Mechanism for deactivation of Rubisco under moderate heat stress

Photosynthesis is particularly sensitive to direct inhibition by heat stress. This inhibition is closely associated with the inactivation of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco). To develop a more complete understanding of the mechanism of inactivation of Rubisco under moderate heat stress, various aspects of the process were examined both in vivo and in vitro. Experiments with isolated Rubisco revealed that the rate of synthesis of the catalytic misfire product, xylulose-1,5-bisphosphate, increased with temperature. Activated Rubisco, produced by reaction with activase at a control temperature of 25°C or by incubation with high CO₂, deactivated when the temperature of the reaction exceeded temperatures that were equivalent to the optimum for activase adenosine triphosphatase (ATPase) activity. Measurements of the activation state of Rubisco in cotton and tobacco leaves showed that Rubisco inactivated within 7 s of imposing a heat stress. Thus, elevated temperature had an opposite effect on the two processes that ultimately determine the activation state of Rubisco, decreasing activase activity but stimulating the catalytic misfire reaction that inactivates Rubisco. These data support a mechanism for the inactivation of Rubisco at high temperature involving an inability of activase to overcome the inherently faster rates of Rubisco inactivation. That the net effect of elevated temperatures on Rubisco activation is similar both in vivo and under controlled conditions in vitro argues for a direct effect of temperature on the activation of Rubisco by activase and against the proposal that the deactivation of Rubisco under moderate heat stress is a secondary consequence of perturbations in the thylakoid membrane.     

Increased heat sensitivity of photosynthesis in tobacco plants with reduced Rubisco activase

High temperature inhibits photosynthesis by several mechanisms including deactivation of Rubisco. The inhibition of photosynthesis by high temperature and its relationship to Rubisco deactivation was studied using tobacco (Nicotiana tabaccum L. cv W38) transformed with a Rubisco activase gene inserted in the antisense orientation and untransformed controls. High temperature (42 °C) reduced photosynthesis in both lines of plants. However, photosynthesis recovered nearly completely in wild-type plants and very little in plants lacking Rubisco activase. The F₀ level of chlorophyll fluorescence decreased and q_N increased in the control plants during heating. In the antisense plants, q_N was always high and F₀ increased slightly during heat stress. NADP-malate dehydrogenase activation was unaffected by heat stress in control plants but was increased in the transgenic plants, consistent with a high redox status in the chloroplast. In wild-type plants, the inhibition of photosynthesis could be explained by a reversible decarbamylation of Rubisco and an acceptor-side limitation imposed on photosynthetic electron transport. However, in the anti-activase plants, carbamylation was low and constant and could not explain how photosynthesis was reduced at high temperature. Because ribulose bisphosphate was saturating at high temperature, the reduction in photosynthesis must have been caused by some impairment of Rubisco function not reflected in measurements of activation state or carbamylation status. This in vivo Rubisco impairment was not relieved upon return to lower temperature. We speculate that the reversible decarbamylation of Rubisco at moderately high temperature may be a protective mechanism by which the plant avoids more serious effects on Rubisco and the rest of the photosynthetic apparatus.This revised version was published online in October 2005 with corrections to the Cover Date.     

Temperature dependence of photosynthesis in Arabidopsis plants with modifications in Rubisco activase and membrane fluidity

Net photosynthesis (Pn) is reversibly inhibited at moderately high temperature. To investigate this further, we examined the effects of heat stress on Arabidopsis plants in which Rubisco activase or thylakoid membrane fluidity has been modified. During heating leaves from 25 to 40 degrees C at 250 ppm CO2 and 1% O2, the wild-type (WT), plants expressing the 43 kDa isoform only (rwt43), and plants accumulating activase 40% of WT (R100) exhibited similar inhibitions in the Pn and Rubisco activation state. Despite better membrane integrity than WT, plants having less polyunsaturation of thylakoid lipids (fad7/8 double mutant) failed to maintain greater Pn than the WT. Plants expressing the 46 kDa isoform only (rwt46) exhibited the most inhibition, but plants expressing a 46 kDa isoform incapable of redox regulation (C411A) were similar to the WT. The null mutant (rca) exhibited a continuous decline in Pn. As measured by fluorescence, electron transport activity decreased concomitantly with Pn but PSII was not damaged. Following a quick recovery to 25 from 40 degrees C, whereas most lines recovered 90% Pn, the rwt46 and rca lines recovered only to 59 and <10%, respectively. As measured by NADP-malate dehydrogenase activation, after an initial increase at 30 degrees C, stromal oxidation in the WT and rwt46 plants did not increase further as Pn decreased. These results provide additional insight into the role of Rubisco activation and activase in the reversible heat inhibition of Pn.     

Association of Rubisco activase with chaperonin-60beta: a possible mechanism for protecting photosynthesis during heat stress

Previous studies have shown that inhibition of photosynthesis by moderate heat stress is a consequence of Rubisco deactivation, caused in part by the thermal instability of Rubisco activase. This involvement of Rubisco activase was confirmed in heat stress and recovery experiments using transgenic Arabidopsis plants. Compared with wild-type plants, photosynthesis, the effective quantum yield of photosystem II, and Rubisco activation were less thermotolerant and recovered more slowly in transgenic Arabidopsis plants with reduced levels of Rubisco activase. Immunoblots showed that 65% of the Rubisco activase was recovered in the insoluble fraction after heat stress in leaf extracts of transgenic but not wild-type plants, evidence that deactivation of Rubisco was a consequence of thermal denaturation of Rubisco activase. The transgenic Arabidopsis plants used in this study contained a modified form of Rubisco activase that facilitated affinity purification of Rubisco activase and proteins that potentially interact with Rubisco activase during heat stress. Sequence analysis and immunoblotting identified the beta-subunit of chaperonin-60 (cpn60beta), the chloroplast GroEL homologue, as a protein that was bound to Rubisco activase from leaf extracts prepared from heat-stressed, but not control plants. Analysis of the proteins by non-denaturing gel electrophoresis showed that cpn60beta was associated with Rubisco activase in a high molecular mass complex. Immunoblot analysis established that the apparent association of cpn60beta with Rubisco activase was dynamic, increasing with the duration and intensity of the heat stress and decreasing following recovery. Taken together, these data suggest that cpn60beta plays a role in acclimating photosynthesis to heat stress, possibly by protecting Rubisco activase from thermal denaturation.     

Phosphorylation pattern of Rubisco activase in Arabidopsis leaves

Rubisco activase (RCA) is an ancillary photosynthetic protein essential for Rubisco activity. Some data suggest that post‐translational modifications (such as reduction of disulphide bridges) are involved in the regulation of RCA activity. However, despite the key role of protein phosphorylation in general metabolic regulation, RCA phosphorylation has not been well characterised. We took advantage of phosphoproteomics and gas exchange analyses with instant sampling adapted to Arabidopsis rosettes to examine the occurrence and variations of phosphopeptides associated with RCA in different photosynthetic contexts (CO₂ mole fraction, light and dark). We detected two phosphopeptides from RCA corresponding to residues Thr 78 and Ser 172, and show that the former is considerably more phosphorylated in the dark than in the light, while the latter show no light/dark pattern. The CO₂ mole fraction did not influence phosphorylation of either residue. Phosphorylation thus appears to be a potential mechanism associated with RCA dark inactivation, when Rubisco‐catalysed carboxylation is arrested. Since Thr 78 and Ser 172 are located in the N and Walker domains of the protein, respectively, the involvement of phosphorylation in protein–protein interaction and catalysis is likely.     

Co-overproducing Rubisco and Rubisco activase enhances photosynthesis in the optimal temperature range in rice

Rubisco limits C₃ photosynthesis under some conditions and is therefore a potential target for improving photosynthetic efficiency. The overproduction of Rubisco is often accompanied by a decline in Rubisco activation, and the protein ratio of Rubisco activase (RCA) to Rubisco (RCA/Rubisco) greatly decreases in Rubisco-overproducing plants (RBCS-ox). Here, we produced transgenic rice (Oryza sativa) plants co-overproducing both Rubisco and RCA (RBCS-RCA-ox). Rubisco content in RBCS-RCA-ox plants increased by 23%–44%, and RCA/Rubisco levels were similar or higher than those of wild-type plants. However, although the activation state of Rubisco in RBCS-RCA-ox plants was enhanced, the rates of CO₂ assimilation at 25°C in RBCS-RCA-ox plants did not differ from that of wild-type plants. Alternatively, at a moderately high temperature (optimal range of 32°C–36°C), the rates of CO₂ assimilation in RBCS-ox and RBCS-RCA-ox plants were higher than in wild-type plants under conditions equal to or lower than current atmospheric CO₂ levels. The activation state of Rubisco in RBCS-RCA-ox remained higher than that of RBCS-ox plants, and activated Rubisco content in RCA overproducing, RBCS-ox, RBCS-RCA-ox, and wild-type plants was highly correlated with the initial slope of CO₂ assimilation against intercellular CO₂ pressures (A:Ci) at 36°C. Thus, a simultaneous increase in Rubisco and RCA contents leads to enhanced photosynthesis within the optimal temperature range.

A simultaneous increase in Rubisco and RCA contents in transgenic rice leads to an enhancement of photosynthesis at moderately high temperatures within the optimal temperature range.     

Directing the evolution of Rubisco and Rubisco activase: first impressions of a new tool for photosynthesis research

During the last decade the practice of laboratory-directed protein evolution has become firmly established as a versatile tool in biochemical research by enabling molecular evolution toward desirable phenotypes or detection of novel structure-function interactions. Applications of this technique in the field of photosynthesis research are still in their infancy, but recently first steps have been reported in the directed evolution of the CO(2)-fixing enzyme Rubisco and its helper protein Rubisco activase. Here we summarize directed protein evolution strategies and review the progressive advances that have been made to develop and apply suitable selection systems for screening mutant forms of these enzymes that improve the fitness of the host organism. The goal of increasing photosynthetic efficiency of plants by improving the kinetics of Rubisco has been a long-term goal scoring modest successes. We discuss how directed evolution methodologies may one day be able to circumvent the problems encountered during this venture.     

Regulation of Rubisco activase and its interaction with Rubisco

The large, alpha-isoform of Rubisco activase confers redox regulation of the ATP/ADP response of the ATP hydrolysis and Rubisco activation activities of the multimeric activase holoenzyme complex. The alpha-isoform has a C-terminal extension that contains the redox-sensitive cysteine residues and is characterized by a high content of acidic residues. Cross-linking and site-directed mutagenesis studies of the C-terminal extension that have provided new insights into the mechanism of redox regulation are reviewed. Also reviewed are new details about the interaction between activase and Rubisco and the likely mechanism of 'activation' that resulted from mutagenesis in a 'Sensor 2' domain of activase that AAA(+) proteins often use for substrate recognition. Two activase residues in this domain were identified that are involved in Rubisco recognition. The results directly complement earlier studies that identified critical residues for activase recognition in the large subunit of Rubisco.     

The effects of Rubisco activase on C4 photosynthesis and metabolism at high temperature

The activation of Rubisco in vivo requires the presence of the regulatory protein Rubisco activase. This enzyme facilitates the release of sugar phosphate inhibitors from Rubisco catalytic sites thereby influencing carbamylation. T(1) progeny of transgenic Flaveria bidentis (a C(4) dicot) containing genetically reduced levels of Rubisco activase were used to explore the role of the enzyme in C(4) photosynthesis at high temperature. A range of T(1) progeny was screened at 25 degrees C and 40 degrees C for Rubisco activase content, photosynthetic rate, Rubisco carbamylation, and photosynthetic metabolite pools. The small isoform of F. bidentis activase was expressed and purified from E. coli and used to quantify leaf activase content. In wild-type F. bidentis, the activase monomer content was 10.6+/-0.8 micromol m(-2) (447+/-36 mg m(-2)) compared to a Rubisco site content of 14.2+/-0.8 micromol m(-2). CO(2) assimilation rates and Rubisco carbamylation declined at both 25 degrees C and 40 degrees C when the Rubisco activase content dropped below 3 mumol m(-2) (125 mg m(-2)), with the status of Rubisco carbamylation at an activase content greater than this threshold value being 44+/-5% at 40 degrees C compared to 81+/-2% at 25 degrees C. When the CO(2) assimilation rate was reduced, ribulose-1,5-bisphosphate and aspartate pools increased whereas 3-phosphoglycerate and phosphoenol pyruvate levels decreased, demonstrating an interconnectivity of the C(3) and C(4) metabolites pools. It is concluded that during short-term treatment at 40 degrees C, Rubisco activase content is not the only factor modulating Rubisco carbamylation during C(4) photosynthesis.     

Enhanced trehalose production improves growth of Escherichia coli under osmotic stress

The biosynthesis of trehalose has been previously shown to serve as an important osmoprotectant and stress protectant in Escherichia coli. Our results indicate that overproduction of trehalose (integrated lacI-Ptac-otsBA) above the level produced by the native regulatory system can be used to increase the growth of E. coli in M9-2% glucose medium at 37 degrees C to 41 degrees C and to increase growth at 37 degrees C in the presence of a variety of osmotic-stress agents (hexose sugars, inorganic salts, and pyruvate). Smaller improvements were noted with xylose and some fermentation products (ethanol and pyruvate). Based on these results, overproduction of trehalose may be a useful trait to include in biocatalysts engineered for commodity chemicals.     

Reductions of Rubisco activase by antisense RNA in the C4 plant Flaveria bidentis reduces Rubisco carbamylation and leaf photosynthesis

To function, the catalytic sites of Rubisco (EC 4.1.1.39) need to be activated by the reversible carbamylation of a lysine residue within the sites followed by rapid binding of magnesium. The activation of Rubisco in vivo requires the presence of the regulatory protein Rubisco activase. This enzyme is thought to aid the release of sugar phosphate inhibitors from Rubisco's catalytic sites, thereby influencing carbamylation. In C3 species, Rubisco operates in a low CO2 environment, which is suboptimal for both catalysis and carbamylation. In C4 plants, Rubisco is located in the bundle sheath cells and operates in a high CO2 atmosphere close to saturation. To explore the role of Rubisco activase in C4 photosynthesis, activase levels were reduced in Flaveria bidentis, a C4 dicot, by transformation with an antisense gene directed against the mRNA for Rubisco activase. Four primary transformants with very low activase levels were recovered. These plants and several of their segregating T1 progeny required high CO2 (>1 kPa) for growth. They had very low CO2 assimilation rates at high light and ambient CO2, and only 10% to 15% of Rubisco sites were carbamylated at both ambient and very high CO2. The amount of Rubisco was similar to that of wild-type plants. Experiments with the T1 progeny of these four primary transformants showed that CO2 assimilation rate and Rubisco carbamylation were severely reduced in plants with less than 30% of wild-type levels of activase. We conclude that activase activity is essential for the operation of the C4 photosynthetic pathway.     

Inhibition of photosynthesis by heat stress: the activation state of Rubisco as a limiting factor in photosynthesis

Although the catalytic activity of Rubisco increases with temperature, the low affinity of the enzyme for CO₂ and its dual nature as an oxygenase limit the possible increase in net photosynthesis with temperature. For cotton, comparisons of measured rates of net photosynthesis with predicted rates that take into account limitations imposed by the kinetic properties of Rubisco indicate that direct inhibition of photosynthesis occurs at temperatures higher than about 30°C. Inhibition of photosynthesis by moderate heat stress (i.e. 30–42°C) is generally attributed to reduced rates of RuBP regeneration caused by disruption of electron transport activity, and specifically inactivation of the oxygen evolving enzymes of photosystem II. However, measurements of chlorophyll fluorescence and metabolite levels at air-levels of CO₂ indicate that electron transport activity is not limiting at temperatures that inhibit CO₂ fixation. Instead, recent evidence shows that inhibition of net photosynthesis correlates with a decrease in the activation state of Rubisco in both C₃ and C₄ plants and that this decrease in the amount of active Rubisco can fully account for the temperature response of net photosynthesis. Biochemically, the decrease in Rubisco activation can be attributed to: (1) more rapid de-activation of Rubisco caused by a faster rate of dead-end product formation; and (2) slower re-activation of Rubisco by activase. The net result is that as temperature increases activase becomes less effective in keeping Rubisco catalytically competent. In this opinionated review, we discuss how these processes limit photosynthetic performance under moderate heat stress.     

Heat sensitivity of Rubisco,Rubisco activase and Rubisco binding protein in higher plants

During the past few years the investigations concerning Rubisco and the changes of its activity and properties at elevated temperature were reconsidered with special reference to the important role of Rubisco activase and Rubisco binding protein. The major changes in Rubisco, Rubisco activase and Rubisco binding protein reported recently are presented in this review. New information on these proteins, including their changes under heat stress conditions, is discussed together with open questions.     

Antisense inhibition of Rubisco activase increases Rubisco content and alters the proportion of Rubisco activase in stroma and thylakoids in chloroplasts of rice leaves

BACKGROUND AND AIMS: Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) activase (RCA) is a nuclear-encoded chloroplast protein that modifies the conformation of Rubisco, releases inhibitors from active sites, and increases enzymatic activity. It appears to have other functions, e.g. in gibberellin signalling and as a molecular chaperone, which are related to its distribution within the chloroplast. The aim of this research was to resolve uncertainty about the localization of RCA, and to determine whether the distributions of Rubisco and RCA were altered when RCA content was reduced. The monocotyledon, Oryza sativa was used as a model species. METHODS: Gas exchange and Rubisco were measured, and the sub-cellular locations of Rubisco and RCA were determined using immunogold-labelling electron microscopy, in wild-type and antisense rca rice plants. KEY RESULTS: In antisense rca plants, net photosynthetic rate and the initial Rubisco activity decreased much less than RCA content. Immunocytolocalization showed that Rubisco in wild-type and antisense plants was localized in the stroma of chloroplasts. However, the amount of Rubisco in the antisense rca plants was greater than in the wild-type plants. RCA was detected in both the chloroplast stroma and in the thylakoid membranes of wild-type plants. The percentage of RCA labelling in the thylakoid membrane was shown to be substantially decreased, while the fraction in the stroma was increased, by the antisense rca treatment. CONCLUSIONS: From the changes in RCA distribution and alterations in Rubisco activity, RCA in the stroma of the chloroplast probably contributes to the activation of Rubisco, and RCA in thylakoids compensates for the reduction of RCA in the stroma, allowing steady-state photosynthesis to be maintained when RCA is depleted. RCA may also have a second role in protecting membranes against environmental stresses as a chaperone.     

Decreased CO2 availability and inactivation of Rubisco limit photosynthesis in cotton plants under heat and drought stress in the field

Heat and drought stresses are often coincident and constitute major factors limiting global crop yields. A better understanding of plant responses to the combination of these stresses under production environments will facilitate efforts to improve yield and water use efficiencies in a climatically changing world. To evaluate photosynthetic performance under dry-hot conditions, four cotton (Gossypium barbadense L.) cultivars, Monseratt Sea Island (MS), Pima 32 (P32), Pima S-6 (S6) and Pima S-7 (S7), were studied under well-watered (WW) and water-limited (WL) conditions at a field site in central Arizona. Differences in canopy temperature and leaf relative water content under WL conditions indicated that, of the four cultivars, MS was the most drought-sensitive and S6 the most drought-tolerant. Net CO₂ assimilation rates (A) and stomatal conductances (gs) decreased and leaf temperatures increased in WL compared to WW plants of all cultivars, but MS exhibited the greatest changes. The response of A to the intercellular CO₂ concentration (A–C_i) showed that, along with stomatal closure, non-stomatal factors associated with heat stress also limited A under WL conditions, especially in MS. The activation state of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) decreased in WL compared to WW plants, consistent with thermal inhibition of Rubisco activase activity. The extent of Rubisco deactivation could account for the metabolic limitation to photosynthesis in MS. Taken together, these data reveal the complex relationship between water availability and heat stress for field-grown cotton plants in a semi-arid environment. Both diffusive (drought-stress-induced) and biochemical (heat-stress-induced) limitations contributed to decreased photosynthetic performance under dry-hot conditions.     

Effect of activase level and isoform on the thermotolerance of photosynthesis in Arabidopsis

Ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) activation decreases under moderate heat stress. This decrease is caused by an impairment of activase function, which is exacerbated by faster rates of Rubisco deactivation at elevated temperatures. To determine if stromal oxidation causes inhibition of activase, transgenic Arabidopsis plants expressing suboptimal amounts of either the redox-regulated 46 kDa alpha- or non-redox regulated 43 kDa beta-isoform of activase were examined. Photosynthesis, as measured by gas exchange and chlorophyll fluorescence, and Rubisco activation were inhibited to a much greater extent by moderately high temperatures in the two transgenic lines expressing suboptimal levels of the individual isoforms of activase compared with wild-type plants or transgenic plants expressing levels of the beta-isoform sufficient for wild-type rates of photosynthesis. Net photosynthesis and Rubisco activation in transgenic plants expressing suboptimal amounts of the beta-isoform of activase from the Antarctic hairgrass were even more sensitive to inhibition by moderate heat stress than in the transgenic plants containing Arabidopsis activase. The results demonstrate that photosynthesis exhibits a similar sensitivity to inhibition by moderately high temperature in plants expressing either of the two different isoforms of activase. Thus, impairment of activase function under heat stress is not caused by oxidation of the redox-sensitive sulphydryls of the alpha-isoform of activase. Instead, the results are consistent with thermal denaturation of activase under moderate heat stress, the effects of which on Rubisco activation would be enhanced when activase levels are suboptimal for photosynthesis.     

Characterization of the regulatory function of the 46-kDa isoform of Rubisco activase from Arabidopsis

Arabidopsis Rubisco activase was recently shown to be regulated by redox changes in the larger (46-kDa) isoform specifically mediated by thioredoxin-f [Zhang and Portis (1999) Proc Natl Acad Sci USA 96: 9438–9443]. Reduction greatly increases the activity of the 46-kDa isoform and the native protein at physiological ATP/ADP ratios. In this study we conducted additional experiments to characterize the regulation of Rubisco activase by thioredoxin-f. The K_m for both ATP hydrolysis and Rubisco activation by the 46-kDa isoform was lowered by 4 to 5-fold after reduction, but the maximum activity was increased by only 10%. Only 0.35 μM thioredoxin-f was required for a half-maximal activity change after a 10 min preincubation and activation with 1 μM was complete after 10 min. Equal amounts of 46-kDa and 43-kDa isoforms were required for a complete inhibition of the Rubisco activation activity after a reduction-oxidation cycle and assay at an ATP/ADP ratio of 3:1, whereas activity was only inhibited by 50% at a 2:1 ratio (43-/46-kDa) of the isoforms. This requirement is consistent with the fact that Arabidopsis normally contains about a 1:1 ratio of the two isoforms at both the mRNA and protein levels. Redox titrations indicated a midpoint potential of −344 mV for the 46-kDa isoform as compared to −342 mV for spinach fructose 1,6-bisphosphatase at pH 7.9, consistent with previous reports indicating that these proteins are co-regulated by light intensity in a similar manner. This revised version was published online in June 2006 with corrections to the Cover Date.     

外源甜菜碱对盐胁迫下枸杞光合功能的改善

研究了外源甜菜碱对盐胁迫下枸杞扦插苗叶片光合功能的影响。结果表明,外源甜菜碱能使盐胁迫下的枸杞叶片叶绿素荧光动力学参数Fo、Fm、Fv、Fv／Fm、Fm／Fo和Fv／Fo增高,使光合色素叶绿素a(Chla)、叶绿素b(Chlb)和类胡萝卜素(Car)含量明显增加,叶绿素a与b的比值(Chla／Chlb)升高,类胡萝卜素与叶绿素的比值(Car／Chl)降低,说明外源甜菜碱有利于植物对光能的捕获、吸收、传递和转换,提高叶片的光合活性,降低盐胁迫对植物的抑制作用。     

Heat stress effects on ribulose-1,5-bisphosphate carboxylase/oxygenase,Rubisco binding protein and Rubisco activase in wheat leaves

Changes in chlorophyll content, ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) binding protein (RBP), Rubisco activase (RA), Rubisco large (LS) and small (SS) subunits, and electrolyte leakage were investigated in wheat leaf segments during heat stress (HS) for 1 h and for 24 h at 40 °C in darkness or in light, as well as after recovery from heat stress (HSR) for 24 h at 25 °C in light. The 24-h HS treatment in darkness decreased irreversibly photosynthetic pigments, soluble proteins, RBP, RA, Rubisco LS and SS. An increase in RA and RBP protein contents was observed under 24-h HS and HSR in light. This increase was in accordance with their role as chaperones and the function of RBP as a heat shock protein.This work was partially supported by Swiss National Science Foundation (Project 31-55289.98).