Synthetic antisense oligodeoxynucleotides to transiently suppress different nuclear- and chloroplast-encoded proteins of higher plant chloroplasts

Selective inhibition of gene expression by antisense oligodeoxynucleotides (ODNs) is widely applied in gene function analyses; however, experiments with ODNs in plants are scarce. In this work, we extend the use of ODNs in different plant species, optimizing the uptake, stability and efficiency of ODNs with a combination of molecular biological and biophysical techniques to transiently inhibit the gene expression of different chloroplast proteins. We targeted the nuclear-encoded phytoene-desaturase (pds) gene, encoding a key-enzyme in carotenoid biosynthesis, the chlorophyll-a/b binding (cab) protein genes and the chloroplast-encoded psbA gene, encoding the D1 protein. For pds and psbA , the in vivo stability of ODNs was increased by phosphorothioate modifications. After infiltration of ODNs into juvenile tobacco leaves we detected a 25-35% reduction in mRNA-level and a ~5% decrease in both the carotenoid content and variable fluorescence (Fv) of photosystem II. In detached etiolated wheat leaves after 8 h of greening the mRNA level, the carotenoid content and Fv were inhibited up to 75, 25 and 20%, respectively. Regarding cab , ODN-treatments of etiolated wheat leaves resulted in an up to 59% decrease in the amount of chl b , 41% decrease of the maximum chlorophyll fluorescence intensity, the cab -mRNA levels was reduced to 66% and the protein level was suppressed up to 85% compared to the control. The psbA -mRNA level and protein levels in Arabidopsis leaves were inhibited by up to 85% and 72%, respectively. To exploit the potential of ODNs for photosynthetic genes we propose molecular design combined with fast, non-invasive techniques to test functional effects. ODNs efficiently knock down the gene expression of the pds gene, resulting in decreased carotenoid contents and diminished photosystem II (PSII) activity; these inhibitory effects were enhanced by PS modification of the antisense ODNs. Regarding the cab gene we show with western blot analysis that a single antisense ODN is capable of inhibiting more than one gene from the same gene family with high efficiency. Using ODNs to inhibit the psbA gene, encoding for the D1 protein, we show for the first time that it is possible to inhibit a chloroplast encoded gene with antisense ODN technology.


INTRODUCTION
6 Sequence-selective inhibition of gene expression is applied extensively for elucidation of complex gene expression patterns or validation of results gained from high throughput genomic experiments such as DNA-arrays (Gleave and Monia, 2005;Rayburn and Zhang, 2008). Antisense ODNs have gained great interest, not only as genomic tools, but also as possible therapeutic entities. These synthetic molecules are capable of interfering with the expression of the targeted gene, thus enabling selective and rational design of genomic or therapeutic agents (Gleave and Monia, 2005;Rayburn and Zhang, 2008). Potentially, antisense ODNs could be used to cure any disease that is caused by the expression of a deleterious gene, e.g. a viral infection, cancer growth or inflammatory diseases. It has been first demonstrated by Zamecnik and Stephenson (1978) that antisense ODNs can be successfully used to inhibit viral replication in cell cultures and since then a number of applications have been reported (Scherer and Rossi, 2003).
One of the major challenges in the application of antisense ODNs is stabilization, since the natural molecular structure of ODNs is exposed to quick endo-and exo-nucleolytic degradation. In order to extend the biological life-span and improve accessibility, a number of chemical modifications were introduced successfully into antisense ODNs (Chan et al., 2006;Lebleu et al., 2008). Most commonly, the stability of ODNs is increased by phosphorothioate (PS) modification (Eckstein, 1985). Phosphorothioated ODNs are the major representatives of first generation DNA analogs that are the best known and most widely used antisense agents to date. In this class of analogs, one of the non-bridging oxygen atoms in the phophodiester bond is replaced by sulfur (Matsukura et al., 1987;Eckstein, 2000;Kurreck, 2003). The PS modification is particularly popular because it provides sufficient stabilization against nucleolytic degradation, whereas the duplex of PS-ODN and RNA is still recognized by cellular RNase H. This is an advantage compared to most other modifications. The cleavage of target mRNA by RNase H is considered an important factor for the activity of antisense ODNs (Stein et al., 1993;Stein, 1995). Equally important is the cellular delivery. A further challenge is to find the optimal target sites inside the RNA, which provide maximum efficacy and minimize off-target effects at the same time.
The application of sequence-selective gene-silencing ODNs has been described for several organisms, including mammalian cells and it is an important emerging therapeutic tool in clinical medicine (Gewirtz et al., 1998;Dagle and Weeks, 2001;Hu et al., 2002;Shi and Hoekstra, 2004;Yang et al., 2004). In spite of the general applicability, the antisense ODN technology has not yet been truly exploited in plant tissue, although it was observed almost two decades ago that plant cell suspension cultures are capable of taking up single-stranded accumulated both in the epidermal and parenchymal cells indicating the transport of ODNs inside the leaves. High resolution confocal laser scanning microscopy showed that the ODNs accumulated both in the cytoplasm and the chloroplasts of parenchyma cells of wheat (Figs. 1D and F) and tobacco (Figs. 1G and I). Chl a fluorescence emission was used to localize the position of the chloroplasts (Figs. 1B,E,H and K). Oligodeoxynucleotides accumulated also in the nucleus (Fig. 1J). These data indicate that ODNs efficiently penetrate through cell walls and plasma membranes.

Phytoene desaturase (pds) as a model gene to test the efficacy of antisense ODNs and the selection of target sequences
To demonstrate the applicability of the antisense ODN technology in plants, we chose the nuclear-encoded phytoene desaturase (pds) and the chlorophyll a/b binding protein (cab) genes as model genes. The pds gene has been used as a marker in several studies on gene silencing in plants (Chamovitz et al., 1993;Kumagai et al., 1995;Lindgren et al., 2003;Tao and Zhou, 2004;Wang et al., 2005;Wang et al., 2009). Phytoene desaturase encodes a keyenzyme in carotenoid biosynthesis, which converts phytoene to the colored ξ -carotene in a two-step desaturation reaction (Bartley and Scolnik, 1995). Low expression of the pds gene results in a general suppression of carotenoid biosynthesis (Wetzel and Rodermel, 1998).
The mRNA molecules form complex secondary and tertiary structures, therefore, in antisense technology, the identification of mRNA sites that can be targeted efficiently is the key to success (Stein, 2001). The percentage of active antisense ODNs is known to vary from one target to the next. A variety of strategies has been developed to successfully design ODNs (Sohail and Southern, 2000). In this work, ODN sequences were designed based on a search 9 for the presence of accessible single-stranded loops in the mRNA secondary structure (Zuker, 2003).
For the targeted pds mRNA, ten different antisense ODNs were designed and synthesized both for wheat and tobacco, with their natural phosphodiester and PS structures. It has been shown that a decrease in the carotenoid content due to the suppression of the pds gene results in a decrease of the Chl content and in a decrease of the photosynthetic efficiency as well (Wang et al., 2010), which is probably due to the instability of the synthesized Chl-binding protein complexes in the absence of carotenoids (Plumley and Schmidt, 1987). Therefore, to select the most efficient ODNs and also to determine the time points where the largest differences are found between the control and the ODN-treated plants, we used the fast Chl a fluorescence (OJIP) transient technique, a non-invasive and very sensitive method for measuring photosynthetic efficiency (Govindjee, 2004;Schansker et al., 2005, Lazár andSchansker, 2009 The 24 h of antisense ODN treatment resulted in 5-10% lower F V values compared to the random-nonsense controls ( Fig. 2A). This was likely the consequence of the decrease in the transcript level of the pds gene, which was suppressed by about 25-35% compared to the random-nonsense control (Fig. 2B). This inhibition led to a 5-10% decrease in the carotenoid content (Fig. 2C), a decrease similar to the one observed for F V . The data also show that the antisense ODN 8 with PS modification (8T) was somewhat more efficient than its natural structure (8N). This relatively small effect of the antisense ODNs on the mRNA level, the carotenoid content and F V value can be explained by the slow turnover of the PDS enzyme and the high background levels of the pigments and photosynthetic complexes. We, therefore, carried out experiments on greening leaves, i.e. during the synthesis of the photosynthetic machinery.

The effects of pds antisense ODNs during greening of wheat leaves
It is well known that etiolated leaves contain no Chl, whereas lutein and violaxanthin can be found at low levels when compared with light-grown plants and they contain only traces of any other carotenoid (Park et al., 2002). When etiolated leaves are exposed to light, Chl biosynthesis starts and the carotenoid content increases (von Lintig et al., 1997). Based on these observations we expected that pds antisense ODNs could efficiently slow down or inhibit the de-etiolation process. Etiolated leaves were detached and their basal part was submerged for 12 h in a 10 µM pds antisense ODN solution in the dark. Subsequently, they were illuminated with white light of 100 µmol photons m -2 s -1 for 24 h. The carotenoid content of etiolated leaves was low compared to green leaves and remained constant during the 12 h incubation in the presence of antisense ODNs in the dark (data not shown).
After 8 h of illumination of wheat leaves, the F V value was 12-30% lower in the antisense ODN-treated plants (Fig. 3A) compared with controls, i.e. random-nonsense ODNs. At the same time, the expression level of the pds gene was suppressed dramatically, by about 50 to 70% (Fig. 3B). This inhibition led to a significant decrease in the carotenoid content: 18-24% depending on the antisense ODN applied ( Fig. 3C; the total carotenoid content of the control was about 0.2 µg mg -1 FW). It is important to note that the differences in all these three parameters were more pronounced in plants treated with PS modified ODNs compared to the ODNs with a natural structure; the largest difference was found between ODNs 11T and 11N.
These data show that pds ODNs slow photomorphogenesis down. Indeed, the Chl content was also lower in the antisense ODN-treated plants (by about 25-42%, depending on the ODN applied, data not shown). This is consistent with the well established fact that the stability of pigment-protein complexes depends on the presence of carotenoids. It is interesting to note that following an 8 h de-etiolation, the Chl a/b ratio was higher (typically between 10 and 16) in ODN-treated plants than in the control plants (Chl a/b ratio around 8), suggesting a much slower synthesis of the outer (chl a and b containing) antennae in the ODN-treated leaves.
This difference vanished after 24 h of illumination, which is in agreement with the results of does not change the Chl a/b ratio.
Seeing the interference of ODN treatments with early greening events, we looked at wheat leaves in a more advanced stage of their photomorphogenesis as well. After 24 h of deetiolation the effect of ODNs on the F V value was less than 10% (Fig. 4A). Concerning the transcript level of the pds gene, the decrease induced by the antisense ODNs was found to be about 25% at this later phase of de-etiolation (Fig. 4B), while these values varied between 50 and 70% after 8 h (Fig. 3B). Also, the reduction in the carotenoid content became considerably smaller ( Fig. 4C; the total carotenoid content was about 0.27 µg/mg FW in the control leaves) compared to leaves that had been treated at an earlier stage of greening. This result is in accordance with the idea that ODNs can only inhibit de novo synthesis.

The effect of chlorophyll a/b binding protein (cab) gene antisense ODNs on antenna development during the greening of wheat leaves
The other model genes were the genes coding for the chlorophyll a/b binding proteins (cabgenes). The CAB proteins (also called light harvesting complexes (LHCs)) form a family of nuclear-encoded thylakoid proteins. At least ten distinct types of CAB proteins have been proteins (often called CP29, CP26 and CP24, respectively) are monomeric proteins linking the LHCII trimers to the reaction center core (Jansson, 1999).
For the inhibition of the cab gene mRNAs of wheat, ten different antisense ODNs were designed and synthesized using the natural (phosphodiester) structure based on the single mRNA-sequence found in the genebank database for wheat (Triticum aestivum). Although all ODNs were able to inhibit the cab genes to some extent, the two most effective ODNs were selected using the fast Chl a fluorescence transients as a selection criterion. The fluorescence data also indicated that the maximum effect of the ODN treatments was obtained after 8 h of greening. Fig. 5A, a photograph of etiolated wheat seedlings treated with CAB ODNs also demonstrates very clear effects after 8 h of greening. After 8 h of illumination, the F M values were 36 to 41% lower in the antisense ODN-treated plants than in the controls (Fig. 5B). At the same time, the expression levels of the cab genes were suppressed by about 59 to 66% ( Fig. 5C). This inhibition led to a significant decrease in the chl contents of the leaves (Chl a of the trimeric LHCII components (Lhcb1, Lhcb2 and Lhcb3 proteins) and for one of the monomeric light harvesting complexes of PSII, the Lhcb6 protein (CP24), as well as for three components of the PSI light harvesting antenna complexes (Lhca1, Lhca2 and Lhca3). In the case of the cab6 antisense ODN, of these complexes the most affected proteins were Lhcb3, Lhcb6, Lhca1 and Lhca2 (66-72% inhibition, Fig. 6). The cab10 antisense ODN inhibited the synthesis of Lhcb3 with a very high efficiency (~88% compared to the control), whereas the other proteins analyzed were less affected. The least affected proteins were Lhcb1, Lhcb2 and Lhca3 for both ODNs tested (Fig. 6).

Effects of antisense ODNs on psbA in Arabidopsis thaliana leaves
As a chloroplast encoded model gene we choose the psbA gene, which encodes the D1 protein

DISCUSSION
Key to identifying the functions of genes and for target validation is the ability to perturb the function of a particular gene in a given biological system. In plant science, several techniques have been applied for assigning the function to genes and proteins that include the creation of mutants, and the knockout or knockdown of certain genes, e.g. by antisense cloning.
In undetached tobacco leaves a clear effect of the pds ODNs could be observed (Fig. 2), though it was not as pronounced as during greening of etiolated wheat leaves. In tobacco leaves carotenoids were already present at the start of the treatment and, therefore, a smaller effect was expected than in the case of de-etiolation. It should be noted that the uptake of the ODNs was not continuous when they were introduced by infiltration into the leaves. Since antisense ODNs can inhibit only the de novo biosynthesis a larger effect is to be expected in leaves in which the pigment content is still increasing, preferably from a low level.
In the case of wheat leaves, the inhibitory effect of pds ODNs on the carotenoid content after 8 h of de-etiolation was striking (Fig. 3C). After 24 h of greening the effects were less pronounced but still noticeable (Fig. 4C), similar to the effects in juvenile tobacco leaves (Fig.   2C). The fact that the effects of antisense ODNs were less pronounced after 24 h of greening compared to 8 h greening is a consequence of the process studied. During greening the light harvesting complexes of PSII and PSI and their associated chlorophylls and carotenoids accumulate. The ODNs targeting the pds and cab genes slow this accumulation process down, but do not prevent it completely. Therefore, the difference between the control and the ODNtreated leaves will gradually disappear. Indeed, in the unreported experiment, in which the greening process was followed for 72 h, the difference disappeared completely.
The inhibitory effect of antisense ODNs on the cab proteins was strong despite the use of only the natural (phosphodiester) structure (Figs. 5, and 6). For example, in the case of the cab10 antisense ODN the protein level of Lhcb3 was inhibited by about 88%. Our results also show a strong differential effect on the different members of the CAB protein family (Fig. 6). This illustrates that it is possible to design ODNs targeting a gene family, which is a clear advantage over creating mutants. For wheat, presently, only one cab sequence is available in the database and it was, therefore, not possible to design antisense ODNs specific for a particular cab gene. With information on the whole wheat genome it would be possible to align the cab genes, which can then be used as a starting point for the design of ODNs that target one gene or several genes that show a high sequence homology. Nevertheless, our data show that already on the basis of a single gene sequence, ODNs could be designed that suppress the expression of several light harvesting proteins during de-etiolation of detached leaves. The third target gene was psbA, which is a chloroplast encoded gene. The inhibition of psbA by ODNs was as strong as the inhibition of the cab genes, suggesting that ODNs are as effective in chloroplasts as in the cytoplasm/nucleus. This experiment also shows that ODNs work as well in Arabidopsis thaliana as in wheat or tobacco.

ODN application and delivery
In the present study it is shown that ODNs dissolved in distilled water are efficiently taken up by and transported in plant leaves (Fig. 1). When testing the effect of different sugars we found no evidence for enhanced uptake (data not shown). This result contrasts with the observation of Sun et al. (2005Sun et al. ( , 2007 that the uptake by barley leaves of the SUSIBA2 transcriptional factor antisense ODNs was promoted by high concentrations of sucrose (200 mM). During our preliminary experiments, we found that wheat, the wheat spring cultivar CY-45 in particular, was sensitive to osmotic stress, which could already be observed in the presence of 200 mM sucrose. This prompted us to make an attempt on sugar-independent uptake of ODNs, which appeared to be an efficient method under our experimental conditions, using the simple techniques described in Materials and Methods.
The application method was adapted to the type of plant used. Tobacco leaves are too large for an efficient passive uptake of ODNs from a solution. In the literature it has been shown that fluorescent fusion proteins (Sparkes et al., 2006) and fluorescent sensors of reactive oxygen species (Hideg et al., 2002) can be injected into the leaf. We show here that this method also works for ODNs ( Fig.1 G-L). Arabidopsis leaves are too small for passive transport or injection, thus we used vacuum infiltration to treat detached leaves with ODNs.
A special feature of the application of ODNs is the timing. The expression of both the pds and the cab genes is under circadian control (http://diurnal.cgrb.oregonstate.edu). Though the circadian rhythm does not play a role yet in etiolated wheat leaves but for the experiments with juvenile tobacco leaves, the same daily schedule has been followed to avoid effects of the circadian rhythm.

Phosphorothioate modification of antisense ODNs
One way to increase the efficiency of the inhibition of gene expression by antisense ODNs is to increase the stability of ODNs by chemical modification. Phosphorothioate modification (Eckstein, 1985;Campbell et al., 1990;Agrawal et al., 1991;Ghosh et al., 1993 , 2000) was used in several studies to protect ODNs against exo-and endonucleases, of which the most active enzyme is the 3'-exonuclease (Uhlmann and Peyman 1990;Hoke et al., 1991;Shaw et al., 1991). Capping of the 3' end, or both the 3' and 5' ends by PS linkages protects the ODN against exonuclease degradation (Shaw et al., 1991;Gillardon et al., 1995), but these PS end-capped ODNs are still subject to endonuclease degradation. The combination of the end-capping technique and protection at internal pyrimidine residues, which are the major sites of endonuclease degradation, has been shown to be more efficient (Uhlmann et al., 2000;Samani et al., 2001).
Phosphorothioate-modified antisense ODNs are highly water-soluble (Levin, 1999), and can efficiently recruit RNase H to cleave the target RNA (Zamaratski et al., 2001). All these properties make PS ODNs superior to the natural (phosphodiester) ODNs in antisense activity (Ravichandran et al., 2004).
In our study we used a combination of end-capping and internal pyrimidine protection by PS linkages that indeed resulted in a significantly higher inhibition in the transcript level compared to the natural structure in the case of the pds gene (Figs. 2B, 3B, 4B). For tobacco leaves the influence of the PS modification could be detected in the case of the 8T-8N ODNs on the mRNA levels. A similar but smaller effect existed for the 9T-9N ODN pair (Fig. 2B).
Studying the wheat pds antisense ODNs with PS-modification the inhibition was about 15% higher compared to the natural structures (Fig. 3B). These data suggest that nucleases are active in plant cells and that PS modification can efficiently protect ODNs. Moutinho et al. (2001a, b) have already used ODNs having PS modifications in the three bases adjacent to each terminus in a pollen tube model system, but to our knowledge our study is the first that demonstrates that PS modification increases the efficacy of ODNs in leaves of higher plants.
New generations of modified ODNs structures, bearing e.g. a 3`-O substitution or locked nucleic acids (LNA), provide higher hybridization efficiencies and cellular stability. We, therefore, anticipate a further increase in the inhibitory efficiency in future studies (Kurreck, 2003).

Mechanism of the knock-down of gene expression by antisense ODNs
The first proven mechanism of antisense ODN inhibition was the so-called hybridizational arrest mechanism in which the ODN sterically blocks the ribosome translocation (Crooke, 1999;Kurreck, 2003). However, the predominant mechanism of inhibition of antisense ODNs For the three studied genes we observe that in all cases the inhibition at the mRNA level is higher than at the protein level. The smallest difference is observed for the psbA gene but this is probably due to the high turnover rate of the D1-protein. However, the complexity of the system that we study does not allow us to discriminate on this basis between the working mechanisms of ODNs that have been proposed in the literature (see above). A strategy that may be used to address the action mechanism of ODNs in plants is to use ODNs that have been modified in such a way that they are no longer a target for RNase H like 2' O-methyl RNA, locked nucleic acid (LNA) or morpholino phosphoroamidate based ODNs (Kurreck, 2003).

Conclusions and perspectives
In this paper we have shown that ODNs can be efficiently transported within leaves of vascular plants and that they reach the chloroplasts in monocotyledonous and dicotyledonous plant species. Antisense ODNs were used to knock down the expression of the nuclearencoded pds and cab genes and the chloroplast encoded psbA gene. We have also shown that the efficiency of the inhibition of a target gene, the pds gene, can be increased by PS modification of the antisense ODNs. We also show, for the first time, that ODNs are efficiently transported into the chloroplasts, and chloroplast-encoded genes can be targeted by antisense ODNs. Antisense ODNs can be designed and synthesized in a very short period of time, no selection markers are required, and handling does not require special care since plants treated with antisense ODNs are not genetically modified. With this approach, the function of essential genes can also be readily studied and fewer pleiotropic effects are expected to occur since the effects of the antisense ODNs develop on the timescale of hours to tens of hours. Although the effect of antisense ODNs is transient and full inhibition might be not possible to achieve, disadvantage in many applications, these features can also be www.plantphysiol.org on August 19, 2017 -Published by Downloaded from Copyright © 2011 American Society of Plant Biologists. All rights reserved. regarded essential in certain important applications. For instance, in order to enhance the hydrogen evolution in green algae Melis et al. (2000) used sulphur deprivation to transiently inhibit the de novo synthesis of the PSII reaction center complex protein D1. An additional advantage is that ODNs can be designed to target multiple genes at the same time. We propose that antisense ODN technology can be used as a high-throughput screening method for studies on the role of genes and proteins with yet unknown function. It is also conceivable that interaction of genes can be studied with the help of antisense ODNs since ODNs can be applied in combination. Photosynthesis is one of the research areas to which this method can be successfully applied. A special advantage of the study of photosynthetic genes is the availability of a range of sensitive, non-invasive techniques, such as time-resolved Chl a fluorescence and fluorescence lifetime measurements (Krumova et al., 2010), from which information can be derived on the operation of PSII and variations in the light-harvesting antenna system, as well as different time resolved absorption measurements at various wavelengths providing information on e.g. the operation of the PSI reaction center, the cytochrome b 6 /f complex or on the ability of the thylakoid membranes to generate and maintain a transmembrane electric field (Junge, 1977;Kramer et al., 2004). These techniques can be applied on small leaf areas and for some of these techniques even on individual cyanobacterial or algal cells and chloroplast elements as small as µm-sized voxels (volumetric pixels).

Design and synthesis of antisense ODNs
Oligodeoxynucleotide sequences were selected based on a search for freely available single stranded loops in the mRNA secondary structure (M. Zuker -Mfold web server for nucleic acid folding and hybridization prediction (http://mfold.rna.albany.edu/?q=mfold/RNA- -Form, Zuker, 2003 For uptake experiments, random non-sense ODNs were labelled with Fluorescein Cyanoethyl Phosphoramidite (6-FAM, Link Technologies; 494 nm excitation, 525 nm green emission) at 5'-position and purified by HPLC.

ODNs used
In the experiments a random nonsense control was used. A nucleotide blast search of the whole GenBank database with this sequence did not yield a single hit of a sequenced angiosperm gene.

19
As controls, a 17-mer random-nonsense PS 5'-G S G S C S GGC S TAAC S GC S TT S C S G S A-3' ODN was used.

Antisense ODN treatment of tobacco leaves
Tobacco plants (Nicotiana benthamiana) were grown in a greenhouse at 20-25°C.
Supplemental light was provided by metal halide lamps for 12 h a day and the light intensity at the plant surface was about 200 µmol photons m -2 s -1 . Expanding leaves of two months old plants were used for the experiments. Antisense ODNs and random nonsense ODNs were dissolved in sterile distilled water at a concentration of 10 µM and infiltrated in intact leaves with the help of a syringe, similarly as it was described for transient plasmid uptake (Sparkes et al., 2006). The ODN treatments were carried out between 8 and 10 am. After 24 h, fast Chl a fluorescence transients were measured and treated leaf discs were cut and stored at -80°C until further use. Beside the random-nonsense ODNs, we always used water controls as well, i.e. treating the leaves with distilled water.

Antisense ODN treatment of wheat leaves
Wheat seeds (Triticum aestivum, genotype CY-45) were sown in pots filled with one part sand and two parts soil. The pots were kept in a growth chamber in total darkness for 11 days at 19°C, 70% humidity. 10 cm long leaf segments of etiolated wheat seedlings were prepared.
The leaves were cut under water to avoid the formation of air bubbles and the lower 1-2 cm part was submerged in 10 µM water solution of pds and cab antisense ODNs, randomnonsense ODNs and pure water in a 1.5 ml Eppendorf tube. The tube was sealed with Parafilm in order to avoid evaporation of solution. After 12 h of incubation in the dark, leaves were illuminated at 100 µmol photons m -2 s -1 for 24 h. Based on preliminary experiments using fast Chl a fluorescence transients (see below), we chose the 8 h and 24 h illumination times, after which fluorescence measurements were performed and leaf segments were harvested and stored at -80°C until further use. Four independent repetitions were made for each ODN. were measured and leaf discs were cut and stored at -80°C until further use. Eight independent repetitions were made for each ODN.
Wheat leaves photographs were taken with Olympus Camedia C7070 wide zoom digital camera.

Fast Chl a fluorescence (OJIP) measurements
Fluorescence measurements were carried out at room temperature with a Handy-PEA instrument (Hansatech Instruments Ltd, UK). Leaf samples were illuminated with continuous red light emitted by three LEDs (3,500 µmol photons m -2 s -1 , 650 nm peak wavelength; the spectral half-width was 22 nm; the emitted light is cut off at 700 nm by a near-infrared short-pass filter). The first reliably measured point of the fluorescence transient is at 20 µs, which was taken as F 0 . The length of the measurements was 1 s. To study the effects of antisense ODNs on the photosynthetic machinery, the maximum fluorescence (F M ) and the variable fluorescence (F V = F M -F 0 ) were used.

Quantitative real-time reverse transcription-PCR (QRT-PCR)
Total RNA was isolated from wheat, tobacco and Arabidopsis leaves using TRI reagent

Carotenoid and chlorophyll content determination
Analysis of total carotenoid in plant tissue was performed in N,N-dimethylformamide (DMF) (Inskeep et al., 1985). The OD-values were measured at 663, 646 and 470 nm using a