Is plastidic glutamine synthetase essential for C₃ plants? A tale of photorespiratory mutants, ammonium tolerance and conifers

Summary

Agriculture faces the considerable challenge of having to adapt to a progressively changing climate (including the increase in CO₂ levels and temperatures); environmental impact must be reduced while at the same time crop yields need to be maintained or increased to ensure food security. Under this scenario, increasing plants’ nitrogen (N) use efficiency and minimizing the energy losses associated with photorespiration are two goals of crop breeding that are long sought after. The plastidic glutamine synthetase (GS2) enzyme stands at the crossroads of N assimilation and photorespiration, and is therefore a key candidate for the improvement of crop performance. The GS2 enzyme has long been considered essential for angiosperm survival under photorespiratory conditions. Surprisingly, in Arabidopsis GS2 is not essential for plant survival, and its absence confers tolerance towards ammonium stress, which is in conflict with the idea that NH₄ ⁺ accumulation is one of the main causes of ammonium stress. Altogether, it appears that the ‘textbook’ view of this enzyme must be revisited, especially regarding the degree to which it is essential for plant growth under photorespiratory conditions, and the role of NH₄ ⁺ assimilation during ammonium stress. In this article we open the debate on whether more or less GS2 is a desirable trait for plant productivity.

Introduction

The glutamine synthetase/glutamate synthetase (GS/GOGAT) cycle is the pathway for the incorporation of inorganic nitrogen (N) into organic molecules. Glutamine synthetase catalyzes the conversion of glutamate (Glu) and ammonium (NH₄ ⁺) into glutamine (Gln). Then, GOGAT produces two molecules of Glu from Gln and 2‐oxoglutarate (Fig. 1). In seed plants, the GS family is composed of the cytosol‐localized GS1 and the plastid‐localized GS2. In a recent study, it was proposed that GS1 is divided into two evolutionary lineages, named according to their sequence and functional similarity to the gymnosperm GS1s: GS1a‐like and GS1b‐like (Valderrama‐Martín et al., ²⁰²²). GS1b is found in all seed plants and GS1a is present exclusively in gymnosperms and basal angiosperms. GS2 is present in all seed plants except conifers and gnetales (Valderrama‐Martín et al., ²⁰²²). In general, GS1 is encoded by multiple genes, whereas GS2 is, in diploid species, generally encoded by a single gene. For instance, the Arabidopsis thaliana genome harbors five GS1 genes (GLN1‐5) and one gene for GS2 (GLN2). Interestingly, some diploid species, such as Medicago truncatula, possess a second GS2 gene that is exclusively expressed in seeds (Seabra et al., ²⁰¹⁰). Although different GS isozymes exhibit specific functions, GS1 is generally considered to govern primary NH₄ ⁺ assimilation and its reassimilation during N remobilization and translocation. For GS2, its main role is the reassimilation of NH₄ ⁺ released during photorespiration and the assimilation of NH₄ ⁺ derived from nitrite reduction in the plastids (Bernard & Habash, 2009; Thomsen et al., ²⁰¹⁴; Hirel & Krapp, ²⁰²¹). Nitrogen is the major nutrient which limits crop productivity, and research on GS has therefore been extensive, with the aim of improving plants’ N use efficiency (NUE) via GS overexpression strategies (James et al., ^2018a; Amiour et al., ²⁰²¹). Among the different isozymes of GS, the genetic manipulation of GS2 has provided contrasting results that have always been interpreted in reference to its function in photorespiration (Table 1). In light of recent results obtained with GS2 mutants, the best strategy for the biotechnological use of GS2 (i.e. increasing or decreasing its expression to improve crop productivity) is a topic of ongoing debate.

Fig. 1.

Schematic illustrating the basics of ammonium assimilation in photosynthetic cells and a summary of the reported effects of the absence of GS2 (in mutants/conifers). Purple solid lines show pathways related to GS2 activity. Red dashed lines show pathways related to GS1 activity. Black solid lines show common pathways. 1 and 2 refer to photorespiratory and nonphotorespiratory conditions, respectively.

Table 1.

Catalogue of mutant and transgenic plant phenotypes engineered for higher or lower GS2 expression.

Species	Transgenic plant/mutant	Growth conditions	Phenotype	Reference
Hordeum vulgare cv Maris Mink	KO (azide)	Control conditions	Lethal	Wallsgrove et al. (1987) Blackwell et al. (1988)
		NPC	Similar to WT
H. vulgare cv Maris Mink	KO (GS2) × KO (Fd‐GOGAT) (azide)	Control conditions	Lethal	Blackwell et al. (1988)
		NPC	Similar to WT
Nicotiana tabacum	Overexpression 35S:NtGS2	High‐intensity light	Increased tolerance	Kozaki & Takeba (1996)
	Co‐suppression 35S:NtGS2		Increased sensitivity
Oryza sativa cv Kinuhikari	Overexpression 35S:OsGS2	Salinity	Increased tolerance	Hoshida et al. (2000)
		Cold stress	Increased tolerance
	Co‐suppression 35S:OsGS2	Salinity	Increased sensitivity
N. tabacum line SR1	Overexpression rbcS:NtGS2	Control conditions	Increased growth	Migge et al. (2000)
Lotus japonicus	KO (EMS)	Control conditions	Lethal	Orea et al. (2002)
		NPC	Similar to WT
Brassica napus	Reduced expression	Control conditions	Similar to WT	Husted et al. (2002)
	35S: antisense BnGS2
N. tabacum line SR1	Co‐suppression 35S:PsGS2	Control conditions	Reduced growth and chlorosis	Oliveira et al. (2002)
Arabidopsis thaliana Col‐0	Overexpression 35S:DvGS2 1	Control conditions	Increased growth	Zhu et al. (2014)
		Low N	Increased growth
O. sativa cv Zhongua 11	Co‐overexpression	Phosphinothricin	Enhanced resistance	Sun et al. (2005a)
	pOsAct1:PsGS1 + pZmUbi1:PsGS2
O. sativa	Co‐overexpression	N deficiency	Enhanced growth	Sun et al. (2005b)
cv. Zhongua 11	pOsAct1:PsGS1 + pZmUbi1:PsGS2
Triticum aestivum	Co‐overexpression	Phosphinothricin	Enhanced resistance	Huang et al. (2005)
	pOsAct1:PsGS1 + pZmUbi1:PsGS2
L. japonicus	KO (EMS)	Drought in NPC	Increased sensitivity	Díaz et al. (2010)
O. sativa cv Zhongua 11	Co‐suppression 35S:OsGS2	Control conditions	Reduced growth and chlorosis	Cai et al. (2010)
N. tabacum cv Xanthi	Overexpression rbcS:AtGS2	Low‐N condition	Increased growth	Wang et al. (2013)
	Co‐overexpression rbcS:AtGS2 + rbcS:Dof1;7		Increased growth relative to rbcS:AtGS2
T. aestivum cv Ji5265	Expression of GS2 allele from Xiaoyan 54 pTaGS2‐2Ab:TaGS2‐2Ab	High‐N field trial	Increased growth and yield	Hu et al. (2018)
		Low‐N field trial	Increased growth and yield
O. sativa cv Nipponbare	Co‐overexpression with GS1;1 pOsAct1:OsGS2 + pOsAct2 OsGS1;1	Osmotic stress	Increased tolerance	James et al. (2018b)
		Salinity	Increased tolerance
		Drought	Increased tolerance
		Phosphinothricin	Enhanced resistance
N. tabacum cv K326	Overexpression SP: TaGS2	Control conditions	Similar to WT	Wei et al. (2018)
		N starvation	Similar to WT
A. thaliana Col‐0	Knocked‐out (T‐DNA)	Normal air	Reduced growth and chlorosis	Ferreira et al. (2019)
		Salinity	Increased tolerance
		NPC	Reduced growth
N. tabacum cv K326	Overexpression SP: TaGS2	Drought	Increased tolerance	Yu et al. (2020)
A. thaliana Col‐0	Co‐suppressed 35S:AtGS2	Control conditions	Reduced growth, no chlorosis	Hachiya et al. (2021)
	Knocked‐out (T‐DNA)	Control conditions	Reduced growth, no chlorosis
		Ammonium stress	Increased tolerance

At, Arabidopsis thaliana; Bn, Brassica napus; Dv, Dunaliella viridis; EMS, ethyl methanesulfonate mutagenesis; KO, knockout; NPC, non‐photorespiratory condition; Nt, Nicotiana tabacum; Os, Oryza sativa; Ps, Pisum sativum; SP, super promoter; Ta, Triticum aestivum; WT, wild‐type; Zm, Zea mays.

¹ DvGS2 is a homologue of the GLN2 gene from the green algae Chlamydomonas reinhardtii and thus corresponds to a different evolutionary lineage than GS2 from seed plants (Valderrama‐Martín et al., ²⁰²²).

Is GS2 necessary for plant survival? A tale of photorespiration and photorespiratory mutants

Mutants lacking GS2 were first isolated in barley by screening a large ethyl methanesulfonate (EMS)‐mutagenized population in search of photorespiratory mutants (Blackwell et al., ^{1987, 1988}; Wallsgrove et al., ¹⁹⁸⁷). GS2 mutants showed severe stress symptoms, such as chlorosis and necrosis of the leaves, and finally died when grown under normal air conditions. Indeed, these mutants lacked the ability to reassimilate the NH₄ ⁺ lost during photorespiration and apparently died not because of the toxic buildup of NH₄ ⁺ but rather because of the drain on the organic nitrogen pool (Wallsgrove et al., ¹⁹⁸⁷). In agreement with this explanation, the mutants grew normally under a CO₂‐enriched atmosphere, where photorespiration is suppressed, leading the authors to conclude that GS2 is only necessary for plant survival when photorespiration is active. The finding that a plastidial enzyme was responsible for the reassimilation of NH₄ ⁺ produced in the mitochondria by the breakdown of glycine was puzzling. Indeed, due to the high flux of the photorespiratory route, a transporter that permits the import of photorespiratory NH₄ ⁺ into the chloroplast is likely needed. However, in spite of many efforts to identify such a transporter, it has not yet been found (Kuhnert et al., ²⁰²¹).

Extensive screening for photorespiratory mutants was also carried out in Arabidopsis, but mutants lacking GS2 were not isolated (Somerville & Ogren, 1982). Indeed, while mutants from other Arabidopsis photorespiratory enzymes, like Fd‐GOGAT, were identified, Arabidopsis GLN2 mutants were missing for > 35 years after the original screenings. This was especially surprising since GS2 mutants with a severe photorespiratory phenotype were also isolated in the model legume Lotus japonicus (Orea et al., ²⁰⁰²). The hypotheses that have been put forward in various attempts to explain the Arabidopsis GS2 enigma included, among others, a possible lethal phenotype of this mutation, and the presence of compensatory cytosolic GS activity (reviewed by Lam et al., ¹⁹⁹⁶). Finally, an Arabidopsis GLN2 mutant was described by Ferreira et al. (2019). The mutant showed lower growth under normal air (photorespiratory active conditions), but surprisingly did not show the strong and finally lethal phenotype observed in other species, and was even able to complete its life cycle. These results were also confirmed by Hachiya et al. (2021). Altogether, these observations break a long‐standing paradigm and demonstrate that GS2 function is not essential for plant survival in normal air, since, unlike L. japonicus and barley, GS2 absence in Arabidopsis is not lethal.

Photorespiration is an energetically expensive process that may have a strong impact on crop yields. At present, photorespiration is estimated to reduce US soybean and wheat production by up to 36% and 20%, respectively (Walker et al., ²⁰¹⁶). In spite of the differences in the phenotypes reported for GS2 mutants from different species (Fig. 1; Table 1), there is no doubt that this enzyme has a central role in photorespiration. In a future climate change scenario, elevated CO₂ will probably reduce photorespiratory rates (Walker et al., ²⁰¹⁶), and it is possible that high levels of GS2 might not be strictly necessary. The introduction of different photorespiratory bypasses that avoid NH₄ ⁺ release from glycine breakdown have been successful in increasing plant yield (Shen et al., ²⁰¹⁹; Cavanagh et al., ²⁰²²). Although bypassing GS2 will certainly reduce photorespiratory ATP losses, it will also prevent the protective role of this cycle under stress conditions. How photorespiration will evolve in the future and whether more or less GS2 is a desirable trait for plant productivity are still open questions.

Arabidopsis GLN2 mutants reveal that plastidic shoot NH₄ ⁺ assimilation may be detrimental to plants’ ammonium tolerance

Although NH₄ ⁺ is an essential intermediate for N incorporation into biomolecules, when plants are exposed to high soil NH₄ ⁺ concentrations they often display stress symptoms that include growth retardation and leaf chlorosis. When severe, these symptoms may even lead to plant death (Britto & Kronzucker, 2002). Obtaining crops that exhibit optimal performance under ammonium nutrition is of great interest, since boosting ammonium‐based nutrition, as opposed to nitrate‐based nutrition, has the benefit of mitigating NO₃ ^– leaching and N₂O greenhouse gas emission (Subbarao & Searchinger, 2021). The cause of ammonium stress‐derived symptoms is multifactorial and includes, among other factors, oxidative stress, pH alterations, energetic trade‐offs and cation imbalance. In general, the excessive NH₄ ⁺ accumulation in tissues has been conventionally accepted to be the main trigger underlying plants’ sensitivity to ammonium stress. In agreement with this idea, the promotion of NH₄ ⁺ assimilation has been generally shown to act as a tolerance‐promoting mechanism, and mutants defective in NH₄ ⁺ assimilation showed enhanced sensitivity towards ammonium nutrition, such as the Arabidopsis gln1;1:gln1;2, gln1;2, and gln1;2:gln1;3 GS1 mutants (Guan et al., ²⁰¹⁶; Konishi et al., ²⁰¹⁷), the rice OsGS1;1 knockout mutant (Kusano et al., ²⁰¹¹), and Arabidopsis mutants defective in NADH‐GOGAT (Konishi et al., ²⁰¹⁴). Surprisingly, Arabidopsis GLN2 mutant plants, in addition to being able to survive under normal air conditions, were also more tolerant to ammonium stress than wild‐type plants (Hachiya et al., ²⁰²¹). In agreement with our understanding of GS2 function, GLN2 mutant plants accumulated very high quantities of NH₄ ⁺. Obviously, these observations are in conflict with the idea that NH₄ ⁺ is a toxic molecule when present at high concentrations, and they indicate that NH₄ ⁺ assimilation, rather than NH₄ ⁺ accumulation, may be responsible for the sensitivity of Arabidopsis to ammonium nutrition. Likewise, Poucet et al. (2021) also reported that NH₄ ⁺ accumulation in the leaves of tomato plants grown under ammonium nutrition was dependent on leaf phenological stage and was not correlated to their reduced growth compared to leaves of plants grown with nitrate (NO₃ ⁻). Hachiya et al. (2021) described the phenotype of GLN2 mutants in relation to the prevention of shoot acidification associated with plastidic proton (H⁺) release during excessive NH₄ ⁺ assimilation by GS2. Indeed, this acidification did not occur in gln1;2:gln1;3 mutant plants (Hachiya et al., ²⁰²¹). These observations lead to a number of key questions which need to be answered in order to advance our understanding of plants’ metabolic adaptation to ammonium stress. First, it appears that the localization of NH₄ ⁺ assimilation – plastidic in the shoot vs cytosolic in the root – is associated with different functions, in terms of coping with ammonium stress. Among other observations, the contrasting response between GLN1 and GLN2 mutants (i.e. GLN1 mutants are sensitive to ammonium stress) implies the existence of a root‐specific mechanism to efficiently deal with H⁺ release that is yet to be elucidated. In addition, it might be possible that the toxic effect of NH₄ ⁺ overaccumulation is dependent on its subcellular localization, a hypothesis that needs to be explored. Finally, it should not be forgotten that NO₃ ^– assimilation, but not NH₄ ⁺, depends on the reducing power exported from the chloroplast through the malate valve driven by the flux through the photorespiratory pathway (Shi & Bloom, 2021). If photorespiratory levels are reduced under a future climate scenario, plants that use NH₄ ⁺ as a primary N source might have an advantage over plants that depend on NO₃ ⁻. Should this be the case, a reduction in GS2 levels (less need for photorespiratory capacity and probably increased ammonium tolerance) potentially appears to be a win–win approach.

Can gymnosperms shed light on the need for GS2?

Conifers and gnetales lack GS2 (Valderrama‐Martín et al., ²⁰²²) but they have high photorespiratory rates (Hanawa et al., ²⁰¹⁷). As an alternative to GS2, gymnosperms possess GS1a, which, although cytosolic, performs a GS2‐like function, namely photorespiratory NH₄ ⁺ reassimilation (Cánovas et al., ²⁰⁰⁷). Indeed, as is the case with GS2, GS1a is mainly expressed in photosynthetic tissues, and its gene expression is light‐dependent (Cantón et al., ¹⁹⁹⁹; Valderrama‐Martín et al., ²⁰²²). This fact, together with the findings regarding GLN2 mutants in Arabidopsis (Ferreira et al., ²⁰¹⁹; Hachiya et al., ²⁰²¹), demonstrate that GS subcellular location is not essential for plant survival, but probably for the coordination of GS gene expression with photosynthesis and photorespiration, as is the case for GS1a in conifers. In fact, Arabidopsis GLN2 mutants exhibit increased expression of GLN1;2 and GLN1;3 (Ferreira et al., ²⁰¹⁹).

Primary forest soils are generally acidic, with low nitrification rates, and NH₄ ⁺ is therefore the dominant form of N. In agreement, most conifers take up NH₄ ⁺ preferentially and are tolerant to ammonium nutrition (Cui & Song, 2007; Kronzucker et al., ¹⁹⁹⁷). However, recent findings have demonstrated that mature conifer trees can also assimilate NO₃ ^– efficiently in natural conditions (Zhou et al., ²⁰²¹). As described in the previous section, the absence of GS2 confers upon Arabidopsis tolerance to ammonium stress. Since the reduction of NO₃ ^– takes place in the plastids, we can hypothesize that when NO₃ ^– reduction is low, a plastidic GS2 is not essential. This hypothesis is in agreement with the enhanced ammonium tolerance observed in Arabidopsis GLN2 mutants (Hachiya et al., ²⁰²¹). Thus, the absence of GS2 in gymnosperms might be an evolutionary mechanism to promote growth in NH₄ ⁺‐rich habitats. This hypothesis must be studied through the analysis of chloroplast pH control in conifers and the production of conifer transgenic lines overexpressing a plastidic GS in different photorespiratory and nutritional conditions.

Is GS2 overexpression a strategy to increase yield?

Increasing plastidic GS expression has been shown to generally be beneficial for plant growth (Table 1) by improving, among other traits, those related to N use (Zhu et al., ²⁰¹⁴; Hu et al., ²⁰¹⁸). In addition, GS activity and expression are known to be enhanced by a number of abiotic stresses (Bernard & Habash, 2009; James et al., ^2018b). In particular, GS2 overexpression lines have generally shown increased tolerance to abiotic stresses (Table 1). One hypothesis that may explain the beneficial effect of GS2 overexpression is related to photorespiration acting as an electron‐sink that would dissipate excessive reducing power, thus protecting against the production of reactive oxygen species and conferring photoprotection (Kozaki & Takeba, 1996; Betti et al., ²⁰¹⁶; James et al., ^2018b). Alternatively, abiotic stresses are known to increase processes that promote NH₄ ⁺ release, such as protein degradation. Thus, a higher GS activity would be beneficial for the synthesis of osmolytes such as proline or polyamines. In agreement with this idea, L. japonicus GS2 mutant plants showed low proline synthesis and drought sensitivity (Díaz et al., ²⁰¹⁰), and rice plants co‐overexpressing OsGS1;1/OsGS2 accumulated osmolytes and showed enhanced tolerance to drought and salinity (James et al., ^2018b). Again, Arabidopsis appears to be an exception, and in a study by Ferreira et al. (2019), Arabidopsis GLN2 mutants displayed enhanced tolerance to salinity, but the metabolic adjustment associated with GS2 absence was not studied in this work. However, Hachiya et al. (2021) observed proline accumulation in Arabidopsis GS2 mutants under ammonium stress, which might explain the phenotypes reported by Ferreira et al. (2019). However, it remains unclear why proline levels in the context of GS2 expression levels are species‐dependent.

What to breed for? More or less GS2 activity?

Climate‐resilient crops are needed to maintain agricultural productivity, meaning that crops adapted to constantly increasing atmospheric CO₂ conditions that are also able to deal with higher temperatures and water scarcity are required. In addition, the impact of agriculture on the environment should be minimized, and one aim should be a reduction in the loss of N, which is often > 50% of the amount of applied N (Socolow, 1999; Coskun et al., ²⁰¹⁷). Thus, breeding crops with a higher NUE is desirable. In addition, the use of ammonium‐based nutrition combined with nitrification inhibition is of great interest, but this approach demands that crop plants are better adapted to the use of NH₄ ⁺ as their main source of N (Coskun et al., ²⁰¹⁷; Marino & Moran, ²⁰¹⁹; Subbarao & Searchinger, ²⁰²¹). GS2 stands at a crossroads of different processes, such as photorespiration, N assimilation and stress tolerance; therefore, breeding to alter GS2 expression levels represents a promising but also debatable strategy, especially in light of the conflicting data obtained for GS2 mutants and overexpression lines that have been summarized in this article (Table 1). Indeed, GS2 overexpression has generally been considered to be a means by which higher NUE can be achieved (Zhu et al., ²⁰¹⁴; Hu et al., ²⁰¹⁸). However, recent reports in Arabidopsis showed enhanced growth of GS2 mutants under ammonium nutrition, and this finding, together with the fact that ammonium‐tolerant conifers lack GS2, suggests that plastidic GS2 absence may represent a benefit for plant performance when grown under ammonium nutrition. The future rise in atmospheric CO₂ concentrations should reduce photorespiration, which is assumed to be positive for plant productivity (Walker et al., ²⁰¹⁶). However, elevated temperatures and increasingly frequent episodes of drought will also influence photorespiration (Betti et al., ²⁰¹⁶), making it difficult to predict how photorespiratory activity will change in the future. These puzzling results open a debate regarding the interest of breeding for more or for less GS2 in order to optimize crop yield and quality. Interestingly, it appears that in conifers and Arabidopsis photorespiratory NH₄ ⁺ could be assimilated by alternative enzymes, such as cytosolic GS. In this sense, engineering crops with slightly modified levels of this enzyme may be a promising approach. Besides, another promising strategy to be explored is breeding for increased GS1 expression while minimizing GS2 expression. Altogether, more research is needed to fully understand GS2 function in different plant species. For example, the role of GS2 under ammonium nutrition must be studied in species other than Arabidopsis in order to observe whether this phenotype is exclusive to Arabidopsis or conserved in other species from the Brassicaceae and in other plant families. Obviously, a rapid approach could be to grow the available GS2 mutants or co‐suppressed lines under nonphotorespiratory conditions with NH₄ ⁺ as the main source of N. While it is evident that there are several layers of complexity to the GS2 puzzle due to the involvement of this enzyme in several key processes, we believe that the recent advances pave the way to newly promising strategies for crop improvement by engineering GS2 expression.

Author contributions

DM conceived the article; DM, RAC and MB wrote the manuscript.