Journal list menu

Volume 580, Issue 2 p. 597-602
Short communication
Free Access

Calcium ions are involved in the delay of plant cell cycle progression by abiotic stresses

Toshio Sano

Corresponding Author

Toshio Sano

Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs Platz 2, 97082 Würzburg, Germany

Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Chiba 277-8562, Japan

Corresponding author. Fax: +81 471 36 3706.Search for more papers by this author
Takumi Higaki

Takumi Higaki

Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Chiba 277-8562, Japan

Search for more papers by this author
Koichi Handa

Koichi Handa

Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Chiba 277-8562, Japan

Search for more papers by this author
Yasuhiro Kadota

Yasuhiro Kadota

Department of Applied Biological Science, Tokyo University of Science, Yamazaki, Noda, Chiba 278-8510, Japan

Search for more papers by this author
Kazuyuki Kuchitsu

Kazuyuki Kuchitsu

Department of Applied Biological Science, Tokyo University of Science, Yamazaki, Noda, Chiba 278-8510, Japan

Search for more papers by this author
Seiichiro Hasezawa

Seiichiro Hasezawa

Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Chiba 277-8562, Japan

Search for more papers by this author
Anja Hoffmann

Anja Hoffmann

Institut für Pharmazeutische Biology, Universität Würzburg, Julius-von-Sachs Platz 4, 97082 Würzburg, Germany

Search for more papers by this author
Jörg Endter

Jörg Endter

Lehrstuhl für Biotechnologie, Universität Würzburg, Am Hubland, Würzburg, Germany

Search for more papers by this author
Ulrich Zimmermann

Ulrich Zimmermann

Lehrstuhl für Biotechnologie, Universität Würzburg, Am Hubland, Würzburg, Germany

Search for more papers by this author
Rainer Hedrich

Rainer Hedrich

Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs Platz 2, 97082 Würzburg, Germany

Search for more papers by this author
Thomas Roitsch

Thomas Roitsch

Institut für Pharmazeutische Biology, Universität Würzburg, Julius-von-Sachs Platz 4, 97082 Würzburg, Germany

Search for more papers by this author
First published: 04 January 2006
Citations: 25

Abstract

Higher plants respond to environmental stresses by a sequence of reactions which include the reduction of growth by affecting cell division. It has been shown that calcium ions plays a role as a second messenger in mediating various defence responses under environmental stresses. In this study, the role of calcium ions on cell cycle progression under abiotic stresses has been examined in tobacco BY-2 suspension culture cells. Using synchronized BY-2 cells expressing the endogenous calcium sensor aequorin as experimental system, we could show that oxidative and hypoosmotic stress both induce an increase of intracellular calcium and cause a delay of the cell cycle. The inhibitory effect of these abiotic stress stimuli on cell cycle progression could be mimicked by increasing the intracellular calcium concentration via application of an external electrical field. Likewise, depletion of calcium ions in the culture medium suppressed the effect of the stimuli tested. These results demonstrate that calcium signalling is involved in the regulation of cell cycle progression in response to abiotic stress.

1 Introduction

Plants are exposed to a variety of abiotic and biotic environmental stresses, such as cold, draught, high light and pathogen infection. To cope with the various stimuli a cascade of reactions is initiated that includes immediate posttranslational responses, gene regulations as well as long term adaptations. One of the adaptive responses is the reduction of growth by repression of cell division [1] to preserve the limited energy to the mother cell and to avoid heritable damage.

The analyses of signal transduction cascades involved in environmental stress responses of plants has revealed that the increase in cytoplasmic calcium concentration ([Ca2+]cyt) serves as second messenger to mediate defense responses [2]. For example, oxidative stress induces a transient increase of the [Ca2+]cyt lasting for one to two min [3, 4] prior to the induction of defense responses. During plant cell cycle progression, oxidative stress is known to impair G1/S transition to slow down DNA replication and to delay the initiation of mitosis [5]. Treatment with the proteinaceous elicitor cryptogein also induced [Ca2+]cyt increase [6] and caused cell cycle arrest in G1 and G2 phase before the induction of cell death [7]. However, the relation of the [Ca2+]cyt increase induced by stresses and the reduction of cell division is not yet understood. For plant cell cycle studies, the tobacco suspension culture line BY-2 cell has proven to be most suitable because this cell line grows fastest among the known plant cell culture lines and can be highly synchronized using the DNA polymerase α-inhibitor aphidicolin [8].

In the present study, we examined the effect of two abiotic stress stimuli, oxidative and hypoosmotic stress, respectively, on cell cycle progression of the tobacco BY-2 suspension culture line as a model system. To generate oxidative stress, we used potassium permanganate (KMnO4). It is a powerful oxidation agent that selectively reacts with the double bond of pyrimidine bases of DNA and results in DNA cleavage [9]. This chemical is widely used in conformational studies of DNA. In addition, the permanganate oxidation reaction on DNA can be carried out under in vivo conditions without interference of other metabolites [10]. Both abiotic stress stimuli were shown to delay cell cycle progression and to elicit transient [Ca2+]cyt increases. The role of calcium in controlling the stress dependant cell cycle inhibition was substantiated by the findings that voltage mediated increase of the [Ca2+]cyt was able to mimic the inhibition by the stress stimuli while EGTA mediated depletion of calcium in the medium suppressed their effects on cell cycle progression.

2 Materials and methods

2.1 Plant materials and culture conditions

A transgenic tobacco BY-2 cell line (Nicotiana tabacum L. cv. Bright Yellow 2) that constitutively expresses apoaequorin protein [11] was maintained by weekly subculture in a modified Linsmaier and Skoog medium supplemented with 2,4-D (LSD medium), in which KH2PO4 and thiamine HCl were increased to 370 and 1 mg/l, respectively. To this basal medium, sucrose and 2,4-D were supplemented to 3% and 0.2 mg/l, respectively, and the pH was adjusted to 5.8 before autoclaving [8]. The cell suspension was cultured on a rotary shaker at 130 rpm and 25 °C in the dark.

Cell synchrony was induced by treatment with 5 μg/l aphidicolin (Sigma Chemical Co., St. Louis, MO, USA) essentially as described in [8]. During the aphidicolin treatment, the cell culture was incubated with 1 μM coelenterazine (Dojindo, Kumamoto, Japan) in darkness for 16 h to reconstitute functional aequorin, as a calcium-sensitive photoprotein. After 24 h of aphidicolin treatment, the cell culture was washed with 1 l of a medium containing 3% sucrose, 1.5 mM CaCl2, 0.1% MES and 0.2 mg/l 2,4-D (pH 5.8) on a glass filter and further incubated in this medium. The cell culture was divided into 2–4 portions and the application of the stimuli, Ca2+ luminescence measurement, and cell cycle analysis were carried out as described below.

2.2 Cell cycle and cell death analysis

The mitotic index (MI) was determined with a fluorescence microscope after staining the nuclei with 1 μM of SYTOX (Molecular Probes Inc., Eugene, OR, USA).

For flowcytometric analysis, cells were fixed with 100% ethanol. After rehydration in Galbraith's buffer (45 mM MgCl2, 30 mM Na-citrate, 20 mM MOPS, and 1 g/l Triton X-100, pH 7.0) [12], cells were treated with 20 μg/l RNase A (Sigma) and 10 μg/l propidium iodide (Sigma) for 1 h at room temperature. Cytometrical analysis was performed on 5 × 103 cells with a laser scanning cytometer (LSC101, Olympus, Tokyo, Japan).

Cell death was determined after by staining the cells with 0.05% Evans Blue (Sigma) as described in [7].

2.3 Application of abiotic stimuli and measurement of Ca2+

Oxidative stress was applied by addition of KMnO4 to a final concentration of 100 μM into the medium. Hypoosmotic stress was applied by addition of one volume of water to the culture medium. To apply an electrical field to suspension culture cells, they were transferred into an electroporation cuvettes (1 mm gap) positioned in a luminometer (Sirius FB12, Berthold, Bad Wildbad, Germany). Voltage was applied by a pulse generator (Type-214B, Hewlett–Packard) for 30 s; a current of 10 mA was observed at 80 V. For EGTA pretreatment, cells were washed with and resuspended in medium without Ca2+. The total amount of reconstructed aequorin was determined by measurement of luminescence from the cells after addition of 20% of 1 M CaCl2 in 20% ethanol and every measured aequorin luminescence was normalized in relation to the total aequorin luminescences in the corresponding samples.

3 Results

3.1 Oxidative stress mediated delay of cell cycle progression

Among the abiotic stimuli, we first examined the effect of oxidative stress on the cell cycle progression of tobacco BY-2 cells. Aphidicolin treatment arrested the cell cycle at the G1/S boundary and aphidicolin removal re-started the synchronized cell cycles from the S phase [8]. When we monitored the cell cycle progression by determination of the MI after fluorescent staining of the nuclei, the MI started to increase about 6 h after aphidicolin removal, reached the maximum level at 8 h and then decreased showing the cell cycle progression from mitosis to G1 phase (Fig. 1 A). When we applied KMnO4 (100 μM) as oxidative stress 1 h after aphidicolin removal, almost all of the cells were arrested and did not enter mitosis (Fig. 1A).

figure image
Effect of oxidative stress on cell cycle progression and [Ca2+]cyt. (A) Cell cycle progression was monitored by mitotic index after aphidicolin removal in control cells (closed circles) and cells treated with KMnO4 (100 μM, closed triangles) and cells pretreated with EGTA (100 μM) for 5 min and treated with KMnO4 (100 μM, open triangles) 1 h after aphidicolin removal. Data show a representative from three independent experiments. (B) Average traces of [Ca2+]cyt increase in KMnO4 treated cells (closed triangles) and cells pretreated with EGTA (100 μM) for 5 min and treated with KMnO4 (100 μM, open triangles). Vertical bars represent S.E. (n = 3).

Since oxidative stress has been shown to induce an increase of [Ca2+]cyt [3], the effect of KMnO4 on [Ca2+]cyt was measured using aequorin-transformed BY-2 cells. Fig. 1B shows a fast, transient and sharp increase of the [Ca2+]cyt 20–30 s after KMnO4 application that declined very rapidly. To ensure the specificity of this measurement, wild type BY-2 cells not transformed with aequorin were reconstituted with coelenterazine in the same way as the transgenic culture. This control experiment revealed that no comparable light emission could be observed (data not shown). To further determine the relation between the observed transient increase of the [Ca2+]cyt and plant cell cycle progression, we depleted Ca2+ in the culture medium and chelated the remained calcium ions with EGTA. After pretreatment of the suspension with EGTA (100 μM), the KMnO4-induced [Ca2+]cyt increase was repressed to about 40% of the control culture value (Fig. 1B). In addition, by pretreatment with EGTA, the cells entered mitosis even upon the KMnO4 treatment (Fig. 1A), which was confirmed by the flowcytometric analysis (Fig. 2 A). Addition of 100 μM EGTA to the control medium only marginally reduced the maximum value of MI obtained 8 h after aphidicolin (see Fig. 4A), supporting the specificity of the KMnO4 effect.

figure image
Effect of oxidative stress on cell cycle progression and cell death. (A) Cell cycle progression after oxidative stress monitored by flowcytometry. DNA histograms were determined by LSC analysis at the indicated times after aphidicolin removal in control cells (control) and cells treated with KMnO4 (100 μM, +KMnO4) and cells pretreated with EGTA (100 μM) for 5 min and treated with KMnO4 (100 μM, EGTA + KMnO4) 1 h after aphidicolin removal. The relative DNA content of 2c and 4c were determined by comparing the value of 7 d-old stationary phase cells (not shown). Data show a representative result of three independent experiments. (B) Cell death 24 h after oxidative stress was determined by Evans Blue assay. Vertical bars represent S.E. (n = 3).
figure image
Effect of hypoosmotic stress on cell cycle progression and [Ca2+]cyt. (A) Cell cycle progression was monitored after aphidicolin removal in control cells (closed circles), cells treated with EGTA (1 mM, open circles), hypoosmotic stress for 10 min (closed diamonds) and cells pretreated with EGTA (1 mM) for 5 min and treated with hypoosmotic stress for 10 min (open diamonds) 1 h after aphidicolin removal. EGTA was removed after application of the hypoosmotic stress by exchange the medium to the LSD medium. (B) Average traces of [Ca2+]cyt increase in hypoosmotic stress treated cells (closed diamonds) and cells pretreated with EGTA (1 mM) for 5 min and treated with hypoosmotic stress (open diamonds). Vertical bars represent S.E. (n = 3).

Since the KMnO4-induced oxidative stress inhibited the cells to enter mitosis, the mode of cell cycle progression was monitored by flowcytometry. In the control condition, the cell cycle progressed from S to G2 phase 2 h after aphidicolin removal and mitosis at 8 h (Fig. 2A). In the KMnO4-treated cells, the cell cycle stage was still in S phase 2 h after aphidicolin removal and did not progress further. Determination of cell death by Evans Blue staining revealed a higher proportion of dead cells in the KMnO4-treated samples compared to the control incubations. In the EGTA-pretreated cells, cell cycle arrest by oxidative stress was suppressed and the proportion of the dead cells decreased (Fig. 2A and B).

3.2 Voltage mediated calcium increase delayed cell cycle progression

Oxidative stress is known to elicit a variety of defense responses of plants such as H2O2 production, activation of MAP kinases, induction of defense-responsive genes and cell death [13]. To dissect the effect of oxidative stress mediated [Ca2+]cyt increase on cell cycle repression from calcium independent responses, the [Ca2+]cyt has been modulated by voltage treatment. It has been shown previously that it is possible to mimic calcium dependant processes in tomato suspension culture cells, such as MAP-kinase activation and defense gene induction, by the application of a voltage pulse and that voltage treatment acts via activation of voltage-gated calcium permeable channels [14]. When we applied a voltage pulse to the aequorin-transformed BY-2 cells, an immediate, transient increase of the [Ca2+]cyt was observed (Fig. 3 B). Since the voltage stimulus can be handled periodically and quantitatively, we applied a voltage of 80 V for 30 s to mimic the KMnO4-induced increase of [Ca2+]cyt in shape and size. Application of this voltage pulse to synchronized BY-2 cells in the S phase delayed the cell cycle progression by 1 h without a decrease of the MI value (Fig. 3A). Chelating calcium ions in the medium by EGTA pretreatment (100 μM) reduced the voltage-dependant increase of [Ca2+]cyt to about 35% of the control culture (Fig. 3B) and suppressed the delay of the start of mitosis (Fig. 3A). These data further support the conclusion that the oxidative stress induced delay of cell cycle progression is mediated by a transient [Ca2+]cyt increase.

figure image
Effect of voltage stimulus on cell cycle progression and [Ca2+]cyt. (A) Cell cycle progression was monitored after aphidicolin removal in control cells (closed circles), cells applied with voltage stimulus (closed squares) and cells pretreated with EGTA (100 μM) for 5 min and applied with voltage stimulus (open squares) 1 h after aphidicolin removal. The data shows a representative from three independent experiments. (B) Average traces of [Ca2+]cyt increase in voltage applied cells (closed squares) and cells pretreated with EGTA (100 μM) for 5 min and applied voltage stimulus (open squares). Voltage stimulus was applied at 80 V for 30 s. Vertical bars represent S.E. (n = 3).

3.3 Hypoosmotic stress mediated delay of cell cycle progression

Hypoosmotic stress reversibly increases cell turgor pressure and cell volume [15]. This mechanical stress is assumed to induce [Ca2+]cyt via slow-anion channels [16], that have been suggested to be involved in sensing the degree of hypoosmolarity [17].

Hypoosmotic stress was tested as a second abiotic stress stimulus which was imposed by diluting the incubation medium with water (1:1) and thus decreasing the osmolarity from 220 to 110 mosmol. When we examined the effect of hypoosmotic stress on cell cycle progression, treatment for 10 min delayed the start of mitosis by 1 h (Fig. 4 A). The hypoosmotic stress also induced a [Ca2+]cyt increase in the BY-2 cells, but the signature of the [Ca2+]cyt increase was different from that of the oxidative stress described above. [Ca2+]cyt started to increase only after 1 min after the application of the stress and lasted for about 3 min (Fig. 4B). The peak value of the [Ca2+]cyt increase seemed to reduce only about 10% compared to the treatment with KMnO4.

Upon hypoosmotic stress, EGTA treatment was also able to suppress the [Ca2+]cyt increase to 25% of the control value (Fig. 4B), whereas 1 mM was required and 100 μM EGTA was shown not to be sufficient (data not shown). Likewise, the EGTA pretreatment also suppressed the delay of the start of mitosis upon the hypoosmotic stress (Fig. 4A).

Since the effect of hypoosmotic stress can be easily and completely elevated by an additional exchange of the medium by standard medium, we examined the effect of the duration of treatment by this stress on cell cycle progression. Hypoosmotic treatment for 10 min delayed the cell cycle (Fig. 5 A). Continuous presence of this stress also delayed the cell cycle by the same extent as the 10 min treatment. In contrast, the treatment for a short period of 30 s had no substantial effect on cell cycle progression (Fig. 5A). When we measured the [Ca2+]cyt increase upon the treatment for 30 s, also no significant [Ca2+]cyt increase was observed (Fig. 5B), indicating the requirement for a minimum period of stress perception and further supporting the correlation between stress dependant changes in [Ca2+]cyt and cell cycle progression.

figure image
Effect of the duration of hypoosmotic stress treatment on cell cycle progression and [Ca2+]cyt. (A) Cell cycle progression was monitored in control cells (closed circles), cells treated with hypoosmotic stress for 30 s (open circles), for 10 min (closed diamonds) and continuously (open diamonds) 1 h after aphidicolin removal. The data shows a representative from three independent experiments. (B) Average traces of [Ca2+]cyt increase in cells treated with hypoosmotic stress continuously (closed diamonds) and for 30 s (open circles). Hypoosmotic stress was removed by exchange the medium to the LSD medium. In the latter case, [Ca2+]cyt measurement was started after removal of the hypoosmotic stress. The vertical bar represents S.E. (n = 3).

Cell cycle progression upon hypoosmotic treatment was further monitored by flowcytometry. After aphidicolin removal, cell cycle progressed from S to G2 phase at 2 h, stayed G2 phase by 6 h and entered mitosis in control cells. In the hypoosmotic-stress treated cells, the cell cycle progressed to G2 phase like the control cells but entering or progression of mitosis was delayed since reappearance of the G1 phase cells was delayed (Fig. 6 , 9 h). The cell cycle of the cells treated by the hypoosmotic stress only for 30 s progressed normally (Fig. 6). Upon hypoosomotic treatment, the proportion of dead cells did not increase significantly irrespective to the duration of the treatment (data not shown).

figure image
Cell cycle progression after hypoosmotic stress monitored by flowcytometry. DNA histograms were determined by LSC analysis at the indicated times after aphidicolin removal in control cells (control) and cells treated with hypoosmotic stress for 30 s (HypoO 30 s), for 10 min (HypoO 10 min) and continuously (HypoO Continu) 1 h after aphidicolin removal. The relative DNA content of 2c and 4c were determined by comparing the value of 7 d-old stationary phase cells (not shown). Data show a representative result of three independent experiments.

4 Discussion

4.1 Abiotic stress delayed cell cycle progression in a Ca2+-dependent manner

In this study, we showed that two abiotic stress stimuli, oxidative and hypoosmotic stress, delayed the cell cycle progression of tobacco BY-2 cells. Cell cycle arrest in S phase and induction of cell death by KMnO4 application (Fig. 2) supposed DNA damage by its oxidation in vivo. The negative effect of oxidative stress induced by a quinone, 2-methyl-1,4-naphthoquinone (menadione), on plant cell cycle progression [5] is supporting the results of the present study. Hypoosmotic stress was previously shown to cause an increase in cell turgor pressure and subsequently cell volume in a brackish water charaphyte [15]. In this study, we showed hypoosmotic stress delayed the cell cycle progression from G2 phase to mitosis in tobacco BY-2 cells but did not induce cell death (Fig. 6).

The two abiotic stresses applied to aequorin-transformed tobacco BY-2 cells induced a transient [Ca2+]cyt increase. This finding is supported by related observations in other systems [3, 17]. In addition, reduction of the [Ca2+]cyt increase suppressed the cell cycle delay even upon these stress application. Thus, oxidative and hypoosmotic stress apparently delayed the cell cycle in a Ca2+-dependent manner.

It was possible to mimic the KMnO4-induced calcium signature by applying a defined electrical field. The finding that the oxidative stress-independent modulation of [Ca2+]cyt by voltage application also resulted in a delay of cell cycle progression supports the role of calcium as second messenger in regulating this process. However, the effect of the oxidative stress stimulus on BY-2 cell cycle delay was stronger than the effect of the voltage stimulus although the level of the [Ca2+]cyt increase was almost the same. The voltage stimulus is applied only as short, single transient [14], while oxidative stress persisted causing a number of long term defense responses [13]. Thus, the more dramatic effect of the oxidative stress could be due to long term stress effects while voltage stimulation seems to avoid secondary effects. In a previous study it has been shown for tomato suspension cultured cells that [Ca2+]cyt mediated processes such as MAP kinase activation and defense gene activation could be elicited by the application of a voltage pulse [14]. The data obtained with the BY-2 suspension culture cells support an involvement of de- or hyper-polarization activated voltage-gated calcium channels.

Hypoosmotic stress also caused cell cycle delay and induced [Ca2+]cyt increase in tobacco BY-2 cells. However, the mode of the [Ca2+]cyt increase was different from that of the oxidative stress. The oxidative stress was suggested to activate voltage-dependent Ca2+-permeable channels [18, 19] while the hypoosmotic stress is supposed to activate voltage-independent Ca2+-permeable channels [6, 19]. However, since both the oxidative stress and the hypoosmotic stress delayed the cell cycle progression, the [Ca2+]cyt increase seems to be the trigger regardless of the nature of the Ca2+-permeable channel involved.

Both continuous hypoosmotic stress and a single, transient for only 10 min caused cell cycle delay. This suggests that the stimuli that delay cell cycle are not necessary to persist in the culture. This finding is supported by voltage stimulated [Ca2+]cyt increase. This stimulus, that was also transiently applied and did not persist in the medium, also delayed the cell cycle. In contrast, the finding that the application of hypoosmotic stress for a very short period (30 s) neither was sufficient to elicit a significant [Ca2+]cyt nor to exert an effect on cell cycle progression further supports the conclusion that a pronounced [Ca2+]cyt increase is involved in the stress dependant cell cycle delay.

Preliminary experimental evidence suggest that the degree of [Ca2+]cyt increase in response to both oxidative and hypoosmotic stress and voltage application is dependent on the stage of the cell cycle. Whereas the magnitude of the response was higher in the G1 and G2 phase, the cells responds less sensitive in the S and M phase (Sano and Roitsch, unpublished observations). Such a cell cycle dependent response is supported by the finding that in tobacco BY-2 cells, cryptogein treatment in G1 and S phase induced a higher level of [Ca2+]cyt increase and cell death than in G2 and M phase [7] and the observation that the G1 phase is more sensitive than the S phase to oxidative stress induced cell death [5]. However, the significance and mechanism of the cell cycle dependency of the stress mediated response remains to be further elucidated. The [Ca2+]cyt of plant cells increases in response to a variety of environmental challenges [2]. In tobacco BY-2 cells, oxidative stress in S phase arrested the cell cycle in S phase (Fig. 2A) which was suggested to result in DNA damage or a delay in DNA replication [5]. Cryptogein treatment in S phase induced [Ca2+]cyt increase, arrested the cell cycle in G2 phase and suppressed gene expression of cyclin A and B [7]. Therefore, the [Ca2+]cyt increase in S phase may be a signal that indicates the DNA damage and regulates cell cycle progression by modulating cyclin expression. In animal cell systems, cell cycle progression is known to be accompanied by a transient increase of [Ca2+]cyt [20]. Thus the transient increase of [Ca2+]cyt seems to have opposite roles in plant and animal cell cycle progression.

Acknowledgements

We are grateful to Dr. T. Furuichi of Nagoya University for a gift of aequorin-transformed tobacco BY-2 cell line, Prof. A.J. Trewavas of University of Edinburgh for permission to use aequorin cDNA-containing materials, Dr. V. Soukhoroukov for advice with the current application, Dr. Dirk Becker for critical reading of the manuscript, and Mr. Walid Waheed Ali for technical assistance. The authors thank for the financial support by the Alexander-von-Humboldt foundation to T.S. and DFG to R.H.