Dynamics of hemoglobins during nodule development, nitrate response, and dark stress in Lotus japonicus

Abstract Legume nodules express multiple leghemoglobins (Lbs) and non-symbiotic hemoglobins (Glbs), but how they are regulated is unclear. Here, we study the regulation of all Lbs and Glbs of Lotus japonicus in different physiologically relevant conditions and mutant backgrounds. We quantified hemoglobin expression, localized reactive oxygen species (ROS) and nitric oxide (NO) in nodules, and deployed mutants deficient in Lbs and in the transcription factors NLP4 (associated with nitrate sensitivity) and NAC094 (associated with senescence). Expression of Lbs and class 2 Glbs was suppressed by nitrate, whereas expression of class 1 and 3 Glbs was positively correlated with external nitrate concentrations. Nitrate-responsive elements were found in the promoters of several hemoglobin genes. Mutant nodules without Lbs showed accumulation of ROS and NO and alterations of antioxidants and senescence markers. NO accumulation occurred by a nitrate-independent pathway, probably due to the virtual disappearance of Glb1-1 and the deficiency of Lbs. We conclude that hemoglobins are regulated in a gene-specific manner during nodule development and in response to nitrate and dark stress. Mutant analyses reveal that nodules lacking Lbs experience nitro-oxidative stress and that there is compensation of expression between Lb1 and Lb2. They also show modulation of hemoglobin expression by NLP4 and NAC094.


Introduction
Hemoglobins are widespread in archaea, bacteria, and eukaryotes, showing an outstanding diversity of structures and functions (see reviews by Vinogradov et al., 2005;Hoy and Hargrove, 2008;Becana et al., 2020).Symbiotic hemoglobins seem to be restricted to the nodules of legumes and actinorhizal plants.Their main function is to transport O 2 in the cytosol of infected cells to the N 2 -fixing bacteria at a steady low concentration compatible with nitrogenase activity and bacterial respiration (Appleby, 1984;Ott et al., 2005;Wang et al., 2019).The symbiotic hemoglobins of legumes (leghemoglobins; Lbs) are present in nodules at millimolar concentrations as a mixture of multiple isoproteins or components encoded by different genes (Larrainzar et al., 2020).Plants also express non-symbiotic hemoglobins (phytoglobins; Glbs) at (sub)micromolar concentrations in virtually all organs, from roots and leaves to flowers and fruits (Hill, 2012).Glbs are categorized into three classes according to phylogeny and biochemical properties.Class 1 Glbs have an extremely high O 2 affinity and are therefore unlikely to transport and deliver O 2 .They contribute to the survival of plants under hypoxic conditions and to the homeostasis of nitric oxide (NO), a key signaling gas molecule with a plethora of functions (Domingos et al., 2015).Class 1 Glbs are also critical for the onset and functioning of the legume-rhizobia symbiosis by modulating NO levels (Shimoda et al., 2005;Fukudome et al., 2016;Berger et al., 2020).Class 2 Glbs show homology with Lbs, moderate O 2 affinity, and NO-scavenging activity (Hebelstrup and Jensen, 2008;Smagghe et al., 2009).Class 3 Glbs share homology with the 'truncated' hemoglobins of prokaryotes, exhibit low O 2 affinity, and have unknown functions (Vieweg et al., 2005).
The major aim of this work was to investigate the regulation of the whole set of hemoglobins in a leguminous plant during nodule development and in response to nitrate and dark stress conditions.For this purpose, it was essential to individualize all Lb and Glb mRNAs and proteins.We chose the model legume Lotus japonicus because it is more amenable for this study than Medicago truncatula, which contains 12 Lbs and five Glbs (Berger et al., 2020;Larrainzar et al., 2020), and also because we wanted to investigate two unusual Glbs, Glb2-1 and Glb2-2, tentatively assigned to class 2 Glbs.This assignment is based on a number of differences of those Glbs with respect to genuine Lbs, albeit they do not seem to be either typical class 2 Glbs like those of Arabidopsis, tomato, and sugar beet (Becana et al., 2020).The genome of L. japonicus contains nine hemoglobin genes that encode three Lbs, two class 1 Glbs, two class 2 Glbs, and two class 3 or 'truncated' Glbs.The first part of our study was focused on the regulation of hemoglobins by deploying mutants that differ in Lb composition (Wang et al., 2019).The second part of the study examined the expression of all the hemoglobins in nodules induced to senesce by treating plants with nitrate or exposing them to dark stress.For this we used wild-type (WT) and mutant plants defective in the transcription factor NODULE INCEPTION-LIKE PROTEIN 4 (NLP4), previously known as NITRATE UNRESPONSIVE SYMBIOSIS 1 (NRSYM1), or in the transcription factor NAC094.These transcription factors were selected because NLP4 is related to the nitrate response of the symbiosis (Nishida et al., 2018) and NAC094 is associated with nodule senescence (Wang et al., 2023).Here, we show that hemoglobins of L. japonicus are distinctly regulated during development as well as in the nitrate and dark stress responses of nodules.Furthermore, our results with loss-of-function mutants demonstrate that Lb-deficient nodules experience nitro-oxidative stress and provide insights into how the expression of the various Lb and Glb isoproteins is modulated by NLP4 and NAC094.

Plant material and growth conditions
Single (lb3), double (lb13 and lb23), and triple (lb123) lb mutants of Lotus japonicus ecotype MG-20 were generated using CRISPR/Cas9 [clustered regularly interspaced palindromic repeats (CRISPR)/CRISPRassociated protein 9] genome editing (Wang et al., 2019).Mutant plants of ecotype MG-20 deficient in NLP4 or NAC094 were isolated by chemical mutagenesis (Nishida et al., 2018) or by CRISPR/Cas9 (Wang et al., 2023), respectively.Seeds were germinated and seedlings were grown in plates following standard protocols (Villar et al., 2021) in a cabinet with 23 °C/21 °C (day/night), 120-140 μmol photons m −2 s −1 , and a 16 h photoperiod.Seedlings were then transferred to vermiculite-containing pots, nodulated with Mesorhizobium loti strain MAFF303099 (1.5 × 10 8 cells per seedling), and grown in cabinets, being watered twice a week with N-free B&D nutrient solution (Broughton and Dilworth, 1971).For nitrate treatment, plants at 4 weeks post-inoculation (wpi) were supplied for 2 d with B&D solution containing 0 (control), 0.5, 5, or 10 mM KNO 3 as indicated in each experiment.For dark stress, nodulated plants were exposed to continuous darkness, for the number of days indicated in each experiment, in growth cabinets under otherwise identical conditions.Nodules were harvested in liquid nitrogen and stored at -80 °C until use.

Gene expression
Total RNA was extracted from nodules using the RNAqueous Total RNA isolation kit and was treated with DNase I (Thermo Fisher Scientific).cDNA was synthesized with MMLV reverse transcriptase (Promega).Quantitative reverse transcription-PCR (qRT-PCR) analyses were performed using a QuantStudio3 Real-Time PCR System (Applied Biosystems) as described (Rubio et al., 2019).Gene identifiers, primer sequences, and amplification efficiencies are provided in Supplementary Table S1, and gene exon-intron composition and primer positions are given in Supplementary Fig. S1.The amplification efficiency for each set of primers was calculated by analysis of standard curves obtained from serial template dilutions (Pfaffl, 2001).Normalized relative quantities (NRQ) were calculated using the geometric means of two reference genes (Rubio et al., 2019), either LjUbiquitin and LjATP synthase or LjUbiquitin and LjeIF4A.The selection of the reference gene pairs was made based on their stability in each experiment.

Immunoblots
Immunoblots were carried out on 15% native-PAGE or 15% SDS-PAGE gels following standard protocols.The primary antibodies were generated against common bean Lba and L. japonicus Glb1-2, Glb2-1, and Glb3-2, and were affinity purified (Sainz et al., 2013(Sainz et al., , 2015)).Ferritin antibody was generated against Arabidopsis ferritin 1 and was kindly provided by Brigitte Touraine (University of Montpellier, France).The secondary antibody was goat anti-rabbit IgG conjugated with horseradish peroxidase.Further details are indicated in the figure legends.Immunoreactive proteins were detected by chemiluminescence using the SuperSignal West Pico kit (Pierce) and images were acquired using a ChemiDoc MP Imaging System (Bio-Rad).

Detection of reactive molecules
ROS and NO were localized in nodule sections (~85 µm) by cytochemical staining.Superoxide radical production was detected using nitroblue tetrazolium (NBT).Nodule sections were immersed in staining buffer (50 mM potassium phosphate, pH 7.8, 0.25 mM NBT) for 1 h at room temperature in the dark.For H 2 O 2 detection, nodule sections were stained with 3,3'-diaminobenzidine (DAB) in buffer (10 mM Tris-HCl, pH 7.4, 1 mg ml -1 DAB) for 5 min in the dark at room temperature.For NO detection, nodule sections were incubated with 10 µM 4-amino-5-methylamino-2',7'-difluorofluorescein diacetate in buffer (10 mM Tris-HCl, pH 7.4, 10 mM KCl) for 30 min at room temperature in the dark.To test inhibitors and/or scavengers, nodule sections were pre-incubated for 1 h (ROS) or 30 min (NO) in the dark, then subjected to the same staining protocol as described above but keeping the inhibitors during the incubation period.Specifically, we examined the effects of diphenyleneiodonium (100 µM DPI) on NBT staining; KI (10 mM) on DAB staining; and 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (1 mM cPTIO) on NO staining.DPI is an inhibitor of NADPH oxidase, KI is a H 2 O 2 scavenger; and cPTIO is an NO scavenger (Dunand et al., 2007;Sandalio et al., 2008).Images of ROS were acquired using a DM750 (Leica) light microscope and images of NO with an LSM880 (Zeiss) confocal microscope with excitation at 488 nm and detection at 498-549 nm.

Determination of antioxidant enzyme activities
Enzymes were extracted from nodules in a medium consisting of 50 mM potassium phosphate buffer (pH 7.8), 0.1 mM EDTA, 1% (w/v) soluble polyvinylpyrrolidone, and 0.1% (v/v) Triton X-100.Nodule extracts were centrifuged at 13 000 g for 10 min at 4 °C, and superoxide dismutase (SOD), catalase, and ascorbate peroxidase (APX) activities were determined in the supernatants using standard protocols.Total SOD activity was measured following the inhibition of the reduction of ferricytochrome c by superoxide radicals generated by a xanthine-xanthine oxidase system (Rubio et al., 2001).The SOD isoforms were resolved on native-PAGE gels stained with NBT and were identified by in-gel differential inhibition with 3 mM KCN (inhibitor of CuZnSOD) and 5 mM H 2 O 2 (inhibitor of CuZnSOD and FeSOD).Further identification of SOD isoforms was based on previous studies (Rubio et al., 2007).The activity of each isoform was assessed by calculating the relative proportions by gel densitometry and referring those values to the total SOD activity of extracts.Catalase activity was assayed by following the decomposition of H 2 O 2 at 240 nm (Aebi, 1984), and APX activity by following ascorbate oxidation at 290 nm (Asada, 1984).All enzyme activities were measured at 25 °C within a linear range.

Search for nitrate-responsive elements and double nitrateresponsive elements in hemoglobin promoters
The promoters of L. japonicus hemoglobin genes were searched using MEME (https://meme-suite.org/meme/).To sensitively detect nitrateresponsive elements (NREs) in the target promoters, they were analyzed along with nitrite reductase (NiR) genes of 38 different plant species (-500 bp to ATG) that were previously shown to contain canonical NREs.The sequence list can be found in Bailey et al. (2006) and Jiang et al. (2021).To detect double nitrate-responsive elements (dNREs) in the L. japonicus hemoglobin promoters, they were analyzed along with 12 Lb genes of M. truncatula that were previously shown to harbor single copies of these elements.The sequence list can be found in Jiang et al. (2021).For searching, the minimum and maximum sites per motif used were set at 21-60 bp and 24-50 bp for the dNRE and NRE analyses, respectively.The detected motifs were confirmed as positive matches through TOMTOM (MEME suite) and by comparison with published sequences (Jiang et al., 2021).

Gene profiling reveals distinct regulation of each Lb and Glb during the development and nitrate response of nodules
We first quantified the expression levels of all L. japonicus hemoglobin genes during nodule development.Nodules were harvested at 2 wpi (young), 4-6 wpi (mature), and 8-10 wpi (senescent).The data of this and other experiments involving qRT-PCR were calculated as NRQ.To this end, it was critical to design gene-specific primers because the three Lb genes share extremely high nucleotide sequence identity in the coding regions: 99% between Lb1 and Lb2, 87% between Lb1 and Lb3, and 87% between Lb2 and Lb3.This problem was circumvented by designing one of the primer pairs in the 5'-untranslated region (UTR) of Lb2 and in the 3'-UTR of Lb1 (Supplementary Fig. S1).The hemoglobin genes display distinct expression levels and patterns (Fig. 1).In young nodules, there was a difference of ~8000-fold in the mRNA levels of hemoglobin genes between the most (Lb3) and least (Glb2-2) expressed.Based on the order of magnitude of expression (NRQ for young nodules in parentheses), we can distinguish four groups of hemoglobins: Lbs (13-73); Glb1-1, Glb2-1, and Glb3-1 (0.5-1.4); Glb1-2 and Glb3-2 (0.03); and Glb2-2 (0.009).The same groups were observed, albeit with different ranges, when considering the NRQ in senescing nodules.These observations are in line with the high Lb concentrations usually observed in legume nodules (Berger et al., 2020;Larrainzar et al., 2020), as well as with the higher expression of Glb1-1 and Glb3-1 relative to their Glb1-2 and Glb3-2 counterparts in nodules of L. japonicus grown in hydroponics (Bustos-Sanmamed et al., 2011).
The Lb3 gene was by far the most highly expressed among all hemoglobin genes and did not vary during nodule development (Fig. 1).In contrast, Lb1 expression was drastically reduced from young to mature nodules, remaining thereafter at a low level, whereas Lb2 expression decreased gradually with advancing age.There were only slight or no changes in expression of the genes encoding class 1 and 3 Glbs, except Glb3-1, which was reduced ~3-fold in senescent nodules.Moreover, Glb2-1 and Glb2-2 behaved like Lb2, with a progressive reduction in expression from young to senescent nodules (Fig. 1).We also investigated the response of hemoglobin gene expression in nodules of plants that had been growing without nitrate until 4 wpi and then supplied with low or high nitrate for 2 d (Fig. 2).Nitrate concentrations were selected for comparison because 0.5 mM nitrate is non-inhibitory for nodulation, whereas 5 mM and 10 mM nitrate induce nodule senescence (Nishida et al., 2018;Wang et al., 2023).Nitrate decreased expression of Lbs and class 2 Glbs; however, it increased expression of class 1 Glbs gradually from 0.5 mM to 10 mM and of class 3 Glbs at 0.5 mM and 5 mM, returning to the control value at 10 mM.Interestingly, a decline in Lb and Glb2 mRNAs was already apparent at 0.5 mM nitrate relative to nodules without nitrate.Based on the transcriptional response to this low concentration, the most nitrate-sensitive Lb gene was found to be Lb1, followed by Lb2 and Lb3, with respective decreases of 62, 53, and 34% in mRNA levels.No major differences in the nitrate response were observed between the two members within each class of hemoglobin genes (Fig. 2).The differential expression of hemoglobin genes in response to nitrate delineates a clear-cut separation of Lbs from class 1 and 3 Glbs.However, as occurred with nodule development (Fig. 1), this was not the case for class 2 Glbs, which behave like Lbs (Fig. 2).

Lb levels are precisely regulated through transcriptional feedback
To gain further insight into regulation of hemoglobin genes, we characterized mutants deficient in one, two, or all three Lbs (Fig. 3).For this study, it was essential to quantify each Lb mRNA by qRT-PCR and each Lb protein on immunoblots of native gels.This allowed us to verify the primer specificities and the knockout mutations, as well as to uncover a compensatory phenomenon by which the double mutants overexpress their single functional Lb.Thus, we observed an increase of Lb1 expression (mRNA and protein) in the lb23 mutant and of Lb2 expression (mRNA and protein) in the lb13 mutant relative to the WT nodules (Fig. 3).We then phenotyped the nodulated mutants at 4 wpi by measuring growth parameters (Supplementary Fig. S2).At this stage, all plants exhibited an intense green color of leaves, with the exception of lb123, which appeared slightly less green, suggesting the start of nitrogen deficiency.The single and double mutants showed decreases in the growth of the shoot (length and weight) and leaves (number and weight) and, in the case of lb13, also of the root (length and weight) (Supplementary Fig. S2).The triple mutant showed decreases in all of these parameters except root length.Interestingly, lb3, lb13, and lb123 had lower nodule weights per plant than the WT, but lb123 produced more nodules, probably in an attempt to overcome the lower nitrogen fixation rate observed in this mutant (Wang et al., 2019).
The next step was to determine the expression profiles of Glbs (mRNAs and proteins) in nodules (Fig. 4A, B).No major changes were found except for the lb123 mutant, which showed a lower expression of Glb1-1 (mRNA and protein) and Glb2-2 (mRNA) and a higher expression of Glb3-1 (mRNA and protein) relative to the WT (Fig. 4A, B).The decrease in Glb2-2 mRNA could not be tested at the protein level because Glb2-2 is expressed at levels ~100-fold lower than Glb2-1 (Figs 1, 2, 4).Unexpectedly, we observed in nodule extracts two close immunoreactive bands with the Glb3-2 antibody (Fig. 4B).This antibody recognizes Glb3-1 and Glb3-2 as they share 83% amino acid identity.However, the two protein bands are still present in glb3-2 knockout mutants but almost disappear in glb3-1 knockout mutants (Supplementary Fig. S3), indicating that both bands correspond to Glb3-1.These might have originated by alternative splicing or posttranslational modification.The enhanced content of both proteins matched the Glb3-1 mRNA profile in the lb123 mutant but not in the lb13 and lb23 mutants, which showed enhanced protein content with no change in the transcript level (Fig. 4A, B).

Lb-deficient mutants show alterations in the production of reactive molecules and antioxidant defenses leading to nitro-oxidative stress
We next investigated whether Lb deficiencies cause nitrooxidative stress in nodules by localizing the production of ROS and NO by using cytochemical staining.The production of superoxide radicals, H 2 O 2 , and NO increased consistently in the nodules of lb123 but not in the nodules of the single and double mutants, indicating that the presence of a single Lb is sufficient to avoid nitro-oxidative stress (Fig. 5).To confirm this, we measured some key antioxidant enzyme transcripts and activities in all the mutant nodules (Fig. 6A, B).The transcript level of each SOD gene was quantified by qRT-PCR (Fig. 6A) and the SOD isoforms were in-gel assayed for activity with NBT (Fig. 6B).To calculate the activity of each isoform, densitometric values of abundance were referred to the total SOD activity of nodule extracts.Interestingly, we found major changes in the SOD isoform composition of lb3 and lb123 nodules, with an apparent compensation in expression (mRNAs and enzyme activities) of the two cytosolic isoforms.Thus, in the two mutants, there was a decrease in FeSODc and a concomitant increase in CuZnSODc, while the total SOD activity remained constant.Significant changes were also detected in the FeSODc mRNA level of lb13 and in the mitochondrial MnSOD activity of lb123 (Fig. 6A, B).We next measured the activities of the two major H 2 O 2 -scavenging enzymes of nodules.Catalase is a peroxisomal enzyme and APX occurs as several isoforms; however, the APX activity assayed in our study corresponds to the cytosolic isoform because it is by far the most abundant in nodules and, unlike the other isoforms, is stable when ascorbate is omitted from the extraction medium (Dalton et al., 1987).We found significant decreases of catalase activity in the lb123 mutant and of APX activity in the lb3 and lb123 mutants (Fig. 6C).
Because the mRNA and enzyme activity profiles of CuZnSOD and FeSOD were drastically affected in lb3 and lb123 nodules, we examined the expression of three key genes associated with metal homeostasis and senescence: copper chaperone for SOD (CCS), cysteine protease 1 (CYP1), and ferritin (Fig. 7A, B).CCS is required for Cu delivery to CuZnSOD, and the expression of CCS and CuZnSOD is regulated by Cu availability (Cohu et al., 2009).CYP1 is a senescence-associated gene (SAG) and a useful marker of nodule senescence (Wang et al., 2023).Ferritin stores Fe in a safe form avoiding Fenton reactions, is regulated by Fe availability, and is responsive to oxidative stress (Ravet et al., 2009).We found that the three genes were highly up-regulated in the lb123 nodules, with CYP1 being the most notorious case because its mRNA was undetectable in the WT and all other mutant nodules (Fig. 7A, B).Notably, the lb3 mutant nodules also showed some changes in ferritin mRNA and protein (Fig. 7B) but these were less pronounced than in the lb123 mutant.However, the lb3 nodules showed a similar production of ROS and NO compared with the WT nodules (Fig. 5), and the CCS and CYP1 mRNA levels were not affected (Fig. 7A).Taking together the results shown in Figs 5-7, we conclude that nodules of lb123, and to a lesser extent lb3, experience nitro-oxidative stress and alterations of metal homeostasis.

NAC094 is a marker of nodule aging and stressinduced senescence and is required for CYP1 activation even in the absence of nitrate
The contrasting patterns of the various Lbs and Glbs in nodules during aging or after nitrate supply prompted us to further investigate their regulation by means of two lossof-function mutants, nlp4 and nac094, of transcription factors.(B) Immunoblots of SDS-PAGE gels loaded with 10, 20, or 30 μg of protein per lane for Glb1, Glb2, and Glb3, respectively.Proteins were detected with affinity-purified polyclonal antibodies raised against L. japonicus Glb1-2 (αGlb1), Glb2-1 (αGlb2), and Glb3-2 (αGlb3).The primary and secondary antibodies were used at dilutions of 1:1000 and 1:80 000, respectively.The immunoblots are representative of gels loaded with proteins from three different biological replicates per genotype.The molecular mass (kDa) is indicated on the left.
We chose them because NLP4 is a master regulator of the symbiotic nitrate response in M. truncatula (Lin et al., 2018) and L. japonicus (Nishida et al., 2018) and because NAC094 is a positive regulator of nitrate-induced nodule senescence (Wang et al., 2023).Prior to hemoglobin profiling, we determined the responses of both transcription factors in nodules under our experimental setup.First, we quantified NLP4 and NAC094 mRNAs during nodule development of WT plants (Supplementary Fig. S4A) as well as in mature nodules of lb mutants (Supplementary Fig. S4B).No change in the NLP4 mRNA level was found under such conditions.In sharp contrast, NAC094 was highly up-regulated in aging nodules, especially at 10 wpi, and in the lb123 mutant, although its expression was unaffected in the single and double lb mutants.Second, we quantified NAC094 mRNA in nlp4 nodules.Following supply of plants at 4 wpi with low or high nitrate for 2 d, the expression level of this gene remained constant in the nlp4 mutant (Supplementary Fig. S5).Because NAC094 is up-regulated by high nitrate through NLP4 (Wang et al., 2023), our results may be interpreted in terms of a superior symbiotic tolerance of nlp4 plants to nitrate (Lin et al., 2018;Nishida et al., 2018).Subsequent experiments included dark stress for comparison with high nitrate because both treatments induce nodule senescence (Escuredo et al., 1996).Plants were grown without nitrate until 4 wpi and then exposed to continuous darkness for 1-5 d under otherwise identical environmental conditions (Fig. 8).In non-stressed nodules, the NAC094 mRNA was virtually undetectable.In contrast, dark stress led to a progressive up-regulation of NAC094 in the WT and nlp4 mutant.Nevertheless, the expression profiles were not identical in the two genetic backgrounds (Fig. 8), which suggests an interplay between NLP4 and NAC094.To confirm that NAC094 expression was linked to nodule senescence in a nitrate-independent manner, we profiled the senescence marker gene CYP1.We found that its expression level followed a similar trend to NAC094 but was lower in The production of superoxide radicals was localized using nitroblue tetrazolium staining, which is reduced by superoxide to insoluble blue formazan.A control with 100 µM diphenyleneiodonium (DPI) was included to test the involvement of NADPH oxidases in superoxide generation.The production of H 2 O 2 was detected by the oxidation of 3,3'-diaminobenzidine to a brown precipitate by endogenous peroxidases.A control with 10 mM potassium iodide (KI), which oxidizes H 2 O 2 , was included to ascertain genuine H 2 O 2 production.The production of NO was localized using 4-amino-5-methylamino-2',7'-difluorofluorescein diacetate.This cell membrane-permeable probe is hydrolyzed by intracellular esterases and reacts with N 2 O 3 derived from NO in the presence of O 2 to yield a highly fluorescent triazole (green).A control with 1 mM 2-(4-carboxyphenyl)-4,4,5,-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO), an NO scavenger, was included to ascertain genuine production of NO.For all three panels, note the higher staining intensity in the nodule sections of the lb123 mutant and, notably, a lower intensity in those of the lb3 mutant.For each ROS and NO, three or four nodules per genotype with two sections per nodule were examined.The whole experiment was repeated three times for ROS and twice for NO.Scale bars, 150 µm.Fig. 6.Expression of several major antioxidant enzymes in nodules of leghemoglobin mutants (lb) of L. japonicus.Nodulated plants were grown without nitrate until 4 wpi.(A) Transcript levels of superoxide dismutase (SOD) genes are expressed in normalized relative quantities (NRQ) using LjUbiquitin and LjATP synthase as reference genes.(B) Total SOD and SOD isoform activities are expressed in units per gram of nodule fresh weight.One SOD unit is defined as the amount of enzyme that inhibits the reduction of ferricytochrome c by 50%.(B) also shows an in-gel SOD activity assay (negative image of the gel) revealing the various isoforms: mitochondrial MnSOD (MnSODm), bacteroid MnSOD (MnSODb); FeSOD (two bands: the upper most intense band is the cytosolic isoform; the lower faint band is the plastid isoform); CuZnSODc, cytosolic CuZnSOD; and CuZnSODp, plastid CuZnSOD.(C) Catalase and ascorbate peroxidase (APX) activities are given, respectively, in mmol of H 2 O 2 decomposed and µmol of ascorbate oxidized per min and gram of nodule fresh weight.For all panels, data are means ±SE of 3-7 biological replicates, and asterisks indicate significant differences relative to the WT based on Student's t-test (*P<0.05;**P<0.01;***P<0.001).
nlp4 than in the WT after 4 d and 5 d of darkness (Fig. 8).In connection to this finding, we observed that, whereas CYP1 is strongly activated in WT nodules after 3 d of dark stress, the activation is suppressed in the nac094 mutant (Supplementary Fig. S6).We conclude that NAC094 is a general key marker of nodule senescence because it is drastically up-regulated not only in lb123 nodules but also during developmental and stress-induced senescence.Our results also indicate that NAC094 is required for activation of CYP1 in dark-stressed nodules and that this activation also occurs in the absence of nitrate.

NLP4 and NAC094 modulate hemoglobin expression in a gene-specific manner in response to nitrate
We next profiled expression of Lb and Glb genes under nitrate or dark stress in the nlp4 and nac094 mutants.For clarity and statistical analysis, the results obtained with the two mutants are represented in separate figures but they use the same data of the WT plants.This representation facilitates comparison of each mutant with the WT as well as the response of each genotype to nitrate or dark stress.We found higher mRNA levels of genes encoding Lbs and class 2 Glbs in the nlp4 mutant than in the WT at low (0.5 mM) and high (10 mM) nitrate concentrations, consistent with the relative nitrate insensitivity of nlp4 (Fig. 9).In contrast, expression of Glb1-2 and Glb3-1 displayed similar levels and nitrate responses in both genotypes.Expression of Glb1-1 and Glb3-2 increased with nitrate in the WT but not in nlp4.Of these four genes, only Glb3-2 showed a significantly lower expression in nlp4 than in WT plants (Fig. 9).These results, together with those of Fig. 2, not only underline a distinct regulation and sensitivity of individual Lb and Glb genes in response to increasing nitrate, but also evidence transcriptional modulation of most hemoglobin genes by NLP4.The exceptions seem to be Glb1-2 and Glb3-1, which display similar expression trends in the nodules of WT and nlp4 plants.There were also distinct effects of dark stress on hemoglobin gene expression.Thus, in the WT nodules, the expression of Glb1-1 increased and that of Glb3-2 remained unaffected, whereas the expression of all other hemoglobin genes was suppressed or drastically reduced (Fig. 9).However, comparison of control (0 N) and dark-stressed nodules between the WT and nlp4 reveals some differences in the two backgrounds: Glb1-1 (up-regulated the WT), Glb1-2 (down-regulated in the WT), and Glb3-2 (up-regulated in nlp4) (Fig. 9).
A similar expression profiling of Lb and Glb genes was conducted in the nac094 mutant (Fig. 10).Only a few differences were observed between the nodules of WT and nac094 mutant plants with nitrate or dark stress.In response to low nitrate, consistent differences were found in the expression of Lb3 and Glb2-1 (up-regulated by ~58% in nac094).We also detected a major difference of Glb2-2 expression (up-regulated by ~92% in nac094) with low and high nitrate (Fig. 10).These results indicate that NAC094 modulates the expression of the genes coding for Lb3 and class 2 Glbs, although to variable extents depending on the presence of non-inhibitory or inhibitory nitrate concentrations.

Conserved NREs and dNREs were identified in hemoglobin genes of Lotus japonicus
The results shown in Figs 2, 9, and 10 indicate that the expression of hemoglobin genes in nodules of L. japonicus is altered in response to different concentrations of nitrate.Promoters of NiR genes were previously shown to contain NREs that serve as direct targets for activation of their expression by NLPs (Konishi and Yanagisawa, 2010), whereas dNREs were subsequently identified in the promoters of Lb genes and are required for high expression of Lbs in nodules (Jiang et al., 2021).Analysis of the Lb and Glb promoter sequences using MEME revealed the presence of NREs in the promoters of Glb1-1, Glb1-2, and Glb2-1 and of dNREs in the promoters of Lb genes and Glb2-1 (Fig. 11).However, NRE or dNRE motifs were not detected in the promoters of Glb2-2, Glb3-1, and The sequences and positions of the NRE and dNRE motifs are provided in Supplementary Tables S2 and  S3, respectively.

Discussion
Hemoglobins are encoded by multigene families and occur as complex mixtures of isoproteins in plant tissues (see reviews by Hoy and Hargrove, 2008;Hill, 2012;Becana et al., 2020).Early studies reported that the proportions of Lb isoproteins in soybean and pea varied with nodule age (Fuchsman and Appleby, 1979;Uheda and Syōno, 1982;Sarath et al., 1986) and that the abundance of two Lbs in mungbean responded oppositely to application of high nitrate (Becana and Sprent, 1989).At that time, the limited availability of genomic sequences, model legumes, and mutants precluded comprehensive and detailed studies of hemoglobin gene regulation.Here we present such a study encompassing the full complement of L. japonicus hemoglobins.To this end, we used mutants deficient in Lbs or in transcription factors associated with nitrate sensitivity and nodule senescence (Nishida et al., 2018;Wang et al., 2019Wang et al., , 2023)).
A relevant finding of our study was the contrasting transcript profiles of the three Lb genes during nodule development (Fig. 1), which suggests a major contribution of Lb1 early in nodule development and a more general role for Lb3 throughout the life span of nodules.This further indicates that Lb3 is regulated, to some degree, independently of Lb1 and Lb2, despite them all having dNREs.These results contrast with previous studies with L. japonicus (Wang et al., 2019) and soybean (Du et al.,  2020) in which no major differences were seen in Lb expression profiles during nodule development.The discrepancies may be ascribed to the use of different legume species and plant growth conditions; for example, plants were fed on 0.5 mM nitrate by Wang et al. (2019) while no nitrate was employed in this study.We also found considerable variations in the sion of Glb genes during nodule development.Genes of class 2 Glbs behaved like Lb1 and Lb2 whereas those of class 1 Glbs and Glb3-2 remained essentially unaffected.Notably, Glb3-1 was down-regulated in old nodules but the expression of the orthologous gene of M. truncatula increased by 2-fold under similar conditions (Berger et al., 2020), pointing out differences between legumes in hemoglobin regulation.
We also observed contrasting responses of each hemoglobin gene to nitrate.In particular, it was puzzling that Lb gene expression was so sensitive to a nitrate concentration commonly used to promote nodule formation.This was most evident in the Lb1 and Lb2 mRNA levels, which were reduced by ~58% at 0.5 mM nitrate (Fig. 2).The distinct nitrate responses place in one group the genes of Lb and the unusual class 2 Glbs (down-regulated with 0.5-10 mM) and in another group the genes of class 1 Glbs (up-regulated with 5-10 mM) and class 3 Glbs (up-regulated with 0.5-5 mM).To facilitate comparisons among Lbs and the three Glb classes of L. japonicus, we have compiled some important biochemical properties of the proteins along with findings of the present work in Table 1.Although Glb2-1 and Glb2-2 have homology with Lbs and typical class 2 Glbs, they also show some differences.On the one hand, unlike Lbs, which do not contain cysteine (with the exception of Lb1 of Aeschynomene evenia; Quilbé et al., 2021), Glb2-1 and Glb2-2 contain cysteine residues and occur in nodules at much lower concentrations (especially Glb2-2) than Lbs.Also, Glb2-1 shows a hexacoordinate heme in the ferric form, does not complement the lb123 triple mutant, and its deficiency in non-nodulated plants delays growth and causes alterations in the leaf metabolome (Villar et al., 2021).On the other hand, Glb2-1 and Glb2-2 show homology with Lbs, are mainly but not exclusively expressed in nodules, and their expression is strongly down-regulated by nitrate and dark stress, as is the case of Lbs (Table 1).In a recent phylogenetic analysis of ~100 sequences of Lbs, class 1 Glbs, and class 2 Glbs (Becana et al., 2020), we observed that Glb2-1 clusters with M. truncatula Lb3 (MtLb3; Medtr1g090810) and that they shared 66% identity and 81% similarity in amino acid sequences.After phenotypic examination of a knockout mutant and biochemical characterization of the protein, we concluded that MtLb3 an atypical Lb (Villar et al., 2021).It will be worthwhile to investigate whether similar 'unusual' Lbs or Glbs with high amino acid identity with L. japonicus Glb2-1 or Glb2-2 occur in other legumes.
Previously we had observed an enhanced production of superoxide and H 2 O 2 in lb123 mutant nodules (Wang et al., 2019).Here, we show that those nodules undergo nitrooxidative stress as evidenced by an increased production of ROS and NO (Fig. 5), as well as by major alterations in antioxidant defenses (Fig. 6) and expression of marker genes (Fig. 7).
The finding that NO accumulates in the absence of Lbs is a most surprising and novel finding because plants used for cytochemical studies were grown in the absence of nitrate, whereas the only demonstrated pathways of NO production in nodules, namely plant and bacterial nitrate reductases and the mitochondrial electron transport chain, require nitrate (Horchani et al., 2011).An alternative pathway involving an NO synthaselike activity reported in lupine nodules (Cueto et al., 1996) was never verified by direct measurements of NO, nor has a plant enzyme responsible for such activity been identified; hence, we cannot ascertain that the observed accumulation of NO originated through this pathway.We conclude that there exists a nitrate-independent pathway for NO generation that becomes conspicuous in lb123 nodules.Why is the NO not scavenged to avoid nitro-oxidative stress?Based on our results, we propose that this is due to the lack of Glb1-1 and Lbs (Figs 3, 4) because these hemoproteins are able to convert NO to nitrate in vitro through their NO dioxygenase activities (Herold and  , 2005;Rubio et al., 2019).This may also happen in vivo as down-regulation of Glb1-1 by 3-fold caused a moderate increase of NO in mature nodules of M. truncatula (Berger et al., 2020).Here we found that the lb123 nodules seem to behave as knockout mutants of Glb1-1 because this protein became almost undetectable on immunoblots (Fig. 4B).However, Lbs may also be important NO scavengers, taking into account that they exhibit NO dioxygenase activity and are present at much higher concentrations than Glb1-1.

Puppo
The lb123 mutant nodules also showed some alterations related to Fe homeostasis and/or metabolism.This is supported by two observations.First, ferritin, located in nodule plastids, is transcriptionally activated by oxidative stress (Ravet et al., 2009) and by an increase of intracellular free Fe that may be mediated by NO signaling during nodule senescence (Chungopast et al., 2017).We found that both NO and ferritin (mRNA and protein) accumulate in lb123 which is fully consistent with that report (Figs 5,7).Second, SODs are regulated by the availability of their metal cofactors (Alscher et al., 2002;Moran et al., 2003;Cohu et al., 2009).We found in the mutant nodules that FeSODc is replaced by CuZnSODc, which, together with the up-regulation of CCS, suggests some alterations in the availability of Fe and Cu.Also of interest, catalase and APX activities decreased markedly in the lb123 mutant nodules, which may explain the accumulation of H 2 O 2 .According to a previous RNA-seq analysis of lb123 nodules (Wang et al., 2019), the transcript levels of APX are not affected.This leads us to conclude that the decline of activities is probably a result of enzyme degradation that takes place during senescence, as evidenced by the activation of CYP1 and NAC094 Fig. 11.Detection of nitrate-responsive elements (NREs) and double nitrate-responsive elements (dNREs) in the promoters of L. japonicus hemoglobin genes using MEME.(A) Sequence logos of the detected NREs and dNREs.(B) Location of NREs and dNREs in hemoglobin promoters.NREs (red boxes) were detected in the promoters of L. japonicus Glb1-1, Glb1-2, and Glb2-1 by analyzing them with a large set of nitrite reductase (NiR) promoters from different plant species as described in the Materials and methods.dNREs (cyan boxes) were detected in the promoters of L. japonicus Lb genes and Glb2-1 by analyzing them with promoters of 12 Lb genes of Medicago truncatula.No NRE or dNRE motifs were detected in the promoters of L. japonicus Glb2-2, Glb3-1, and Glb3-2.The gene identifiers for M. truncatula NiR, Glycine max NiR, M. truncatula Lb4, and M. truncatula Lb11 are Medtr4g086020, Glyma.07G212800,Medtr1g090820, and Medtr5g081030, respectively.(Fig. 7A; Supplementary Fig. S4B) (Wang et al., 2023).An unexpected observation is that the lb3 mutant nodules display several alterations resembling those of lb123 but in an attenuated form.However, the double mutants also lacking Lb3, namely lb13 and lb23, do not show those alterations.We cannot offer an explanation for this, except that the overexpression of Lb2 and Lb1, respectively, compensates at least partially for the deficiency of Lb3.
In our comparison among L. japonicus hemoglobins, the detection of NREs and/or dNREs in their gene promoters is especially relevant (Fig. 11; Table 1).The NREs have been identified in the promoters of many well-known nitrate-regulated genes, such as NiR (Konishi and Yanagisawa, 2010), whereas the dNREs are required for activation of Lb expression in nodules (Jiang et al., 2021).Indeed, the expression of the three Lb genes, all of which feature dNREs, was negatively correlated with nitrate concentration (Fig. 2).We found NREs in the promoters of Glb1-1 and Glb1-2, and both genes showed progressively higher levels of expression in nodules with increasing nitrate concentrations.The expression of these genes was not, however, affected in nlp4 nodules (this work), suggesting that some other NLP regulates them.However, Glb2-1 also contained an NRE in its upstream sequence but it was repressed, rather than induced, by nitrate.This might be explained by our finding that the Glb2-1 promoter also has a dNRE (Supplementary Tables S2, S3).In fact, the discovery of both NRE and dNRE in close proximity in the promoter of Glb2-1 provides further support for our conclusion that Glb2-1 is neither a genuine Lb nor a genuine class 2 Glb (Larrainzar et al., 2020;Villar et al., 2021;this work).Because all known Lbs, except those of A. evenia (Quilbé et al., 2021), were recruited for symbiotic function through evolution of class 2 Glbs, we propose that Glb2-1 is an intermediate gene between Lbs and class 2 Glbs.It is thus not surprising that Glb2-1 also performs non-symbiotic functions, as proven by the altered phenotype of non-nodulated glb2-1 mutants (Villar et al., 2021) and its expression in flowers (Lotus Expression Atlas, https://lotus.au.dk/).However, Glb2-1 is consistently down-regulated even with 0.5 mM nitrate, much like Lb2 (Fig. 2), which suggests that dNRE is prevalent over NRE in its nitrate response.Interestingly, Glb2-2, a close homolog of Glb2-1 (63% identity and 77% similarity in amino acid sequences), harbored neither of these elements.Our data show that the expression level of Glb2-2 is two orders of magnitude lower than that of Glb2-1 (Table 1), demonstrating that they are subject to different transcriptional regulation.We failed to detect NREs or dNREs in the promoters of Glb3-1 and Glb3-2, even though both genes showed increased expression in nodules treated with 0.5 mM or 5 mM nitrate (Fig. 2).This could be due to either NREs present outside of the sequence analyzed, or a regulation independent of NLPs.

Fig. 1 .
Fig. 1.Expression of hemoglobin genes during development and senescence of L. japonicus nodules.Transcript levels were normalized using LjUbiquitin and LjeIF4A as reference genes.Data are given in normalized relative quantities (NRQ) and are means ±SE of 3-5 biological replicates.Different letters indicate significant differences among means according to Duncan's multiple range test (P<0.05).

Fig. 2 .
Fig.2.Expression of hemoglobin genes in response to nitrate in L. japonicus nodules.Nodulated plants were grown without nitrate until 4 wpi and then were treated with 0, 0.5, 5, or 10 mM KNO 3 for 2 d.Control plants supplied with 10 mM KCl for 2 d in place of nitrate were also included.Transcript levels were normalized using LjUbiquitin and LjeIF4A as reference genes.Data are given in normalized relative quantities (NRQ) and are means ±SE of 3-7 biological replicates.Asterisks indicate significant differences compared with the control (0 N) based on Student's t-test (*P<0.05;**P<0.01;***P<0.001).

Fig. 3 .
Fig. 3. Expression of the three leghemoglobins (Lbs) in nodules of Lb mutants (lb) of L. japonicus.Nodulated plants were grown without nitrate until 4 wpi.The knockout lb mutants were generated by CRISPR/Cas9 and are deficient in one (lb3), two (lb13 and lb23), or all three (lb123) Lbs.(A) Transcript levels were normalized using LjUbiquitin and LjATP synthase as reference genes and are given in normalized relative quantities (NRQ).Data are means ±SE of 3-5 biological replicates.Asterisks indicate significant differences with respect to the WT plants based on Student's t-test (*P<0.05;**P<0.01;***P<0.001).(B)Immunoblots of native-PAGE gels were used to resolve the three Lbs.Gels were loaded with 25 μg of protein per lane and challenged with an affinity-purified polyclonal antibody raised against common bean Lba.The primary and secondary antibodies were used at dilutions of 1:1000 and 1:80 000, respectively.Each immunoblot is representative of gels loaded with proteins from three different biological replicates per genotype.The right side of (B) shows the densitometric determination (ImageJ) of Lb amounts on gels, which were normalized with respect to those of the WT.Data are means ±SE of three biological replicates, and asterisks denote significant differences based on Student's t-test (**P<0.01).The hyphens indicate that the protein is not detectable.

Fig. 5 .
Fig. 5. Localization of ROS and NO in nodules of leghemoglobin mutants (lb) of L. japonicus.Nodulated plants were grown without nitrate until 4 wpi.The production of superoxide radicals was localized using nitroblue tetrazolium staining, which is reduced by superoxide to insoluble blue formazan.A control with 100 µM diphenyleneiodonium (DPI) was included to test the involvement of NADPH oxidases in superoxide generation.The production of H 2 O 2 was detected by the oxidation of 3,3'-diaminobenzidine to a brown precipitate by endogenous peroxidases.A control with 10 mM potassium iodide (KI), which oxidizes H 2 O 2 , was included to ascertain genuine H 2 O 2 production.The production of NO was localized using 4-amino-5-methylamino-2',7'-difluorofluorescein diacetate.This cell membrane-permeable probe is hydrolyzed by intracellular esterases and reacts with N 2 O 3 derived from NO in the presence of O 2 to yield a highly fluorescent triazole (green).A control with 1 mM 2-(4-carboxyphenyl)-4,4,5,-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO), an NO scavenger, was included to ascertain genuine production of NO.For all three panels, note the higher staining intensity in the nodule sections of the lb123 mutant and, notably, a lower intensity in those of the lb3 mutant.For each ROS and NO, three or four nodules per genotype with two sections per nodule were examined.The whole experiment was repeated three times for ROS and twice for NO.Scale bars, 150 µm.

Fig. 7 .
Fig. 7. Expression profiles of genes associated with senescence and metal homeostasis in nodules of leghemoglobin mutants (lb) of L. japonicus.(A) Transcript levels of copper chaperone for CuZnSOD (CCS) and cysteine protease (CYP1).(B) Transcript and protein levels of ferritin.For both panels, transcript levels are expressed in normalized relative quantities (NRQ) using LjUbiquitin and LjATP synthase as the reference genes.Data are means ±SE of 3-4 biological replicates.For the immunoblot, proteins (10 μg per lane) were resolved on SDS-PAGE gels and ferritin was detected with an antibody raised against Arabidopsis ferritin 1.The primary and secondary antibodies were used at dilutions of 1:5000 and 1:40 000, respectively.The immunoblot shows the molecular mass (kDa) on the left and the values of the densitometric scan (ImageJ) below each lane.Data are means ±SE of three biological replicates, normalized to the values of the WT nodules.For both panels, asterisks indicate significant differences relative to the WT according to Student's t-test (**P<0.01;***P<0.001).

Fig.
Fig. Profiles of NAC094 and CYP1 transcripts in nodules of WT and nlp4 mutant plants of L. japonicus during dark stress.Nodulated plants grown without nitrate until 4 wpi were exposed to prolonged darkness for 1-5 d.Transcript levels are expressed as normalized relative quantities (NRQ) using LjUbiquitin and LjeIF4A as reference genes.Data are means ±SE of 3-12 biological replicates.Means denoted by different uppercase black letters (for treatment comparisons in WT plants) or lowercase green letters (for treatment comparisons in nlp4 plants) are significantly different according to Duncan's multiple range test (P<0.05).For each treatment, asterisks indicate significant differences between the WT and nlp4 based on Student's t-test (*P<0.05;**P<0.01).

Fig. 9 .
Fig. 9. Expression of leghemoglobin (Lb) and phytoglobin (Glb) genes in nodules of nlp4 mutants of L. japonicus in response to nitrate or dark stress.Nodulated plants were grown without nitrate until 4 wpi and then were treated with 0, 0.5, or 10 mM KNO 3 for 2 d, or were exposed to continuous darkness for 3 d.Transcript levels are given in normalized relative quantities (NRQ) using the geometric means of LjUbiquitin and LjeIF4A.Data are means ±SE of 3-11 biological replicates.Means denoted by different black uppercase letters (for treatment comparisons in WT plants) or green lowercase letters (for treatment comparisons in nlp4 plants) are significantly different from each other according to Duncan's multiple range test (P<0.05).For each treatment, asterisks indicate significant differences between the WT and nlp4 based on Student's t-test (*P<0.05;**P<0.01;***P<0.001).

Fig. 10 .
Fig. 10.Expression of leghemoglobin (Lb) and phytoglobin (Glb) genes in nodules of nac094 mutants of L. japonicus in response to nitrate or dark stress.Nodulated plants were treated as indicated in Fig. 9. Transcript levels are given in normalized relative quantities (NRQ) using the geometric means of LjUbiquitin and LjeIF4A.Data are means ±SE of 3-11 biological replicates.Means denoted by different black uppercase letters (for treatment comparisons in WT plants) or orange lowercase letters (for treatment comparisons in nac094 plants) are significantly different from each other according to Duncan's multiple range test (P<0.05).For each treatment, asterisks indicate significant differences between the WT and nac094 based on Student's t-test (*P<0.05).
a Number of amino acid residues.bProtein molecular mass in kDa including the heme group.cd Plant organ showing maximal expression.eHemoglobin classification based on spatial disposition of α-helices relative to the heme: canonical (3/3-fold) or truncated (2/2-fold) tertiary structures.f Main known function, which may not be the single one.g Expression levels in mature nodules (4 wpi) are given as normalized relative quantities.h Up-regulation (↑ moderate; ↑↑ strong), no change (=), or down-regulation (↓ moderate; i Presence of nitrate-responsive elements (NREs) or double nitrate-responsive elements (dNREs) in the gene promoters.n.d., not detected.