Rubisco synthesis, turnover and degradation: some new thoughts on an old problem
The life cycle of most higher plant species (including major cultivated crops such as maize, rice and wheat) can be roughly divided into two main developmental phases. In the first phase, during vegetative growth, young developing roots and leaves behave as sinks and efficiently absorb and assimilate inorganic nitrogen (N) for amino acid synthesis. These amino acids are mainly used to synthesize proteins involved in photosynthesis. They are also used to build basic cellular structures required for plant development as well as proteins and enzymes involved in the functioning and regulation of other metabolic pathways. In the second phase, the plant enters into the remobilization stage during which leaves, stems and roots behave as a source for N and export amino acids for the formation of newly developing leaves and/or storage organs.
Storage organs such as seeds, fruits, tubers, bulbs and trunks are involved in plant survival and reproduction (Masclaux et al., 2001). Since Neolithic times, many of these storage organs were used for both animal and human consumption. This eventually led to plant breeding to increase the production of harvestable material (yield, resistance to biotic and abiotic stresses) or for improving its nutritional and biotechnological quality (lipids, carbohydrate and protein contents). A better understanding of the physiological and genetic control of inorganic N assimilation and N recycling from proteins during plant growth and development would greatly enhance the chances of achieving genetic improvement. Thus, identifying the regulatory mechanisms controlling these metabolic processes would allow the selection of target traits for breeding new genotypes or new cultivars exhibiting a better nitrogen use efficiency (NUE, the capacity to produce a supplement of yield for each added unit of N fertilizer) and consequently enhancing carbon translocation in storage organs (Below et al., 2000). In turn, such a strategy would be beneficial both for the rationalization of crop fertilization and for the improvement of end product quality (Good et al., 2004).
‘Irving & Robinson have questioned the role of Rubisco as a leaf storage protein.’
The dual role of Rubisco
In C3 plants, one of the main components of the photosynthetic machinery is the enzyme Ribulose-1,5-bisphospahte carboxylase/oxygenase (Rubisco, E.C. 4.1.139). This enzyme catalyses two competing reactions, photosynthetic CO2 fixation and oxygenation of the RuBP substrate. In C4 plants, the situation is different because phosphoenol-pyruvate-carboxylase (PEPc, E.C. 18.104.22.168) is the enzyme primarily responsible for CO2 fixation in the mesophyll cells. This evolutionary characteristic of C4 plants is a way to limit the loss of carbon (C) during photorespiration, which remains very low during the transfer of CO2 to Rubisco in the bundle sheath cells (Edwards et al., 2001). Despite this difference, in both types of plants Rubisco accounts for 15–35% of total N and around 30–60% of total shoot soluble proteins, which constitute a large pool of stored N. In wheat, once the flag leaf is fully developed, Rubisco represents approximately 60% of total soluble proteins. The amount of protein starts to decrease around 8–10 d after anthesis, a time that corresponds to the beginning of the grain-filling period (Peoples et al., 1980). In maize, a decrease in the amount of Rubisco in the ear leaf is already visible around 15 d after silking (Smart et al., 1995), whereas in rice it occurs just after the completion of leaf expansion (Mae et al., 1983). Interestingly, in N-starved plants, it was shown that the decrease in the amount of Rubisco during leaf senescence is higher compared to plants grown under suboptimal N-feeding conditions (Tercé-Laforgue et al., 2004), suggesting that Rubisco is a source of N when there is a shortage of soil inorganic N.
By virtue of its relative amount within the pool of soluble proteins, the Rubisco enzyme is one of the most important sources of N that can be potentially remobilized and translocated to growing or storage organs. Understanding the mechanisms that control its synthesis and degradation is therefore of major importance in the control of the plant's N economy. A number of studies, mainly performed on cereals, was undertaken to identify mechanisms that trigger Rubisco synthesis and turnover in vivo, and to examine the influence of various biotic and abiotic stresses on its rate of degradation (Makino et al., 1984; Cheng et al., 1998; Esquivel et al., 2000). In this issue of New Phytologist, Irving & Robinson (pp. 493–504) extensively discuss the potential and pitfalls of these different studies. In particular, these authors have highlighted the fact that both the biosynthetic and proteolytic rates of cells within the leaf may be different, that the mathematical models developed to describe Rubisco turnover using 15N-labeling require a larger set of data, and that the developmental stage of the leaf from life to death has to be taken into account. Due to these uncertainties, Irving & Robinson have questioned the role of Rubisco as a leaf storage protein. Similar conclusions were previously drawn by Esquivel et al. (2000), following the observation that the rate of Rubisco protein degradation was different in C3 and C4 plants and according to either sulfur or N deprivation.
The dynamic model of Rubisco turnover: from theory to physiology and agronomy
Rubisco protein turnover during the life span of the leaf life was studied by a number of authors in both C3 (Peterson & Huffaker, 1975; Mae et al., 1983) and C4 grasses (Pernollet et al., 1986; Esquivel et al., 2000). In particular, by labeling leaf proteins with 15N-isotopes, Mae et al. (1983) showed that part of the Rubisco protein pool was subjected to constant turnover even at a late stage of leaf senescence. However, the model proposed by these authors to describe Rubisco synthesis and turnover during the life span of rice leaves was apparently not fully satisfactory, owing to a number of experimental assumptions and oversimplifications. This prompted Irving & Robinson to reanalyze the data sets obtained from rice and other cereals such as wheat and barley, and to propose a new mathematical model to estimate Rubisco content, synthesis and degradation at any point in the lifespan of the leaf and for any duration. One of their main conclusions was that Rubisco concentration must be exerted by adjustment of biosynthetic activity rather than by proteolysis because it is a first-order kinetic process. This can be mainly explained by a decrease in the availability of N during senescence that would be directed to other sinks rather than the leaves, thus leading to a decrease in the synthesis of both Rubisco transcripts and protein. However, Irving & Robinson pointed out that their model is not applicable to dead or dying leaves, in which it was shown that there is no direct relationship between the N status and the onset of tissue death (Hirel et al., 2005). They also mentioned that the model might not be applicable to describe Rubisco turnover in C4 species, because more N is invested in growth and grain filling rather than in leaf proteins. This also explains why, when compared with C3 plants, C4 plants have a lower N content per unit of leaf area (Hirel & Lemaire, 2006). The case of nongraminoid plants mentioned by Irving & Robinson, which apparently does not fit to the model, is also interesting as it would be a way for the plant to enhance leaf N content by increasing leaf thickness. Such a strategy is apparently adopted by sorghum stay-green genotypes, although this species belongs to graminoids (Borrell et al., 2001), which allow the plant to use ‘luxury’ N stored in additional leaf cell layers, particularly under water stress conditions (A. Borrell, Hermitage Research Station, Warwick, Australia, pers. comm.).
Despite these species-specific limitations, the conclusions of Irving & Robinson raise the question about the origin of N that is translocated to the seeds (which represent a major sink in both C3 and C4 cereals). A few 15N-labeling experiments performed either under controlled growth conditions (Cliquet et al., 1990) or in the field (Hirel & Limami, 2003; Kichey et al., unpublished) have clearly demonstrated that a large part of the N that is accumulated in the leaf is further remobilized and translocated to the seeds independently of postflowering N absorption. Depending on the genotype or on the photosynthetic type of the plant, remobilized N accounts for about 50% of the seed N content in maize (Martin et al., 2005) and 80% in wheat (Le Gouis et al., 2000).
More recently, results obtained from 15N-labeling in maize have been further analyzed, and a model describing 15N-fluxes within the plant after silking (i.e. during grain filling) has been proposed (Gallais et al., 2006). In this study, it was demonstrated that there is a protein turnover within the plant in such a way that more than 50% of the absorbed N allocated to the stover is used for the synthesis of new proteins, the rest being translocated to the grain. During the progression of leaf senescence, a decrease in the protein turnover is observed and the N originating from protein hydrolysis is then mostly used for grain filling and is in fact the cause of the decrease in the protein turnover.
If one considers the two sets of data obtained by Irving & Robinson and by Gallais et al. (2006), it is therefore likely that both Rubisco protein hydrolysis and turnover coexist within the plant and that the organic N released during these two processes is used by newly developing leaves or storage organs. Moreover, the occurrence of Rubisco degradation in senescence cannot be questioned, because it was clearly demonstrated that the protein is cleaved during natural senescence (Kokobun et al., 2002). In addition, it is well known that a large panel of proteolytic activities is induced during leaf senescence in both C3 (Masclaux et al., 2000) and C4 (Feller et al., 1977) plants, concomitantly with the decrease in the amount of leaf soluble proteins. The regulation of Rubisco synthesis and turnover under stress conditions, for example when N is limiting (Tercé-Laforgue et al., 2004), needs also to be considered, because the plant will have to reutilize the N stored into protein for its survival when limited amounts of nitrate or ammonium are available in the soil. The half-life of foliar protein, estimated to be around 7 d, is considerably shortened under stress conditions (Davies, 1982), which once again supports the hypothesis that Rubisco protein turnover slows down when N remobilization is required to ensure plant growth and development under adverse environmental conditions.
The enzyme Rubisco is the most abundant protein in the world (Ellis, 1979) and therefore constitutes a large reserve of N, together with N2 present in the atmosphere and mineral and organic N present in the soil (Morot-Gaudry & Touraine, 2001). Despite this unique characteristic, the number of investigations aimed towards understanding its synthesis and degradation are relatively limited compared with the studies performed on both understanding and improving symbiotic N2 fixation and mineral N acquisition by higher plants. Although the model proposed by Irving & Robinson has several limitations with regards to its physiological reality in certain species and at certain stages of leaf development, it is important in that it renews the debate on the importance of N use and recycling by the plant – a process which does not seem to be as simple as described in the literature. This new concept of considering the regulation of plant N invested in Rubisco therefore opens a new way of thinking and therefore a better understanding of the regulation N use efficiency by plants that can be further used for agronomic applications. One of the ways towards improving our knowledge on this biological process would be to study the genetic variability of Rubisco accumulation and degradation in relation to protease activities, yield and grain protein content using a quantitative genetic approach, similar to that which has already been developed to study the genetic basis of N uptake, assimilation an recycling in both model and crop species (Gallais & Hirel, 2004; Yang et al., 2004). Performing 15N-labelling experiments either under controlled or field conditions offers one of the most reliable ways for undertaking such approach because it provides the possibility of considering plant N economy, not only at the whole plant level but also at the organ and cellular level during cereal growth and development (Hirel et al., 2005; Kichey et al., 2005, 2006). Including these data into agronomic modeling approaches appears to be another challenge for the future.