Ensifer meliloti denitrification is involved in infection effectiveness and N2O emissions from alfalfa root nodules

Alfalfa is one of the most valuable forage crops in temperate climate zones. Ensifer meliloti, the endosymbiont of alfalfa, contains all the denitrification genes but the capacity of alfalfa root nodules to produce nitrous oxide (N2O) is not known. In this work, N2O emissions as well as the influence of bacteroidal denitrification on nodulation competitiveness and N2O release from alfalfa nodules has been investigated. Medicago sativa cv. Victoria plants were inoculated with E. meliloti 1021, a periplasmic nitrate reductase (Nap) defective mutant, a Nap overexpressing strain and a nitrous oxide reductase defective mutant. Plants were grown in the presence of different nitrate and copper treatments and subjected to flooding during one week before harvesting. N2O production by the nodules was analysed by using gas chromatography. Methyl viologen-dependent nitrate reductase (MV+-NR), nitrite reductase (MV+-NIR) and nitrous oxide reductase (N2OR) enzymatic activities were measured in isolated bacteroids. Alfalfa root nodules produce N2O in response to nitrate and flooding. Overexpression of Nap improved nodulation competitiveness and induced N2O emissions from nodules. Copper is required for an effective symbiosis as well as triggered a reduction of N2O production due to the induction of the N2OR and a reduction of NIR activities in the bacteroids. Alfalfa root nodules emit N2O. Nap is involved in nodulation competitiveness and in N2O emissions by the nodules. Bacteroidal N2OR and NIR activities are modulated by Cu and may be considered as effective targets for the mitigation strategies of N2O emissions derived from alfalfa crops.


Introduction
Nitrous oxide (N 2 O) is a potent greenhouse gas and also represents an important ozone layer depleting factor (Ravishankara et al. 2009). According to various studies (Rockström et al. 2009;Steffen et al. 2015) and the last IPCC report (IPCC 2022), agriculture and livestock farming are the main human activities contributing to the increase of the anthropogenic N 2 O levels in the atmosphere, while fossil fuel combustion, biomass burning and water treatment cause less impact to the environment in terms of N 2 O emissions. In fact, the non-synchronized application of synthetic nitrogen fertilizers to crops contributes to intensify the biological release of N 2 O, and has been considered as the main factor causing the remarkably rapid increase of atmospheric N 2 O concentration occurred over the last century. Therefore, a better understanding of the pathways involved in N 2 O formation in agricultural soils is essential for the reduction of these emissions to the atmosphere. In this context, several processes involved in N 2 O formation have been proposed, being nitrification and denitrification the main sources (Butterbach-Bahl et al. 2013;Torres et al. 2016).
Legumes and soil bacteria collectively termed "rhizobia" establish a symbiotic relationship characterized by the formation of new root organs, called nodules, where biological nitrogen fixation takes place (Oldroyd and Downie 2008). Nevertheless, nodulation requires complex chemical and physiological signalling interactions between both partners of the symbiosis (Poole et al. 2018). Either the plant or the endosymbiont are responsible for nodulation competitiveness. In agriculture, microbial interactions are part of a multicomponent equation involving plant genotype, environment and plant and soil microbiomes (Onishchuk et al. 2017). Following this line, genetic features influencing competitiveness are involved in rhizosphere colonization, establishment of an effective symbiosis, or even in plant growth promotion or prevention of the growth of other bacterial cells (reviewed by Mendoza-Suárez et al. 2021). Once nodules are formed, rhizobia differentiate into bacteroids inside the nodules, acquiring the capacity of fixing N 2 through nitrogenase activity. However, as nitrogenase is inhibited by oxygen, symbiotic nitrogen fixation (SNF) requires a microaerophilic environment, thus changes in oxygen concentrations from normoxic to microoxic (hypoxic) levels during nodule formation and maturation are required (Rutten and Poole 2019).
Inoculation of legumes with rhizobia has been considered as an environmental-friendly agricultural practice recommended all over the world as part of a strategy for N 2 O mitigation (Torres et al. 2016). However, it has been reported that legume crops also contribute to N 2 O production by providing N-rich residues for decomposition Sánchez and Minamisawa 2019), or directly by some free-living or symbiotically-associated rhizobia that are able to denitrify (Bedmar et al. 2005Hirayama et al. 2011).
Denitrification is a sequential respiratory process in which nitrate (NO 3 − ) or nitrite (NO 2 − ) is reduced to N 2 , releasing nitric oxide (NO) and nitrous oxide (N 2 O) as intermediates, through four steps sequentially catalyzed by a periplasmic (Nap) or a membrane-bound nitrate reductase (Nar), a Cu-containing (NirK) or a cytochrome cd 1 -containing nitrite reductase (NirS), a c type (cNor), quinol-dependent (qNor) or Cu-containing nitric oxide reductase (Cu A Nor) and a nitrous oxide reductase (Nos). Reviews covering the biochemistry and physiology of denitrification have been published elsewhere (Bueno et al. 2012;Kraft et al. 2011;Richardson 2011;Salas et al. 2021;Torres et al. 2016;van Spanning et al. 2005;2007). The importance of denitrification in N 2 O reduction lies in the fact that the Nos enzyme is the only known sink of N 2 O and, consequently, scientific research has focused on it as a key enzyme in N 2 O mitigation strategies (Richardson et al. 2009). In this context, it has been reported that inoculation of soybean with N 2 O-reducing strains of Bradyrhizobium can mitigate N 2 O emissions (Akiyama et al. 2016;Itakura et al. 2013;Woliy et al. 2019). However, several of the most agronomical interesting rhizobial species do not contain Nos, being unable to reduce N 2 O to N 2 and emitting great levels of N 2 O as a consequence (Sánchez and Minamisawa 2019;Torres et al. 2016).
N 2 O emissions from agricultural soils can be influenced by environmental factors such as nitrate or N-derived species, oxygen-limiting conditions, pH or copper concentration among others (Liu et al. 2022;Richardson et al. 2009). With respect to legumes, Tortosa et al. (2015) reported significant increases of N 2 O emissions from soybean root nodules in response to nitrate and flooding, being the denitrification performed by Bradyrhizobium diazoefficiens bacteroids the main contributor to N 2 O release. In addition to oxygen and nitrate, another environmental factor that contributes to N 2 O emissions in soils is copper (Cu) (Liu et al. 2022). In fact, a variety of studies with free-living rhizospheric denitrifying microorganisms have demonstrated that nosZ expression or Nos synthesis and activity decreased when Cu was a limiting nutrient, resulting in notable increases in N 2 O emissions (Felgate et al. 2012;Pacheco et al. 2022;Sullivan et al. 2013). Recent studies have shown that Cu addition to the plant growth nutrient solution reduced statistically N 2 O emissions from soybean nodules (Tortosa et al. 2020).
Alfalfa (Medicago sativa L.) is one of the most valuable forage crops and the most productive forage legume in zones with temperate climate . This crop does not require nitrogen fertilization, since it is able to fix up to 463 kg of N 2 per ha and year, which is mainly used for the synthesis of their own protein. The rest of the nitrogen provided by the nodules is incorporated into the following crops through the remnant plant residues (roots and harvest rests). Because of that reason, alfalfa is also cultivated in crop rotations to improve nitrogen soil enrichment and biomass production of the following crops .
Ensifer meliloti is a rhizobial species which symbiotically associates with plants of the genera Medicago, Melilotus and Trigonella, and possesses all the denitrification genes: napEFDABC, nirK, norECBQD and nosRZDFYLX, encoding Nap, NirK, cNor and Nos, respectively (Torres et al. 2014). Nevertheless, this bacterium has been considered as a partial denitrifier due to its incapacity to grow with nitrate or nitrite as respiratory substrate under anoxic conditions (Bueno et al. 2015;Torres et al. 2011Torres et al. , 2014. In fact, a recent study reported that nap expression was significantly lower respect to the rest of the denitrification genes when E. meliloti cells were incubated anoxically (Torres et al. 2014). This limitation in the induction of nap genes may be responsible of the incapacity of E. meliloti to respire nitrate in an anoxic environment. In fact, overexpression of nap genes recovered the capacity of E. meliloti to grow anaerobically and to produce N 2 O under free-living conditions (Torres et al. 2018). However, the capacity of the symbiotic forms of E. meliloti to produce N 2 O has not been explored so far.
Taking in consideration all this background, the aim of this study was to investigate the capacity of alfalfa root nodules to produce N 2 O in response to nitrate, flooding and Cu availability. The involvement of E. meliloti denitrification in the symbiotic interaction with alfalfa plants and in N 2 O emissions from the nodules has also been explored.

Plant growth conditions
Alfalfa (Medicago sativa, cv. Victoria) seeds were surface-sterilized by immersion in 2.5% HgCl 2 for 9 min. Then, seeds were washed with sterile distilled water and germinated on filter paper discs in Petri dishes in darkness for 2-3 days at 30 ºC. Alfalfa plants were grown using a modified Rigaud and Puppo nutrient solution (1975)  The NS was supplemented with 1, 2, 3, 4 or 10 mM KNO 3 . The standard Cu concentration of NS was 0.2 mg•l − 1 (0.8 µM). For studies of the effect of Cu on symbiosis, NS was supplemented with a higher Cu concentration of 5 mg•l − 1 (20 µM), used in previous studies from our group (Tortosa et al. 2020). Plants in tubes or pots were placed into growth chambers from the Greenhouse and Growth Chamber Service (GGCS) (EEZ, Granada, Spain) with the following parameters: night/day temperature, 24/20 ºC; photoperiod, 16/8 h; photosynthesis photon flux density of 403 µmol photons•m − 2 •s − 1 .

Plant experimental setting
For nodulation kinetics assays, a methodology described by Torres et al. (2013) was used. Basically, germinated seeds were transferred into 43-ml autoclaved glass tubes containing 10 ml water and kept in darkness for approximately 24 h. Then, these tubes were placed into the growth chamber and, after 5 days, water was replaced with 10 ml of NS containing a cell suspension of approximately 10 8 CFU•ml − 1 . NS was supplemented with 1, 2, 3 or 4 mM KNO 3 . Anoxic conditions were achieved by sparging NS with N 2 gas before adding the inoculum. Finally, tubes were incubated for 30 days, and nodule number was daily counted.
For competitivity assays, seeds were germinated as for nodulation kinetics experiments. After 5 days, water was replaced with 10 ml of NS supplemented with 3 mM KNO 3 and seeds were inoculated individually with strains 4002, 4004, 4002-GUS3, or 4004-GUS3 as controls. To determine the competence for nodulation a mix in 1:1 proportion of 4004 and 4002-GUS3 or 4004-GUS3 and 4002 was used at a cellular density of approximately 10 5 CFU•ml − 1 . Tubes were incubated for 30 days in the controlled environmental chamber with the parameters enumerated above and nodules were revealed with X-Gluc (0.53 mg•ml − 1 ) according to Nogales et al. (2006). Nodules produced by 4004 or 4002 strains were revealed after isolation of bacteria from the nodules in plates containing Km and, therefore, carrying the pDS4004 or pDS4002 plasmids.
For plant assays in pots, germinated seeds were transferred into 250-ml autoclaved pots filled with perlite as substrate. The following experiments were carried out ( Supplementary Fig. S1): Experiment 1 ( Supplementary Fig. S1A): in order to study the influence of nitrate and flooding in N 2 O emissions from alfalfa nodules, these pots were placed on 1-l glass jars containing 500 ml of N-free NS. Eight seedlings per pot were inoculated at sowing with a cell suspension of the WT 4004 strain of about 10 8 CFU•ml − 1 . Three sets were established: the first set, with 20 pots, was watered with N-free Translational fusion between pnfeD (coordinates 2993-3345) and gusA in pBI101; Km r García-Rodríguez and Toro (2000) NS, and the second and third sets, with 10 pots each, were watered with NS supplemented with 1 mM or 3 mM KNO 3 , respectively. Plants were watered every two weeks under sterile conditions. Seven days before harvesting (i.e., after 43 days), a nitrate shock of 10 mM KNO 3 was applied to 10 pots from the first set. Additionally, at the same time, half of each set was also subjected to flooding conditions, which were achieved by removing alfalfa plants from the pots and transferring them into a glass jar filled with 900 ml NS, thus nodulated roots were completely submerged. After 50-days growth, plants and nodules were harvested. Experiment 2 (Supplementary Fig. S1B): to investigate the role of Nap in N 2 O emissions from alfalfa nodules, pots were prepared as described above, and 10 pots with 8 seedlings each were inoculated at sowing with a cell suspension of about 10 8 CFU•ml − 1 of the WT 4004, the nap + strain, or the napA mutant strain (denoted as nap − throughout the manuscript). Plants were grown for 50 days, and watered every two weeks under sterile conditions. Seven days before harvesting, the pots were watered with NS supplemented with 10 mM KNO 3 and subjected to flooding conditions as indicated above. After plant growth, the nodules harvested from 5 pots from each treatment were immediately used for N 2 O emission measurements, whereas the nodules harvested from the remaining 5 pots from each treatment were frozen in liquid nitrogen and stored at -80 ºC for further determinations.
Experiment 3 (Supplementary Fig. S1C): this experiment was performed to study the involvement of Cu in plant and nodule physiology as well as in N 2 O emissions from alfalfa nodules. For this goal, pots were placed on glass jars containing 500 ml of N-free NS. Eight seedlings per pot were inoculated at sowing with the WT 1021 or the nosZ mutant strain at a cellular density of about 10 8 CFU•ml − 1 . For pots inoculated with the WT 1021 strain, three different sets of 10 pots each were established: the first set was watered with NS without Cu added, the second set, with NS supplemented with 0.8 µM CuSO 4 •5H 2 O, and the third set, with NS containing 20 µM CuSO 4 •5H 2 O. The 5 pots inoculated with the nosZ mutant were watered only with 0.8 µM Cu NS. Plants were grown for 43 days and plants and nodules from 5 pots from each set inoculated with the WT were harvested. To induce N 2 O production, the remaining 5 pots from each WT set were treated with 10 mM KNO 3 and subjected to flooding for 7 days. Then, nodules were harvested and N 2 O emissions from detached nodules were measured.

Plant physiological analyses
Physiological data such as nodule number (NN), nodule fresh weight (NFW), shoot dry weight (SDW) and plant dry weight (PDW) were recorded and expressed per plant. SDW and PDW were determined after 3 days in an oven at 70 ºC.
Prior to analytical determinations (nitrogen and copper content), dry shoots and roots were ground using an IKA A11 mill to less than 0.5 mm according to Tortosa et al. (2020). Seeds and nodules were dried in an oven at 70 ºC for only one day.
Nitrogen content was analysed in dried and ground shoots of alfalfa plants by the N/C Analysis Service of Estación Experimental del Zaidín (EEZ, Granada, Spain) by using an elemental analyser LECO TruS-pec® CN (LECO, St Joseph, MI, USA). Briefly, the sample was subjected to complete combustion at 950 ºC in the presence of O 2 . Then, all the nitrogen oxides formed were converted into N 2 , and this gas was detected by a thermal conductivity detector. Data were expressed as mg N•g − 1 of dry sample.
Cu concentration was analysed in dry alfalfa seeds before germination, dry nodules as well as in dry and ground roots and shoots using the Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) available at the Instrumental Technical Service of EEZ (Granada, Spain), model PlasmaQuant® PQ 9000 (Analytik Jena, Jena, Germany). Data were expressed as mg Cu•kg − 1 of dry sample.

Determination of leghemoglobin content in nodules
Leghemoglobin (Lb) concentration in nodules was determined by fluorimetry according to a method described by Tortosa et al. (2020), which was based on the standard method established by LaRue and Child (1979), but was adapted to alfalfa nodules in the present work. Briefly, 0.125-0.13 g of NFW were crushed and homogenised with a pestle in a cold porcelain mortar by adding 6 ml of buffer solution (50 mM Na 2 HPO 4 •2H 2 O/ NaH 2 PO 4 •H 2 O, pH 7.4; 0.02% w/v K 3 Fe(CN) 6 ; 0.1% w/v NaHCO 3 ) and 0.1 g of polyvinyl polypyrrolidone (PVPP). The extract was centrifuged at 12,000x g at 4 ºC for 20 min and 200 µl from the supernatant were transferred into glass tubes containing 3.15 ml of a saturated calcium oxalate solution (66 g•l − 1 ) and subsequently autoclaved at 120 ºC during 30 min. Then, samples were cooled down and measured in a Shimadzu spectrofluorometer (Shimadzu Scientific Instruments, Kyoto, Japan), setting λ = 405 and 600 nm as excitation and emission wavelengths, respectively, and compared to non-autoclaved samples as control. Lb concentrations were expressed as mg Lb•(g NFW) −1 and obtained after extrapolation of the data using a human hemoglobin standard curve built from a stock of 300 mg•l − 1 and including the following concentrations (mg•l − 1 ): 0, 60, 120, 180, 240 and 300.

Nitrous oxide determinations
Detection of N 2 O emissions was performed according to Tortosa et al. (2020), including certain modifications for alfalfa detached nodules. Briefly, harvested nodules from the same pot (0.2-0.3 g) were immediately transferred to a 10-ml glass vial (SUPELCO®) and a volume of 1 ml or 100 µl NS (with the corresponding nitrate and Cu concentration) was added depending if nodules were isolated from flooded or non-flooded plants, respectively. The vials containing nodules were incubated at 30 ºC. N 2 O was detected by an HP 4890 gas chromatography instrument provided with an electron capture detector (ECD) as essentially described by Torres et al. (2014). The column was packed with Porapak Q 80/100 mesh. N 2 was used as the carrier gas at a flow rate of 23 ml•min − 1 . The injector, column and detector temperatures were 125, 60 and 375 °C, respectively. Gas samples were taken from the headspace of the vials after 3 and 6 h incubation and injected manually by using luer-lock gas-tight syringes BD Microlance™ 3. Peaks corresponding to N 2 O were integrated by using GC ChemStation Software (Agilent Technologies, Santa Clara, CA, USA) and the values obtained were used to calculate N 2 O concentration in each sample by extrapolation from a standard curve, performed by using 2% (v/v) N 2 O standard (Air Liquid, Paris, France) and including the following gas volumes: 0, 0.2, 0.4, 0.6, 0.8 and 1 ml. Total N 2 O concentration was determined by taking into account both N 2 O in headspace, and dissolved N 2 O applying Bunsen solubility coefficient (47.2% at 30 °C). N 2 O fluxes from alfalfa detached nodules were expressed as nmol N 2 O•(g NFW•h) −1 .

Bacteroids isolation
For bacteroids isolation, a method described by Mesa et al. (2004) was used. Basically, bacteroids were isolated by homogenizing 0.25 g of alfalfa nodules with a pestle in a cold porcelain mortar with 7.5 ml of extraction buffer (45.5 g•l − 1 D-mannitol dissolved in 50 mM Tris-HCl, pH 7.5) previously added. Then, the extract obtained was filtered through a sterile cheesecloth filter and centrifuged at 250x g at 4 ºC for 5 min to remove nodule debris. Subsequently, the supernatant was centrifuged at 12,000x g at 4 ºC during 10 min and pellets were washed twice and resuspended in 0.5 ml of wash buffer (50 mM Tris-HC l, pH 7.5) prior to biochemical determinations.
The nitrite concentration present in the bacteroids extract was estimated colorimetrically after diazotisation by adding the sulphanilamide/naphtylethylene diamino dihydrochloride reagent (Hageman and Hucklesby 1971) and extrapolating from a standard curve including 0, 20, 40, 60, 80 and 100 µM NaNO 2 from a stock solution of 100 µM.
Determination of nitrate reductase (NR, EC 1.7.99.4), nitrite reductase (NIR, EC 1.7.2.1) and nitrous oxide reductase (N 2 OR, EC 1.7.2.4) activities Methyl viologen (MV + )-dependent nitrate reductase (MV + -NR) and nitrite reductase (MV + -NIR) activities were determined as essentially described by Delgado et al. (2003). The reaction mixtures contained 200 µM methyl viologen, 20-30 µg protein from the cell suspension, 50 µl distilled water, and 10 mM KNO 3 for MV + -NR or 100 µM NaNO 2 for MV + -NIR assays, adding 50 mM Tris-HCl buffer up to reach a final volume of 450 µl in each reaction tube. Before measurements, a 46 mM sodium dithionite solution was prepared freshly (8 mg•ml − 1 in 50 mM Tris-HCl buffer, pH 7.5), transferring 50 µl from it to each reaction tube. After incubation for 20-30 min at 30 °C, the reaction was stopped by vigorous shaking until disappearance of blue color from the samples. Control tubes were prepared as the reaction tubes, but these tubes were shaken vigorously immediately after the addition of dithionite. MV + -NR activity was expressed as nmol NO 2 − produced•(mg protein•min) −1 . MV + -NIR activity was expressed as nmol NO 2 − consumed•(mg protein•min) −1 . Three biological replicates from at least three independent experiments for each treatment were assayed.
Nitrous oxide reductase (N 2 OR) activity was measured as essentially described by Tortosa et al. (2020), setting some modifications for alfalfa nodules. The assay was performed in 10-ml glass SUPELCO® vials, adding 0.15-0.2 mg protein and 60 mM sodium succinate as electron donor. Then, a mixture of 2% (v/v) N 2 O and 98% (v/v) N 2 (Air Liquid) was injected to reach a final concentration of 0.1% (v/v). To achieve anoxic conditions, the vials were gassed with N 2 during 7 min. All the vials were incubated at 30 °C for 1 h. Next, 0.5-ml aliquots were taken from the headspace of each vial. N 2 O measurements and concentration calculations were performed as described above. N 2 OR activity was expressed as nmol N 2 O consumed• (mg protein•h) −1 . Three biological replicates from at least three independent experiments for each Cu condition were used.

Statistical analysis
Data were checked for normal distribution according to Kolmogorov-Smirnov and Shapiro-Wilk tests. For data obtained from plant assays in tubes, inferential statistics were performed by applying parametric ANOVA and a post-hoc Tukey HSD test at p ≤ 0.05 with SPSS software. For data obtained from plant assays in pots, inferential statistics to test null hypothesis were performed by applying non-parametric Kruskal-Wallis test for more than two unpaired treatments. Next, a post-hoc U Mann-Whitney test at p ≤ 0.05 with SPSS software was performed.

Results
Periplasmic nitrate reductase has a role in infectivity and competitiveness for nodulation in alfalfa As a preliminary experiment, we investigated the effect of nitrate on nodulation capacity of E. meliloti by sowing alfalfa seeds in glass tubes and inoculating them with E. meliloti 4004. Plants were grown during 30 days without nitrate or with different nitrate concentrations, ranging from 1mM to 4 mM ( Supplementary  Fig. S2). In these experiments, nodulation was significantly diminished by 4 mM nitrate, while no differences were observed for the rest of nitrate concentrations ( Supplementary Fig. S2).
Next, we were interested in elucidating whether nodulation capacity of E. meliloti was influenced by nap expression. To achieve this goal, seeds were sown in glass tubes with NS supplemented with 3 mM KNO 3 , and were inoculated with the nap + strain or the strain 4004 (WT). Before inoculation, NS was subjected or not to anoxic conditions, as described in material and methods. As shown in Fig. 1a, similar nodulation capacity between the WT or nap + strains inoculated under normal conditions was observed, counting 6 nodules per plant, approximately, after 30 days of plant growth regardless of the strain (p > 0.05). However, the nap + strain showed a significantly major efficiency for nodulation than the WT under anoxic conditions, counting 8 and 6 nodules per plant, respectively, at the end of the experiment (p < 0.05) (Fig. 1a). These results indicated that nap overexpression might promote nodulation when roots are developed in a hypoxic environment. The next step was to analyse the competitiveness of the nap + strain for nodulation. We performed experiments with alfalfa plants grown in glass tubes with 3 mM nitrate and inoculated with a mixture (1:1 ratio) of the WT 4004-GUS3 and the nap + strain (4002), or the WT 4004 and the nap + -GUS3 strain. Plants inoculated only with the WT or the nap + strain (harbouring pGUS3 or not) were used as control of the experiments. Additionally, stability of pGUS3, pDS4002 or pDS4004 was checked in plates as described in Material and Methods. As shown in Fig. 1b, the nap + strain produced 59% of the total number of nodules, while the WT 4004-GUS3 generated the remaining 41%. Moreover, the nap + -GUS3 strain produced 61% of the total number of nodules, while the WT 4004 elicited the remaining 39%. Therefore, these results confirm the previous results on nodulation kinetics (Fig. 1a) and suggest that E. meliloti nap overexpression improves competitiveness for nodulation of alfalfa plants.

Periplasmic nitrate reductase is involved in N 2 O emissions from alfalfa root nodules
To investigate the capacity of alfalfa root nodules to produce N 2 O, seeds were inoculated with E. meliloti 4004 and N 2 O emissions were measured after growing the plants in pots containing NS with different nitrate concentrations. As shown in Fig. 2, N 2 O emissions were not detected in nodules from plants grown without nitrate independently of the application of flooding conditions or not. A weak induction of N 2 O production was observed in non-flooded nodules and in the presence of 1 mM nitrate in the growth medium. The addition of 3 mM to NS or the treatment of the plants with 10 mM nitrate 7 days before harvesting slightly induced N 2 O release from non-flooded nodules compared to 1 mM treatment (Fig. 2). Interestingly, flooding triggered a significant increase in N 2 O emissions from nodules of plants subjected to 1 mM, 3 mM or 7-day 10 mM nitrate, compared to non-flooded nodules (Fig. 2). It is worth mentioning that no significant differences in N 2 O emission levels under flooding conditions were found between 1 and 3 mM nitrate for 50 days, and between 3 mM and the application of a nitrate shock of 10 mM, 7 days before harvesting. However, N 2 O levels from flooded nodules of 7-day treated plants with 10 mM nitrate were significantly higher compared to those from flooded plants grown in the presence of 1 mM (p < 0.05; Fig. 2).

Fig. 1
Nodulation capacity of a nap overexpressing E. meliloti strain (nap+). a Nodulation kinetics of alfalfa plants inoculated with strain 4004 (wild-type, WT, circles) or the nap + strain 4002 (squares) and grown during 30 days with nutrient solution supplemented with 3 mM KNO 3 . Half of the tubes containing nutrient solution were sparged with N 2 gas during 10 min before inoculation (anoxic conditions, white symbols) and the other half were not fluxed with N 2 (oxic conditions, black symbols) b Nodule competition assays. Data represent the percentage of nodules occupied by the nap + , WT-GUS3, nap + -GUS3 or WT strains inoculated separately as control of the experiments or after co-inoculation (ratio 1:1) under anoxic conditions. In a and b, data represent means with standard error bars using a Tukey HSD test at p ≤ 0.05 from three independent experiments assayed by using ten plant replicates Fig. 2 Nitrous oxide emissions from detached nodules elicited by E. meliloti 4004. Alfalfa plants were grown without nitrate (0 mM), with 1 or 3 mM KNO 3 during 50 days, or treated with 10 mM KNO 3 7 days before harvesting. Flooding conditions (black bars) were applied or not (white bars) during one week before harvesting. Data represent means with standard error bars from three independent experiments assayed by using ten pot replicates containing 8 plants each. Lower-case letters indicate comparisons between plants subjected to flooding and nitrate treatments (1 or 3 mM KNO 3 for 50 days or 10 mM KNO 3 for 7 days). Same lower-case letters are not statistically significant according to U Mann-Whitney test at p ≤ 0.05. NFW, nodule fresh weight As shown in Table 2, NFW per plant significantly decreased when plants were grown with 1 or 3 mM KNO 3 for 50 days in comparison to those plants treated with 10 mM KNO 3 during 7 days before harvesting. However, nodule number (NN) per plant was not affected by any nitrate treatment. It is also important to mention that the application of 1 or 3 mM nitrate during 50 days as well as 10 mM nitrate treatment for 7 days caused a major impact on leghemoglobin content in nodules obtaining a decrease of 1.9, 1.7 and 1.5fold respectively, compared to that observed in nodules from plants grown without nitrate. With respect to PDW per plant, the treatment of 3 mM nitrate during 50 days increased PDW significantly compared to plants grown without nitrate or with 1 mM nitrate or treated with 10 mM nitrate, where no differences were observed. These results indicate that the increase in PDW of plants grown with 3 mM is possibly due to the nitrogen uptake by plant roots rather than the SNF, since nodule growth and physiology were severely affected under these conditions compared to those grown without nitrate (Table 2).
To investigate the involvement of Nap in N 2 O emissions, alfalfa plants were grown in pots and, a week before harvesting, they were subjected to 10 mM KNO 3 and flooding conditions, since these were the conditions where the highest N 2 O emission levels were found and nodule biomass was not affected (Fig. 2, Table 2). Plants were inoculated with the WT strain 4004, the nap + , or the nap − strain (Table 1, Supplementary Fig. S1B). As shown in Fig. 3a, MV + -NR activity from bacteroids of the WT strain was 2.4-fold higher compared to that from nap − bacteroids. Inoculation of the plants with the nap + strain induced about 1.8-fold MV + -NR activity of the bacteroids compared to those from plants inoculated with the WT strain. When N 2 O emissions from the nodules were analysed, N 2 O levels decreased 3.5-fold in the nodules produced by the nap mutant compared to the WT nodules. Interestingly, inoculation of the plants with the nap + strain resulted in a large increase of N 2 O emissions (about 6.7-fold) from these nodules compared to those from plants inoculated with the WT strain (Fig. 3b). Collectively, these results indicate that E. meliloti Nap  (Table 3). Supplementary  Fig. S3C shows the experimental setting for these experiments. As shown in Table 3, NFW, SDW, PDW and N content significantly decreased in plants grown without Cu 2+ added compared to those grown in the presence of Cu 2+ . No differences in those parameters were found between plants grown with 0.8 or 20 µM Cu 2+ . Moreover, plants grown without Cu 2+ added displayed a pale green tone in their leaves, while they were dark green in the other treatments (Supplementary Fig. S3). These results indicate a negative effect of Cu-limitation on alfalfa-E. meliloti SNF.
Leghemoglobin values support this idea, since nodules from plants grown without Cu 2+ contained 1.8fold less leghemoglobin than those from plants grown with 0.8 µM Cu 2+ (Table 3). As observed in Table 3, Cu concentration was higher in roots and nodules comparing to shoots, especially in plants grown with 20 µM Cu 2+ , where we found 5.5-fold and 7.6-fold more Cu in nodules and roots, respectively, compared to shoots. According to these results, Cu may be primarily accumulated in the roots, and only a minimal proportion would be transferred to shoot and leaves. Similar results were obtained by using E. meliloti 2011 as WT (data not shown).
In order to elucidate the contribution of Nos to N 2 O emissions from alfalfa nodules, bacteroids were isolated from nodules elicited by the WT 1021 or a nosZ mutant. These nodules were collected from plants grown for 50 days and subjected to 10 mM nitrate and flooding during 7 days before harvesting ( Supplementary Fig. S1C). Bacteroids from the nosZ mutant showed 5.9-fold less N 2 OR activity than bacteroids from the WT strain. On the contrary, N 2 O emission rates by nosZ − nodules were 5.3-fold higher compared to those by nodules from plants inoculated with the WT strain (Table 4). Similar results were obtained by using E. meliloti 2011 as WT (data not shown). These results demonstrate the involvement of Nos in N 2 O reduction in alfalfa nodules. Finally, another set of alfalfa pots was inoculated with the WT strain and grown without Cu added (0 µM), or in the presence of 0.8 µM or 20 µM Cu 2+ . Seven days before harvesting (i. e., after 43 days), plants were treated with 10 mM KNO 3 and subjected to flooding in order to induce N 2 O emissions (Supplementary Fig. S1C). As shown in Table 4, Cu 2+ accumulation in nodules was correlated with the Cu concentration added to NS. Moreover, while MV + -NR activity from the bacteroids was not significantly influenced by Cu availability, MV + -NIR activity was significantly reduced in bacteroids from 0 or 20 µM Cu 2+ treatments comparing with 0.8 µM Cu 2+ . N 2 OR activity increased in parallel with the Cu concentration provided. Conversely, N 2 O emission rate decreased with Cu concentration. These results suggest that environmental Cu concentration plays an essential role in modulating bacteroidal NIR and N 2 OR activities Table 3 Effect of Cu 2 + on nodule fresh weight (NFW), shoot dry weight (SDW), plant dry weight (PDW), leghemoglobin (Lb) content in nodules, nitrogen concentration ([N]) and Cu concentration ([Cu 2+ ]) in shoots, roots and nodules from alfalfa plants inoculated with E. meliloti 1021 Plants were grown for 43 days without nitrate supplied and with Cu 2 + added to nutrient solution (0.8 or 20 µM) or not (0 µM). Cu content in seeds was also determined (17 ± 0.5 mg·Kg − 1 of dry seeds, or 28 ± 0.8 ng per seed). Data represent means from three independent experiments assayed by using ten pot replicates containing 8 plants each. Values in a row followed by the same lower-case letter are not statistically different according to U Mann-Whitney test at p ≤ 0.05; n.d., not detected and consequently in the decrease of N 2 O emissions by alfalfa nodules.

Discussion
E. meliloti is unable to grow under free-living anaerobic conditions with nitrate as sole electron acceptor. This incapacity is due to the very low expression of nap genes compared to nirK, nor and nos denitrification genes (Torres et al. 2014). In fact, an overexpressing nap mutant (nap + ) recovered the ability to grow anaerobically with nitrate as well as the capacity to produce N 2 O (Torres et al. 2018). In this context, it would be appealing to explore if the capacity to grow anoxically from nitrate respiration confers to E. meliloti an advantage to infect and nodulate alfalfa roots under anoxic conditions, as this question has not been yet addressed for endosymbionts. Our results demonstrate that overexpression of Nap increases the fitness of E. meliloti-alfalfa symbiotic interaction, since plants inoculated with the nap + strain showed a higher capacity for nodulation than those inoculated with the WT under low oxygen conditions. Similar to our observations, Lecomte et al. (2021) also found that nap genes play an important role in root colonization efficiency of the plant-associated microorganism Agrobacterium fabrum. Supporting our findings, previous studies demonstrated that B. diazoefficiens nirK or norC mutants showed a reduced ability for nodulation in soybean plants grown with nitrate (Mesa et al. 2004). These authors proposed that denitrification enzymes played a role in nodule formation rather than in nodule function. On the contrary, no significant differences in competitiveness for nodulation were observed between WT or a napA E. meliloti mutant in M. truncatula, suggesting that Nap is not involved in the early steps of the interaction (Ruiz et al. 2022). The apparent discrepancy between both sets of results can be explained by the difference in the strains used. While in Ruiz et al. (2022), a mutant lacking nap was used, in our work we have used a strain overexpressing nap that confers to E. meliloti the ability to grow under anoxic conditions by nitrate respiration in contrast to the WT, which is not able to respire nitrate anoxically. Another important difference between our work and Ruiz et al. (2022) results is that in our experiments nutrient solution was fluxed with N 2 for 10 min before inoculation in order to provoke low oxygen conditions during the first steps of infection. Our results do not invalidate those obtained by Ruiz et al. (2022), since we do not suggest that nap is important for nodulation, but it might help when nap is overexpressed and roots are subjected to low oxygen conditions during the first steps of the interaction. In soybean plants, it has been reported that 4 mM nitrate and flooding conditions induce N 2 O emissions from nodules and that the bacteroidal denitrification is the main process involved (Tortosa et al. 2015). Since E. meliloti lacks the ability to grow and produce N 2 O under free-living anoxic conditions, we were interested in investigating the capacity of E. meliloti to produce N 2 O under symbiotic conditions as well as the involvement of bacteroidal denitrification. To achieve this goal, we cultivated alfalfa plants in the presence of nitrate. In contrast to soybeans, which are very tolerant to nitrate, 4 mM nitrate is excessive for alfalfa plants, since the number of nodules elicited by E. meliloti was strongly diminished under this concentration. Even more, although 1 or 3 mM nitrate present in the growth NS did not affect nodule number, these nitrate levels caused the reduction of nodule biomass. In order to investigate the effect of nitrate in N 2 O emissions, we have selected a 7-day 10 mM treatment together with flooding, since this nitrate treatment did not affect nodule growth and had a small effect on nodule functionality. This parameter was determined by analysing leghemoglobin content of the nodules, which is directly related with their capacity to fix N 2 . While 7-day 10 mM nitrate treatment provoked a slight decrease in Lb content compared to non-nitrate treatment (7.38 mg versus 10.9 mg), addition of 1 or 3 mM nitrate to the nutrient solution during the entire growth period (50 days) severely diminished Lb content (around 6 mg) indicating a more severe effect on nodule fitness. Nitrate and flooding drastically induced N 2 O emissions, as it was previously reported in soybeans (Tortosa et al. 2015). Regarding the involvement of Nap in N 2 O release from the nodules, we found a correlation between MV + -NR activity and N 2 O emissions either in the nap − or the nap + strains, indicating that Nap has an important role on N 2 O released from alfalfa nodules in response to nitrate and flooding. In this context, Brambilla et al. (2020) isolated novel E. meliloti strains which produced lower N 2 O emissions comparing to the model strain 1021 or the commercial strain B399, and reported that all these isolates harboured spontaneous mutations in napC gene, which encodes NapC, a c-type cytochrome required for Nap activity. Our results complement previous studies where the involvement of bacteroidal nitrate reduction in NO synthesis in Medicago truncatula nodules was reported (Horchani et al. 2011).
Cu is an important micronutrient involved in many physiological plant processes (Nagajyoti et al. 2010;Yruela 2009). In the present work, we have shown the negative effect of Cu limitation in symbiotic nitrogen fixation and a consequent incapacity of the nodules to provide all the N demands required by the alfalfa plant. In fact, non-Cu treated plants displayed a pale green tone in shoot and leaves indicating that Cu-limitation would affect chlorophyll synthesis causing a drastic diminution in plant biomass and chlorosis in leaves. Similarly, in a recent study, Printz et al. (2016) reported leaf chlorosis and a lower leaf density in alfalfa plants inoculated with a commercial peat-based inoculant and grown with 3 and 30 nM Cu. Therefore, the present work highlights that an adequate Cu supply is essential for a proper SNF. With respect to Cu accumulation, it is known that the interaction of Cu with amino and carboxyl ligands reduces its translocation to the shoots (Nikolaevna et al. 2016). Supporting this assertion, Printz et al. (2016) found that Cu accumulation in roots was 325-fold higher in the presence of 10 µM Cu comparing to 3 nM Cu (i.e., the highest and the lowest Cu concentrations assayed in that study), whereas this ratio only reached 50-and 22-fold in stems and leaves, respectively. Our results are coherent with these observations, since shoots displayed a significantly lower Cu concentration in both Cu treatments, 0.8 and 20 µM Cu 2+ , comparing to roots and nodules. In a recent study, Tortosa et al. (2020) demonstrated that 20 µM Cu 2+ treatment was the maximal Cu concentration that soybean plants could bear before suffering Cu stress. Our studies in alfalfa support this idea, since SNF was not affected by 20 µM.
Lastly, NR, NIR and N 2 OR activities were analysed in bacteroids from nodules of alfalfa plants grown in the presence of different Cu levels. Since Nap was involved in N 2 O emissions from alfalfa nodules, as discussed above, the influence of Cu on its activity and on the following denitrification enzyme, NirK, was interesting to be explored. Nevertheless, the results obtained showed that MV + -NR activity was not significantly influenced by Cu at the concentration range assayed in the present work. Similarly, previous studies showed that NR activity of B. diazoefficiens bacteroids was not affected by 20 µM Cu 2+ (Tortosa et al. 2020). The notable decrease in MV + -NIR activity observed in bacteroids under 0 µM Cu 2+ comparing with 0.8 µM Cu 2+ highlights that NirK is a Cudependent enzyme. Furthermore, MV + -NIR activity was also significantly diminished in the presence of 20 µM Cu 2+ . In a previous work, Tortosa et al. (2020) found a significant reduction in this activity when soybean bacteroids were subjected to 10, 20 40 and 60 µM Cu. Therefore, a Cu concentration of 20 µM affects NirK function either in bacteroids from soybean or alfalfa nodules. In B. diazoefficiens, NirK requires a cytochrome c 550 (CycA) as an essential electron donor for its activity (Bueno et al. 2008). In this context, it has been reported that biogenesis of cytochromes c can be blocked by an excess of Cu (Durand et al. 2015). Thus, the sensitivity of NirK to high Cu levels in soybean or alfalfa nodules may be due to a negative effect on the periplasmic cytochrome c 550 (CycA) biogenesis. The contribution of Nos to N 2 O reduction in alfalfa nodules has also been demonstrated, since N 2 OR activity was significantly higher in WT bacteroids comparing to nosZ mutant bacteroids. On the contrary, N 2 O emissions were notably reduced in the WT nodules compared to the levels obtained in the nosZ − nodules. Similar results were reported by Tortosa et al. (2015) for soybean nodules, highlighting the role of Nos as the sole enzyme able to reduce N 2 O to N 2 and confirming the role of Nos as a key enzyme in N 2 O mitigation strategies. In this context, it has been proposed that Cu concentration in agricultural soils may importantly affect N 2 O emissions from microbial processes, especially denitrification and nitrification (Li et al. 2019). Furthermore, Tortosa and colleagues (2020) demonstrated that the decrease in N 2 O emissions was concomitant with the increase in the Cu concentration added to NS, and concluded that Cu was a relevant factor involved in N 2 O reduction in soybean nodules. Following the same line, the results displayed in the present work suggest that Cu bioavailability significantly induces N 2 OR activity and, by extension, reduces N 2 O emissions from nodules elicited by E. meliloti.
Taken together, the results from this work suggest a controversial advantage of nap overexpression, since it improves nodulation capacity under oxygen-limiting conditions, but contributes to N 2 O emissions at the same time. In this regard, a strategy for an effective E. meliloti-alfalfa symbiotic interaction and N 2 O mitigation from legume crops might be the selection of inoculants with an adequate Nap expression and high N 2 OR activity. In this context, it has been recently shown that N 2 O emissions from soybean crops can be reduced at the field scale by inoculation with a mixed culture of indigenous strains of B. diazoefficiens isolated from agricultural fields that show high N 2 OR activity levels (Akiyama et al. 2016).

Conclusion
The present work reports for the first time the capacity of alfalfa nodules to emit N 2 O in response to nitrate, flooding and copper limitation. Furthermore, the involvement of E. meliloti Nap in the competence for nodulation and infectivity effectiveness, as well as in N 2 O emissions from alfalfa nodules has also been demonstrated. Finally, we also report the capacity of Cu to modulate bacteroidal NIR and N 2 OR activity and consequently N 2 O emissions from alfalfa nodules.