Biotechnological approaches to develop nitrogen-fixing cereals: A review

Agricultural yields are often limited by nitrogen (N) availability, especially in countries of the developing world, whereas in indus-trialized nations the application of chemical N fertilizers has reached unsustainable levels that have resulted in severe environmental consequences. Finding alternatives to inorganic fertilizers is critical for sustainable and secure food production. Although gaseous nitrogen (N 2 ) is abundant in the atmosphere, it cannot be assimilated by most living organisms. Only a selected group of microorganisms termed diazotrophs, have evolved the ability to reduce N 2 to generate NH 3 in a process known as biological nitrogen fixation (BNF) catalysed by nitrogenase, an oxygen-sensitive enzyme complex. This ability presents an opportunity to improve the nutrition of crop plants, through the introduction into cereal crops of either the N fixing bacteria or the nitrogenase enzyme responsible for N fixation. This review explores three potential approaches to obtain N-fixing cereals: (a) engineering the nitrogenase enzyme to function in plant cells; (b) engineering the legume symbiosis into cereals; and (c) engineering cereals with the capability to associate with N-fixing bacteria.


Introduction
Human population is dramatically increasing. The current world population of 7.6 billion is projected to reach more than 9 billion in 2050 (UN-DESA, 2017) and consequently demand for food is rising. On a global scale, up to 70% of mankind's food depends directly on cereals, mainly rice, wheat, corn and sorghum. According to Food and Agriculture Organization of United Nations (FAO), food production must increase by 70% to feed this larger population, which means an increase of 44 million tons/ year; that is, a 38% increase over the historical average since there are records of agricultural productivity (FAO, 2020).
Nitrogen (N) availability is one of the main limiting factors for crop yield, especially in developing countries, where small farmers often do not have access to inorganic fertilizers. In contrast, the application of chemicals in developed nations has allowed high agricultural production, but has led to accumulation of inorganic nutrients in agricultural soils. Typical N-use efficiencies for wheat, rice, and maize indicate that more than 50% of N applied in fertilizers is lost to the environment (Lahda et al., 2016;Anas et al., 2020), either in the form of nitrous oxides, which are potent greenhouse gases, or as soluble nitrates, which causes eutrophication of aquatic systems and problems for human health due to their accumulation in ground water (Galloway & Cowling, 2002;Townsend et al., 2003;Glendining et al., 2009;Erisman et al., 2015;Stokstad, 2016). Globally, almost 100 million tons of N is deposited every year in terrestrial, freshwater and marine environments, i.e., a three-fold increase over preindustrial levels (Galloway et al., 2008;Rockström et al., 2009). Nutrient excesses are especially large in China, northern India, the USA, and Western Europe, leading to widespread nutrient pollution (Foley et al., 2011).
The use of N fertilizers obtained through the Haber-Bosch process dominates the fertilization of crops on a global scale, producing more than 100 million tons of nitrogen fertilizer per year, but it consumes a great quantity of energy. In fact, the energy-intensive production of nitrogen fertilizers is the greatest consumption of fossil fuels in agriculture, with predictions that this process will consume around 2% of global energy by 2050 (Glendining et al., 2009). For this reason, considering the future menaces of a decline in petroleum reserves and the low efficiency with which cereal crops use chemical nitrogen fertilizers, the search for new alternative sources of nitrogen to reduce agricultural reliance on nitrogen fertilizers is an urgent need. In the past decades one of the strategies to improve assimilation and use efficiency of nitrogen in cereals has been focused on overexpressing important genes for nitrate and ammonium transporters in the roots of rice, maize and wheat, although with different success (reviewed in . In the case of cereal crops under Mediterranean conditions, it has been shown (Ramos et al., 1989;García del Moral et al., 1999) that a foliar application of elemental sulfur could reduce the need for nitrogen fertilizer by 25%, since SO 2 generated would increase the methionine content and in turn the formation of ethylene, thus favoring the survival of a greater number of ears/m 2 and therefore increasing the grain yield.
Although atmospheric nitrogen represents 78% of the air, due to the stability of the triple bond between the two nitrogen atoms, N 2 is inaccessible to eukaryotes, since only a group of bacteria and archaea (called diazotrophos organisms) have developed the ability to fix N 2 organic through the enzyme nitrogenase. This biological nitrogen fixation (BNF) is the major contributor to the N economy of the biosphere, accounting for 30-50% of the total N in crop fields (Ormeño-Orrillo et al., 2013) and represents a promising substitute for chemical N fertilizers (Olivares et al., 2013;Dent & Cocking, 2017;Good, 2018). BNF is thought to be one of the most ancient enzyme-catalyzed reactions (Raymond et al., 2004) and depending on their way of life, diazotrophos organisms can be divided into three groups (Table 1): (1) free living; (2) symbiotic, mainly bacteria living within plant root nodules; and (3) those that live in associative or endophytic relationships with other organisms.
There are three biotechnological approaches currently being explored that could deliver fixed N from these diazotrophos organisms to cereal crops (Beatty & Good, 2011;Oldroyd & Dixon, 2014;Burén & Rubio, 2018). One option focuses on directly engineering the nitrogenase enzyme into organelles of plant cells to create a new N-fixing capability. This is an attractive solution, but it would need solving two great challenges. Firstly, the nitrogenase is a highly complex enzyme and would require the coordinated expression in the plant cell of at least 16 N 2 -fixation (nif) genes (Temme et al., 2012;Li & Chen, 2020). Secondly, nitrogenase activity has high energetic demand and is irreversibly denatured by oxygen from photosynthesis (Seefeldt et al., 2009;Curatti & Rubio, 2014). Thus, expressing functional nitrogenase in chloroplasts requires temporal (day/night) separation of photosynthesis and N fixation by confining nif gene expression only to dark periods (nights) or, alternatively, by spatially restricting nif gene expression to non-photosynthetic tissues such as the cereal roots.
Legumes have evolved the capability to associate with N-fixing bacteria, which are housed inside nodules on its roots and this ability offers the second possible scenario, i.e., engineering a N-fixing symbiosis in cereal roots through transferring the legume-rhizobial interaction. However, the process of N fixation through symbiosis is very complex, involving multiple events and their regulation in both the host and the rhizobia Therefore, engineering a N-fixing symbiosis will requires adapting existing signaling and developmental mechanisms to provide a suitable environment for nitrogenase activity in the new cereal nodules (Oldroyd & Dixon,3 Biotechnological approaches to develop nitrogen-fixing cereals: A review 2014; Rogers & Oldroyd, 2014;Mus et al., 2016;Goyal et al., 2021).
A third approach could involve the enhancement of naturally occurring plant-associated diazotrophs by generating strains that release fixed N to benefit the crop (Beatty & Good, 2011;Geddes et al., 2015;Stokstad, 2016;Mus et al., 2016). Manipulation of soil diazotrophs can potentially provide a short-term solution to reduce the use of synthetic N fertilizers in agriculture.
This review provides comprehensive and updated information on these different approaches to reducing the N fertilizers demand through improving BNF: (1) direct transfer of bacterial nif genes for expressing heterologous nitrogenase in cereals; (2) engineering new symbioses between cereals and N 2 -fixing bacteria in a similar form to the legume-rhizobium symbiosis; and (3) improvement of N 2 -fixing bacterial endophytes naturally associated to cereals.

The nitrogenase enzymatic complex
The molybdenum nitrogenase is a complex enzyme consisting of two proteins (Seefeldt et al., 2009;Hu & Ribbe, 2015;Addo & Dos Santos, 2020) (Fig. 1). The dinitrogenase reductase or NifH protein, also known as the Fe protein, is a homodimer of the nifH gene product that contains one Mg·ATP-binding site in each subunit. The catalytic dinitrogenase component or NifDK, also termed as the MoFe protein, is a heterotetramer of the nifD and nifK gene products. Fe protein accepts electrons from reduced flavodoxin II and acts as obligate electron donor to MoFe protein, whereas substrate reduction takes places at the FeMo-co inside each MoFe subunit. The catalytic process requires 16 moles of ATP for every mol of dinitrogen gas that is converted to 2 moles of ammonia, as well as reduction equivalents that are supplied by the reduced ferredoxin (Seefeldt et al., 2012;Curatti & Rubio, 2014;Taiz et al., 2015;Nag et al., 2019) (Fig. 1). However, the real cost is estimated to be 20-30 ATP, accounting for the production of the nitrogenase complex, the reductive power, and recycling the toxic dihydrogen waste resulting from the process (Lodwig & Poole, 2003).
Nitrogenase contains three metalloclusters. One of them, the iron-molybdenum cofactor (FeMo-co), located within the MoFe protein, provides the catalytic site for N reduction and contains molybdenum, iron, sulphur, a central carbon atom and homocitrate as an organic compound. FeMo-co is one of the most complex heterometal groups known in biology and its biosynthesis requires several nif products. The other two metalloclusters are a single [4Fe-4S] cluster located at the subunit interface of the Fe protein and one P-cluster at MoFe protein.
These metalloclusters are required for inter-protein and intra-protein electron transfer and reduction of N 2 , in a process that is energetically coupled to Mg·ATP hydrolysis (Seefeldt et al., 2012;Hu & Ribbe, 2015;Curatti & Rubio, 2014;Sickerman et al., 2017). Both Fe protein and MoFe protein are irreversibly inactivated by oxygen (Fig. 1). Although all diazotrophs studied to date contain the molybdenum-dependent nitrogenase, a subset of diazotrophs also have alternative nitrogenases, namely vanadium dependent and iron-only nitrogenases, as recently revealed by genomic analysis (Addo & Dos Santos, 2020).

Engineering nitrogenase biosynthesis in eukaryotic cells
This strategy involves transfer of prokaryotic nif genes so that the eukaryotic cell synthesizes its own N 2 -fixing machinery without the need for bacterial interactions. However, this approach faces three major difficulties: (a) the complexity and fragility of nitrogenase biosynthesis; (b) the high sensitivity to O 2 of nitrogenase and many of the accessory proteins and metal clusters needed for maturation of the nitrogenase components; and (c) the difficulty of coupling plant metabolism to supply energy and reducing power for the N fixation process. The transfer of fixation capacity to a non-diazotroph organism was first achieved in 1971, with the transfer of a nif cluster from Klebsiella oxytoca (formerly Klebsiella pneumonie) into Escherichia coli (Dixon & Postgate, 1971). Subsequently, various transgenic bacteria capable of fixing N have been obtained, which have made it possible to determine the minimum set of nif genes, as well as the requirements necessary to engineer a functional nitrogenase in plant cells (Curatti & Rubio, 2014;Burén & Rubio, 2018;Pankievicz et al., 2019). The genetics of N fixation was initially elucidated in the model N 2 -fixing bacteria K. oxytoca, where the nif genes required for the synthesis of nitrogenase are clustered in a 24 kb region of the chromosome (Arnold et al., 1988). In this bacterium, in addition to the nifH, three additional nif genes are required to assemble a functional NifH protein: nifU, nifS and nifM. The NifU and NifS gene products constitute an [Fe-S] cluster assembly machinery specialized in synthesizing clusters for the nitrogenase component proteins, whereas the nifM gene encodes a peptidyl-prolyl cis-trans isomerase required for NifH maturation. The synthesis of the functional NifDK protein requires, in addition to the NifD and NifK gene products, the generation and proper assembly of the FeMo-co metallocluster, which is known to occur outside of apo-Ni-fDK (Dixon & Kahn, 2004;Curatti & Rubio, 2014;Nag et al., 2019;Li & Chen, 2020). To achieve complete Fe-Mo-co synthesis, it is necessary to express, at a minimum, the nifB, nifE, nifN, and nifH genes whose products are required for the synthesis of the Fe-S core of FeMo-co. The Nif proteins that are required to assemble FeMo-co are NifU and NifS, which provide two pairs of [4Fe-4S] clusters to the NifDK polypeptides, and the NifH protein that drives the reductive coupling of each pair of [4Fe-4S] clusters to form the P-clusters (Dixon & Kahn, 2004;Curatti & Rubio, 2014;Li & Chen, 2020). Table 2 lists the different known genes involved in the conformation of the nitrogenase enzyme complex and its function in Klebsiella oxytoca. The genetic components and arrangement of nif genes vary greatly across the diverse range of diazotrophs reflecting the environmental niche and physiology of the particular organisms (Li & Chen, 2020;Jiang et al., 2021). The rapid expansion of microbial genome sequencing in the last few years has offered novel opportunities to reexamine the distribution of N fixation genes among the known diaztrophos and for the identification of a minimum gene set for N fixation. Thus, in a review of 1200 genomes of different species of bacteria and archaea with fully sequenced genomes, Dos Santos et al. (2012) have found that nearly all known diazotrophs contain a minimum of six conserved genes: nifH, nifD,nifK, nifE, nifN, and nifB. The co-occurrence of these six nif genes, known to be essential for N fixation in characterized systems, has led to these authors to propose this minimum gene set as an in silico tool for the identification both of known diazotrophs as those that could be identified in new research.
Another problem added to the complexity of engineering nitrogenase expression is the level and timing of expression for each individual nif gene transferred into the plant cell. Nitrogenase is a very slow enzyme, and a N 2 -fixing cereal plant might require the accumulation of considerable amounts of NifH and NifDK component proteins. In fact, for N fixation up to 20% of the total proteins are dedicated to nitrogenase production and maturation (Batista & Dixon, 2019). This is a burden for the cell, diverting energy from other pathways, causing oxidative stress and excess of intracellular NH 4 + which can be toxic to the plant cell. However, this can be controlled by expressing the nif genes with the help of heterologous promoters and manipulation or fine tuning of the NifL-NifA regulators (Dixon & Kahn, 2004;Curatti & Rubio, 2014;Li & Chen, 2020).
The transfer of nif genes to engineering the nitrogenase complex in eukaryotes was initially studied in yeasts (a  non-photosynthetic organism), where has it been possible to engineer various genes nif from Azotobacter vinelandii to generate an active Fe protein (López-Torrejón et al., 2016) and the successful formation of NifDK tetramer (Burén et al., 2017b), essential first steps in assembling a functional nitrogenase in an eukaryotic cell. However, although all the Nif components have been successfully expressed in yeast cells, the formation of a fully functional nitrogenase complex has not yet been achieved (López-Torrejón et al., 2016;Burén et al., 2017a;Pérez-Gonzalez et al., 2017). In higher plants there are some attempts to express nif genes in chloroplasts and mitochondria. Thus, Ivleva et al. (2016) introduced the nifH gene together with nifM from K. oxytoca into chloroplasts of tobacco plants, generating functional NifH protein, although with low activity. Similarly, Allen et al. (2017) demonstrated the viability of expressing the complete range of biosynthetic and catalytic nitrogenase Nif as mitochondrial proteins in tobacco leaves. Table 3 shows some of the attempts to engineering the nif genes in various eukaryotes organisms.

Engineering nitrogenase for prevent its inactivation by oxygen
Regarding the sensitivity of nitrogenase to O 2 , it is important to note that filamentous cyanobacteria have been able to reconcile oxygenic photosynthesis and N 2 fixation either by spatial or by temporal separation. Spatial separation is achieved by expressing nitrogenase exclusively inside specific cells termed heterocysts without Photosystem II and therefore O 2 is not generated during photosynthesis (Flores & Herrero, 2010). Temporal separation in cyanobacteria is achieved through control by the circadian rhythm (Chen et al., 1998).
In the plant cell the possible subcellular locations for the nitrogenase proteins are plastids and mitochondria . The advantages of nif gene expression in plastids are: (a) they possess the prokaryotic-type transcription and translation machineries that allow the use of bacterial promoters and gene clusters in the form of operons; (b) the high levels of gene expression and protein accumulation that can be achieved due to the high degree of ploidy of the chloroplastid genome and (c) the local production of ATP and reducing power required for nitrogenase function (Dixon et al., 1997;Scharff & Bock, 2014;Liu et al., 2018). However, a major disadvantage would be the energy costs of synthesizing nitrogenase proteins every night, as photosynthetic O 2 will result in their irreversible inactivation during the day (Scharff & Bock, 2014).
The knowledge that eukaryotic mitochondria harbor machineries for [Fe-S] cluster assembly highly similar to those involved in the early steps of nitrogenase metallocluster biosynthesis, has recognized these organelles as possible candidates for the transfer of nif genes to the plant cell (Allen et al., 2017;Liu et al., 2018 operates in the mitochondrial matrix seems because respiration results in O 2 depletion inside this organelle, allowing thus the biosynthesis of O 2 -sensitive proteins. In addition to respiratory protection, mitochondria can provide the ATP and reducing power required for nitrogenase catalytic activity (Dixon & Kahn, 2004;Scharff & Bock, 2014;Liu et al., 2018). However, the current lack of an efficient system for in vivo transformation of mitochondria is currently a major obstacle to the transfer of nif genes to these organelles. The previous considerations exemplify the complexity of the introduction and expression of N fixation in aerobic plant cells.

Engineering legume symbiosis in cereals
A number of species of plants, most notably legumes, have evolved the ability to form intimate N-fixing symbioses with diazotrophs and form specialized organs (nodules). The best studied plant endosymbiosis are those between legumes and rhizobia. Through this mutualistic relationship, plant cells provide nutrients to the bacteria in exchange for ammonia produced by nitrogenase and provide a suitable oxygen-limited environment for N fixation, because inside the nodule the oxygen concentration is maintained at 20 to 40 nM, levels that can support respiration but are sufficiently low to avoid inactivation of nitrogenase (Oldroyd & Dixon, 2014). The most abundant proteins in nodules are oxygen-binding heme proteins called leghemoglobins with a high affinity for oxygen. Leghemoglobines provide a buffer for nodule oxygen inside the nodule and participate to increasing the rate of oxygen transport to the respiring symbiotic bacterial cells (Larrainzar et al., 2020).
The establishment and functioning of this effective symbiosis is dependent on genetic determinants in both plant and bacteria (Oldroyd et al., 2011;Goyal et al., 2021). Accurate mutual recognition in legume-rhizobia systems is accomplished by exchange of biochemical signals. This signaling, the subsequent infection process, and the development of N-fixing nodules involve specific genes in both the host and the rhizobia (Rolfe & Gresshoff, 1988;Oldroyd et al., 2011;Ibañez et al., 2017;Oldroy & Poole, 2019;Goyal et al., 2021). Plant genes specific to nodules are called nodulin genes and rhizobial genes that participate in nodule formation are named nodulation (nod) genes. The nod genes are classified as common nod genes or host-specific nod genes. The common nodA, nodB, nodC and nodD genes are found in all rhizobial strains, whereas the specific-host genes nodP, nodQ, nodF, nodE or nodL differ among rhizobial species and determine the host range of legumes that can be infected. Only one of the nod genes, the regulatory nodD is constitutively expressed and its protein product NodD regulates the transcription of the other nod genes (Oldroyd et al., 2011;Goyal et al., 2021). The plant signals that induce nod gene expression are commonly flavonoids and isoflavonoids (e.g., the flavone 7,4 dihydroxyflavone and the isoflavone genistein) and betaines secreted by the roots. These attractants activate the rhizobial NodD protein (activator) which then induces transcription of the other nod genes. The nod genes, which NodD activates, code for nodulation proteins involved in the biosynthesis of lipochitin oligosaccharides or Nod factors, molecular signals that trigged nodulation-related changes in the host plant (Oldroyd & Dixon, 2014;Pankievicz et al., 2019;Dong & Song, 2020;Goyal et al., 2021). This activates a downstream gene cascade including those involved in nucleoporin, cation channels, early nodule expression, and cytokinin signaling leading to cortical and pericyclic cell divisions. Rhizobia are then entrapped by root hair curling after the Nod factors has been perceived, which results in initiating the formation of infection thread (a tubular structure) that facilitates the penetration of bacteria into root hair and adjacent cortical cells (Venado et al., 2020;Tsyganova et al., 2021).
The fact that most land plants, including cereals, can form arbuscular mycorrhizal associations but are unable to form root nodules that can fix N, could be a way to transfer symbiotic N fixation to non-legume plants. In fact, recent phylogenomic studies suggest that a small set of genes could convert a species in association with arbuscular mycorrhizal fungi into a N-fixing symbiont (Griesmann et al., 2018;Van Velzen et al., 2018). Interestingly, several components of the legume symbiotic signaling (SYM) pathway also play a role in the arbuscular mycorrhizal symbiosis. The key symbiotic signals produced both by rhizobia (Nod factors) and arbuscular mycorrhizal fungus (Myc factors) are lipochitin oligosaccharides in a process that involve related signaling pathways and receptors (Dénarié et al., 1996;Froussart et al., 2016). Since cereals contain the signaling pathway for arbuscular mycorrhizal associations, nodulation could be established in them by engineering the perception of rhizobial signaling molecules to activate this recognition pathway and to form in roots an oxygen-limited nodule for N fixation (Mus et al., 2016;Oldroy & Poole, 2019). However, to achieve such synthetic N-fixing symbiosis there are several factors that must be taken into account (Oldroyd & Dixon, 2014;Mus et al., 2016): (1) optimization of the colonization process; (2) engineering of symbionts synthetic nif clusters optimized for N fixation; (3) engineering of respiratory protection and O 2 -binding proteins to allow aerobic N fixation by symbionts; (4) conditional suppression of ammonium assimilation by symbionts to ensure adequate N delivery to plants; (5) ensured effective uptake of ammonium by plant cells; and (6) optimization of carbon supply from root cells to endosymbiotic bacteria. From the point of view of genetic engineering the main factors lies with the challenge to control the expression of multigene systems and complex coding sequences; with the limited availability of a wide range of well characterized promoter elements in cereals, because the use of the same promoter to express multiple genes in transgenic cells can induce gene silencing; with the construction of large multigene synthetic cassettes; and with the need to develop highly efficient transformation methods for cereals (Rogers & Oldroyd, 2014;Mus et al., 2016), although in the last two decades notable progress has been made in the transformation of different species of cereals (see a recent review in Hensel, 2020).
One potential criticism for moving the legume symbiosis into cereals is the possible reduction in yield caused by the increased demand on photosynthates required to support N fixation. This could be disadvantageous under intensive agricultural systems in the developed world, but not in low-input agricultural systems, where the limits on nutrient availability (and especially N) are the main cause of low agricultural yield, as smallholder farming in sub-Saharan Africa,

Improvement of N 2 -fixation through bacterial endophytes naturally associated to cereals
In this strategy, plant-growth-promoting rhizobacteria (PGPR) with beneficial effects on plant development and already naturally associated with cereals are modified to improve their colonization ability, density, N 2 -fixing capabilities and release of NH 3 produced to plant cells (Stoltzfus et al., 1997;Savka et al., 2002;Santi et al., 2013;Ryu et al., 2020). Such PGPR can be loosely associated in close proximity to the plant root or invade and spread within the plant tissue as endophytic diazotrophs (Santi et al., 2013;Mus et al., 2016). These last may have an advantage over root-surface associative diazotrophs, as they colonize the interior of plant roots and can establish themselves in niches that provide more appropriate conditions for effective N fixation and subsequent transfer of the fixed N to the host plant (Reinhold-Hurek & Hurek, 2011). In wheat, rice and maize, various species of the genera Azoarcus, Acetobacter, Pseudomonas, Azospirillum, Glucenobacter and Azotobacter have been identified as PGPR (Santi et al., 2013;Ryu et al., 2020) exerting significant effects in increasing plant height, root length and dry-matter production not only through N fixation, but also through other physiological effects on the growth and development of the host plant (Rosenblueth & Martinez-Romero, 2006;Saharan & Nehra, 2011;Aasfar et al., 2021). Indeed, most of PGPR produces the auxin indole-3-acetic acid (IAA) together with cytokinins such as isopentyladenosine ([9R]iP), two hormones that stimulates root branching and root elongation which in turn favour the uptake of soil water and minerals with a positive effect on plant growth (Baca & Elmerich, 2007). Other effects of PGPR include the production of siderophores that may stimulate plant growth directly by increasing the availability of iron in the soil surrounding the roots, phosphate solubilization and acquisition of other nutrients like calcium, potassium, iron, copper, magnesium, and zinc (Richardson et al., 2009;Pérez-Montaño et al., 2014;Fukami et al., 2018;Pankievicz et al., 2019). However, since the population density of endophytic bacteria in plant tissues is too low to support adequate N fixation, it is important to design systems that aid greater colonization of diazotrophic endophytes, for instance, by engineering plants to produce a specific metabolite and thus create a "biased rhizosphere" to favor the growth of an introduced diazotroph able to use the novel metabolite (Rosenblueth et al., 2018;Nag et al., 2019). An interesting finding is that some diazotrophs, including Herbaspirillum species, living in mucilage released from the aerial roots of maize landraces from Sierra Mixe, Mexico, can provide up to 82% of the host N at a critical period of the growing season (Van Deynze et al., 2018).
This strategy will involve the identification of appropriate plant and bacterial signals, receptors, and target genes. Rhizopines are a rare group of compounds produced by a few species of rhizobia inside legume nodules and are exuded into the rhizosphere. Among them, the scyllo-inosamine 1 (SIA) and 3-O-methyl-scyllo-inosamine 2 (3-O-MSI) are believed to be suitable as chemical signals between plants and rhizosphere bacteria (Murphy et al., 1995;Gordon et al., 1996;Savka et al., 2013;Goyal et al., 2021) and recently has been reported the successfully transfer of rhizopine biosynthesis genes into barley (Geddes et al., 2019). A step forward in this strategy consists of plants engineered to release an orthogonal chemical signal (as nopaline or octopine) that could be sensed only by a corresponding engineered bacterium with the appropriate sensors, which would have the added benefit of only inducing nitrogenase in the presence of the engineered crop (Ryu et al., 2020).

Conclusions
N 2 -fixing cereals would be an enormous biotechnological challenge that might revolutionize world agricultural systems. The most important difficulties involved in the direct transfer of bacterial nif genes into the cereal are the sensitivity of nitrogenase to O 2 and the complexity and fragility of nitrogenase biosynthesis. Nitrogen fixation is a highly energy demanding process, and so chloroplasts and mitochondria are envisaged as potential subcellular locations sites for N fixation since they can meet the energy requirements for nitrogenase, each one showing advantages and disadvantages. Because nitrogenase is extremely sensitive to oxygen evolved by chloroplasts during photosynthesis, expression of functional nitrogenase in chloroplasts requires temporal (day/night) separation by confining nif gene expression only to dark periods (nights) or, alternatively, by spatially restricting nif gene expression to non-photosynthetic tissues such as the roots. The possibility of developing legume like root-nodule symbioses in cereals arises from contemporary knowledge that cereals contain the signaling pathway to form arbuscular mycorrhizal associations with diverse diazotrophic rhizobia. However, there is still insufficient knowledge about the genetics and microbiology involved in the formation of an oxygen-limited nodule for N fixation in cereal roots. An alternative option is the use of mixed PGPR and N-fixing bacteria to develop cereals with improved root growth and better exploitation of environmental and nutritional resources. To achieve this, it is critical to design systems that aid greater colonization of diazotrophic endophytes to improve the chances that the inoculated diazotroph will selectively colonize the crop plant, because N fixation is highly variable depending on the associated diazotroph and the plant variety,