity were also observed for A. thaliana.
  • • 
    As growth responses are generally most flexible in early developmental stages, the procedure described here is an important step towards standardized protocols for rapid detection of the effects of changes in internal (genetic) and external (environmental) parameters regulating plant growth.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Plant growth is controlled by a complex network of factors (Walter & Schurr, 2005). From the level of gene transcription, via hormonal control, to control by environmental parameters, a multitude of factors affect the short-term dynamics and long-term performance of plant growth. Light intensity and nutrient availability are environmental factors that strongly affect plant growth performance from early developmental stages onwards. At the seedling stage the plant is most vulnerable to environmental stress conditions, and cannot buffer deficiencies of factors fostering growth as reserves are sparse. It is this stage in which growth most directly reflects the interaction of the entity of genes driving growth (Meyer et al., 2004). For precise quantification of alterations of growth, noninvasive growth measurements are indispensable. Destructive measurements (such as determining fresh or dry weight) are time-consuming, hence only a small number of replicates can be observed in any given study. Moreover, it is difficult to detect growth effects in small seedlings, as measurement devices have to be precise.

As analysis of plant growth is an essential step in understanding plant performance and productivity, several approaches have been used throughout the past decade to quantify the projected leaf area of dicotyledonous plants as a proxy for overall plant growth performance (Leister et al., 1999; Barbagallo et al., 2003; El-Lithy et al., 2004; Granier et al., 2006). Such procedures work best with seedlings, as the effects of leaf overlap can be neglected because of the rosette stage in which several interesting model species remain for an extended period. Yet standardized automatic procedures for image evaluation of a high number of individuals are still not common. Recently, a very interesting platform named PHENOPSIS has been established that specializes in phenotyping plant responses to soil water deficit (Granier et al., 2006). There, leaf area is determined automatically, plants are weighed, and defined amounts of water are added per individual pot in an automated way. Unfortunately, it is difficult to assess the precision of PHENOPSIS as no details of image acquisition or analysis have been reported.

Relative growth rates originating from area differences of individual plants at consecutive time steps are not reported in any of the above-mentioned studies. Several systems that have been established for high-throughput evaluation of plant growth work with a low degree of automation: Threshold values for distinction between leaf and background have to be defined on a reiterative basis for individual images (Leister et al., 1999); 20–96 plants are displayed within one image to reduce data handling, which leads to an enormous loss of resolution (Leister et al., 1999; Barbagallo et al., 2003; El-Lithy et al., 2004). Industrial solutions for plant phenotyping have been elaborated by several companies, but their applicability has not yet been demonstrated in the literature. Hence there is a clear need for phenotyping solutions that allow automated but flexible image acquisition, robust and transparent data evaluation, and problem-specific precision and interpretation of the data.

Alterations in light climate can often happen very quickly (e.g. in canopy gaps), and it has recently been shown that plants can react within a short time frame to alterations in the light climate by dynamically changing the growth of individual leaves (Lai et al., 2005) or roots (Nagel et al., 2007). Moreover, differences in light intensity and light duration (daily quantum input), resulting in differences of assimilation rate, have recently been shown in two independent meta-analyses to be the major environmental factor causing differences in relative growth rate (RGR) (Kruger & Volin, 2006; Shipley, 2006). Although photosynthesis and the subsequent production of carbohydrates are ultimately major driving factors determining the growth potential of a plant, and although it has become feasible in the past decade to investigate the dynamics of growth and photosynthesis with high resolution, it is still unclear how fast alterations in light climate are transformed into alterations in plant growth rate (Schurr et al., 2006). To elucidate this question, it is necessary to perform noninvasive growth measurements providing a high temporal resolution.

It was the aim of this study to establish a procedure called GROWSCREEN that integrates the strength of modern standard procedures of single-image processing with an automated setup rapidly to acquire and evaluate high-quality images of plants that had been raised in standard laboratory conditions. Using this approach, we investigated whether it is possible to detect expansion differences occurring within days between populations that were caused by differences in daily quantum input and external nutrient availability, and whether these differences in leaf expansion were indicative of differences in biomass growth.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Plant material and cultivation procedures

Seeds of Nicotiana tabacum (L.) cv. Samsun were germinated either on soil ‘ED73’ (Einheitserde, Balster Einheitserdewerk, Fröndenberg, Germany; N, approx. 250 mg l−1, P2O5, approx. 300 mg l−1, K2O, approx. 400 mg l−1), providing high nutrient availability; or on soil ‘Typ 0’ (Nullerde, Balster Einheitserdewerk, Fröndenberg, Germany; N, approx. 50 mg l−1, P2O5, approx. 80 mg l−1, K2O, approx. 80 mg l−1), providing low nutrient availability. Seedlings exposed to either high or low nutrient availability were watered with either full-strength or 1/20 Hakaphos nutrient solution, respectively (Hakaphos blau, Compo, Münster, Germany). Full-strength solution was prepared according to the manufacturer's recommendations for young flowering plants, and contained 5.4 mmol l−1 N (1.4 mmol l−1 NO3 and 4 mmol l−1 NH4+) as well as all other macro- and micronutrients accordingly. Seedlings were placed in eight 30-well plastic trays, providing each seedling with a pot surface area of 3.5 × 3.5 cm. Trays were placed under fluorescence lamps that provided them with 80 µmol m−2 s−1 PAR (12 : 12 h light : dark). Temperature was set to 23°C (day) and 18°C (night).

At day 10, plants exposed to high and low nutrient availability were grouped into four different populations, respectively: one population (control, 12L) remained at the conditions previously set. One population was exposed to a twofold light intensity for three 2-h periods each day (6HL; 160 µmol m−2 s−1 PAR) and exposed to the control light intensity for the remaining 3 × 2 h; one population was exposed to a twofold higher light intensity throughout the entire day (12HL; 160 µmol m−2 s−1 PAR); and one population was exposed to constant light throughout 24 h (24L; 80 µmol m−2 s−1 PAR). At the end of the treatment (day 14), shoots were harvested and fresh weight was determined. Dry weight was determined after plants had been oven-dried for 48 h at 80°C.

Arabidopsis thaliana (L.) Heynh. wild type (ecotype Landsberg erecta, Ler) and starch-free mutant plants stf1 (Kofler et al., 2000) were stratified for 2 d at 4°C in the dark before germination in soil (ED 73). The plants were grown for 14 d in a photoperiod of 12 : 12 h light : dark at 50–60 µmol m−2 s−1 PAR. stf1 plants germinated 2 d later than wild-type plants; growth was recorded from the time of full emergence of cotyledons onwards. A 55-bp depletion in the gene of plastidic phosphoglucomutase impedes any starch biosynthesis in stf1 (Kofler et al., 2000).

Analysis of root growth and cultivation of plants for this experiment

For analysis of root growth, seeds of N. tabacum ecotype Xanthi were surface-sterilized with sodium hypochlorite solution and sown on sterile agar (1% w/v) in Petri dishes (120 × 120 × 17 mm; five seedlings per dish). The medium contained full-strength Ingestad mineral nutrient solution, providing the plants with approx. 10 mmol l−1 N and all other macro- and micronutrients accordingly (for more details see Ingestad, 1982; Nagel et al., 2007). We used a cultivation procedure that allowed shoots to grow outside the agar-filled Petri dish (Nagel et al., 2007) to measure total leaf area using GROWSCREEN. Therefore seeds were pushed slightly into the agar through five holes (approx. 2 mm diameter) at one side of the closed Petri dishes. After sowing, the Petri dishes were placed at 12 : 12 h light : dark cycles. Control plants (12L) were illuminated with 60 µmol m−2 s−1 PAR over 19 d; high light-treated plants (12HHL) were illuminated with 60 µmol m−2 s−1 PAR for 14 d and thereafter for 5 d with 300 µmol m−2 s−1 PAR. At day 19, leaf area was measured and the number of all primary and lateral roots, as well as the root surface area, was determined. For determination of root area, the translucent agarose gel containing roots was harvested and scanned with a flat-bed scanner. The root area was then determined in a similar way to the method described below for leaf area, using the procedures developed for GROWSCREEN.

Image acquisition

Projected total leaf area of each seedling was determined each day at 14 : 00 h. At this time of day, leaves were spread out almost horizontally. Leaf inclination angles and leaf growth rates vary systematically throughout the day, which makes it necessary that (i) each population has to be investigated at the same time each day; and (ii) populations should not be too large, as any time lag between analysis of two populations reduces the comparability between these populations. For image acquisition, four trays were placed at defined positions below the setup and 120 images were acquired sequentially in a user-defined protocol (Fig. 1).

image

Figure 1. GROWSCREEN. (a) Mechanical setup with two perpendicular, computer-controlled displacement stages that position the camera and the white LED illumination panel on top of seedlings (bar, 10 cm). (b) Some seedlings of Nicotiana tabacum cultivated at high nutrient availability (bar, 1 cm). (c) Raw data (Bayer) image acquired by the camera before Bayer pattern interpolation (bar, 1 mm). (d) Segmented and colour-interpolated mask of (c) (bar, 3 mm).

Download figure to PowerPoint

Images of each seedling were acquired using a colour charge-coupled device (CCD) camera (Scorpion, IEEE1394, 1280 × 960 pixel, Point Grey, Vancouver, Canada) with CCD-sensor ICX274AQ (Sony) and a 25-mm objective lens (Pentax, Hamburg, Germany) mounted with a 5-mm intermediate ring. The principal axis between sensor and object had a length of 500 mm. Each pixel corresponded to a leaf area of 7200 µm2. The optimal camera settings for all plants investigated were: gain 800, exposure 1, brightness 27, white balance 59/79, shutter 355.

The camera was driven to preset positions via two linear displacement stages (Pico-Maxi, Type FMD-LPT80.550.1205-SM, Laser 2000 GmbH, München, Germany) that were mounted perpendicularly to a solid metal stand (Linos Photonics, Göttingen, Germany). During image acquisition, the seedling was illuminated by six clusters of five light-emitting diodes (LEDs) (MCPE145364, 24 V/DC, 0.4 W, white, Conrad Electronic, Hirschau, Germany) that were mounted fixed to the camera on a 100-mm-diameter ring with 300 mm distance to the object. Time for acquisition of 120 images was approx. 20 min (10 s per frame). Image acquisition frequency was restricted by the speed of the linear displacement stages, by summing up 10 individual images for hardware noise suppression, and by a waiting time required to avoid stand vibrations during image acquisition. The software for image acquisition and evaluation was written in C++.

Image processing and determination of total leaf area and RGR

The Bayer pattern of the colour camera images was transformed to red–green–blue (RGB) values for each individual pixel via linear interpolation between two or four neighbouring pixels, depending on the position in the Bayer pattern (Fig. 2). The Bayer pattern is given by the standard arrangement of red-, green- and blue-light-sensitive pixels situated on a CCD chip; there are twice as many more green-sensitive pixels than red- or blue-sensitive pixels. Such a recalculation is a standard procedure performed by the hardware of each consumer camera.

image

Figure 2. Transformation of raw data (Bayer image) via red, green, blue (RGB) space to hue, saturation and value (HSV) space. (a) Each pixel in a Bayer pattern image (top left) has its own colour filter (top right). Gaps in its RGB channels (bottom) are filled by linear interpolation using two or four nearest neighbours. (b) A typical colour region defining foreground (plant leaf area) in the segmentation step is depicted in RGB and HSV space. While complicated in RGB, in HSV it is given by simple intervals.

Download figure to PowerPoint

From this information, the more appropriate parameters hue, saturation and value (H, S, V) were calculated and used for distinction between plant and background. While R, G and B depict pixel brightness at the mean spectral value of the colour filter (distinction between only three colours), H, S and V allow much finer distinction regarding colour without reducing brightness resolution. Hue is associated with the dominant wavelength of the captured light at each pixel position that is extracted from the RGB information at each pixel. Saturation is inversely proportional to the amount of white light mixed with a hue. Value (or brightness of colour) is given by the maximum radiance received at one of the three colour channels of a pixel. All colour transformations were done using Intel's free computer vision library (OpenCV, command: cvCvtColor). As the plants of interest were mainly green, using HSV space instead of RGB allowed for simple and reliable separation of foreground (leaf area) and background. ‘Green’ is simply a hue interval between values of approx. 40 and 85 (in the case of N. tabacum seedlings).

On the background of the selected camera settings, optimal values for H, S and V were: H, 40–85; S, 98–254; V, 30–150 for analysis of N. tabacum; H, 45–85; S, 110–255; V, 1–255 for analysis of A. thaliana. Those parameters were established initially and were then used throughout the entire experiment for each single plant.

Colour image segmentations performed with the above-mentioned values resulted in ‘segmented’ images with small ‘holes’ in the plant area and small ‘objects’ in the background. These small holes were caused by colour noise, soil or dirt particles on plant leaves; artefact objects within the background resulted from irregularities such as algae cover or dirt particles on the substrate. The size of all objects and holes was registered, and objects and holes were removed automatically (holes were filled up) from the segmentation mask when they were < 20 pixels (approx. 0.15 mm2). A batch-file procedure allowed evaluation of 120 mask areas within approx. 20 min.

Pixel values for each plant were stored in an excel file, in which total leaf area and RGR for individual plants as well as population mean values were finally calculated. The RGR was calculated according to:

  • RGR = 1/t × ln(A2/A1)

where t denotes the time between acquisition of areas A1 and A2.

Statistical analysis

The effect of daily quantum input on total leaf area and on RGR was analysed using one-way anova (SigmaStat, Systat Software Inc., Richmond, CA, USA) or a Kruskal–Wallis one-way anova, if the assumptions of normality and variance homogeneity had not been met. Post hoc comparisons of treatment effects were performed within each group using the Tukey adjustment or Dunn's adjustment, respectively. Differences in growth between A. thaliana wild type and the starch-free mutants stf1 were analysed using a t-test.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Detection of projected leaf area was possible over a sufficient period of vegetative development for N. tabacum and A. thaliana (Fig. 3). Leaf area was monitored until 14 d after germination. Total leaf area expanded exponentially throughout this growth stage in A. thaliana and N. tabacum, respectively. As seedling leaf canopies did not change colour markedly throughout this stage and under the given sets of external conditions, there was no need to adapt the parameters hue, saturation and value for optimal segmentation of plant and background. The RGR remained relatively stable between days 10 and 14 in control populations of the experiment with N. tabacum (Fig. 3a). In an experiment with A. thaliana, the total leaf area of lines that were known to reach different final leaf size clearly developed differently during the monitoring period (Fig. 3b). Wild-type plants had a significantly larger total leaf area than stf1 mutants from day 9 onwards (P = 0.001).

image

Figure 3. Increase in total leaf area throughout 14 d after germination (mean ± SD). (a) Total leaf area and relative growth rate for populations 12L (n = 26) and 12HL (n = 25) of Nicotiana tabacum. (b) Total leaf area for wild type (Ler; n = 13) and stf1 mutant (n = 14) of Arabidopsis thaliana. (See Materials and Methods for details of population abbreviations.)

Download figure to PowerPoint

Total leaf area as indicator of growth acclimation

In the central experiment performed with N. tabacum, from day 10 to day 14 after germination, plants were screened at the same time of day. Plants not reaching at least 20% of the total leaf area of the population mean value at day 10 were considered as outliers and were neglected throughout the treatment period. In other experiments, deviation from the average RGR value on a given day of treatment proved to be a better criterion for removal of outliers (data not shown). Yet, even after removal of outliers, variability of projected total leaf area within a given population was relatively high, which made it difficult to decide whether alteration of light intensity affected growth, when total leaf area was taken as a measurement parameter (Fig. 4). Although plants were distributed randomly, mean values of populations of 24–29 individual plants differed by 20% (plants from high nutrient availability) and 7% (plants from low nutrient availability), respectively, at the first day of treatment (day 10). Within each population, variations of > 20% were found throughout the treatment period (median for all populations and all days of measurement, 31%).

image

Figure 4. Total leaf area of Nicotiana tabacum in different light treatments (mean ± SD). (a) Plants grown at high nutrient availability: populations 12L (n = 26), 6HL (n = 29), 12HL (n = 25), 24L (n = 24). (b) Plants grown at low nutrient availability: populations 12L (n = 28), 6HL (n = 27), 12HL (n = 26), 24L (n = 27). (See Materials and Methods for details of population abbreviations.)

Download figure to PowerPoint

From germination to the first day of treatment at day 10, clear differences were established between all populations from high nutrient availability and those from low nutrient availability (average values 38 and 7.5 mm2, respectively; cf. Fig. 4a,b). By the end of the experiment, average total leaf areas of plants from high and low nutrient availability differed sevenfold (145 and 21 mm2, respectively, at day 14).

anova results showed significant differences between populations differing in light exposure only for populations from high nutrient availability (Fig. 4a): at day 10, population 24L was significantly smaller than the control population 12L (P = 0.019) and population 6HL (P = 0.049), while at day 14, population 24L was significantly larger than control population 12L (P = 0.016). No other significant differences were detected between the four populations from high nutrient availability on any day of the experiment. Either differences between populations were not significant, or the β-error was too high. A higher number of replicates would have reduced the β-error, but would also have led to a longer measurement time and hence to reduced reliability of the results, as explained above. The populations from low nutrient availability did not show any significant differences at all (Fig. 4b).

In general, at day 14 there was a tendency for higher leaf area in plants exposed to increased light intensity at high and low nutrient availability, respectively. At day 12, differences between populations were too small to be interpreted in terms of a tendency on the background of the large variation within each population.

RGR of total leaf area as indicator of growth acclimation

Calculation of RGRs of total, projected leaf area showed that the increase of growth with daily quantum input in plants from high nutrient availability was indeed significant (Fig. 5). The typical variation within each population (median for all populations and all measurement intervals) was only 12% for RGR, and was hence almost threefold smaller than for total leaf area.

image

Figure 5. Relative growth rate (RGR) of the total leaf area of Nicotiana tabacum in different light treatments (mean ± SD). (a) Plants grown at high nutrient availability: populations 12L (n = 26), 6HL (n = 29), 12HL (n = 25), 24L (n = 24). (b) Plants grown at low nutrient availability: populations 12L (n = 28), 6HL (n = 27), 12HL (n = 26), 24L (n = 27). x-axis shows dates of first and second area measurement for calculation of RGR. (See Materials and Methods for details of population abbreviations.)

Download figure to PowerPoint

Growth differences between populations from different external nutrient availability were obvious, and were not statistically tested for (cf. Fig. 5a,b): markedly higher growth rates were found for populations from high nutrient availability in comparable light treatments.

At high nutrient availability (Fig. 5a), population 24L grew significantly more quickly than all other populations as early as the first day of treatment (days 10–11; P = 0.001 in all cases), and grew significantly more quickly than populations 12L (P < 0.05) and 6HL (P < 0.05) for the rest of the experiment. Populations 6HL and 12HL grew significantly more quickly than the control population 12L from the first day of treatment onwards (days 10–11; P < 0.05 in all cases). Population 12HL grew significantly more quickly than population 6HL from days 10–14 (P < 0.05).

At low nutrient availability (Fig. 5b), there was no significant difference between populations at the first day of treatment, and the variations within populations were unusually high. Population 24L grew significantly more quickly than the other three populations from the second day of treatment onwards (days 11–12 and days 10–14; P = 0.001 in all cases). No significant differences were found between populations 12L, 6HL and 12HL at any of the time intervals investigated.

This means that, at high nutrient availability, RGR increased with daily quantum input, showing the highest values if light was received in a continuous manner (24L), while at low nutrient availability only the population receiving constant light (24L) grew more quickly than the control population. Moreover, RGR allowed statistically more reliable conclusions than leaf area, as variability in populations was smaller.

Overall, these results show clearly that seedlings of N. tabacum can react towards increased light exposure within 1 d of treatment, and that the dynamics of the reaction towards increased light exposure depends on the nutrient status of the plant.

Is RGR determined for the expansion of projected leaf area indicative of biomass gain?

All seedling shoots of the central experiment were harvested at day 14, and fresh and dry weight were determined. Those measurements show that fresh weight was correlated closely (r2 = 0.9913) and linearly with total projected leaf area (Fig. 6a). Hence it can be concluded that, for plants of the developmental stage investigated, an increase in total leaf area was indicative of an increase in fresh weight. Also, dry weight was linearly correlated with leaf area (Fig. 6b; r2 = 0.9577). Yet, here, regression was not as perfect as for fresh weight. At both nutrient availabilities, plants from population 24L showed higher dry weight per leaf area compared with the other plants. Despite their strong difference in size, plants from high and low nutrient availability showed exactly the same fresh weight and dry weight, respectively, per leaf area (inclination of the fit lines in Fig. 6a,b: 26 mg FW cm−2 and 1.2 mg DW cm−2, respectively).

image

Figure 6. Correlation between total leaf area and shoot biomass in Nicotiana tabacum in different light treatments. (a) Fresh weight; (b) dry weight. Plants with total leaf area < 60 mm2 were exposed to low nutrient availability; all other plants were exposed to high nutrient availability. (See Materials and Methods for details of population abbreviations.)

Download figure to PowerPoint

Root growth analysis

To evaluate the effect of altered light intensity on root growth, plants were raised in a different cultivation procedure in which roots grew in agar-filled Petri dishes (Fig. 7). This experiment showed that root growth was affected probably even more strongly than shoot growth by increasing light intensity: While total leaf area of plants from the high-light treatment 12HHL was only approx. 20% higher than that of low light-treated plants after 5 d treatment, root area and number of root tips per plant differed threefold. These results show that light-induced increases in total leaf area not only are indicative of above-ground biomass gain, but also are correlated to an overall increase in shoot and root growth.

image

Figure 7. Root growth of Nicotiana tabacum in different light treatments. (a) GROWSCREEN applied for analysis of total leaf area of seedlings grown in agar-filled Petri dishes. (b) Petri dish cultivation of seedlings. (c) Raw data (Bayer) image and segmented and colour-interpolated mask. (d) Total leaf area and total root area of plants grown in control conditions (12L; n = 23; mean ± SD) and at high light intensity for the last 5 d of the experiment (12HHL; n = 24; mean ± SD). (e) Number of root tips per plant. (See Materials and Methods for details of population abbreviations.)

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Phenotyping procedures such as GROWSCREEN are capable of detecting light-induced growth acclimation responses within 24 h. RGR is a more appropriate parameter than total leaf area for quantifying this growth response, as RGR is not affected by the development of the plant before the experimental treatment. This has been pointed out by other authors (Leister et al., 1999); nevertheless, the majority of comparable studies still focus on total leaf area instead of RGR (Barbagallo et al., 2003; El-Lithy et al., 2004; Granier et al., 2006). Our results show that, within a population, smaller and larger plants increase total leaf area by a very similar relative amount per day.

Higher light intensity leads to increased shoot growth within a short time if growth is not nutrient-limited. Plants exposed to low nutrient availability were capable of increasing growth only if the photoperiod was extended from 12 to 24 h. Transfer of plants to continuous light also led to the strongest effects in plants exposed to high nutrient availability. It is conceivable that an extension of the photoperiod can increase the growth performance of N. tabacum more pronouncedly than an amplification of light intensity during a 12-h day, as N. tabacum is adapted to grow in a long photoperiod. Yet extrapolations to later developmental stages from results obtained in this study for seedlings should be treated carefully.

Analysis of fresh weight, dry weight and root growth demonstrated that the entire organism – not merely the area of the leaf canopy – was affected by the treatments. Hence optical phenotyping of seedling plants in the rosette stage is a good proxy of biomass growth processes, as also pointed out by Leister et al. (1999). Increased dry weight of plants from increased light exposure at high nutrient availability is probably caused by increased production of carbohydrates. This hypothesis is supported by findings from a large number of species observed in a time frame of weeks to seasons, for which it was shown that daily quantum input and assimilation rate largely determine plant RGR (Kruger & Volin, 2006; Shipley, 2006).

Because of a lack of measurement methods, the immediate reaction of shoot growth towards increased daily quantum influx has not been assessed up to now. Yet a study analysing single leaf growth of two congener Chamaecyparis species showed recently that leaf growth – not only photosynthesis – reacts within days when light intensity is altered (Lai et al., 2005). Rapid reactions of root growth with increased light exposure of shoots have also been observed recently (Nagel et al., 2007). For root growth, it has been shown conclusively that increased carbohydrate availability generated by amplified shoot photosynthesis causes increased growth at elevated light intensity (Freixes et al., 2002; Nagel et al., 2007).

The optical phenotyping procedure described here is not restricted to analysis of a certain species. As highlighted for the example of A. thaliana, GROWSCREEN can be applied easily to species other than N. tabacum. Magnification of the optical system, number and spatial arrangement of plants can be varied on demand. The optimal parameters hue, saturation and value are easily established for each species and can then be applied throughout the entire experiment. For A. thaliana it was shown that growth of starch-free mutants decreased below wild-type growth soon after measurements started. Carbohydrate reserves of the endosperm might have allowed for comparable growth rates during the first days after germination. Yet the absence of starch in stf1 led to a decrease in growth activity between 3 and 9 d after germination. Transitory starch is a major factor driving nighttime growth activity by providing carbohydrate metabolites for cell-wall assembly and other purposes (Walter & Schurr, 2005). The fact that stf1 mutants remain smaller than wild-type plants has been reported previously (Kofler et al., 2000), but the dynamics of stf1 growth behaviour have not yet been analysed. The finding that growth differences occur at a very early stage in plant development confirms results from the literature obtained with other lines of A. thaliana (Meyer et al., 2004).

The experimental design presented in this study may serve as a protocol for a wide range of future applications. It is conceivable that setups and procedures designed along the lines of GROWSCREEN will be applied to study not only ecophysiology, but also effects of agrochemicals or xenobiotica as well as differences between plant lines caused by their varying genetic backgrounds.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are grateful for the technical support of Andreas Fischbach and Tobias Dierig to establish GROWSCREEN. We thank Tomasz Ochman, Amelie Houben and Milaid Stephan for their help with plant cultivation and image acquisition, and Heike Kofler for provision of seeds.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Barbagallo RP, Oxborough K, Pallett KE, Baker NR. 2003. Rapid, non-invasive screening for perturbations of metabolism and plant growth using chlorophyll fluorescence imaging. Plant Physiology 132: 485493.
  • El-Lithy ME, Clerkx EJM, Ruys GJ, Koornneef M, Vreugdenhil D. 2004. Quantitative trait locus analysis of growth-related traits in a new Arabidopsis recombinant. Plant Physiology 135: 444458.
  • Freixes S, Thibaud MC, Tardieu F, Muller B. 2002. Root elongation and branching is related to local hexose concentration in Arabidopsis thaliana seedlings. Plant, Cell & Environment 25: 13571366.
  • Granier C, Aguirrezabal L, Chenu K, Cookson SJ, Dauzat M, Hamard P, Thioux JJ, Rolland G, Bouchier-Combaud S, Lebaudy A, Muller B, Simonneau T, Tardieu F. 2006. PHENOPSIS, an automated platform for reproducible phenotyping of plant responses to soil water deficit in Arabidopsis thaliana permitted the identification of an accession with low sensitivity to soil water deficit. New Phytologist 169: 623625.
  • Ingestad T. 1982. Relative addition rate and external concentration: driving variables used in plant nutrition research. Plant, Cell & Environment 5: 443453.
  • Kofler H, Häusler RE, Schulz B, Gröner F, Flügge UI, Weber A. 2000. Molecular characterisation of a new mutant allele of the plastid phosphoglucomutase in Arabidopsis, and complementation of the mutant with the wild-type cDNA. Molecular and General Genetics 263: 978986.
  • Kruger EL, Volin JC. 2006. Reexamining the empirical relation between plant growth and leaf photosynthesis. Functional Plant Biology 33: 421429.
  • Lai IL, Scharr H, Chavarria-Krauser A, Küsters R, Wu JT, Chou CH, Schurr U, Walter A. 2005. Leaf growth dynamics of two congener gymnosperm tree species reflect the heterogeneity of light intensities given in their natural ecological niche. Plant, Cell & Environment 28: 14961505.
  • Leister D, Varotto C, Pesaresi P, Niwergall A, Salamini F. 1999. Large-scale evaluation of plant growth in Arabidopsis thaliana in non-invasive image analysis. Plant Physiology and Biochemistry 37: 671678.
  • Meyer RC, Törjek O, Becher M, Altmann T. 2004. Heterosis of biomass production in Arabidopsis. Establishment during early development. Plant Physiology 134: 18131823.
  • Nagel KA, Schurr U, Walter A. 2006. Dynamics of root growth stimulation in Nicotiana tabacum in increasing light intensity. Plant, Cell & Environment 29: 19361945.
  • Schurr U, Walter A, Rascher U. 2006. Functional dynamics of plant growth and photosynthesis – from steady-state to dynamics – from homogeneity to heterogeneity. Plant, Cell & Environment 29: 340352.
  • Shipley B. 2006. Net assimilation rate, specific leaf area and leaf mass ratio: which is most closely correlated with relative growth rate? A meta-analysis. Functional Ecology 20: 565574.
  • Walter A, Schurr U. 2005. Dynamics of leaf and root growth – endogenous control versus environmental impact. Annals of Botany 95: 891900.