EP2673639A2

EP2673639A2 - Method to screen compounds influencing plant cell growth and production

Info

Publication number: EP2673639A2
Application number: EP12702286.1A
Authority: EP
Inventors: Pierre Hilson; Dominique Audenaert; Guillaume QUEVAL; Rudy Vanderhaeghen; Leen VERMEERSCH
Original assignee: Universiteit Gent; Vlaams Instituut voor Biotechnologie VIB
Current assignee: Universiteit Gent; Vlaams Instituut voor Biotechnologie VIB
Priority date: 2011-02-08
Filing date: 2012-02-07
Publication date: 2013-12-18
Also published as: WO2012107451A2; WO2012107451A3

Abstract

The present invention relates to a method for screening compounds influencing plant cell growth and/or production, wherein said method is using fluorescent fusion proteins as reporter proteins and said method is based on a plant protoplast monolayer. The invention relates further to compounds isolated using said method.

Description

METHOD TO SCREEN COMPOUNDS INFLUENCING PLANT CELL GROWTH AND PRODUCTION

To date, genetic screens conducted in plant species rely essentially on the examination of mutant collections generated through mutagenesis by physical or chemical agents, and chance insertion of transposon or T-DNA sequences. The genotypic and phenotypic analysis of thousands of plant individuals remains a laborious and cumbersome process. Furthermore, phenotypes observed in such screens may not inform us about the function of the targeted genes because they result from permanent - possibly lethal - defects and can be indirect. To speed up the screening, automatic systems for high throughput phenotypic analysis of genotypes have been described (Reuzeau et al., 2006). A similar system for identifying plants having different traits differing from a normal population has been disclosed in WO010971 1. However, screening of the effect of compounds on the plant physiology is not described, and growth of the whole plant is still needed, implying a huge investment in labor, space and equipment.

Another approach for screening compounds is disclosed in WO02068606: specific genes that are essential for plant growth are isolated, and compounds that bind those polypeptides are isolated. Alternatively individual plant cells or a whole plant, transformed with a gene encoding such essential protein is grown in presence and absence of a compound, and the amount of messenger RNA encoding the protein or the amount of protein itself in the cell is measured. In this case, a compound with a specific working mechanism can be isolated, but even when plant cells are used, the method is extremely labor intensive and not suited for high throughput analysis.

In order to solve these problems Ma et al (2002) proposed a high throughput screening for herbicides base on green algae suspension cultures. However, this screening system is purely based on the growth of the algae, and due to the physiological differences between plants and green algae suspension cultures, the results cannot easily be extrapolated. WO2004003225 disclosed a method for high throughput screening of plant growth regulators, based on cultures of photomixotrophic cells, including plant cells. However, this method is based on a single measurement of the viability of the cells using 2;3;5-triphenyltetrazolium chloride, and is not giving information on the influence of compounds on cell division, growth or product formation. Moreover, the correlation of growth inhibition caused by herbicides for plants cells grown in photomixotrophic conditions compared with seedlings is rather low (Sato et al., 1987). There is still a need for a simple, automatable high throughput screening method designed to identify the physiological effect of compounds on plants or plant cells.

Surprisingly we found that a screening platform developed to study the cellular phenotypes of plant protoplasts cultured in microwells allowed high-throughput screening of the effect of compounds on plant cell and plant growth. The first operational system developed with this platform enabled the tracking of Arabidopsis cell proliferation. For this purpose, a cell line was constructed expressing a transgene coding for a histone 2B-YFP fusion protein that marks the nucleus and chromosomes in resting and dividing cells. Protoplasts were prepared from cell suspension aggregates after enzymatic digestion of their outer walls. The protoplasts were incubated in isotonic buffered medium and seeded homogeneously across the flat bottom of 96-well plates, forming a monolayer at a density low enough for the cells to be isolated. Because protoplasts are round and relatively large (20 to 40 μηι in diameter) and their nucleus can be positioned at any height within the cell, the imaging of the YFP signal marking the chromatin requires the integration of multiple optical sections in the Z axis for up to 300 time points, with time intervals as short as 15 minutes and over periods of up to 6 days. An algorithmic detection of division events in H2B-YFP Arabidopsis cells was implemented and the monolayer system was validated to study in vitro culture parameters. Furthermore, protocols have been optimized for the high-throughput screening of small compounds with the aim to discover growth-promoting molecules. The bioactivity of candidate molecules identified in a 480-compound library has been confirmed in secondary assays.

The method described here to measure at a large scale the proliferation of isolated Arabidopsis cells has many applications. First, it is ideally suited for the identification of chemicals that act as growth regulators or that impinges in any way on growth regulatory mechanisms by controlling cell proliferation. Second, molecules that show significant bioactivity specific to plant organs may be used to further characterize or discover processes involved in the growth control of economically relevant plant species. Any compound that promotes cell proliferation may be a valuable tool to promote crop production in agricultural conditions or, more directly, to optimize in vitro culture media. In the latter case, novel mitogens may be used to induce the proliferation of protoplasts, so far recalcitrant to in vitro culture, resulting in the formation of microcalli leading to plant regeneration. Alternatively, mitogens may help accelerate the growth of cell cultures raised in industrial fermentors for the synthesis of valuable chemical, thereby shortening batch cycles and reducing production cost. Third, because the method readily identifies molecules toxic to plant cells, it may result in the discovery of novel herbicide leads. Finally the method can be broadened with applications that combine the analysis of cell phenotypes in the monolayer system via high content screening with the introduction, in the plant protoplasts, of genetic gain- or loss-of-function perturbations encoded in nucleic acids. A first aspect of the invention is a method to screen compounds influencing plant cell growth and/or plant cell production, said method comprising (a) expressing a fluorescent fusion protein in a plant cell (b) making protoplasts of the cells (c) culturing the cells under conditions allowing the formation of a cell monolayer (d) contacting the cells with a compound (e) analyzing a representative number of cells of said monolayer with a fluorescent microscope at least 4 time points and (f) calculating the mean fluorescent fusion protein intensity in at least one specific area of the cell.

A "compound" means any chemical or biological compound, including simple or complex organic and inorganic molecules, peptides, peptido-mimetics, proteins, antibodies, carbohydrates, nucleic acids or derivatives thereof. "Influencing" as used here can be positive or negative, thus increasing or decreasing growth and/or production. Preferably, it is an increase of growth and/or production. "Production" as used here can be any production. In one preferred embodiment, production is biomass production. In another preferred embodiment, production is the production of a specific compound, such as a secondary metabolite of said cell. "A plant cell" as used here can be any plant cell, either a cell as part of a complete plant or a cell in a cell culture. Preferably, said cell is a cell in cell culture, even more preferably said cell is an Arabidopsis cell. Preferably, said fluorescent fusion protein is a recombinant protein, endogenous to said cell, fused to a fluorescent polypeptide moiety. Preferably said fluorescent polypeptide moiety is derived from a green fluorescent protein, blue fluorescent protein, cyan fluorescent protein, orange fluorescent protein, red fluorescent protein or yellow fluorescent protein as known to the person skilled in the art. In one preferred embodiment, said fluorescent polypeptide moiety is derived from Yellow fluorescent protein (YFP), even more preferably said fluorescent fusion protein is a fusion between Histone 2B and YFP.

A representative number of cells as used here means that at least 1 % of the cells of the monolayer, preferably at least 5%, more preferably at least 10%, more preferably at least 20%, more preferably at least 30%, more preferably at least 40%, more preferably at least 50%, more preferably at least 60%, more preferably at least 70%, more preferably at least 80%, most preferably at least 90% of the cells of the monolayer is analyzed. A specific area of the cell, as used here, can be the nucleus, the membrane, or a cell organelle such as, but not limited to a vacuole, a plastid or a mitochondrion. In a preferred embodiment, the specific area is the nucleus.

The density in the monolayer should be low enough to allow analysis of individual cells.

Because protoplasts are round and relatively large (30 to 60 μηι in diameter), the microscopic analysis of the cells in the monolayer should be carried out in a way that not only one two dimensional x-y axis section is analyzed, but multiple sections, covering the whole z- axis of the cell, for each time point. Preferably, intervals between time points are short, preferably not more than 24 hours (h), even more preferably not more12 h, even more preferably not more than 6 h, even more preferably not more than 1 hour, most preferably not more than 15 minutes. Preferably at least 4 time points are analyzed, more preferably at least 10 time points, more preferably at least 50 time points, more preferably at least 100 time points, most preferably at least 200 time points. Another aspect of the invention is a compound, isolated by the method according to the invention. Preferably, said compound is a compound selected from the list from figure 10 (Dvs3_1 , Dvs3_9, Dvs3_12, Dvs3_16, Svs3_17, Dvs3_18, Dvs3_21 , Dvs3_26, Dvs3_31 , Dvs3_37, Dvs3_38, Dvs3_40, Dvs3_43, Dvs3_44, 3-E5 and 4-C1 1). Preferably, said compound is N-[5-(2-chlorophenyl)-1 H-1 ,2,4-triazol-3-yl]-4,6-dimethyl-2-pyrimidine (indicated as 4-C1 1):

or a derivative thereof. Preferred derivatives are compounds wherein the CI is replaced by another halogen and/or wherein said halogen or an additional is placed in an ortho (6), meta (3 or 5) and/or para (4) position of the aromatic ring. Still another preferred derivative is a compound wherein one or both methyl groups are replaced by another short alkyl group (up to pentyl), possibly in combination with the variation in position and kind of halogen as described above. A preferred derivative is N-[5-(4-chlorophenyl)-1 H-1 ,2,4-triazol-3-yl]-4,6-dimethyl-2- pyrimidine. Most preferably said compound is N-[5-(2-chlorophenyl)-1 H-1 ,2,4-triazol-3-yl]-4,6- dimethyl-2-pyrimidine.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1. Schematic overview of the analysis of cell proliferation in response to culture medium composition.

Figure 2. Successive changes in nuclei area and YFP intensity mark division events.

Figure 3. Experimental validation for the algorithmic tracking of mitotic events. A; variation of the results in function of the date of measurement (B) validation of the method by the use of known cell cycle inhibitors.

Figure 4. Mitotic rates in function of cell density. Figure 5. Mitotic rates in function of auxin and cytokinin concentration.

Figure 6. Mitotic rates in function of the structure and concentration of auxinic compounds.

Figure 7. DMSO effect on cell division and viability.

Figure 8. Illustration of results obtained the primary compound screen. Conditions are described in the text; 1-12 and A-H indicate the position in the 96 well plate.

Figure 9. Mitogenic activity of compounds confirmed in secondary assays and calculated with time lapse movie analysis. (A) Nuclei numbers in each individual well. (B) Mitotic rate evolution in time.

Figure 10. Chemical structure of compounds with a confirmed mitogenic activity on protoplasts. Figure 11. Mitogenic compounds affect growth of cell-suspension cultures.

Figure 12. Micrographs of seedlings treated with 4-C1 1.

Figure 13. Micrographs taken with a differential interference contrast microscope of plants illustrated in figure 12.

Figure 14. Illustration of the different phenotypes caused by compound treatment.

Figure 15. Structures of the compounds 3-H2, 4-A11 and 6-G5.

EXAMPLES Material and methods to the examples Arabidopsis cell line

The cell culture used in this study was originally obtained from Arabidopsis thaliana (Ler ecotype), and has been renamed PSB-L after adaptation to our facility. The cell line constructed for the tracking of mitotic division events was transformed by co-cultivation with an Agrobacterium tumefaciens strain that carried a binary T-DNA vector containing a transgene coding for a translation fusion between an Arabidopsis Histone 2B (AT5G22880) and the Yellow Fluorescent Protein (H2B-YFP; Boisnard-Lorig et al, 2001). The H2B-YFP is associated with chromatin and marks the cell nuclei.

Cell culture

Cells were grown in 100 mL Murashige & Skoog Mediumwith Minimal Organics (MSMO) containing 0.5 mg.L^"1 NAA and 0.05 mg.L^"1 kinetin with a 16h photoperiod, at 21°C. Every week, cultures were diluted by the addition of 10 mL cell culture to 90 mL fresh MSMO medium. Protoplasts isolation

Protoplasts were isolated from 3-day old Arabidopsis cell cultures. After cell wall digestion with cellulase (15 mg.mL^"1) in 0.4 M mannitol, 5 mM MES pH 5.7, for 3 to 4 h at 25°C, isolated cells were filtered through 40 μηι mesh and incubated overnight in the dark at room temperature for partial cell wall regeneration in MSMO containing 0.35 M mannitol. Dead cells were removed by sucrose gradient (0.5 M) decantation at 100 g for 10 min. Cell density was estimated with a cell counter (Countess, Invitrogen), and cell culture was diluted to a standard density of 5.10⁴ cells per ml. Diluted cells were seeded in a 96-microwell CC2 plate (100 μΙ_ per well; Nunc) and compounds were added. Seeded plates were let to settle on a flat surface to ensure that all cells were in contact with the well bottom, thereby forming a monolayer. The microwell plates were incubated overnight in the dark at room temperature. Cells were imaged for at least 50h.

Image acquisition

Time lapse movies

Images of Arabidopsis monolayer cultures were acquired with the Scan^AR high-content screening station (Olympus). YFP fluorescence was recorded with the fluorescein isothiocyanate parameters, the 10 X objective, following an autofocus, and a15 ms exposure. As plant cells are round and large objects, nuclei can be positioned at any height within the cell. To visualize all nuclei, 10 optical sections were integrated taken at an interval of 5 μηι in the Z axis. Time lapse movies were reconstructed from image series with a temporal resolution of 30 min, recorded for up to 150 h. In optimal conditions, cells divide in 60-90 min, so that 2 to 3 images were obtained capturing mitotic figures of each division event. Such a time lapse is compatible with the acquisition of 4 images from each of the 96 wells in a plate, covering about 10% of the total well surface.

End point analysis

An alternative approach was developed for larger screens. Using the 4 X objective, 12 images captured the entire surface of each of the 96 wells in a plate in 23 min., with only one autofocus in the central position of each well and one channel recording.

Nuclei segmentation and tracking

Time lapse movies were processed with the Scan^AR analysis software suite. After background correction using a rollerball algorithm, nuclei were identified, localized and counted with an edge segmentation algorithm. For each tracked nuclei, position, area, total and mean intensity were computed over time, yielding traces that represent area and intensity variation per nuclei over time (Fig. 2). A decrease in YFP signal area followed by a peak in mean YFP signal intensity corresponded to nucleus condensation, while a drop in the total YFP signal intensity corresponded to karyokinesis. These two events marked the beginning and end of mitotic events respectively. The absolute count of such events leads to the calculation of the division density across time and of the approximate duration of each mitosis. For large screens, cell proliferation (and indirectly cell death) was more simply estimated by subtracting the number of nuclei at the end and the start of the experiment.

Analysis of compound effects on the growth of plant cell suspension cultures

Seven-day-old cultures were diluted with the addition of 1 .5 mL cells to 13.5 mL fresh MSMO medium containing 0.5 mg.L^"1 NAA, 0.05 mg.L^"1 kinetin and the compound of interest. Cultures were grown in Erlenmeyer culture flasks for 11 days with a 16h photoperiod (PAR = 73 μηιοΙ/ΓΤΐ2.3), at 21°C and 130 rpm. Eleven days after inoculation, the cells were harvested, washed with 10 mL fresh hormone-containing MSMO medium, resuspended in 50 mL medium, and subsequently grown for another 7 days. Growth was determined at several timepoints as indicated in the graphs. The OD₆oonm was measured three times for two technical replicates of two dilutions of 200 at an OD in the linear range, with a VERSAmax tunable microplate reader (Molecular Devices, Sunnyvale, CA). The fluorescence was measured from the same samples with a FLUOstar OPTIMA fluorescence plate reader equipped with the FLUOstar optima software (excitation filter, 485 nm; emission filter, 520 nm; gain, 2000; BMG LABTECH GmbH, Ortenberg, Germany). The sedimented cell volume was determined from 10 mL of culture sedimented for 40 min in a 12 mL falcon tube.

Example 1 : Reproducibility of the method

The proliferation of the cells in the monolayer depends on multiple factors, among which the cell density and the phytohormones provided in the fresh culture medium. In our standard protocol, 5.10³ cells were seeded per well in 100 μ\- of MSMO medium (5.10⁴ cells per ml) supplied with auxin (NAA, 0.5 mg.L^"1) and cytokinin (kinetin, 0.05 mg.L^"1). In such conditions, about half the cells divided within the next 50h. Little variation was observed between different experiments over the course of 15 months (Fig. 3A). Example 2: Effect of known cell cycle inhibitors

Four known mitotic inhibitors have been tested (hydroxyurea, aphidicolin, propyzamide and MG132) to assess the robustness of the algorithm designed for the detection of the division events. Images of the Arabidopsis cell monolayer in the absence or presence of the inhibitors were analyzed both with the software and manually. Aphidicolin and MG132 significantly decreased the number of mitotic events at low concentration (5 μΜ), whereas propyzamide and hydroxyurea required higher concentration (100 μΜ and 5 mM respectively) to induce a significant decrease (Fig. 3B).

In these conditions, mitotic events were manually tracked on 19 series of 200 time lapse images. Comparison between the manual annotation and the divisions detected by the software showed similarity rates from 39 to 88 % depending on the treatments of the cells in each well (median = 69%) and false positive rates from 16 to 100 % (median =36%). We note that extreme values (low similarity and high false positives) were obtained when no division occurred and the software detected even fewer divisions (Table 1). The reproducibility of the cell division phenotype was verified across the different well positions in a 96-well plate. The plate was seeded with cells in standard conditions, half of the wells containing a cell cycle inhibitor (propyzamide, 200 μΜ) (Fig. 3B). Anova calculation between the columns or the rows of the plate gave no significant P-values for untreated cells, and variable P-values for treated cells (for both nucleus and mitosis numbers). Z-factors calculated between the two conditions showed the highest values (>0.4) for long kinetics (72h or 96h tracking) and large area imaging (with 4X objective) resulting in a higher number of cells being tracked.

To assess the effect of cell density on proliferation, monolayers have been imaged after seeding microwells with increasing cell concentrations, from 2.5.10⁴ to 15.10⁴ cells per ml_^"1. Results show that mitotic rates peak for middle range concentrations (3.5.10⁴ to 8.5.10⁴ cell.mL^"1) and decrease dramatically with lower and higher concentrations (Fig. 4).

Example 3: Positive effect of phytohormone combinations

Auxins and cytokinins have long been known to control the formation of microcalli starting from in vitro cultured cells (Murashige and Skoog, 1962). To verify that the same effects can be reproduced in the monolayer configuration, we calculated the mitotic rate of cells treated with a series of synthetic auxin (NAA; 0.1 to 10 mg.L^"1) and cytokinin (kinetin; 0.05 to 5 mg.L^"1) concentrations. The maximal mitotic rates were observed at concentrations close to those commonly used for optimal cell proliferation in liquid culture (0.5 mg.L^"1 NAA and 0.05 mg.L^" ¹ kinetin) confirming that the isolated fixed protoplasts within the monolayer respond to these phytohormones in a way similar to the corresponding cell suspensions (Fig.5). Table 1. Comparison of the mitotic event counts obtained with manual annotation and automated algorithmic tracking.

c c0 T

IX- X. Auxin

Multiple chemicals display auxin-like activity in plants. Some are synthesized in plant tissues, others can only be fabricated synthetically. These auxinic compounds have distinct properties, in part because they differ in their ability to penetrate cell membranes or tissues, and because AFB auxin family receptors have distinct affinities for different auxinic compounds (Savaldi- Goldstein et al, 2008). To test the respective effect of such compounds on protoplast proliferation, monolayers were treated with eight different auxinic molecules in a wide concentration range (from 0.05 to 135 μΜ), in combination with kinetin (0.05 mg.L^"1). The tested auxins can be classified in three major groups according to their associated mitotic rate (Fig. 6). The first group contains a single compound, IBA that yielded only few cell divisions in our experimental setting, regardless of concentration. IBA's poor activity may reflect the fact that IBA is not active on its own but has to be metabolized into IAA to induce cell division in vitro. Auxins in the second group are NAA, IAA, TIBA, 2,4-D, and PAA. These molecules yielded fewer cell divisions for the highest concentrations, while the mitotic rate peaked with mid-range concentration. IAA, NAA and TIBA showed the strongest repression for high concentrations. The third groups included Picloram and Dicamba that exhibited high activity across most of the concentration range, with somewhat lower mitotic rates for the lowest concentration. Interestingly, the maximal mitotic rates were similar for all tested auxins, except IBA and TIBA.

Cytokinin

In auxin-cytokinin gradient experiments, an inhibitory effect on mitosis was detected for kinetin at 5 mg.L^"1 , but not at 0.5 mg.L^"1. Cell proliferation was compared in the presence of kinetin or BAP within a concentration range from 0.0001 to 10 mg.L^"1 (0.44 nM to 44 μΜ), in combination with NAA. No differences in the mitotic rate were detected for both species at concentrations below 2.5 mg.L^"1 (1 1 μΜ), but cell proliferation was inhibited by higher concentrations of either kinetin or BAP.

Example 4: Pilot chemical library screen

In our experimental set-up, mitotic rates can be measured following treatment of thousands of cells in a small volume (100 μί). This format is ideally suited for screening compound libraries with the aim to identify molecules with mitogenic activities on plant cells.

In preparation of chemical screening campaigns, we tested the effect of DMSO on cell proliferation because compound libraries are dissolved in this neutral solvent. At a concentration inferior to 1 %, DMSO did not affect cell division. Cell division inhibition was significant at higher concentrations and cell death was observed at concentration higher than 2% (Fig. 7). All chemical assays described below were performed with a concentration of 0.5 or 1 % DMSO.

A library of 480 compounds randomly selected from our main chemical library of 12,000 synthetic organic molecules (DIVERSet™, ChemBridge Corporation, San Diego, California, USA) was screened for mitogenic activity at 100 μΜ and 0.5% DMSO, in four distinct growth conditions. In condition A, protoplasts were seeded at optimal density (5.10⁴ cell.mL^"1) with optimal hormone concentrations in the culture medium (0.5 mg. L^"1 NAA, 0.05 mg. L^"1 kinetin). This assay aimed at identifying compounds that further enhance the mitotic rate associated with the optimal combination of auxin and cytokinin. In condition B, protoplasts were seeded at optimal density but without added hormones, with the aim to identify compounds that display hormone-like activity or that bypass the need for auxin and cytokinin possibly acting downstream of the corresponding signalling pathway(s). Third, protoplasts were seeded at low density (3.5.10³ cell.mL^"1) with optimal hormone concentrations, conditions in which cell division is not observed in the monolayer. In condition C, assay aimed at identifying compounds that compensate for the low density inhibition, possibly replacing the activity of unknown mitogen(s) potentially produced in denser cultures. In condition D, protoplasts were seeded at high density (2.10⁵ cell.mL^"1) with optimal hormone concentrations, conditions in which cell division is also inhibited. This assay aimed at identifying compounds that compensate for the high density inhibition, possibly blocking the activity of unknown mitogenic inhibitor(s) potentially produced in dense cultures. Candidate molecules where identified that promote or inhibit cell proliferation as illustrated in Figure 8.

Condition A

After 72-hour incubation, protoplasts divided with a global rate of approximately 0.46. Twelve compounds induced cell divisions with higher rates, from 0.54 to 1 .07 (Fig. 8A). The induction followed two distinct profiles, with rates either decreasing or remaining constant over time.

Condition B

Without phytohormones, cells normally do not divide and a significant fraction of the cell population dies, yielding an aberrant calculated mitotic rate of -0.24 on average. Nevertheless, two compounds sufficiently promoted division to yield positive mitotic rate: one of them yielded a rate of 0.40 and was already identified in Condition A; the second yielded a rate of 0.23 (Fig. 8B). Condition C

Dilution caused higher levels of variability because a smaller number of cells were taken into account, thereby hampering the interpretation of results obtained at low density. Twenty compounds may induce cell division in this condition, one being common with the first two conditions (Fig. 8C).

Conditions D

In the initial experiment, we failed to reach high inhibitory cell density, seeding wells at only half of the desired cell number. The mitotic rate for untreated cells in this case was about 0.5, almost identical to Condition A. Three compounds distinct from those identified above increased mitotic rates, from 0.56 to 0.66. Cell death

Several compounds markedly decreased protoplast viability (Fig. 8). They can be grouped in two classes according to cell death kinetics. Molecules in the first class were highly and immediately toxic and lead to the death of most cells within the initial overnight incubation following compound addition. Treatment with molecules in the second class resulted in gradual decrease of living cell number. Compounds with the highest toxicity in all four batches were compared to those deleterious to yeast (Saccharomyces cerevisiae). Seventeen chemicals showed detectable toxicity specifically towards Arabidopsis protoplasts but not for yeast.

Secondary screen for positive effectors of cell proliferation

Fourteen compounds flagged as potential mitogens in the primary 480-compound screen have been characterized further. Their effect on cell division was studied for three concentrations (25, 50 and 100 μΜ) through the imaging of protoplasts seeded at optimal density (10⁴ cell.mL^" ¹). The initial data consisted in high resolution time lapse movies collected for 150 h following treatment (4 frames per well corresponding to ca. 400 cells each, imaged every 30 min), and processed automatically with the cell division tracking algorithm to calculate the corresponding mitotic rates. Each compound at each concentration was tested in triplicate.

Among those included in the secondary screen assays, 5 compounds significantly increased the monolayer mitotic rate, either in the presence or in the absence of NAA and kinetin (Conditions A and B) (Fig. 9). These molecules could be classified in three groups according to the magnitude and the timing of their mitogenic effect. 4-C11 and 3-E5 induced proliferation whether hormones were supplied or not in the liquid culture medium (up to 2-fold increase in mitotic rate). The deviation between controls and treated monolayers was not observed in the first 50 h incubation but became increasingly significant in the following 100 h period, when the mitotic rate of control cells declines. 4-A11 and 3-H2 (structures in Figure 15) induced proliferation only in the presence of phytohormones, to a lesser extent but according to the same dynamics as 4-C11 and 3-E5. Finally, 6-G5 (structure in Figure 15) showed an immediate and highly positive effect in the absence of phytohormones (4-fold increase in mitosis number in the first 50 h) sustained throughout the experiment (Fig. 9). The structure of this compound is related to auxin, thereby possibly explaining its bioactivity.

Finally, the same five molecules were tested for their potential mitogenic effect on high cell density monolayers (2.10⁵ cell.mL^"1). 6-G5 and 3-E5 were the only compounds that increased mitotic rate in the absence and presence of the phytohormones, respectively.

Example 5: High-throughput chemical library screen for positive effectors of cell proliferation

As the results of the pilot screen were promising, the 12,000 compounds large Diverset3 library of ChemBridge was screened. This library contains drug-like compounds with maximum pharmacophore diversity. The primary screening was performed in condition A, thus with optimal cell density and phytohormone concentration, by assessing the nuclei fold change over 72h in a concentration of 50 μΜ compound and 0.5% DMSO. As a large set of compounds can be retested for their mitotic division rate in a short time frame, 400 compounds that possibly promoted cell density within either 48 or 72 h (Z-score≥ 2 and counterselection for aberrant high scores caused by toxicity which decreased the start amount of nuclei) were retained for a secondary screen.

In the secondary screen, the compounds of interest were identified by taking into consideration the contribution of cell division to the increase in nuclei fold change. For this purpose, the nuclei fold change was retested and the mitotic rate was calculated for the same compounds in the same conditions from time lapse movies. Selected compounds promoted cell division without affecting cell viability.

As a first selection, 32 compounds were considered as valuable hits with a Z-score≥ 3 for both the mitotic rate and the nuclei fold change, following a 3-day treatment of the Arabidopsis protoplast monolayer. Interestingly, one of these hits, Dvs3_21 , differs from compound 4-C1 1 only in the position of a halogen (CI, Fig. 10), validating the robustness of the platform and screening procedures. Furthermore, 17 additional compounds were selected based on an apparently high Z-score for the nuclei fold change in both the primary as secondary screens. Fourteen out of these 49 hits were confirmed in a time-lapse confirmation screen at 50 μΜ, their structure is indicated in Fig. 10 (compounds indicated in bold). Compounds that enhanced protoplast proliferation in the monolayer system are good candidates to enhance growth of transformed or fused protoplasts in in vitro culture procedures used for example in horticultural breeding. Example 6: Effect of selected compounds on the growth of suspension cultures

Suspension culture cells differ from protoplasts because they are encased in a rigid cell wall and they form large cell clumps. Therefore, we tested whether compounds that act as positive effectors of protoplast proliferation also had similar bioactivity on suspension culture growth. Seven-day-old cultures were resuspended in fresh medium containing one of the sixteen confirmed compounds at 50 or 100 μΜ indicated in Fig.10. Growth was assessed by measuring the optical density, sedimented cell volume and fluorescence emitted by the H2B- YFP nuclei. As such both the biomass and the amount of viable cells was assessed.

Six compounds, Dvs3_17, Dvs3_18, Dvs3_21 , Dvs3_37, Dvs3_38 and Dvs3_43, yielded higher H2B-YFP nucleus fluorescence signal at the time when the control cultures reached their growth plateau (Fig. 11). Cultures treated with compound Dvs3_43 showed a prolonged lag-phase, but also yielded a higher fluorescence signal compared to the control at a later time point. Three of these, Dvs3_21 (100 μΜ), Dvs3_37 (50 and 100 μΜ) and Dvs3_38 (100 μΜ) also increased the OD. The sedimented cell volume was comparable to or lower than the control. The different trends observed for the three growth criteria -OD, sedimented cell volume and H2B-YFP nucleus-fluorescence- might indicate that certain compounds increased cell viability or cell division without affecting cell mass. Smaller cells and multinucleated cells may have a larger nuclei to biomass ratio. In addition, cell clump formation was affected by the compounds, which might alter sedimentation and light scattering, thereby influencing the sedimented cell volume and OD measurements, respectively.

In a follow up experiment, we noted that increased viability, measured 11 days after inoculation and treatment, as indicated by higher fluorescence, correlated with a remarkable subsequent enhancement of growth following dilution of the culture with fresh medium. Eleven days after inoculation, the cells were harvested, washed with fresh medium and resuspended to obtain a five-fold dilution of the culture and 125-fold dilution of the compound. In parallel experiments, cells derived from control treatments led to a limited increase in OD, sedimented cell volume and fluorescence within a 4-day time frame, while at least two out of the three parameters were dramatically higher when inoculation occurred with cells showing a higher fluorescence at the time of dilution (Dvs3_17 and Dvs3_21 at 50 μΜ, Dvs3_18 and Dvs3_43 at both concentrations and Dvs3_38 at 100 μΜ).

Example 7: Effect of selected compounds on the growth of Arabidopsis seedlings

In the monolayer configuration, compounds were selected for their mitogenic activity on Arabidopsis protoplasts. One of the benefits of screening individual cells is that cellular responses are examined in a simplified environment. To further characterize the bioactivity of the compounds of interest on plant growth, we investigated their biological activity in Arabidopsis seedlings. As described above, sixteen compounds identified from the Diverset3^{I M} library and two from the 480 subset Diverset™library were confirmed to enhance cell proliferation of cells seeded at a density 5.10⁴ cells. mL^"1 , in the presence of phytohormones (0.5 mg.L^"1 NAA, 0.05 mg.L^"1 kinetin). To assess their effect in planta, Arabidopsis seedlings carrying a CYCB1; 1 ro::GUS promoter- re porter transgene were treated with these compounds of interest. The transcriptional activity of the cyclinB1; 1 promoter is routinely used as an indicator of cell division in Arabidopsis plants and is associated with mitotically active tissues (Ferreira et al., 1994).

Seedlings were transferred into liquid ½ MS medium containing a tested compound at either 50 or 100 μΜ (0.5% DMSO) eight days after stratification, then incubated for 48 h. In eight- day-old seedlings some lateral root primordia already emerged while others are being formed, therefore enabling the study of the compound effects on both the initiation and outgrowth of lateral root primordia. GUS-staining patterns revealed distinct effects on leaf as well as root development.

As an example, micrographs of seedlings treated with 50 or 100 μΜ of compound 4-C1 1 for 48 h are provided in figure 12 and 13. They illustrate aberrant growth in the elongation and maturation zone of the primary root, possibly explained by radial, instead of longitudinal cell expansion. Furthermore, compared to DMSO treated controls, root hairs were swollen and short, and appeared much closer to the primary root tip, indicating that the activity of the root meristem was inhibited. However, root hairs in the mature part of the primary root did not have an abnormal phenotype. More lateral root primordia were found on treated seedlings versus controls, and occasionally occurred side by side, on directly opposite sides, or close to the primary root tip, in configurations never observed in control roots. Furthermore, lateral root outgrowth was hampered by the compound treatment, and the outer cell layers of the emerged primordia were swollen, while the inner tissues appeared similarly organized compared to controls.

Several of these features were also observed after treatment with other compounds. Table 2 provides an overview of the phenotypes caused by each compound, and figure 13 illustrates each phenotype. Inhibition of root outgrowth was often correlated with the swelling of root tip, outer tissues of the lateral root and lateral root primordia, themselves coincident with ectopic or stronger GUS staining. These observations suggests that some of the selected compounds directly or indirectly alter lateral root initiation and/or root outgrowth, while others do not affect root development and merely slow down the overal growth of the seedling. However, the high incidence of compounds with strong root growth effects probably reflects the specificity and selectivity of the initial screening protocol designed for the identification of positive effector of isolated plant cells in vitro. Table 2. The in planta effects of compounds with a mitogenic activity on protoplasts

(1) M, medium, similar to control; S and XS, respectively smaller and very small compared to the control

(2) N, normal, similar to the control; F, faint; S, strong; NO, no staining visible

(3) N, normal, similar to the control; S, swollen tissues; MES, maturation and elongation zone swollen

(4) C, lateral root primordia close to each other; VC, very close; I, inner tissues stained in stretches

REFERENCES

Boisnard-Lorig C, Colon-Carmona A, Bauch M, Hodge S, Doerner P, Bancharel E, Dumas C, Haseloff J, Berger F (2001) Dynamic analyses of the expression of the HISTONE::YFP fusion protein in arabidopsis show that syncytial endosperm is divided in mitotic domains. Plant Cell 13, 495-509.

Ferreira P, Hemerly A, de Almeida Engler J, Bergounioux C, Burssens S, Van Montagu M, Engler G, Inze D (1994) Three discrete classes of Arabidopsis cyclins are expressed during different intervals of the cell cycle. Proc Natl Acad Sci USA 91 : 11313-11317 - Ma J, Wu L, Wang S (2002) A quick, simple, and accurate method of screening herbicide activity using green algae cell suspension cultures. Weed sci 50, 555-559.

Murashige T, Skoog F (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 15, 473-497.

- Reuzeau C, Frankard V, Hatzfield Y, Sanz A, Van Camp W, Lejeune P, De Wilde C, Lievens K, de Wolf J, Vranken E, Peerbolle R and Broeckaert W. (2006). Traitmill™: a functional genomics platform for the phenotypic analysis of cereals. Plant genetic resources: characterization and utilization 4, 20-24.

Sato F, Takeda S, Yamada Y (1987) A comparison of effects of several herbicides on photoautortrophic, photomixotrophic and heterotrophic cultured tobacco cells and seedlings. Plant cell reports 6, 401-404.

- Savaldi-Goldstein S, Baiga TJ, Pojer F, Dabi T, Butterfield C, Parry G, Santner A, Dharmasiri N, Tao Y, Estelle M, Noel JP, Chory J (2008). New auxin analogs with growth- promoting effects in intact plants reveal a chemical strategy to improve hormone delivery. Proc Natl Acad Sci U S A. 105, 15190-15195.

Claims

1. A method to screen compounds influencing plant cell growth and/or plant cell production, said method comprising (a) expressing a fluorescent fusion protein in a plant cell (b) making protoplasts of the cells (c) culturing the cells under conditions allowing the formation of a monolayer (d) contacting the cells with a compound (e) analyzing a representative number of cells of said monolayer with a fluorescent microscope at least 4 time points and (f) calculating the mean fluorescent fusion protein intensity in at least one specific area of the cell.

2. The method according to claim 1 , wherein said fusion protein is a histone protein.

3. The method according to claim 2, wherein said specific area of the cell is the nucleus.

4. A compound influencing plant cell growth and/or production, isolated with the method according to any of the claims 1-3.

5. The compound of claim 4, whereby said compound is selected from the group consisting of Dvs3_1 , Dvs3_9, Dvs3_12, Dvs3_16, Svs3_17, Dvs3_18, Dvs3_21 , Dvs3_26, Dvs3_31 , Dvs3_37, Dvs3_38, Dvs3_40, Dvs3_43, Dvs3_44, 3-E5 and 4-C11 as indicated in figure 10.

6. The compound of claim 5, wherein said compound is

or a derivative thereof.

7. The compound according to claim 6, wherein said derivative is