Impacts of autophagy on nitrogen use efficiency in plants

ABSTRACT

Crop productivity greatly depends on nitrogen (N) fertilization. Large inputs of N fertilizer impact on both the farmer and the environment. Accordingly, improving N use efficiency (NUE) for crop productivity is important for sustainable agriculture. Much plant nitrogen is allocated into the chloroplasts in leaves to constitute proteins involved in photosynthesis, and remobilization of N from senescent leaves greatly affects crop productivity. Autophagy, a highly conserved system used to degrade intracellular components in eukaryotes, is responsible for degrading chloroplasts and their proteins during leaf senescence. In this paper, we review recent studies establishing that autophagy is a key for maintaining high NUE during vegetative growth and for efficiently remobilizing N to grains.

KEY WORDS:

• Autophagy-related gene
• chloroplast
• leaf senescence
• nitrogen remobilization
• protein degradation

1. Introduction

Nitrogen (N) is an abundant, essential nutrient that limits plant growth and productivity in both natural and agricultural ecosystems. Plants can assimilate inorganic N into amino acids, which are used to synthesize important nitrogenous compounds such as proteins, DNA, and chlorophylls. A large input of N fertilizer is required to maintain high crop yields, but incurs high costs for both the farmer and the environment (Good et al. 2004; Brauer and Shelp 2010). Since the production of N fertilizer is energy-intensive, its application is the most costly input cost for many crops (Rothstein 2007). The production and excessive use of N fertilizers contributes greatly to the depletion of stratospheric ozone and global warming (Wuebbles 2009). In addition, the typical N compound in fertilizers, ammonium is mobile and easily oxidized in the soil. The recovery from such conventional N fertilizers by crops is generally low and most of N is lost in the environments (Mae and Shoji 1984; Raun and Johnson 1999). Therefore, N pollution from agriculture leads to the degradation of downstream water quality and eutrophication of coastal marine ecosystems, the development of photochemical smog, and rising global concentrations of nitrous oxide, a powerful greenhouse gas (Vitousek et al. 2009).

Engineering N use efficiency (NUE) is essential for sustainable and productive agriculture (Good et al. 2004; Mae et al. 2006; Vitousek et al. 2009; Masclaux-Daubresse et al. 2010; Kant et al. 2011; Makino 2011; Xu et al. 2012). Although NUE can be defined in several ways, the simplest definition for crop plants is the yield (grain, fruit or forage) per unit of N available in the soil (Kant et al. 2011) or from N application (Mae et al. 2006). In molecular physiology studies using model plants, a physiological NUE index is expressed as the fresh or dry matter produced per N content or N concentration in the plant or the shoot (Good et al. 2004; Masclaux-Daubresse et al. 2010; Kanno et al. 2017). All plant processes involving N usage affect NUE, namely uptake, translocation, assimilation, recycling, and remobilization. The remobilization of N from senescing leaves to growing organs is particularly important in NUE (Masclaux-Daubresse et al. 2010). Recent studies have shown that autophagy, a eukaryotic process whereby cellular components are degraded in bulk, is crucial for N remobilization in Arabidopsis (Guiboileau et al. 2012), maize (Li et al. 2015a), and rice (Wada et al. 2015).

2. N distribution in plants

Much N is distributed to the leaves during the vegetative growth stage (Schulze et al. 1994). In C3 plants, approximately 80% of the total leaf N is allocated to chloroplasts, mainly as photosynthetic proteins (Makino and Osmond 1991; Makino et al. 2003). Around 70% of chloroplast N is present in the stroma, and the rest is in the thylakoid membrane (Makino et al. 2003). The carbon-fixing enzyme Rubisco (ribulose-1,5-bisphosphate carboxylase/oxygenase) is the most abundant protein, accounting for 27% of the leaf nitrogen in rice (Makino et al. 2003). The second most abundant protein is the light-harvesting chlorophyll a/b protein of photosystem II (LHCII) in the thylakoid membrane, accounting for 7% of the leaf N in rice (Makino et al. 2003). In the C₄ plant maize, Rubisco accounts for 9% and LHCII accounts for 8%; the lesser amount of Rubisco allows this plant to invest more N in other thylakoid components such as photosystem I and ATP synthase (Makino et al. 2003).

3. N remobilization during leaf senescence

As most plants are sessile, efficient use and recycling of assimilated N is important for survival and fitness under given environments. Senescence, the final developmental stage, is a form of programmed cell death in plant organs. Leaf senescence is a specialized form that aims to remobilize various nutrients, particularly N, from source leaves to growing vegetative or reproductive sinks via the phloem.

Chloroplasts are the major reservoirs of assimilated N and thus the primary target for N remobilization. In rice, Rubisco is actively degraded soon after the full expansion of the leaves (Mae et al. 1983). Under N-sufficient conditions, Rubisco accounts for 34% of remobilized N from leaves (Makino et al. 1984). The degradation of LHCII occurs a little after the degradation of Rubisco during leaf senescence (Hidema et al. 1992). Chlorophyll and its primary catabolites are highly phototoxic and risk causing premature cell death. Thus, chlorophyll is catabolized into safe compounds prior to the degradation of LHCII and other chlorophyll-binding proteins, and it is the main cause of leaf color change. Tetrapyrroles of chlorophylls contain N atoms accounting for 1–2% of total leaf N (Makino and Osmond 1991). However, unlike amino acids derived from protein degradation, chlorophyll catabolites are not remobilized, and are permanently stored in the vacuoles of senescent leaves (Hörtensteiner and Kräutler 2011). As levels of the inner components decrease, chloroplasts gradually shrink and transform into gerontoplasts, in which thylakoid membranes disintegrate and plastoglobules accumulate, and the cellular population of chloroplasts concomitantly declines (Krupinska 2006). Unlike chloroplasts, mitochondria are not a major source of N remobilization, and remain active until the later stages of leaf senescence (Thomas and Stoddart 1980). This is probably because mitochondrial maintenance is essential for cell viability, including the export of N compounds from source leaves.

4. Mechanisms of autophagy

Autophagy is a system for the bulk degradation of intracellular components. It is evolutionarily conserved in eukaryotes including fungi, animals, and plants (Bassham et al. 2006; Nakatogawa et al. 2009; Yoshimoto 2012). The mechanism of autophagy has been particularly well studied in yeast species such as Saccharomyces cerevisiae and Pichia pastoris. Two morphologically distinct forms of autophagy, microautophagy and macroautophagy, have been observed. Macroautophagy is considered the major form, and is referred to simply as ‘autophagy’.

Under some stress conditions such as nutrient starvation, intracellular components such as proteins and organelles are sequestered into a double-membraned vesicle called an autophagosome, an organelle specific to autophagy. Autophagosomes are transported into lytic compartments, namely the vacuole in fungi and plants, or lysosomes in animals, where autophagosomes and their contents are degraded by a variety of resident hydrolases. Derived nutrients such as amino acids are returned to the cytosol and are reused as building blocks for protein synthesis, or as sources of respiratory substrates.

Professor Yoshinori Ohsumi was awarded the 2016 Nobel Prize for Physiology or Medicine for his discoveries of the mechanisms of autophagy. Together with colleagues, Ohsumi’s yeast forward genetic experiments identified AuTophaGy-related genes (Atgs) that participate in autophagic processes, and revealed their functions (Takeshige et al. 1992; Tsukada and Ohsumi 1993; Mizushima et al. 1998; Ichimura et al. 2000). To date, 41 Atgs (Atg1–41) have been identified in S. cerevisiae (Yao et al. 2015). Of these, 15 ‘core’ Atgs (Atg1–10, 12–14, 16, and 18) are commonly required for various autophagy pathways including starvation-induced typical macroautophagy (Nakatogawa et al. 2009). Core Atgs are highly conserved in all eukaryotes, including plants, with a few exceptions. Core Atg products (Atgs) constitute the central autophagy machinery and are fundamental for the biogenesis of autophagosomes. Most other Atgs are involved in selective autophagy, which targets limited substrates for degradation.

Numerous excellent reviews have summarized the detailed functions of the core Atgs (Nakatogawa et al. 2009; Mizushima et al. 2011; Li and Vierstra 2012; Noda and Inagaki 2015). The core Atgs constitute functional units, namely the Atg1 kinase complex (Atg1 and 13), the autophagy-specific phosphatidylinositol 3-kinase (PI3K) complex (Atg6 and 14), the Atg9 and Atg2–Atg18 complex, two ubiquitin (Ub)-like conjugation systems, the Atg12 conjugation system (Atg5, 7, 10, 12, 16) and the Atg8 lipidation system (Atg3, 4, 7, and 8).

The Atg1 kinase complex is involved in autophagy induction. Its assembly is regulated by target of rapamycin (TOR) complex 1 (TORC1), which senses cellular nutrient status. The PI3K complex consists of ATG6, ATG14, VPS34 and VPS15, and decorates autophagosomal membranes with phosphatidylinositol 3-phosphate. Atg9 is the only transmembrane protein among the core Atgs, and helps to deliver lipids to the expanding autophagosomal membranes. The Atg2–Atg18 complex interacts with Atg9, but its detailed functions are unknown. Both the Atg12 conjugation system and the Atg8 lipidation system are required for covalent conjugation of Atg8 to phosphatidyl ethanolamine (PE). The ubiquitin-like protein Atg8 is a marker protein of the autophagosome, and is tethered to the autophagosomal membrane by lipidation. This process is important for the expansion and closure of the autophagosome.

5. Autophagy in plants

While microscopic studies have long suggested the existence of autophagy in plants, the identification of Atgs in yeast has opened the door to the development of molecular studies of plant autophagy (Bassham et al. 2006; Yoshimoto et al. 2010; Yoshimoto 2012). Orthologs of yeast core Atgs were first identified in the Arabidopsis genome (Doelling et al. 2002; Hanaoka et al. 2002). The Arabidopsis genome contains almost all the core ATGs, some of which consist of multigene families (Yoshimoto 2012). Recent molecular studies have indicated that the core machinery of Arabidopsis ATG (AtATG) functions in the same way as it does in yeast (Yoshimoto et al. 2004; Thompson et al. 2005; Xiong et al. 2005; Suttangkakul et al. 2011). Yeast Atg orthologs have also been comprehensively identified in the genomes of Chlamydomonas reinhardtii (Avin-Wittenberg et al. 2012), rice (Oryza sativa) (Xia et al. 2011), maize (Zea mays) (Li et al. 2015a), barley (Hordeum vulgare) (Avila-Ospina et al. 2016), tobacco (Nicotiana tabacum), and foxtail millet (Setaria italica) (Li et al. 2016). Plant genomes commonly lack ATG14, suggesting that plants have developed alternative genes to complement the functions of ATG14.

Knockout mutants of Arabidopsis core ATGs (Atatgs) have defective autophagy activity. Whereas Atatg mutants can complete their life cycles under favorable nutrient and growth conditions, they cannot survive for long periods under N or carbon depletion conditions (Hanaoka et al. 2002; Thompson et al. 2005; Phillips et al. 2008). Atatg mutants show premature senescence and death of leaves, with accelerated losses of chlorophyll and some chloroplast proteins (Doelling et al. 2002; Hanaoka et al. 2002). This observation led to the view that other proteolytic systems besides autophagy are responsible for degrading chloroplasts and proteins, which are the major source of N for recycling (Levine and Klionsky 2004; Bassham et al. 2006).

Our previous review described the roles of autophagy in chloroplast degradation (Ishida et al. 2014). Briefly, there are two distinct pathways of chloroplast-targeted autophagy: entire-organelle autophagy (‘chlorophagy’), and piecemeal-type autophagy via Rubisco-containing bodies (RCBs). These two pathways are partially responsible for decreases in size and number of chloroplasts during senescence (Wada et al. 2009). The mechanisms of Rubisco degradation in senescent cereal leaves such as wheat (Triticum aestivum) and barley have long been studied as a hallmark of nitrogen remobilization. In well-developed leaf mesophyll cells, almost all of the proteolytic activity against Rubisco takes place in the large vacuoles (Wittenbach et al. 1982). Indeed, electron microscopy of wheat leaves under dark-induced senescence revealed that chloroplasts are located within the mesophyll cell vacuoles (Wittenbach et al. 1982). This pioneer study led to the proposal that sequential degradation of chloroplasts within the vacuole serves as the major pathway for chloroplast protein degradation in senescent leaves. Vacuole-incorporated chloroplasts can conveniently be visualized by light or laser-scanning confocal microscopy; however, these chloroplasts are not observed in individually dark-induced senescent leaves of Atatg mutants (Wada et al. 2009). A recent study showed that chlorophagy is not only responsible for recycling, but also for organelle quality control; chloroplasts photodamaged by exposure to UV-B or strong visible light were eliminated by autophagy (Izumi et al. 2017).

Clearly, chlorophagy is not the only pathway responsible for Rubisco degradation during leaf senescence. Many studies have shown that Rubisco levels decline much faster than chloroplast population size during senescence (Martinoia et al. 1983; Mae et al. 1984). Originally identified in senescent wheat leaves by immuno-electron microscopy with anti-Rubisco antibodies, RCBs were initially proposed to degrade Rubisco outside of chloroplasts (Chiba et al. 2003). RCBs are autophagic bodies containing chloroplast stromal proteins, but they do not contain thylakoid membranes or proteins (Chiba et al. 2003; Ishida and Yoshimoto 2008; Ishida et al. 2008). According to our quantitative analyses, RCB-type autophagy is responsible for at least 40% of Rubisco degradation during dark-promoted senescence (Ono et al. 2013). There is another piecemeal autophagy pathway of chloroplasts via ATG8-INTERACTING PROTEIN1 (ATI1) bodies which contain a stromal proteins marker as cargo (Michaeli et al. 2014). ATI1 specifically interacts with some plastid-localized proteins but does not interact with Rubisco, suggesting that ATI bodies are not much important to Rubisco degradation during senescence.

6. Effects of autophagy deficiency on N remobilization and productivity

Arabidopsis studies have revealed the physiological importance of autophagy on N remobilization and productivity (Guiboileau et al. 2012; Guiboileau et al. 2013). Vegetative growth and seed production is reduced in Atatg mutants under both low N and high N conditions. In the vegetative stage, the N concentration in rosette leaves of Atatg mutants is higher than in the wild-type plants (Guiboileau et al. 2013), indicating that autophagy is critical to maintain NUE at this stage. ¹⁵N-labeling and tracing experiments show that the efficiency of N remobilization into seeds is reduced in Atatg mutants (Guiboileau et al. 2012).

Maize autophagy-deficient mutants, harboring the core ATG ZmATG12 inactivated by insertion of the UniformMu transposon, have recently been isolated and physiologically analyzed (Li et al. 2015a). Under low-N conditions, Zmatg12 mutants show a delayed growth stage, and reduced leaf elongation and ear development. Under high N conditions, Zmatg12 mutants have a normal phenotype, fertility and growth rate, but reduced grain yields. Applying ¹⁵N-labeling experiments previously reported in Arabidopsis to these maize mutants, the authors revealed that the efficiency of N remobilization into grains is reduced in Zmatg12 mutants, as well as the Arabidopsis core atg mutants, and that this may decrease grain yields. Differences were found in the harvest index of maize and Arabidopsis atg mutants under high-N conditions. Zmatg12 mutants produce the same total dry matter and have a reduced harvest index compared to wild-type plants, whereas Atatg5 mutants have lower total biomass and seed yield, and the same harvest index as the wild type. These data indicate that autophagy is not critical for vegetative growth under high-N conditions in maize. Although autophagy contributes to energy production for Arabidopsis growth at night (Izumi et al. 2013), this might not be the case in maize.

Rice knockout mutants of core ATGs have recently been isolated and characterized (Kurusu et al. 2014). An Osatg7 mutant (Osatg7-1), in which OsATG7 is disrupted by insertion of the Tos17 retrotransposon, lacks autophagy activity (Kurusu et al. 2014; Izumi et al. 2015). Surprisingly, unlike autophagy-defective Arabidopsis and maize mutants, Osatg7-1 exhibits complete sporophytic male sterility irrespective of nutrient conditions. Osatg7-1 shows delayed growth of young panicles, incomplete pollen maturation, reduced pollen germination activity, and limited anther dehiscence. A T-DNA-insertional mutant of another core ATG, OsATG9, also has a completely sterile phenotype, indicating that autophagy is critical in reproductive developmental processes in rice (Kurusu et al. 2014).

Because core atg mutants are completely sterile, it is not known how much autophagy contributes to N remobilization at the ripening stage of rice. However, analysis of Osatg7-1 showed that, in rice, autophagy deficiency affects growth and N usage in the vegetative stage (Wada et al. 2015). Compared to the wild-type plants, Osatg7-1 has reduced shoot and root growth under N-sufficient conditions. The mutant grows in the same way as the wild-type plants under high-N conditions, but is much smaller than the wild-type plants under low-N conditions. This growth response pattern suggests that the reduced growth of the mutants is largely related to N nutrition. In hydroponic systems, Osatg7-1 absorbs the same amount of N from the medium, but shows a higher N concentration, and thus has a lower usage index (UI) compared to the wild-type plants (Wada et al. 2015). UI, which is used to estimate NUE in the vegetative stage, is calculated by dividing the shoot dry weight by shoot N concentration (Good et al. 2004; Brauer and Shelp 2010).

Like Arabidopsis and maize atg mutants, Osatg7-1 displays early visible leaf senescence, but the N concentration remains high in the senescent leaves (Wada et al. 2015). Soluble proteins such as Rubisco, and a variety of organelles remain at a high level; thus the degradation of these components seems to be widely suppressed in senescent leaves of Osatg7-1. ¹⁵N pulse–chase analysis revealed the suppression of N remobilization from senescent leaves to newly expanding leaves in Osatg7-1. The reduction of N available for newly developing tissues in Osatg7-1 is the likely cause of its reduced leaf area, tillers, and total biomass. Limited leaf development in Osatg7-1 may decrease the photosynthetic capacity at the level of the whole plant.

7. Approaches for manipulating autophagy activity in plants

The disruption of autophagy has negative influences on many important agricultural traits such as vegetative biomass, yields, N remobilization efficiency, and UI. Additionally, in rice, it causes complete male sterility. Thus, it is worthwhile examining whether increasing autophagy activity, which corresponds to the amounts of cytoplasmic degradation substrates transported into the vacuole, correspondingly improves such traits – especially plant NUE under N-limited conditions.

Autophagy activity is regulated at both the transcriptional and post-transcriptional level (Jin and Klionsky 2014). Little is known about which transcriptional factors are specifically involved in the expression of core ATGs in plants. Post-transcriptionally, TORC1 serves as an upstream negative regulator of the Atg1 complex in yeast autophagy (Nakatogawa et al. 2009). Under nutrient-rich conditions, TORC1 hyperphosphorylates Atg13, promoting dissociation of the Atg1 kinase complex. Under starvation conditions, TORC1 is inactivated, Atg13 is rapidly dephosphorylated, the Atg1 complex is formed, and autophagy is induced. In Arabidopsis, TOR kinase serves the same function as in yeast, regulating induction of autophagy (Liu and Bassham 2010). The limiting step of autophagosome formation, in which core ATGs are involved after activation by the ATG1 kinase complex, remains unknown in plants.

The size and number of autophagosomes are the two main mechanistic factors that regulate autophagy activity; in yeast, these are in turn regulated by the transcriptional regulation of Atg8 and Atg9 (Xie et al. 2008; Jin et al. 2014). The ubiquitin-like protein Atg8 has a central role in autophagosome formation. Atg8 levels determine the size of autophagosomes in yeast (Xie et al. 2008; Bartholomew et al. 2012). The functions of many of the core Atgs are related to the production of Atg8-PE. Atg8 controls phagophore expansion during autophagosome formation, promotes hemifusion of the autophagosomal membrane (Nakatogawa et al. 2007), and recruits specific cargo into autophagosomes (Noda et al. 2010). In yeast, the levels of both Atg8 mRNA and protein become dramatically elevated soon after the onset of nitrogen starvation (Jin and Klionsky 2014).

In plants, ATG8 is encoded by a large multigene family: Arabidopsis has nine ATG8 genes (AtATG8a–i) (Yoshimoto et al. 2004), maize has five (ZmATG8a–e) (Li et al. 2015a), and the rice genome has five to seven ATG8 gene candidates (Chung et al. 2009; Xia et al. 2011). The biological significance of ATG8 multigene families in plants is currently not known, and it is not clear whether different ATG8 proteins have different cargo specificities, or whether they are redundant.

Several studies have demonstrated the effects of ATG8 overexpression on growth and stress tolerance in plants, including positive effects in Arabidopsis (Slavikova et al. 2008). Expression of GFP-AtATG8a-HA fusion under a constitutive 35S promoter enhanced growth under both N-sufficient and N-deficient conditions, and also enhanced tolerance of carbon limitation induced by whole-plant treatment with darkness. In Arabidopsis, heterologous expression of GmATG8c from soybean (Glycine max) under a 35S promoter conferred tolerance to N deficiency, and increased vegetative growth and yield under favorable conditions (Xia et al. 2012), while heterologous expression of SiATG8a from foxtail millet conferred tolerance to both N starvation and drought stress (Li et al. 2015b). Positive effects of SiATG8a overexpression on tolerance to N starvation have also been confirmed in rice (Li et al. 2016). However, the extent to which the overexpression of ATG8 influences autophagy activity or cargo specificity has not yet been elucidated in plants.

Of the core Atgs, Atg9 is the only transmembrane protein, and the only one considered to be a membrane carrier (Noda et al. 2000). Levels of Atg9 are correlated with the numbers of autophagosomes; overexpression of Atg9 leads to an increased accumulation of autophagic bodies in yeast (Jin et al. 2014). Arabidopsis has a single ATG9 gene (Hanaoka et al. 2002). Unlike other core ATGs that are single copy genes, ATG9 is not strictly essential for autophagy; however, it certainly affects autophagy activity (Shin et al. 2014). The effects of ATG9 overexpression on autophagy activity have not yet been examined in plants.

8. Outlook

Recent molecular physiological studies using several model plants have established that autophagy is a key to maintaining both high NUE during the vegetative growth stage, and N remobilization efficiency into grains, which strongly influence yields. Manipulating autophagy activity is a potential target for improving NUE, especially under N-limited conditions. To evaluate this hypothesis, further details are required about the regulatory mechanisms of plant autophagy at the transcriptional and post-transcriptional levels.

Acknowledgment

The authors would like to thank all collaborators and laboratory members who worked on our autophagy projects.

Additional information

Funding

This research was supported by KAKENHI Grants-in-Aid for Scientific Research [Grant nos. 15H04626 for H.I. and 16H06379 for A.M.].