Oxygen, a key signalling factor in the control of seed germination and dormancy

Abstract Oxygen is a major factor of seed germination since it allows resumption of respiration and subsequent metabolism reactivation during seed imbibition, thus leading to the production of reducing power and ATP. Most studies carried out in the 60s to 85s indicate that oxygen requirement depends on the species and is modulated by environmental factors. They have also demonstrated that the covering structures mainly inhibit germination by limiting oxygen supply to the embryo during imbibition through enzymatic oxidation of phenolic compounds by polyphenol oxidases (catechol oxidase and laccase) and peroxidases. Recent use of oxygen-sensitive microsensors has allowed to better characterize the oxygen diffusion in the seed and determine the oxygen content at the level of embryo below the covering structures. Here, I will also highlight the major data obtained over the last 30 years indicating the key role of oxygen in the molecular networks regulating seed germination and dormancy through (1) the hormonal balance (ethylene, ABA and GA), the hormone-signalling pathway and, in particular, the ABA sensitivity, (2) the emerging role of mitochondria in ROS production in hypoxia and (3) the involvement of the N-degron pathway in the turnover of proteins involved in seed tolerance to hypoxia.

, my scientific work concerned the following main topics: (1) Physiological, biochemical and hormonal regulation of seed germination and dormancy as related to external factors, in particular temperature and oxygen; (2) Regulation of ethylene biosynthesis and involvement of this plant hormone in germination and breaking of dormancy; (3) Physiological and biochemical mechanisms involved in the response of seeds and seedlings to abiotic stresses (dehydration, chilling, high temperatures); (4) Ageing and storage of orthodox and recalcitrant seeds and (5) Evaluation and improvement of seed quality.
As oxygen was always a key factor of the scientific work that I have done in relation to seed photosensitivity (Corbineau, 1983;Corbineau and Côme, 1985), metabolism, energy charge (EC) and the pentose-phosphate pathway, expression and breaking of seed dormancy (Corbineau and Côme, 1995), I have chosen to talk about oxygen as a key signalling factor in the control of seed germination and dormancy for this specific J. Derek Bewley, Career Lecture. Seed germination is subjected to a very precise regulation, the complexity of which originates both in the environmental factors and the seeds themselves. Water, oxygen and temperature are the three essential factors regulating seed germination and dormancy expression (Bewley and Black, 1994;Corbineau and Côme, 1995;Baskin and Baskin, 1998;Finch-Savage and Leubner-Metzger, 2006;Finch-Savage and Footitt, 2017). Water uptake is required in order to allow resumption of seed metabolism among which respiration, protein and RNA synthesis, plant hormone (in particular abscisic acid, gibberellins and ethylene) biosynthesis and signalling pathways (Leopold and Vertucci, 1989;Bewley and Black, 1994;Bradford, 1995;Hadas, 2004;Benech-Arnold et al., 2006;Bradford et al., 2008;Holdsworth et al., 2008;Nonogaki, 2010Nonogaki, , 2014Nonogaki, , 2017Nonogaki, , 2019Nonogaki et al., 2010).
Oxygen is essential for respiration but its requirement for seed germination depends on the species, depth of dormancy and temperature Côme et al., 1988;Corbineau and Côme, 1995;Bradford et al., 2007Bradford et al., , 2008. In anoxia/hypoxia, the production of ATP by fermentation allows the maintenance of EC in the seeds Côme and Corbineau, 1989). However, partial or total oxygen deprivation may be beneficial to germination of various aquatic species or some dormant seeds and complete deprivation of oxygen can break dormancy of some seeds Côme et al., 1991;Corbineau and Côme, 1995;Baskin and Baskin, 1998).
In the present paper, major data obtained from 60 to 85s that mainly take the form of descriptive physiology (1) investigating the sensitivity of seeds to oxygen supply and the role played by surrounding structures in the diffusion of oxygen from the medium to the embryo (Côme, 1967(Côme, , 1970Côme and Tissaoui, 1973;Corbineau, 1992, 1998;Corbineau and Côme, 1995) and (2) measuring the respiration activity and energy metabolism during germination of non-dormant and dormant seeds incubated in low oxygen supply Côme and Corbineau, 1989;Corbineau and Côme, 2003) are highlighted.
Diffusion of oxygen from the atmosphere, in the soil and in the seed The soil atmosphere composition Partial pressure of oxygen depends on soil structure and compaction (Hadas, 2004). In agricultural soils, oxygen content has been found to vary from 19 to 21%, thus indicating that it is not limiting if the soil is well drained (Richard and Boiffin, 1990). However, the oxygen level can decrease to 10% when a crust is present at the soil surface leading to reduced exchange between the atmosphere and the soil (Richard and Guerif, 1988a,b;Dekker, 2000;Hadas, 2004). In flooded soil when the pore system is filled with water (Gambrell et al., 1991), the oxygen level can be less than 1% since there is a 10 4 reduction in oxygen diffusion in water relative to air, and oxygen solubility in water decreases with increasing temperature being 14.56 mg l −1 at 0°C, 11.25 mg l −1 at 10°C, 9.09 mg l −1 at 20°C and 7.49 mg l −1 at 30°C.
The level of CO 2 does not usually exceed 0.5-1% in drained soil, but can increase up to 5-8% in flooded soils and silt loams when decomposition of organic matter occurs (Buyanovsky and Wagner, 1983;Nakayama and Kimball, 1988).

Seed coat structures: barriers to oxygen diffusion to the embryo
Embryo covering structures include nucellus and testa and endosperm in some species. The pericarp is fused to the seed testa in the case of caryopses, and it is not fused to the testa in the case of achenes. They are associated with various other structures among which hulls consisting of glumellae or glumes depending on the Poaceae species, and dead perianth tissues, for example in beet. They inhibit seed germination by controlling seed water uptake and oxygen supply to the embryo (Bewley and Black, 1994;MacGregor et al., 2015). Surrounding structures impact oxygen diffusion and are responsible for a reduction of oxygen diffusion to the embryo that depends on their thickness and their biochemical properties. Oxygen goes through the seed coat by dissolving in the water they contain during seed imbibition, but since oxygen solubility is low, little oxygen reaches the embryo. Imbibed coats become then a barrier to oxygen diffusion since they constitute a continuous wet layer surrounding the embryo (Côme, 1970(Côme, , 1982Côme and Tissaoui, 1973;Corbineau, 1992, 1998).
Many studies have demonstrated that phenolic compounds, in particular flavonoids, secondary metabolites derived from the phenylpropanoid pathway (Lepiniec et al., 2006;Debeaujon et al., 2007) reinforce coat-imposed dormancy. For example, Arabidopsis seeds accumulate flavonols, and proanthocyanidins during development, the brown colour being conferred by proanthocyanidin after oxidation. This pathway has been characterized using transparent testa tt(g) mutants that are affected in seed coat pigmentation (Debeaujon et al., 2007;North et al., 2010); mutants lacking PAs germinate faster than the corresponding wild-type brown seeds. Flavonoids, particularly condensed tannins, are known to regulate water uptake (Debeaujon et al., 2000;MacGregor et al., 2015), however, their impact on oxygen diffusion is unknown. In wheat and rice, red grains show a more intense dormancy than the white grain lines and thus have a higher resistance to preharvest sprouting (PHS) (Cohn and Hughes, 1981;Flintham, 2000), this trait has provided breeders with an easy character to select cultivars with higher dormancy even if 'red gene' contribution to dormancy is small (Warner et al., 2000;Himi et al., 2012).
In cereals, the covering structures result in the inability of the grains to germinate at temperature higher than 10-15°C. In barley, the glumellae have the main role in dormancy (Lenoir et al., 1983(Lenoir et al., , 1986Corbineau and Côme, 1996), while in oat both glumellae with pericarp, and probably testa participate to dormancy Seed Science Research . In naked grains such as sorghum and wheat, dormancy results from the pericarp, testa and endosperm (Steinbach et al., 1995). The surrounding structures inhibit germination by reducing oxygen supply to the embryo due to enzymatic and/or non-enzymatic oxidation of the phenolic compounds (Côme, 1967(Côme, , 1980Lenoir et al., 1986;Fig. 1). The inhibitory effect of the covering structures increases with increasing temperature since the fixation of oxygen in the surrounding structures increases with temperatures (Lenoir et al., 1986;Lecat et al., 1992). In barley, for example, more than 50% of the total oxygen uptake by the grain results from oxidation of phenolic compounds in the glumellae, whereas it is less than 20% at 10-15°C Fig. 1). More recently using fibre-optic O 2 micro-sensor, Hoang et al. (2013) demonstrated that the oxygen content in the embryo below the glumellae is around 15.8% at 15°C, temperature that allows germination, but only 0.3% at 30°C, temperature at which grains remain dormant.
However, afterripening that breaks dormancy has no effect on the phenolic compound composition nor on the polyphenol oxidase activity, but delays in time the oxygen uptake by the glumellae allowing a good oxygen supply to the embryo during the first hours of imbibition (Lenoir et al., 1983(Lenoir et al., , 1986Corbineau et al., 1986).

Oxygen requirement for germination of non-dormant seeds
Oxygen requirements for germination have been reviewed by different authors (Côme, 1982;Mayer and Poljakoff-Mayber, 1989;Côme and Corbineau, 1992;Bewley and Black, 1994;Corbineau and Côme, 1995;Bradford et al., 2007Bradford et al., , 2008. However, it is not always indicated whether seeds under investigation are dormant or not, although seed sensitivity to oxygen depends on the depth of dormancy (Corbineau and Côme, 1995;Bradford et al., 2007Bradford et al., , 2008 and other environmental factors as temperature, light or water potential of the medium. For most species, seeds fail to germinate when deprived of oxygen, and restricted oxygen availability prevents seedling growth. However, several aquatic species like Alisma plantago (water plantain), Cynodon dactylon (bermuda grass), Echinochloa turnerana (barnyard grass), Leersia oryzoides (rice cut-grass), Thypha latifolia (cattail), Trapa natans (water caltrop) and Zizania aquatica (wild rice) germinate better under reduced oxygen concentrations than in air (review by Corbineau and Côme (1995)). Al-Ani et al. (1982,1985) classified seeds in two groups according to their response to low oxygen concentration. Group I corresponds to fatty seeds (cabbage, flax, lettuce, radish, soybean, sunflower, turnip) the germination of which is completely inhibited when oxygen concentration is close to 2% (Fig. 2). Seeds of group II are mainly starchy seeds among which cereals (barley, maize, oat, rice, sorghum, wheat) and pea, and they did not germinate in atmosphere containing less than 1% oxygen (Fig. 2). Figure 2 indicates also that 50% of the seed population germinate in 1-3 and 7-8% oxygen for groups II and I, respectively.
Oxygen requirement for germination depends on environmental factors including temperature, osmotic pressure of the medium and light (review by Corbineau and Côme (1995)). For many species, the sensitivity to oxygen deprivation decreases with decreasing temperature, probably because oxygen solubility in water increases with decreasing temperature. In addition, sunflower (Smok et al., 1993) and tomato seeds (Corbineau and Côme, 1995) require more oxygen for their germination when placed on a medium with low water potential. For example, in the case of tomato, 50% of the population germinate within 7 d on water at 25°C in around 5% oxygen, but germination requires at least 10% oxygen when seeds were incubated on a PEG-6000 solution at 0.5 MPa. Continuous white light also reinforces the seed responsiveness to hypoxia in case of negative photosensitive seeds of Amaranthus caudatus (Gutterman et al., 1992) and Bromus rubens . Table 1 indicates that 2-3% oxygen in the atmosphere is sufficient for the germination of isolated embryos. Isolated embryos from non-dormant seeds require less oxygen for germinating than the corresponding intact seeds indicating that even in nondormant seeds, the seed coats inhibit the germination by reducing the oxygen supply to the embryos (Table 1; Corbineau and Côme, 1995;Côme and Corbineau, 1998).

Relationships between seed dormancy and responsiveness to oxygen
For some species, dormant seeds do not germinate in air (21% oxygen), but their germination capacities are improved in hypoxia. Table 2 shows the sensitivity to oxygen of Oldenlandia corymbosa L. seeds as a function of oxygen in dormant seeds (at harvest) and non-dormant ones (after breaking of dormancy by a chilling treatment) (Corbineau and Côme, 1980a). In dormant seeds, 91-95% of the population easily germinate in a narrow range of oxygen (5-7%) when the non-dormant seeds germinate Effects of temperature on the oxygen flux through the coats (pericarp + glumellae) in dormant barley grains incubated at 15°C (left) and 30°C (right). Increase in temperature results in a decrease in oxygen solubility in water during seed imbibition and in an increase in oxygen uptake through oxidation of phenolic compounds. As a result, the embryo receives less oxygen at 30°C than at 10-15°C. Adapted from Côme and Tissaoui (1973), Lenoir et al. (1983Lenoir et al. ( , 1986, Côme et al. (1988), Côme and Corbineau (1992) and Corbineau and Côme (1995). (B) Oxidation of phenolic compounds in the seed coats. The phenolic compounds are first oxidized to quinones by polyphenol oxidases (PPO) and then the quinones undergo non-enzymatic oxidative polymerization. From Côme and Corbineau (1992) and Corbineau and Côme (1995).

128
F. Corbineau in air and in a wide range of oxygen concentration from 5 to 21%. Similar behaviour was observed in dormant apple embryos ; at harvest, dormant apple embryos do not germinate in air at 15°C, but their germination is improved at lower oxygen concentration. Release of dormancy enables them to germinate in a large range of oxygenation . However, in the majority of species, seeds are more sensitive to oxygen deprivation when they are dormant than the nondormant ones. This influence of dormancy on oxygen sensitivity has been demonstrated with cereals like barley (Corbineau and Côme, 1980b;Benech-Arnold et al., 2006), wheat (Corbineau et al., 1981) and oat (Lecat et al., 1992), sunflower (Gay et al., 1991;Côme, 1992, 1995) and Douglas fir (Pseudotsuga menziesii) (Corbineau et al., 2002). Figure 3 shows, in barley, the changes in oxygen sensitivity of the intact grains (Fig. 3A) and isolated embryos (Fig. 3B) incubated on water at 20°C at harvest (i.e. dormant grains) and during afterripening (grains dry-stored for 5 and 9.5 months at 25°C). At harvest in 2000, less than 50% of the grain population can germinate at 20°C in air (21% oxygen) and no germination occurred in atmosphere containing less than 3% oxygen (Fig. 3A). Dry storage of dormant grains for 5 and 9.5 months at 25°C results in breaking of dormancy allowing germination of all the grain population in air and reducing the oxygen concentration at which 50% of the Fig. 2. Oxygen concentration of the atmosphere below which no seed germinated (O 2 0% in red) or which allowed 50% germination (O 2 50% in black). From Corbineau (1995) and Corbineau and Côme (1995). Means of 4 replicates. Modified from Corbineau and Côme (1995) and Côme and Corbineau (1998). Means of 3 replicates. Modified from Côme (1980a, 1980/81).
population germinate to 4.8% and 0.4% (Fig. 3A). By contrast, isolated embryos from dormant grains at harvest germinate in atmosphere containing at least 5% oxygen (Fig. 3B, curve 1) demonstrating that the covering structures are involved in oxygen diffusion (c.f. Oxygen requirement for germination of non-dormant seeds). Afterripening also resulted in an improvement of embryo germination in hypoxia: 100% of the population germinate in 4 and 0.5% oxygen, respectively, after 5 and 9.5 months of dry storage at 25°C (Fig. 3B, curves 2 and 3). Using a population-based threshold model, Bradford et al. (2008) have calculated a median base (or threshold) O 2 [(Ox(b) (50)]. The model predicted that 36.3% oxygen would be required for 50% germination of intact dormant grains at harvest in 2000, while the germination of isolated embryos was predicted to require only 0.123% oxygen to achieve 50% germination. For the intact grains, the Ox(b)(50) value was 2.87% and 0.30% after 5 and 9.5 months of afterripening, respectively (Bradford et al., 2008). Environmental conditions during seed production influence their sensitivity to oxygen. Hence, the Ox(b)(50) is dependent of the year of harvest and was decreased to 15.7% for grains obtained in 2002 compared to 36.3% for those obtained in 2000 ( Fig. 3A and Table 3).

Specific effect of hypoxia and anoxia on breaking of dormancy and induction of secondary dormancy
Oxygen is required for germination, however partial or total oxygen deprivation is not always harmful at least if it is not too prolonged. Temporary anaerobiosis applied to imbibed seeds may promote germination in dormant seeds of some species (Lonchamp and Gora, 1979;Côme et al., 1985) and may substitute for cold stratification Corbineau and Côme, 1995). For example, incubation in anoxia (pure nitrogen) for 2 or 3 weeks breaks dormancy in apple embryos , Oldenlandia corymbosa seeds , sunflower achenes (Corbineau and Côme, 1995) and Xanthium pennsylvanicum (Esashi et al., 1976). In the case of apple embryos, this beneficial effect of anoxia increases with temperature .
In contrast, incubation of seeds in anoxia or hypoxia, conditions which inhibit seed germination, induces secondary dormancy in barley (Hoang et al., 2013), Xanthium pennsylvanicum (Esashi et al., 1978), Viola spp., Veronica hederofolia and Veronica persica (Lonchamp and Gora, 1979), lettuce (Vidaver and Hsiao, 1975) and Brassica napus (Pekrun et al., 1997). In Oldenlandia corymbosa seeds, responsiveness to light depends also on the temperature and on the oxygen tension of the atmosphere. Seeds remain responsive to white light at 10 and 20°C, but lose their responsiveness during incubation in darkness at temperature of at least 30°C; this phenomenon has been observed in many photoblastic seeds (Karssen, 1967;Rollin, 1970) and is accelerated at low oxygen concentration in Oldenlandia corymbosa (Corbineau and Côme, 1985).

Effects of hypoxia on ABA and GA metabolism and sensitivity
The hormonal regulation of dormancy and germination is well established (Bewley, 1997;Finch-Savage and Leubner-Metzger, 2006;Finkelstein et al., 2008;Holdsworth et al., 2008;Graeber et al., 2012;Shu et al., 2016;Nonogaki, 2017). Oxygen uptake by the glumellae during the first 12-14 h of imbibition appears to be related to ABA metabolism in the embryo, as a transient increase in ABA content occurs in dormant grains and not in non-dormant ones (Benech-Arnold et al., 2006). The presence of glumellae alters the expression of genes involved in ABA synthesis (HvNCED1, NCED2), ABA catabolism (HvABA8'OH1) and signalling (HvABI5, HvVP1 and HvPKABA) (Mendiondo et al., 2010). During imbibition, the decrease in embryonic ABA level is slower in hypoxia than in air, but this effect is independent of the regulation of HvABA8'OH1 expression (Mendiondo et al., 2010;Hoang et al., 2013). Incubation of embryos of primary dormant grains in the presence of exogenous ABA at various concentrations shows that hypoxia increased embryo sensitivity to ABA by twofold (Fig. 4), and this effect was more pronounced at 30°C than at 20°C (Benech-Arnold et al., 2006), suggesting that hypoxia imposed by glumellae increases embryo sensitivity to ABA. ABA delayed and reduced seed and embryo germination; using a population-based threshold model, Bradford et al. (2008) show that ABA concentration up to 10 μM had relatively little effect on O 2 sensitivity of seeds. However, at higher concentration, the value of Ox b 50 increased from 4.8% in the absence of ABA to 12 and 14.5% in the presence of 100 and 1000 μM ABA, respectively (Table 3). Incubation of embryos in the presence of GA 3 results in a decrease in Ox b 50 (Table 3) from 18.4% on water down to 8.2% in the presence of 1000 μM GA.
Hypoxia at 15°C which induced secondary dormancy in barley was related to an increased expression of ABA synthesis genes (HvNCED2) only after 3 d and a strong regulation of gibberellin (GA) metabolism genes with promotion in HvGA2ox3 (64-fold compared with grains in air) and inhibition of HvGA3ox2 and HvGA20ox1 after 1 d (Hoang et al., 2013). In addition, embryo sensitivity to ABA increases in hypoxia and seeds became insensitive to GA at low oxygen (Benech-Arnold et al., 2006). However, hypoxia did not alter the embryo sensitivity to ABA and GA which was similar in both primary and hypoxia-induced secondary dormant grains (Hoang et al., 2013).

Effect on fermentation
Although EC has been cited as being essential in the regulation of germination (Raymond and Pradet, 1980;Raymond et al., 1985), dormancy in oat does not result from an inability to synthesize ATP from ADP and AMP (Côme and Corbineau, 1989). During seed incubation at 30°C, a temperature at which dormant grains cannot germinate, embryos from dormant and non-dormant grains exhibited a similar increase in the ATP/ADP ratio and in the EC [(ATP + 0.5 ADP)/(ATP + ADP + AMP)] as long as seeds had not germinated . In anoxia (nitrogen) or hypoxia (1% oxygen), the immediate response at the level of mitochondria corresponded to a decrease in ATP synthesis and in the EC within 10 min (Lecat, 1987; Table 4), the secondary responses resulted in an increase of EC that was stabilized after 1 to 3 h due to fermentation and accumulation of ethanol in starchy seeds (Tables 4 and 5) Raymond et al., 1985;Lecat, 1987).
Ethanol is the major product of fermentation in higher plants whether they are tolerant to anoxia or not Ricard et al., 1994). Late responses leading to tolerance to oxygen deprivation were associated with a decrease in ADH activity and a change in synthesis of malate, succinate, lactate and shikimate resulting from amino-acid metabolism instead of ethanol Bui et al., 2019). In higher plants, L-lactate is often produced prior to ethanol in the first min after the transfer to anoxia, and alanine is the third major fermentation product in plants. The adenylate EC under anoxia depends on the seed reserves; in group I (lipid seeds including lettuce, sunflower, radish, turnip, cabbage, flax and soybean), germination does not occur in oxygen pressure below 1 kPa and the EC is below 0.4 while it remains at a higher value (0.70-0.82) in seeds from group II (i.e. starch seeds as rice, wheat, maize, sorghum and pea) for which germination still occurs at 0.1 kPa (Table 5; Al-Ani et al., 1985).
In contrast, change in non-adenylic triphosphate nucleotides (NTP) content within the embryo during seed incubation at 30°C depended on the dormancy state . Thermo-dormant grains lost their ability to synthesize NTP from adenine after transfer from 30 to 20°C; the capacity of synthesis of NTPs seems then to be a good marker of dormancy in oat seeds, as in buds of Helianthus tuberosus L. tubers (Le Floc'h and Lafleuriel, 1981), of Fraxinus excelsior L. (Lavarenne et al., 1982), Quercus spp. (Barnola et al., 1986) and Prunus persica L. (Balandier et al., 1993).
In addition, modulation of EC using NaN 3 and polyethyleneglycol (PEG) has demonstrated that the induction of thermodormancy in oat seeds is an active phenomenon requiring an EC higher than 0.5 . Exogenous GA strongly stimulated germination of oat seeds at high temperature and in hypoxia (Lecat et al., 1992) but has no effect on EC.

Involvement of the N-end rule pathway
The ubiquitin-proteasome system (UPS) degrades the proteins by targeting specific signals known as degrons. Among the degradation signals, N-terminus is the one targeted at the N-terminal residues rather than the sequence elements of the protein (Bachmair et al., 1986). The N-end rule pathway has three branches: the Arg/N-end rule pathway, the Ac/N-end rule pathway and the Pro/N-end rule pathway (Lee et al., 2016;Nguyen et al., 2018Nguyen et al., , 2019Dong et al., 2018, Zhang et al., 2018a. The Arg/N-end rule pathway identifies specific N-degrons such as aromatic residues (Phe, Trp, Tyr) by PROTEOLYSIS 1 (PRT1) and basic residues (Arg, His, Lys) by PRT6 in plant. Tertiary destabilizing residues (Asn, Gln and Cys) are first modified, either enzymatically (by deamidation of Asn or Gln) or chemically (by oxidation of Cys), to generate a secondary destabilizing residue (Asp, Glu and oxidized Cys). Studies reviewed by Licausi (2011Licausi ( , 2013, Voesenek and Bailey-Serres (2015) and Loreti and Perata (2020) indicate that the sensory system to hypoxia relies on an oxygen-mediated branch of the N-end rule pathway for protein degradation acting on a specific clade of ethylene-responsive transcription factor (ERF-VII). Both the N-end rule pathway and ERF-VII are components of the oxygen signalling pathway, in particular involved in responses to low oxygen stress (Gibbs et al., 2011(Gibbs et al., , 2015Licausi, 2011Licausi, , 2013Licausi et al., 2011Licausi et al., , 2013Hartman et al., 2021). Group VII ERF transcription factors are either constitutively expressed and/or differentially transcriptionally regulated in response to different signals including low oxygen (Bailey-Serres et al., 2012;Gibbs et al., 2014aGibbs et al., ,b, 2015Voesenek and Bailey-Serres, 2015;Loreti and Perata, 2020;Hartman et al., 2021). Four of the five ERFs, HYPOXIA RESPONSIVE1 (HRE1 and HRE2) and RELATED TO APETALA 2.2 (RAP 2.2 and RAP 2.12) are implicated in the regulation of hypoxia-responsive genes. Under normoxia, ERFs are degraded via the N-end rule pathway of proteolysis. When oxygen becomes limiting, the degradation of the ERFs by the N-end rule pathway is inhibited because of the reduction of oxygen-mediated Cys oxidation (Bailey-Serres et al., 2012;Gibbs et al., 2015;Loreti and Perata, 2020;Hartman et al., 2021).

Concluding comments
From 1960 to 1990, seed research took the form of descriptive physiology that highlights the sensitivity of seeds to the environment. The main research topics were developed at that time in France by the team of Dr A. Pradet and Dr P. Raymond at INRA (Bordeaux) and the team of Prof D. Côme at CNRS (Meudon) and the University Pierre et Marie Curie (Paris) were on (1) the oxygen requirement for the seed germination as a function of dormancy (Figs 2 and 3; Tables 1 and 2), (2) the involvement of the seed coat in dormancy due to the interference with oxygen diffusion to the embryo through phenolic compounds oxidation (Fig. 1) and (3) the seed/seedling metabolism in response to oxygen deprivation (Tables 4 and 5). In some ways, physiological and biochemical mechanisms have become clearer, but these data do not allow to explain the signalling mechanisms involved between oxygen, ROS and hormones in the regulation of seed germination and dormancy. It was then necessary to consolidate the interfaces among physiology, biochemistry, ecophysiology, cell and molecular biology, and genetics in order to better understand the seed behaviour and its plasticity in various environmental constraints. Access to 'omic' tools, genetics and micro-measurements with microsensors have allowed an enormous advancement over the last 35-60 years. Figure 5    Modified from Pradet and Raymond (1983) and Raymond et al. (1985).

F. Corbineau
Perata (2020) and Hartman et al. (2021)), regulation of dormancy by hormone synthesis and sensitivity, NO and ROS production and communication between organelles and cell compartments (Feurtado and Kermode, 2007;Footitt et al., 2011;Nonogaki, 2017Nonogaki, , 2019Bailly and Merendino, 2021;Jurdak et al., 2021). In addition, ethylene which breaks dormancy (Corbineau et al., 2014;Wang et al., 2018Wang et al., , 2021Wang, 2019) was shown to increase the response to hypoxia through enhanced stability of specific transcription factors (ERFs) and is involved in mitochondria retrograde regulation (MRR) or signalling (MRS) in which alternative oxidase (AOX) is likely to have an essential role (Nonogaki, 2014;Bailly and Merendino, 2021;Hartman et al., 2021;Jurdak et al., 2021). These recent data point out the role of mitochondria as a sensor organelle that regulates ROS production, and the control of oxygen sensing by the Cys branch of the N-degron pathway. Under hypoxia ATP synthesis is reduced but ERF-VII are stabilized and migrated to the nucleus and then activated the transcription of hypoxia-responsive genes (Gibbs et al., 2011(Gibbs et al., , 2014bLicausi et al., 2011Licausi et al., , 2013Pucciariello and Perata, 2017;Loreti and Perata, 2020;Hartman et al., 2021). Better understanding of the ROS trafficking and interorganelle communication will bring novel insights in the regulation of germination and seedling emergence by the environmental conditions, and in particular by oxygen (Nonogaki, 2014(Nonogaki, , 2017(Nonogaki, , 2019Bailly and Merendino, 2021). Despite the progress of the last 30 years, numerous questions remain under debate. For example, it would be essential to characterize the ROS targets and the oxidative modifications using proteomic and/or transcriptomic approaches. The better understanding of the role of ROS in seed germination and dormancy also requires to determine the dynamics of ROS generation at the subcellular level, where ROS are produced and sensed, and how ROS level is regulated through detoxication systems in which the covering structure could be involved. It would be of particular interest to investigate how nuclear ROS may regulate seed dormancy by inducing specific transcriptomic effect. Fig. 5. Main data obtained after the 85-90s concerning the network between hormones and ROS production involved in the dormancy regulation, and the involvement of hypoxia in stabilization of ERF group VII and tolerance to oxygen deprivation. The left part of the panel highlights the involvement of hypoxia in ROS and NO production in mitochondria and the stabilization of ERFs because of the absence of oxygen. ERFs migrate then to the nucleus and activate the transcription of hypoxia-response genes important for the tolerance to hypoxia. The right part of the panel indicates the effects of hypoxia on ABA and GA balance (ABA degradation and GA synthesis) and sensitivity, and the regulation of dormancy. The relative level of ABA to GAs is a major determinant of the depth of dormancy, in addition C 2 H 4 which improves germination of dormant seeds contributes to a decrease in ABA level, in modification in ABA and GA sensitivity, and on ROS production at the level of mitochondrion, and the stabilization of ERFs. NO induces a decrease in ABA level correlated with the regulation of ABA 8 ′ -hydroxylase expression. ABA, abscisic acid; GA, gibberellins; ERF, ethylene response factor; RAP2.12, related to Apetala 2 12; ROS, reactive oxygen species; PDC, pyruvate decarboxylase; HRE, hypoxia responsive; HRA, hypoxia-response attenuator1; HB1, haemoglobin 1 ; PCO, plant cysteine oxidases. Modified from Borisjuk et al.