Exogenous hydrogen sulphide alleviates nodule senescence in Glycine ma x‑ Sinorhizobium fredii symbiotic system

Aim Hydrogen sulphide (H 2 S), as a signalling molecule, regulates the process of plant growth and development. H 2 S is involved in the establishment of the legume and rhizobial symbiotic relationship, but little is known about the role of H 2 S during nodule senescence. Method We studied the function of H 2 S during nodule senescence in the Glycine ma x-Sinorhizobium fredii symbiotic system. Physiological and biochemical methods were used to modulate nitrogen fixation, nodule structure, senescence-related protein and gene expression. Results Our results indicated that H 2 S decreased the senescent nodule number, increased nitrogenase (Nase) activity and ureide content, and further promoted the accumulation of NO 3− -N, NH 4+ -N, and total N, which caused the accumulation of biomass in the late steps of symbiosis. H 2 S alleviated nodule senescence by maintaining the normal nitrogen fixation structure of nodules. Meanwhile, the expression levels of the NifH and NodGS proteins and the N metabolic enzyme activities were enhanced by H 2 S during nodule senescence. Moreover, H 2 S stimulated the expression of nitrogen metabolism-related genes in the nodules. Finally, senescence-related gene expression levels were regulated by H 2 S during nodule senescence.

(Nase) activity and ureide content, and further promoted the accumulation of NO 3 − -N, NH 4 + -N, and total N, which caused the accumulation of biomass in the late steps of symbiosis. H 2 S alleviated nodule senescence by maintaining the normal nitrogen fixation structure of nodules. Meanwhile, the expression levels of the NifH and NodGS proteins and the N metabolic enzyme activities were enhanced by H 2 S during nodule senescence. Moreover, H 2 S stimulated the expression of nitrogen metabolism-related genes in the nodules. Finally, senescence-related gene expression levels were regulated by H 2 S during nodule senescence. Conclusion These findings suggested that H 2 S prolonged the lifespan of N 2 fixation by delaying nodule senescence, which will be important for increasing soybean yield and agricultural sustainability development.

Introduction
Legume yields are significantly increased due to their symbiotic association with rhizobia. Therefore, legumes play an important role in sustainable agriculture (Fox et al. 2007). Legumes interact with rhizobia to form a unique root organ, the nodule, in which the rhizobia reduce atmospheric N 2 to ammonia (NH 3 ) and differentiate into bacteroids (Puppo et al. 2005). The late stage of nodule development, including determinate and indeterminate nodules, modulates their fate and function, namely, nodule senescence. Nodule senescence starts from the nodule centre, and slowly extends to the periphery in determinate nodules such as in Glycine max and Lotus japonicus (Puppo et al. 2005). During nodule senescence, the color of the nitrogen fixation zone is converted from pink to green, suggesting that functional leghemoglobin is degraded (Puppo et al. 2005). In other words, a decrease in leghemoglobin causes the loss of nitrogen fixation activity during natural nodule senescence. In addition, at the ultrastructural level, in determinate nodules, the loose plasma membrane separates from the cell wall and the electron density of the cytoplasm becomes progressively lower. In addition, the cell wall is damaged, the bacteroids are dissolved, and the peribacteroid membrane is degraded (Puppo et al. 2005;Timmers et al. 2000; Van et al. 2006).
Proteolytic activity plays an important role in the process of nodule senescence, which can help to remove misfolded or modified proteins (Pladys and Vance 1993). When this enzyme activity isn't controlled, the key regulators and symbiosome membrane proteins are influenced, and nitrogen fixation is ultimately inhibited. Cysteine protease (CP; EC 3.4.22) mainly catalyzes cysteine to a nucleophile during proteolysis. Cysteine proteases in most plants are members of the papain (C1) and legumain (C13) families, and recently, in some plants, calpains (family C2, calcium-dependent proteases) and metacaspases (family C14) have been found (Grudkowska and Zagdańska 2004). Of all the cysteine proteases, the plant papain-like cysteine proteases (PLCPs) have been studied the most. The expression levels of cysteine protease genes, such as GmCYSP1, have been shown to be induced during nodule senescence in G. max nodules (Alesandrini et al. 2003;Oh et al. 2004). Moreover, nodule senescence is regulated by transcription factors, such as MtATB2, MtNAC969, MtbHLH2 and MtNAC920 (Deng et al. 2019). The expression levels of MtATB2, encoding a bZIP transcription factor, was shown to obviously increase in senescence nodules (de Zélicourt et al. 2012).
Moreover, phytohormones, including jasmonic acid (JA), ethylene, and abscisic acid (ABA), positively modulate nodule senescence (Guinel 2015;Puppo et al. 2005; Van et al. 2006). For instance, ABA can induce nodule senescence through regulating the expression levels of senescence-associated genes (SAGs) (Fukudome et al. 2018;González et al. 2001). In addition, the ethylene and JA have been found to positively induce the expression levels of ETHYLENE RESPONSE FACTOR (ERF) and ACC synthase (ACS) genes, and JA biosynthesis (Tatiana 2017; Van et al. 2006). Interestingly, GAs and nitric oxide (NO) positively influence nodule senescence (Cam et al. 2012;Fukudome et al. 2018;Serova et al. 2019). For instance, previous studies showed that a decrease in nitrogen fixation and early nodule senescence is caused by an increase in endogenous NO levels in M. truncatula, which is closely related to bacterial flavohaemoglobin (hmp) gene expression (Cam et al. 2012). During nodule senescence, NO production is related to the N 2 -fixing zone or at the boundary between the N 2 -fixing and senescence zones (Baudouin et al. 2006;Cam et al. 2012). Fukudome et al. (2018) reported that overexpression of the phytoglobin LiGlb1-1 delayed nodule senescence by decreasing NO levels in roots and nodules in L. japonicus. Therefore, the regulation of nodule senescence is complex, but whether other signalling molecules are involved in nodule senescence remains unclear.
Hydrogen sulfide (H 2 S) acts as a gaseous signalling molecule at lower concentrations and can participate in many biological processes in animals (Li et al. 2006;Wang 2002). Recently, many studies have revealed the physiological function of H 2 S in plants (Du et al. 2019;Liu et al. 2016;Scuffi et al. 2018). For instance, H 2 S regulates seed germination, root growth and development, and adaptive responses to environmental stresses (Aroca et al. 2021;Baudouin et al. 2016;Chen et al. 2018Chen et al. , 2011Zhang et al. 2009). Interestingly, H 2 S plays a vital role in plant tissue senescence. For example, during dark-induced senescence, H 2 S delays leaf yellowing by maintaining energy status and increasing antioxidative capacity in spinach seedlings (Hu et al. 2015). H 2 S delays flower senescence by changing antioxidative activities (Zhang et al. 2011). Additionally, during the postharvest of horticultural products H 2 S significantly regulates senescence-related gene expression levels (Huo et al. 2018).
Interestingly, our recent studies found that H 2 S plays a vital role in legume-rhizobia interactions (Zou et al. , 2019. For instance, H 2 S promotes the establishment of the legume and rhizobia relationship, and endogenous H 2 S not only accumulates during the first step of the interaction but also is producted in the mature nodules ). Moreover, exogenous H 2 S promoted soybean growth, nodulation, nitrogenase (Nase) activity, Nase component (NifH) protein expression, and symbiosisrelated gene expression in the G. max-Sinorhizobium fredii symbiotic system (Zou et al. 2019). Our previous research showed that H 2 S was generated in the nitrogen-fixing zone of soybean nodules, and that the deletion of the cystathionineγ-lyase (CSE) gene in S. fredii (ΔCSE) led to an obvious decrease in H 2 S production in both the free-living rhizobia and soybean nodules ). In addition, using physiological and element stoichiometry methods, It was shown that during N deficiency-induced senescence in soybean, H 2 S and rhizobia regulated nitrogen (N) assimilation and remobilization , and H 2 S interacted with rhizobia to modulate the growth rate and nutrient resorption efficiency in soybean seedlings (Chen et al. 2022).
However, the role of H 2 S in the late stages of symbiosis is still unclear. In this work, we address this issue using the legume G. max and S. fredii. We analysed the effect of exogenous H 2 S addition on the soybean growth phenotype, senescent nodule proportion, Nase activities, ureide content, total N, NH 4 + -N, and NO 3 − -N contents, nodule ultramicrostructure, nitrogen fixation-related protein expression, N metabolic enzyme activities, N metabolism-and senescence-related gene expression at the late stage of symbiosis (from 3 to 8w) in soybean nodules. Our results suggested that H 2 S played a vital role in nodule senescence and that exogenous H 2 S addition could significantly delay nodule senescence and further increase the lifespan of N fixation. This research will offer a possible solution to increase soybean production by delaying nodule senescence and enhancing N fixation efficiency.

Plant materials and H 2 S treatments
Glycine max (Zhonghuang13) seeds were surface sterilized with 75% ethyl alcohol for 30 s and then washed with 50% sodium hypochlorite for 4 min. The sterilized seeds were washed with ddH 2 O 6 times. Seeds were placed on a plate with 1% agar for 48 h in the dark at 28 °C. Nitrogen-free nutrient solution (400 ml) was added to the growth medium (700 ml) of vermiculite and perlite (1:1 v/v). The nitrogen-free nutrient solution was prepared according to the method of Zou et al. (2019). The growth medium was placed into a polypropylene planting bag. One seedling was placed into a bag. Seedlings were grown in a controlled growth chamber with a relative humidity of 80%, a light/dark regime of 16/8 h, a temperature of 27 °C, and photosynthetically active radiation of 280 μmol/m 2 /s. Soybean seedlings of 10-day-old were separated into four groups: (1) control plants without 100 μM NaHS and rhizobia; (2) seedlings treated with 100 μM NaHS (S); (3) seedlings inoculated with Sinorhizobium fredii Q8 (Q8); and (4) seedlings inoculated with rhizobia and treated with 100 μM NaHS (Q8 + S). In the Q8 and Q8 + S groups, seedlings of 10-day-old were inoculated with 10 ml of a rhizobial suspension (the optical density at OD 600 nm was 0.05). After 4 weeks of rhizobial treatment, seedlings in the S and Q8 + S groups were treated with 100 μM of 10 ml NaHS solution every 3 days. To maintain a steady humidity and ionic concentration, sterile nitrogen-free nutrient solution (50 ml) was added to each bag every 7 days. The seedlings were harvested at 5 weeks (w), 6w, 7w, and 8 w, and half of the samples were immediately frozen in liquid nitrogen and stored at -80 °C while the other half was dried to a constant weight for dry matter determination.

Biomass and chlorophyll content determination
The total dry weight, shoot and root dry weights of soybean were measured with an analytical balance. The chlorophyll content was determined with a SPAD-502 Plus leaf chlorophyll metre (Konica Minolta, Kumamoto, Japan). Fifteen leaves from 5 soybean plants per treatment were measured for chlorophyll content at 10:00 A.M.

Nodule number and senescent nodule proportion assessment
The nodule number of soybean seedlings was counted at 5, 6, 7, and 8 w. A binocular microscope was used to observe the nodules. When at least half of its length was green, the nodule was considered senescent. The ratio (× 100) of the senescent nodule number over the total number of nodules per root was used to calculate the proportion of senescent nodules.

Nitrogenase (Nase) activity assay
The acetylene reduction assay (ARA) was used to determine the Nase activity (Fishbeck et al. 1973). Nitrogen fixation was determined on fresh nodule tissue in a 10 ml rubber-capped airtight glass bottle, filled with a mixture of acetylene and air (v:v = 1:100). Bottles were incubated at room temperature for 2-3 h. Gas samples (250 μl) were analysed in a gas chromatography system (Agilent Technologies, La Jolla, USA). For each treatment, the height of the ethylene peak on the chromatogram relative to the length of the reaction was used to measure the amount of ethylene produced. Standard curves were produced with pure ethylene, which were used to calibrate the results of the gas chromatography.

Ureide content determination
Dried roots and nodules for each treatment and for each age were homogenized. The homogenates were placed in a boiling water bath for 20 min. After cooling, the solutions were centrifuged at 12,000 g for 20 min. The supernatants were measured following the instructions from the manufacturer (Michy Biolory, M0216A). In this method, allantoin is transformed to allantoate by alkaline hydrolysis and determined by a special absorption peak at 535 nm. At least three replicates were included for all analyses. NO 3 − -N, NH 4 + -N, and total N content determination For NO 3 − -N determination, a series of standard KNO 3 solutions were used to prepare a standard curve. Dried leaf, stem, root, and pod samples (0.05 to 0.1 g) were ground with liquid nitrogen and boiled in 1 ml of double distilled H 2 O for 30 min, and the supernatant (0.1 ml) was added to 0.4 ml of 5% salicylic acid in sulfuric acid (5 g salicylic acid dissolved in 100 ml 98% concentrated sulfuric acid). After 20 min at 25 C, an 8% NaOH solution (9.5 ml) was added, and the absorbance at 410 nm (OD 410 ) was measured and compared with the standard curve to estimate the NO 3 − concentration. For NH 4 + -N determination, a series of ammonia standard solutions were used to prepare a standard curve. Dried leaf, stem, root, and pod samples (0.05 to 0.1 g) were ground with 5 ml of 10% acetic acid and double distilled H 2 O was added to 50 ml. A 2 ml aliquot of ammonia standard and 2 ml of supernatant were added to 0.1 ml of 1% ascorbic acid, followed by mixing 3 ml of ninhydrin solution containing 50 mM ninhydrin, 10% propanol, 30% butanol, and 60% ethanediol. The absorbance at 570 nm (OD 570 ) was measured and compared with the standard curve to estimate the NH 4 + content. The total N contents were measured by the Kjeldahl method. First, 0.2 g of dry sample was ground into powder, and then placed into a digestive tube. Five millilitres of concentrated H 2 SO 4 was added, followed by shaking and mixing. Subsequently, the mixture was digested in a microwave digestion system (Labtec™ Line, FOSS, Denmark) at 365 °C, and every 30 min, seven to eight drops of 30% H 2 O 2 were added. Finally, the digestion solution was diluted to a constant volume with distilled water, and then an automatic Kjeldahl apparatus (Kjeltec™ 8400, FOSS, Denmark) was used to measure the N content.
Light microscopic observation of the nodule structure Nodules were harvested at 5, 6, and 7 w. The nodule structure observations were analysed according to the method of Chou et al. (2016). Nodules were fixed with formalin:acetic acid:alcohol stationary liquid (70% ethanol:methanol:acetic acid = 16:1:1, v/v/v) for 3 days at 4 C. Fixed nodules were then dehydrated with a graded ethanol series of 50% (30 min), 80% (1 h), 95% (1 h), and 100% (1 h × 2), and every step was performed twice at room temperature. Nodules were treated with 1% sarranine (v/v) and then infiltrated with xylene:ethanol (1:1), and then 100% xylene, and each step was repeated twice at room temperature for 1 h. Next, nodules were infiltrated with liquid filtered paraffin (Paraplast) at 56 C for 2 days, including four changes of paraffin. Finally, the nodules in solidified paraffin were cut into 8-μm slices on a microtome. Slices containing the whole nodule structure were then deparaffinized with xylene (1 h), xylene:ethanol (1:1, 5 min), a descending ethanol series (100%, 95%, 80%, 70%, 50%, and 30%, 5 min each), and distilled water (5 min). After deparaffinization, nodule slices were stained with 0.5% toluidine blue for 5 min. Finally, after treatment with distilled water, 95% ethanol, and xylene, the nodule structures were observed under an optical microscope (Olympus, Tokyo, Japan) and images were captured. For each age, nine nodule sections from three nodules for each treatment were observed and analysed.

Electron microscopy observation of the nodule micromorphology
Nodules harvested at 5, 6, and 7 w were observed using a transmission electron microscope (FEI, Czech Republic, USA) to assess the micromorphological differences during nodule senescence between NaHStreated nodules and controls. Sample preparation was prepared according to the method of Chou et al. (2016). For each age, nine nodule sections from three nodules for each treatment were observed and analyzed.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting assays
Total protein of samples (0.5 g) was extracted according to the method of Chen et al. (2011), and then the protein concentrations were measured with a BCA protein quantification kit (Genstar, Beijing, China).
Proteins (20 μg for each sample) were separated by SDS-PAGE using 12% (w/v) acrylamide gels and transferred to a polyvinylidene difluoride (PVDF) membrane for 2 h for Western blotting analysis. Five percent skim milk powder was used to block the membrane overnight. The protein blots were probed with primary antibodies for Nase iron-containing protein (NifH) (1:1000, AS01021A, Agrisera, Vännäs, Sweden) and ornodulin/glutamate-ammonia ligase-like protein (NodGS) (1:500, AS153030, Agrisera, Vännäs, Sweden) at 4 °C overnight. Then the blots were washed three times using phosphatebuffered saline with Tween-20 (PBST) solution containing 150 mM NaCl, 50 mM TRIS-HCl (pH 8.0), and 0.05% Tween-20(v/v), and then incubated with secondary antibody (antirabbit immunoglobulin G horse-radish peroxidase-conjugated, 1:5000 dilution; Sungen, Tianjin, China) at 4 °C overnight. An internal control was Actin (1:5000, AS132640, Agrisera, Vännäs, Sweden). Finally, the blots were washed three times with phosphate-buffered saline-Tween and imaged with a Molecular Imager Gel Dox XR System (Bio-Rad, Hercules, CA, USA). ImageJ software was used to estimate the protein levels. The comparative optical density was used to determine the relative amount of protein expression, with the expression of actin used as an internal control. At least three replicates were included for Western blotting assays.
N metabolic enzyme activity determination Glutamate synthetase (GOGAT), glutamine synthetase (GS), nitrate reductase (NR), urease (UE), and nitrite reductase (NiR) activities in nodules for each treatment and for each age were measured with a respective test kit (Comin, Nanjing, China), following the manufacturer's recommendations. Fresh nodule samples (0.1 g) were thoroughly grounded on ice and extracted with respective enzyme buffers (1 ml). After centrifugation at 8,000 g at 4 ℃ for 10 min, the supernatants were used for enzyme activity assays. The absorbance value was measured by a spectrophotometer. The activity of UE was measured by the absorbance value of NH 3 − -N produced from urea hydrolysis at 578 nm by indigo blue colorimetry. One unit of UE activity was defined as the production of 1 µg NH 3 − -N per gram of nodule per minute. NiR activity was determined by the consumption rate of NO 2 − which is involved in the diazotization reaction and produce purplish red compounds at 240 nm. The activity of NiR was expressed as the reduction of 1 μmol NO 2 − per gram of nodule per hour. The activity of GS was assayed by the product of γ-glutamyl isohydroxamic acid formed from glutamine at 540 nm. One unit of GS activity was defined as the production of 1 μmol γ-glutamyl isohydroxamic acid per gram of nodule per hour. The activities of GOGAT and NR were assayed by monitoring the rate of NADH oxidation at 340 nm. Consumption of 1 nmol NADH per gram of nodule per minute is defined as a unit of enzyme activity. At least three replicates were included for all analyses.
Total RNA isolation, reverse transcription, and gene expression analysis Quantitative real-time RT-PCR (qRT-PCR) was used to investigate the effect of H 2 S on senescence-and N fixation-related gene expression during nodule senescence in soybean. Senescence genes included GmCYP1, GmCYP2,GmVPE,GmSrs8,GmELF,GmCYS1,GmbZIP,GmACS, and GmSAN1B in nodules and GmCAT , GmSOD, and GmCP in roots. N fixation genes were GmLEGH, GmNIR, and GmGS in nodules and roots. N metabolism genes were: GmGOGAT , GmNR, GmRubisco LSU, GmGDH, and GmUPS in nodules. GS isoform genes were GmG-S1alpha, GmGS1beta1, GmGS1beta2, GmGS1gamma, GmGS1gamma2, GmGS2, and GmGS nodule isozymes. The detailed gene information is listed in Supplementary  Table S1. Nodules or roots (0.5 g) were ground with liquid nitrogen containing 2% polyvinylpyrrolidone. The TaKaRa MiniBEST Plant RNA Extraction Kit (TaKaRa, Beijing, China) was used to extract total RNA according to the manufacturer's instructions. One percent agarose gel electrophoresis was used to determine RNA integrity. An Epoch microplate spectrophotometer (BioTek, Winooski, VT, USA) was used to measure the RNA concentration. Total RNA were transcribed into first-strand cDNA using TaKaRa PrimerScript™ RT Master Mix (TaKaRa, Beijing, China). qRT-PCR was performed with a QuantStudio 6 Flex real-time PCR system (Thermo Fisher, Carlsbad, CA, USA) and SYBR Premix Ex Taq II (TaKaRa, Beijing, China). The qRT-PCR programs are described in Supplementary Table S2. The sequences of  the primers used in the present study are listed in Supplementary Table S1. bZip and actin were used as internal control genes to normalize the quantification of transcript levels in soybean. The mRNA transcription abundance value was expressed as 2 −ΔΔCt (Livak and Schmittgen, 2001).
Statistical analysis SPSS 19.0 (SPSS Inc., Chicago, IL, USA) was used to carry out statistical significance, and the results are expressed as the mean values ± standard error. Student's t-test was used at a significance level of P < 0.05. Columns labelled with different letters indicate significant differences at P < 0.05. The significance level of the differences between the control and treatment groups is indicated by asterisks * P < 0.05, * * P < 0.01, and * * * P < 0.001.

Exogenous addition of an H 2 S donor alleviates N-deficiency caused by soybean plants senescence
In both inoculated and non-inoculated groups, at 8w the total dry weight and shoot dry weight were increased by NaHS, but there was no obvious effect on root dry weight ( Fig. 1A-D). Moreover, the rhizobial inoculation plants showed higher total dry weight and shoot dry weight than the non-inoculated plants but did not show a changed root dry weight (Fig. 1A-D). Interestingly, in both the inoculated and non-inoculated groups, the chlorophyll content was slightly increased in response to NaHS during nodule senescence (6-8 w) (Fig. 1E). Taken together, these results demonstrated that NaHS alleviated N-deficiency caused soybean plant senescence, especially in the late stage of symbiotic nodule senescence.
Exogenous addition of H 2 S donor affects nodule numbers, senesced nodules proportion, ARA activity, and ureides content during nodule senescence To determine the effects of H 2 S on nodule senescence in soybean plants, we counted the nodule number Fig. 1 Effect of H 2 S on growth phenotype (A), total dry weight (B), shoot dry weight (C), root dry weight (D), and chlorophyll content (D) of soybean seedlings during nodule senescence. Soybean plants were harvested at 5, 6, 7, and 8 week (w) after treatment. Control: without NaHS and rhizobia; S: 100 μM NaHS; Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. Three independent experiments were conducted. Values are means ± standard error (n = 60). Columns labelled with different letters indicate significant differences at P < 0.05 ◂ of each harvested soybean root at 5, 6, 7, and 8 w.
Our results show that the addition of NaHS clearly increased the nodule number at 6-8 w ( Fig. 2A). Moreover, we estimated the proportion of senescent nodules and found NaHS decreased this proportion at 8 w, but not at other time points was examined (Fig. 2B). We also analysed the acetylene reduction (ARA) activity during nodule senescence. Our results showed that at 5, 6, and 8 w NaHS sharply increased ARA activity (Fig. 2C). Interestingly, from the 5w, the ARA activity sharply declined, suggesting that nodules started the senescence process (Fig. 2C). In addition, during nodule senescence, NaHS significantly increased nodule ureide content at 8 w and root ureide content at 5, 6, and 8 w (Fig. 2D, E). These results suggested that H 2 S alleviated nodule senescence and maintained high N fixation capacity in the soybean-rhizobia symbiotic system. Fig. 2 Effect of H 2 S on nodules number (A), senescent nodules proportion (B), ARA activities (C), nodule ureides content (D), and root ureides content (E) during soybean nodule senescence. Nodules were harvested at 5, 6, 7, and 8 week (w) after treatment.. Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. Three independent experiments were conducted. Values are means ± standard error (n = 60). The significant level of the difference between control and treatment is indicated by an asterisk * P < 0.05, * * P < 0.01, and * * * P < 0.001 Exogenous addition of an H 2 S donor promotes NO 3 − -N, NH 4 + -N, and total N contents in different tissues of soybean plants To confirm the effect of H 2 S on N fixation capacity during nodule senescence, we analysed the NO 3 − -N, NH 4 + -N, and total N contents in different tissues of soybean plants. In leaves, rhizobial inoculation increased the NO 3 − -N, NH 4 + -N, and total N contents compared with those in non-inoculation plants, and NaHS slightly promoted N accumulation in leaves during nodule senescence (Fig. 3A-C). Meanwhile, the NO 3 − -N, NH 4 + -N, and total N contents in leaves of all treatment groups sharply declined with the ageing of symbiotic nodules ( Fig. 3A-C). In addition, from 5 to 7w, the addition of NaHS promoted the NO 3 − -N content in stems of both the inoculated and non-inoculated groups (Fig. 3D), but the NH 4 + -N content was only increased by NaHS in stems of the inoculated plants (Fig. 3E). The total N content in stems increased to varying degrees in the Q8 + S-treated plants compared with the Q8-treated plants, but in non-inoculated plants NaHS did not significantly influence the total N content of stems (Fig. 3F). Generally, the NO 3 − -N, NH 4 + -N, and total N contents in stems of all treatment groups continuously declined with the ageing of the symbiotic nodules ( Fig. 3D-F). In addition, at most checkpoints in both the inoculated and non-inoculated groups, a higher NO 3 − -N content in roots was observed after the addition of NaHS (Fig. 3G), whereas NaHS did not influence the NH 4 + -N content and total N content in roots in either the inoculated and non-inoculated groups ( Fig. 3H-I). Moreover, rhizobial inoculation significantly promoted the NO 3 − -N, NH 4 + -N, and total N contents in roots compared with the non-inoculated plants during nodule senescence, and NaHS also effected these parameters (Fig. 3G-I). Besides, NaHS increased the NO 3 − -N, NH 4 + -N, and total N contents in pods compared with those in plants treated with Q8 alone during the late stage of symbiotic N fixation (5-7w) (Fig. 3J-L). Altogether, the above results showed that during the late stage of symbiotic N fixation NaHS alleviated nodule senescence by increasing the NO 3 − -N, NH 4 + -N, and total N contents in different plant tissues.

Exogenous addition of an H 2 S donor alleviates nodule senescence as determined by ultrastructural analysis
To determine whether exogenous addition of H 2 S could cause structural changes in nodules during the late stage of symbiotic N fixation, we assessed the nodule structure using a light microscope. Our results showed that at 5w the paraffin sections of nodules did not exhibit any obvious difference and the internal infected cells were closely arranged between Q8 and Q8 + NaHS-treated nodules (Fig. 4A, B). However, at 6 w, the cavities in the senescence center were observed in the Q8 nodule (Fig. 4C), and at 7 w further development of the ageing area resulted in the destruction of root nodule tissue (Fig. 4E), whereas the Q8 + NaHS-treated nodule exhibited a senescence centre at 7 w (Fig. 4F). In addition, transmission electron microscopy (TEM) was then used to observe the microscopic structure of infected cells and bacteroids in nodules during the late stage of symbiotic N fixation. At 5w polyhydroxybutyrate (PHB) particles were increased, and the bacteroids were closely arranged. The two groups were not significantly different at 5w (Fig. 5A, B). However, at 6w, the decaying bacteroids appeared, and there were voids in the vesicles after bacteroid degradation in the Q8 nodules (Fig. 5C), but in the Q8 + NaHS nodules senescence characteristics were not obvious (Fig. 5D). At 7 w, the vesicle space became larger in the Q8 nodules (Fig. 5E). Meanwhile, decaying bacteroids appeared in the Q8 + NaHS nodules (Fig. 5F). Notably, at 7 w, the number of spherical bacteroids that shrank from rod-shaped bacteroids due to ageing significantly increased in the Q8 nodules compared to the Q8 + NaHS nodules (Fig. 5G, H). The above results indicated that exogenous addition of an H 2 S donor maintained normal nodule structure and further delayed nodule senescence at the late stage of symbiotic N fixation.
H 2 S promotes nitrogen fixation-related protein expression and N metabolic enzyme activities during nodule senescence To further confirm the regulatory function of H 2 S on nodule senescence, we analysed the Nase iron-containing protein (NifH) and nodulin/glutamate-ammonia ligase-like protein (NodGS) expression levels in soybean nodules using Western blotting (Fig. 6A-D). Our results showed that NaHS promoted NifH protein expression, and at 5 and 6 w the levels of NifH protein in the Q8 + NaHS nodules was increased 2 ~ threefold compared with that in the Q8 nodules. At 7-8 w  E, and F), root (G, H, and I), and pod (J, K, and L) of soybean during nodule senescence. Nodules were harvested at 5, 6, and 7 week (w) after treatment.. Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. Three independent experiments were conducted. Values are means ± standard error (n = 4). Columns labelled with different letters indicate significant differences at P < 0.05 NaHS also enhanced the expression level of NifH protein (Fig. 6A, B). In addition, at 5, 6, and 8 w H 2 S also increased the protein expression levels of NodGS in the Q8 + NaHS nodules (Fig. 6C, D). We determined the activity of key enzymes involved in N metabolism (Fig. 6E-I). NaHS significantly decreased glutamate synthetase (GOGAT) activity at 5-7 w (Fig. 6E). Glutamine synthetase (GS) activity was reduced by NaHS at 6-7 w but increased at 8 w (Fig. 6F). Nitrate reductase (NR) activity under NaHS nodules were inoculated with S. fredii Q8 at 5, 6, and 7 w, respectively; (B, D, and F): nodules were inoculated with S. fredii Q8 and treated with 100 μM NaHS at 5, 6, and 7 wpi, respectively. Number of spherical or rod-shaped bacteroids in A, C, and E (G), and B, D, and F (H). For each age, nine nodule sections form three nodules for each treatment were observed and analyzed. Values are means ± standard error (n = 9). PHB, β-polyhydroxybutyrate granules; JM, capsule; B, rhizobia; SB, senescence bacteria; F, cellulose; the arrow indicates the vesicle. Bar = 2 μm treatment increased at 5 and 8 w and decreased at 6-7 w (Fig. 6G). NaHS increased urease (UE) activity and nitrite reductase (NiR) activity at 5 w (Fig. 6H,   I). Additionally, the expression levels of genes encoding N metabolism enzymes, including GmGOGAT , GmNR, GmRubisco LSU,GmGDH,GmUPS,and Fig. 6 Western blot analysis of nase iron-containing protein (NifH) (A) and ornodulin/glutamate-ammonia ligase-like protein (NodGS) (C) of soybean nodules. Nodules were harvested at 5-8 week (w) after treatment.. Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. Relative expression level is shown as the ratio of NifH:Actin (B) and NodGS:Actin (D) using Quantity One software. H 2 S affected the activity of enzymes associated with N metabolism, which include glutamate synthetase (GOGAT) (E), glutamine synthetase (GS) (F), nitrate reductase (NR) (G), urease (UE) (H), and nitrite reductase (NiR) (I) and the expression levels of genes encoding N metabolism enzymes (J) during nodule senescence. At least three replicates were included for Western blotting assays.. Values are means ± standard error (n = 9). The significant level of the difference between control and treatment is indicated by an asterisk * P < 0.05, * * P < 0.01, and * * * P < 0.001 different GmGS isoforms, were altered during nodule senescence ( Fig. 6J; Fig. S1). The expression levels of GmGOGAT and GmUPS were decreased by the addition of NaHS at 6-8 w. GmNR expression was notably enhanced by NaHS at 5-6 w. GmRubisco LSU expression was increased at 7 w and GmGDH was highly expressed at 8 w under NaHS treatment. Additionally, our results showed that at 5-8 w, the expression level of GmLEGH in the Q8 + NaHS nodules was higher than that in the Q8 nodules (Fig. 7A). In addition, NaHS increased the expression level of GmNIR in nodules at the late stage of nodule senescence (7-8 w) (Fig. 7B). The expression levels of GmGS were enhanced by NaHS in the inoculated plants from 6 to 8 w (Fig. 7C). The above results suggested that exogenous addition of an H 2 S donor alleviated nodule senescence through enhancing the expression levels of N fixation-related proteins in nodules.

Exogenous addition of an H 2 S donor affects senescence-related gene expression during nodule senescence
To further study the mechanism by which H 2 S decays nodule senescence, we determined the expression profile of several senescence-related genes in nodules, including GmCYP1, GmCYP2,GmVPE,GmSrs8,GmELF,GmCYS1,GmbZIP,GmACS,and GmSAN1B. Our results showed that at the late stage of nodule senescence (6-8 w), NaHS significantly inhibited the expression levels of GmCYP1 and GmCYP2, but at 5 w, the expression levels of these two genes were increased by NaHS (Fig. 8A-B). NaHS also slightly inhibited the expression level of GmVPE and GmSrs8 at 6 and 8w during nodule senescence (Fig. 8C-D). Additionally, GmELF was expressed at lower levels in the NaHStreated plants at the late stage of nodule senescence (7 and 8 w) (Fig. 8E). NaHS slightly increased the gene expression level of GmCYS1 at 5 w, but at 6-8 w, it was not changed (Fig. 8F). In contrast, at 5 and 6 w, the gene expression abundance of GmbZIP in the Q8 nodules was much higher than that in the Q8 + NaHS nodules, whereas at 6 and 7 w, this gene expression pattern was opposite to that at 5-6w (Fig. 8G). NaHS inhibited the expression levels of the GmACS and GmSAN1B genes to varying degrees in soybean nodules at 5-8 w (Fig. 8H-I). Additionally, we analysed and GmGS (C) in soybean nodules. Nodules were harvested at 5-8 week (w) after treatment.. Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. The relative mRNA level of each gene was normalized to the mRNA of Gmactin. Data are presented as means ± standard error (n = 9). The significant level of the difference between control and treatment is indicated by an asterisk * P < 0.05, * * P < 0.01, and * * * P < 0.001 the senescence-related genes in soybean roots, including GmCAT , GmSOD, and GmCP. The expression levels of GmCAT in the Q8 + NaHS roots were higher than those in the other three groups (Fig. 9A), but the expression levels of GmSOD in the roots weren't obviously different among the four treatments (Fig. 9B). In addition, the gene expression levels of GmCP in the Q8 + NaHS roots were lower than those in the Q8 roots at the late stage of nodule senescence (Fig. 9C). Overall, these results indicated that exogenous addition of an H 2 S donor alleviated nodule senescence, which was closely related to the senescence gene expression levels in nodules and roots.

Discussion
H 2 S alleviates N-deficiency caused by soybean plant senescence H 2 S plays a vital role in the establishment and maintenance of the symbiosis between soybean and rhizobia (Zou et al. , 2019. However, the role played by H 2 S in nodule senescence is still unclear. Interestingly, Zhang et al. (2020) found that H 2 S and rhizobia effectively alleviated the leaf senescence in the developmental stage under N deficiency conditions. Simultaneously, our results showed that H 2 S Nodules were harvested at 5-8 week (w) after treatment.. Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. The relative mRNA level of each gene was normalized to the mRNA of Gmactin. Data are presented as means ± standard error (n = 9). The significant level of the difference between control and treatment is indicated by an asterisk * P < 0.05, * * P < 0.01, and * * * P < 0.001 delayed nodule senescence, increased Nase activities and ureide content, further increased the lifespan of N fixation and ultimately alleviated plant senescence ( Fig. 1-3). For instance, nitrogenase activity was greatly increased in the presence of H 2 S during the first 5 w, which was consistent with the increase in chlorophyll content during the first 5 w (Fig. 1E and  2C). Moreover, at the late stage of symbiosis, H 2 S significantly increased shoot dry weight, total dry weight, and chlorophyll content in the inoculated plants (Fig. 1), suggesting that H 2 S delays plant senescence by alleviating N deficiency and further promoting biomass accumulation..
When soil is deficient in N, nodules will be formed, which will provide N for plant growth and development by biological nitrogen fixation (BNF) (Puppo et al. 2005). Here, the total N, NO 3 − -N, and NH 4 + -N contents in different tissues were increased by rhizobial inoculation, and exogenous H 2 S addition also promoted N accumulation in legume-rhizobia symbiosis (Fig. 3). Interestingly, H 2 S significantly strengthened N accumulation in pods, suggesting that at the late stage of symbiosis, H 2 S may promote N transfer from roots or leaves to pods. Consistent with our results, H 2 S has been shown to increase the N content in soybean under Fe deficiency or N deficiency conditions by affecting N transport enzyme activities (Chen et al. 2018;Zhang et al. 2020). Therefore, H 2 S enhances N absorption and remobilization by delaying nodule senescence, ultimately alleviating plant senescence and facilitating plant growth.
H 2 S regulates the nodule senescence phenotype of soybean Nodule senescence changes the nodule structure, leads to leghemoglobin degradation, causes a reduction in N fixation, and ultimately inhibits the accumulation of shoot dry weight (Puppo et al. 2005). Nase is an important enzyme for symbiotic nitrogen fixation, and previous studies have reported that nodule senescence inhibits Nase activity (Chungopast et al. 2014;Shelp and Ireland 1985). Our results showed that at the late stage of symbiosis, H 2 S delayed nodule senescence by increasing the Nase activities, ureide content, and nodule number in soybean, and at 8 w the senescent nodule proportion was decreased by H 2 S (Fig. 2), indicating that H 2 S is able to play an important signalling role in nodule senescence. These results and GmCP (C) in soybean root. Soybean roots were harvested at 5-7 week (w) after treatment.. Control: without NaHS and rhizobia; S: 100 μM NaHS; Q8: soybean seedlings inoculated with Sinorhizobium fredii Q8 strain, and Q8 + S: soybean seedlings inoculated with S. fredii Q8 and treated with 100 μM NaHS. The relative mRNA level of each gene was normalized to the mRNA of Gmactin. Data are presented as means ± standard error (n = 9). Columns labelled with different letters indicate significant differences at P < 0.05 were contrary to the function of NO in M. truncatula that provokes nodule senescence (Cam et al. 2012). During the nodule senescence process, visible changes occur when the colour of the N fixation tissue of the nodule changes from red to green (Matamoros et al. 1999). Even in determinate nodules, the senescence phenomenon is not uniformly distributed in the nodule; therefore ultrastructural differences between senescing and mature adjacent infected cells were compared (Matamoros et al. 1999;Puppo et al. 2005). For example, at 6 w after rhizobial inoculation, cavities in the senescence centre were observed in the Q8 nodule (Fig. 4C), and at 7 w, further development of the ageing area resulted in the destruction of root nodule tissue (Fig. 4E), whereas the Q8 + NaHStreated nodule exhibited a senescence centre at 7 w ( Fig. 4F), suggesting that H 2 S delayed senescence centre emergence and inhibited nodule senescence. A previous study demonstrated that less electron-dense and numerous vesicles appear was observed in the cytoplasm of senescent soybean nodule cells (Puppo et al. 2005). For instance, we observed the microscopic structure of infected cells and bacteroids in nodules during nodule senescence by TEM. Our results showed that at 6w, decaying bacteroids appeared, and there were voids in the vesicles after bacteroid degradation in the Q8 nodules (Fig. 5C), but in the Q8 + NaHS nodules, the characteristics of senescence were not very obvious (Fig. 5D). Since the bacteroids represent the functional nitrogen-fixing bacteria, this result represents a higher nitrogen-fixing capacity supporting the higher activity of ARA (Fig. 2D). In other words, at 6 w, H 2 S caused bacteroids to keep higher nitrogen-fixing activities, and the bacteriods began to senescence if H 2 S was lacking. However, at 7 w, the number of spherical bacteroids significantly increased, and the vesicle space became larger in the Q8 nodules (Fig. 5E). Meanwhile, at 7 w, decaying bacteroids appeared in the Q8 + NaHS nodules (Fig. 5F). Hence, the above results provided evidence that H 2 S delayed the appearance of nodule senescence characteristics in soybean.
H 2 S regulates N fixation-related protein and gene expression during the senescence process of symbiotic nodules H 2 S regulates various physiological processes through changing gene or protein expression levels in plants (Filipovic and Jovanović 2017;Liu et al. 2015;Zou et al. 2019). Therefore, we hypothesized that H 2 S modulated the protein and gene expression involved in the N metabolism processes during nodule senescence. First, we analysed the protein expression of the component of Nase, which contains two metalloproteins. One protein contains the active site for substrate reduction with a molybdenum-iron (MoFe) cofactor, and the other protein is the iron (Fe)-containing protein. The Fe protein is a 64-kDa dimer of the nifH gene product. Our results indicated that during nodule senescence, NaHS promoted the abundance of the NifH protein ( Fig. 6A-B). In addition, we determined the protein expression of NodGS, a nodulin/ glutamine synthase-like protein, which is important for N fixation in legumes (Doskočilová et al. 2011). Interestingly, nodulins are highly expressed in mature nodules and are key for nodule activity during nodule senescence (Nap and Bisseling 1990). We found that the protein level of NodGS was obviously increased by NaHS at other stages of symbiosis except for 7 w in nodules (Fig. 6C-D), suggesting that H 2 S may regulate nodule Nase activity by enhancing the symbiosis-related protein expression, and further promoting nitrogen fixation and delaying nodule senescence in legumes.
N metabolism plays an essential role in plant growth and development . Our results showed that during nodule senescence, the N metabolic enzyme activities including GOGAT, GS, NR, UE, NiR were affected by H 2 S, which was consistent with the N content in soybean ( Fig. 6E-I). Leghemoglobin (Lb), as one of the main functional components, provides an optimal environment for the N fixation process in nodules (Zou et al. 2019). Here, NaHS increased the expression of the GmLEGH gene at 5-8w during nodule senescence (Fig. 7A), suggesting that H 2 S may provoke a positive alteration in the redox environment during nodule senescence, which benefits higher N fixation. Combined with the results of NifH western blot analysis and Nase acticity, these results showed that at the late stage of symbiosis, NaHS prolonged the N fixation capacity by increasing the expression levels of the GmLEGH gene and NifH protein in nodules. Moreover, nitrite reductase (NIR) is a primary N assimilation enzyme. During nodule senescence, the expression level of GmNIR was lower than that in younger nodules, and H 2 S increased this gene expression abundance, suggesting that H 2 S could delay nodule senescence by regulating the transcriptional level of GmNIR. Cytosolic glutamine synthetase (GS) is not only a core multifunctional enzyme, but also a key enzyme in N assimilation and remobilization because it plays a vital role in ammonium fixation (Veliz et al. 2017). The expression level of GmGS was strongly induced by H 2 S compared with the control plants at the late stage of symbiosis ( Fig. 7B-C), suggesting that N assimilation and reductase were enhanced by H 2 S and further promoted N utilization at nodule senescence stage in legumes. Therefore, we concluded that H 2 S is needed for the high expression levels of GmLEGH, GmNIR, and GmGS at the late stage of symbiosis, and through which increases to N fixation and assimilation and delays nodule senescence in soybean.
H 2 S affects senescence-related gene expression in the nodules and roots of soybean The early senescence of nodules is characterized by a rapid decline in N 2 fixation, coincident with the onset of pod filling. Similar to the leaf and flower senescence, nodule senescence is characterized by a decrease in soluble proteins (Puppo et al. 2005). Protease activity appears to increase markedly in the cytoplasm of infected nodule cells during senescence (van Wyk et al. 2014). Importantly, thiol-type protease activities play a key role in senescing soybean nodules (Malik et al. 1981), alfalfa nodules (Pladys and Vance 1993), and french bean (Pladys and Rigaud 1991). Moreover, the expression of the cysteine protease (CP) gene has been reported in senescing pea nodules (Kardailsky and Brewin 1996). In the present study, in 7-8-weekold nodules, the expression levels of GmCYP1 and GmCYP2, which encode cysteine protease genes, were higher than that those in 5-6-week-old nodules, but NaHS decreased the senescence-induced these two genes expression in nodules (Fig. 8A-B). Our results suggested that H 2 S delayed nodule senescence and increased N 2 fixation by inhibiting cysteine protease gene expression in nodules. Similarly, previous studies found that CPs were highly expressed during M. truncatula nodule senescence (Fedorova et al. 2002). Moreover, a group of vacuolar CPs, collectively called "vacuolar processing enzymes" (VPE-type), in determinate soybean crown nodules was found during nodule senescence (Roberts et al. 2012). VPE cysteine proteases Fig. 10 Schematic model of the mechanisms underlying the effects of H 2 S on nodule senescence in G. max-S. fredii symbiotic system contribute to senescence and programmed cell death (PCD) (Puppo et al. 2005). Similarly, VPEs might play an important role in the activation of cysteine proteases and are predominantly transcribed in senescent nodules (van Wyk et al. 2014). Here, H 2 S slightly decreased the GmVPE gene expression abundance (Fig. 8C), suggesting that H 2 S inhibited cysteine protease activity, alleviated the degradation of bacteroid and nodule cells, and ultimately delayed nodule senescence and prolonged nodule N fixation capacity. Similarly, transcription regulators are involved in modulating the process of nodule senescence (Deng et al. 2019;van Wyk et al. 2014). Our results found that H 2 S inhibited the expression levels of GmSrs8 and GmELF at the late stage of symbiosis, but increased the GmbZIP expression levels (Fig. 8D, E, G), suggesting that the H 2 S function among the three transcription regulators was different in the process of nodule senescence. These results indicated that H 2 S regulated nodule senescence by affecting the expression of transcription regulators. Cystatin plays an important role and modulates cysteine protease activity during nodule development and senescence (van Wyk et al. 2014). Here, we found that H 2 S significantly increased the expression level of the cystatin gene (GmCYS1) at 5 w, but didn't affect it at the late stage (Fig. 8F). The possible reason was that many cystatins are involved in regulating nodule development and senescence, and GmCYS1 may function in nodule development, but not nodule senescence in soybean plants. In addition, previous studies showed that ethylene contributes to the senescence of the symbiotic nodule, and the expression abundance of 1-aminocyclopropane-1-carboxylate (ACC) synthase was down-regulated in the process of nodule senescence (Serova et al. 2019;Tatiana 2017). In contrast, our results showed that at the late stage of symbiosis, the expression level of GmACS was increased compared with that of the early nodules, and that H 2 S inhibited this gene expression in the process of nodule senescence (Fig. 8H). Webb et al. (2008) reported that they isolated three senescence-associated nodu-lin1 (SAN1) genes, and during nodule senescence, the expression levels of SAN1 were increased. Antioxidant enzyme activities are altered during nodule senescence (Puppo et al. 2005). Here, the expression levels of GmCAT , GmSOD, and GmCP were affected during nodule senescence in soybean roots, and H 2 S also affected these three genes expression levels in soybean (Fig. 9). Taken together, this evidence indicated that H 2 S can regulate senescence-related gene expression levels and further affect nodule senescence and N 2 fixation in soybean nodules.

Conclusion
This study demonstrated the potential role of H 2 S in symbiotic nodule senescence in the G.max-S.fredii symbiosis system (Fig. 10). Our results showed that the alleviation of the symbiotic nodule senescence due to H 2 S prolonging nodule Nase activity and decreasing the proportion of senescent nodules. During nodule senescence, NaHS obviously increased the NH 4 + -N, NO 3 − -N, and total N contents in different soybean tissues. Moreover, H 2 S alleviated nodule senescence by maintaining nodule normal nitrogen fixation structure. Meanwhile, the expression levels of NifH and NodGS proteins were enhanced by H 2 S during nodule senescence in soybean. NaHS also stimulated the N metabolic enzyme activities and the expression of nitrogen metabolism-related genes in soybean nodules. Finally, senescence-related gene expression levels were regulated by H 2 S during nodule senescence. Taken together, these results indicated that H 2 S delayed nodule senescence and prolonged the lifespan of N 2 fixation, which will be important for increasing soybean yield and agricultural sustainability development.