Introduction

Leaf senescence, as the final stage of leaf development, is an important biological process and affected by intrinsic factors, such as organ age and environment conditions (like light, temperature and plant growth regulators)1,2,3,4,5. Lesion mimic mutant, termed as spotted leaf (spl) in rice, is often associated with early leaf senescence. Hence it provides a tool for dissecting molecular mechanisms about leaf senescence. For example, rapid leaf senescence occurs in rls1 (rapid leaf senescence 1) mutant, and RLS1 encodes an uncharacterized NB-ARM (nucleotide binding site-armadillo domain) protein regulating programed cell death (PCD) during leaf senescence6. SPL28, encoding a clathrin-associated adaptor protein complex 1, initiates leaf senescence through regulating vesicular trafficking7.

Besides displaying necrotic spots, lesion mimic mutants spontaneously activate plant defense response. Previous study reported defense response such as the resistance to bacterial blight inoculation was enhanced in 21 lesion mimic mutants isolated from IR64 rice mutant populations8, and similarly the resistance to rice blast fungus inoculation was increased in five lesion mimic mutants: spl5-2, spl12, spl13, spl14, and spl159. Along with elevated resistance, expressions of some pathogenesis-related (PR) genes were induced in rice lesion mimic mutants. For example, PR1 (pathogenesis-related 1) and PBZ1 (probenazole-inducible 1) were expressed abundantly in spl11, spl18 and HM47 mutants10,11,12,13. Hence, studying lesion mimic mutants will help us to elucidate the mechanism associated with leaf senescence and defense responses.

Glutamate synthase (GOGAT) and glutamine synthetase (GS) are known as key enzymes in the process of inorganic nitrogen assimilation14,15,16. Through coordinated action of GOGAT and GS, the inorganic nitrogen is transformed to organic nitrogen17. Ammonia, the main form of inorganic nitrogen, is assimilated though GS/GOGAT cycle to produce glutamine (Gln) and glutamate (Glu), both of which constitute the main form of organic nitrogen. In higher plants, there are two kinds of GOGAT18, viz. NADH-GOGAT (EC 1.4.7.1) and Fd-GOGAT (EC 1.4.1.14), using NADH and ferredoxin (Fd) as the electron donors respectively. NADH-GOGAT was found mainly in non-photosynthetic tissues such as roots and seeds19, while Fd-GOGAT plays a major role in reassimilating the large amounts of ammonia derived from photorespiration20. There are two expressed Fd-GOGAT genes in Arabidopsis, viz. AtGLU1 and AtGLU2. Analysis of gls mutants revealed that the major role of AtGLU1 is to reassimilate the large amounts of ammonia derived from photorespiration in leaves, while AtGLU2 functions in the assimilation of primary nitrogen in roots20. In another study, authors reported Arabidopsis Fd-GOGAT encoded by GLU1 is dual targeted to the chloroplasts and mitochondria, and plays an important role in photorespiration, largely consistent with the previous study21. In rice, Fd-GOGAT was revealed to locate mainly in mesophyll cells of leaves, chloroplast-containing cross-cells of grain pericarp, as well as in the apical meristem through immunocytological analysis22. Nevertheless, the function of Fd-GOGAT in leaf senescence and defense response was largely unknown in rice.

Here, we report a new spotted leaf mutant (spl32) derived from an indica cultivar 9311 by radiation. The spl32 mutant is covered with necrotic spots in leaves from tip to base in seedling stage, and spotted leaves emerge from bottom to top of the plant. Positional cloning indicated that SPL32 encodes an Fd-GOGAT protein. We further identified the mutant was subjected to oxidative stress due to suppressed reassimilation system of ammonia, thereby inducing increased resistance to Xanthomonas oryzae pv. Oryzae (Xoo). Our results reveal that Fd-GOGAT involves leaf senescence and defense response in rice.

Results

Phenotypic analyses of the spl32 mutant

Leaves of the spl32 mutant remained largely the same as the wild type before four-leaf stage (Fig. 1A). The necrotic spots were initiated from tip of older leaves in five-leaf stage (Fig. 1B) and then spread through whole leaf (Fig. 1C). From tillering to heading, the necrotic spots became more serious, accompanied by shorter culm and smaller grain, compared to the wild-type 9311 (Fig. 1D,E). Agronomic traits, such as tiller number, seed setting rate, and thousand kernel weight, were all significantly decreased in the spl32 mutant (Table 1).

Figure 1: Phenotypic comparison of the wild-type 9311 and spl32 plants.
figure 1

(A,B,C) Phenotypes of 15 (A), 23 (B), 31 (C) days old wild-type 9311 and spl32 seedlings. The white arrows indicate the fourth real leaves of the wild-type 9311 and spl32 plants. The inserts show enlargement of the fourth real leaves of the wild-type 9311 and spl32 mutant (AC). (D) Phenotypic comparison of the wild-type 9311 and spl32 mutant at heading stage. The insert shows flag leaves of the wild-type 9311 and spl32 plants. (E) Comparison of the wild-type 9311 and spl32 seeds (dehulled). Bars = 5 cm in A, B and C, 15 cm in D and 5 mm in (E). (F) Determination of the chlorophyll contents of the fourth leaves of the wild-type 9311 and spl32 mutant at three different developmental stages: 15, 23, 31d after sowing. Chla, Chlorophyll a; Chlb, chlorophyll b; Car, total carotenoids. Data represent means ± SD of three independent measurements. Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05. FW, fresh weight.

Table 1 Comparison of agronomic traits between wild-type 9311 and spl32 mutant.

Environmental conditions could induce formation of lesion mimic symptom23, thus a shading experiment was conducted here. There was no difference between shaded leaf and non-shaded leaf in the wild type. Remarkably, non-shaded leaf in the spl32 mutant showed obvious lesion mimic phenotype, in contrast to shaded leaf showing no symptom (Supplemental Fig. S1). Thus the formation of lesion is induced by light in the spl32 mutant.

According to the severity of lesion spots in the spl32 mutant, we divided the phenotype into three stages, i.e. before appearance of spots (the first stage), spots emerge (the second stage), and serious spots present (the third stage). Chlorophyll contents were examined in the three stages. The result showed that the contents of chlorophyll a and chlorophyll b were both decreased during the second and third stages in the spl32 mutant relative to the wild-type 9311(Fig. 1F).

Chloroplast changes in leaves of the spl32 mutant

Transmission electron microscopy was used to compare ultrastructure of chloroplasts in the 4th leaves of wild-type and spl32 seedlings at the all three stages. The mutant leaf had same chloroplast morphology as the wild type at the first stage (Fig. 2A,B), that is, chloroplasts were well-developed with rich lamellae and a small number of osmiophilic bodies. Thus chloroplast development was supposed to be unaffected in the spl32 mutant. At the second stage, the lamellar structure of chloroplasts began to collapse and osmiophilic bodies were obviously increased in the mutant (Fig. 2C,D). Then the shrinking chloroplasts started to disintegrate and gradually little complete thylakoid structure was observed in the spl32 mutant. At the third stage, number and size of osmiophilic bodies in the mutant chloroplasts were significantly increased, relative to the wild type (Fig. 2E,F). This result suggested that chloroplasts in the spl32 mutant are normal during initial development but disintegrated early in later stages.

Figure 2: Transmission electron microscopy analyses.
figure 2

The wild-type 9311(A,C,E) and spl32 (B,D,F) mutant at 15d (A and B), 23d (C and D), 31d (E and F) after sowing. Bars = 5 μm. Thy: thylakoid lamellae; OB: osmiophilic body.

Gene expression of leaf senescence in the spl32 mutant

Leaf senescence is controlled by a number of genes. Among them, a senescence-induced gene, STAY GREEN (SGR), was reported to regulate chlorophyll degradation24. Analysis by quantitative RT-PCR (qRT-PCR) revealed that SGR transcript was dramatically upregulated in the spl32 mutant at tillering stage (Fig. 3A). Additionally, many transcription factors and senescence-associated genes (SAGs) were reported upregulated to induce leaf senescence25,26. To confirm senescence occurred in the spl32 mutant, we performed gene expression of OsWRKY23, OsWRKY72, Osl43 and Osl85 using qRT-PCR (Fig. 3A). The result showed that expression levels of all four genes were notably increased in the mutant at tillering stage (Fig. 3A), in agreement with the symptom of early leaf senescence.

Figure 3: Gene expression analyses.
figure 3

The expression analysis of two senescence-associated transcription factors (OsWRKY23 and OsWRKY72), two senescence-associated genes (Osl43 and Osl85) SGR (A) and photosynthesis-associated genes (B) at tillering stage. Data represent means ± SD of three independent measurements. Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05. The accession numbers were listed as follows: OsWRKY23 BAG98560.1; OsWRKY23 BAG98549.1; Osl43 AAK82986; Osl85 AAL65398; SGR AAW82954.1; rbcS AAR19268.1; lhcA AAB65793.1; lhcB (Arabidopsis thaliana), AAA32760; rbcL NP_039391.1; psaA AJC99402.1; psbA AJC99383.1; petD AJC99432.1; ndhA AJC99458.1; atpA AJC99399.1.

A previous study showed photosynthesis-associated gene expressions were decreased when leaf senescence appeared5. Thus nine photosynthesis-associated genes were selected for analysis by qRT-PCR (Fig. 3B). Among them, three (rbcS, lhcA, and lhcB) belong to nuclear-encoded genes and the others (rbcL, psaA, psbA, petD, ndhA, and atpA) are chloroplast-encoded ones. The result revealed that all photosynthesis-related genes were down-regulated in spl32 mutant leaves (Fig. 3B).

Map-based Cloning of the SPL32 gene

Genetic analysis from reciprocal crosses between 9311 and the spl32 mutant showed that the lesion mimic phenotype was controlled by a single recessive nuclear locus (Supplemental Table S1). Molecular analysis of an F2 population and its derived F2:3 population from the cross spl32 × 02428 (a japonica rice variety) restricted the SPL32 locus between the markers RM118 and RM172 on chromosome 7 (Fig. 4A). Twelve insertion/deletion (InDel) markers were developed between the two markers (Supplemental Table S2). The SPL32 locus was further narrowed down to a 66-kb region (Fig. 4B), containing eight putative open reading frames (ORFs) (Fig. 4C). Sequencing all ORFs indicated a single base mutation (G into A) in the junction between the third exon and intron in Fd-GOGAT (Os07g0658400) (Fig. 4D) led to alternative splicing in mRNA. Thus a 57-bp sequence was missing in the complementary DNA (cDNA) in the mutant (Fig. 4F). A newly developed InDel marker zf-1 was used to distinguish the wild type from spl32 mutant genotypes by the size of amplified cDNA fragment (Supplemental Table S3). The cDNA of spl32 mutant has two forms: one is supposed to encode a truncated protein (ΔFd-GOGAT) (a strong band) lacking 19 amino acids; and the other one encodes a whole Fd-GOGAT protein (indicated by a weak band), just like wild type (Fig. 4F). Therefore the spl32 mutant is an Fd-GOGAT knockdown mutant.

Figure 4: Cloning of SPL32 gene and phylogenetic analysis of Fd-GOGAT homologs.
figure 4

(A) The SPL32 locus was mapped to the long arm of chromosome 7 between the markers RM118 and RM172. (B) Mapping of the SPL32 locus between markers zzy-21 and zzy-23 based on bacteriophage P1-derived artificial chromosome (PAC) or bacterial artificial chromosome (BAC) clone sequence (PAC1, P0047B07; BAC2, OJ1477_F01; PAC3, P0496C02). ‘n’ represents homozygous recessive individuals derived from an F2 population and an F2:3 population from the cross spl32 × 02428. The number of recombinants is indicated below the map. (C) the SPL32 locus was narrowed down to a 66-kb region; there are 8 putative open reading frames (ORFs). (D) The gene structure of 9311. The junction of the third exon and intron had a single base mutation (G into A) in ORF4 of spl32 mutant. (E) Complementation analysis of the spl32 mutant. The wild-type 9311 plant and homozygous recessive mutant derived from the cross spl32 × Nipponbare transformed with pFd-GOGAT show normal green leaves, whereas the spl32 mutant and homozygous recessive mutant of F2 population transformed with pΔFd-GOGAT show necrotic spots in leaves. (F) PCR analysis with primers Zf-1 showed the difference in the Fd-GOGAT cDNA in the wild-type 9311, spl32 mutant, ΔFd-GOGAT transgenic plants and Fd-GOGAT transgenic plants. Alternative splicing occurs in the coding region of Fd-GOGAT in spl32 mutant. (G) Phylogenetic analysis of Fd-GOGAT. Fd-GOGAT is most closely homologous to maize and sorghum Fd-GOGATs. The accession numbers were listed as follows: SiFd-GOGAT (Solanum lycopersicum), XP_004234830.1; StFd-GOGAT (Solanum tuberosum), XP_006363768.1; RcFd-GOGAT (Ricinus communis), XP_002526914.1; GmFd-GOGAT (Giycine max), XP_006576787.1; CsFd-GOGAT (Cucumis sativus), XP_004136778.1; OsFd-GOGAT (Oryza sativa), NP_001060520.1; ZmFd-GOGAT (Zea mays), NP_001105693.1; SbFd-GOGAT (Sorghumbicolor), XP_002463318.1; AtGLU1 (Arabidopsis thaliana), NP_850763.1; AtGLU2 (Arabidopsis thaliana), NP_181655.1.

To verify identity of Fd-GOGAT, we constructed the plasmid pFd-GOGAT, containing a 5-kb fragment with only Fd-GOGAT coding region of 9311. We also constructed the plasmid pΔFd-GOGAT, containing the 57 bp-deleted fd-gogat coding region of the mutant. Then, we introduced the two constructs into homozygous recessive individuals selected from the cross spl32 × Nipponbare. All transgenic lines containing pFd-GOGAT rescued the lesion mimic phenotype, whereas 10 independent lines transformed with vector pΔFd-GOGAT didn’t (Fig. 4E,F). The result indicated that the 57-bp missing in Fd-GOGAT cDNA was responsible for the necrotic spots in the spl32 mutant. Phylogenetic analysis showed that rice Fd-GOGAT was most closely related to maize and sorghum Fd-GOGATs (Fig. 4G).

Expression analyses and subcellular localization of Fd-GOGAT protein

To investigate expression pattern of Fd-GOGAT in rice, we analyzed Fd-GOGAT expression in different tissues by qRT-PCR. We found that Fd-GOGAT was constitutively expressed in leaf, root, panicle and other tissues analyzed. The Fd-GOGAT was most pronounced in young leaves at seedling stage while relatively weak in roots (Fig. 5A), consistent with a previous study that Fd-GOGAT mainly expressed in photosynthetic tissues16. To detect expression levels of Fd-GOGAT at the three different developmental stages in the seedling period, we chose the fourth leaves of the spl32 mutant and 9311 at 15, 23, and 31 days after sowing (DAS) for qRT-PCR analysis. As the development of lesion spots, Fd-GOGAT expression showed a decreasing trend. At 31 DAS when the fourth leaves of spl32 mutant were filled with necrotic spots, Fd-GOGAT almost couldn’t be detected in the mutant (Figs 1C and 5B), indicating that the more severe necrotic spots, the lower expression of Fd-GOGAT in the spl32 mutant.

Figure 5: Expression analysis and subcellular localization of Fd-GOGAT protein.
figure 5

(A) The expression of Fd-GOGAT in different tissues. Leaf was selected from seedling of 30 days after sowing of the wild-type 9311; root, flag leaf, panicle, and leaf sheath were selected from the wild-type 9311 at heading stage. (B) The expression of Fd-GOGAT at three different developmental stages. The fourth leaves of 15, 23, 31 days after sowing in wild-type 9311 and spl32 mutant were selected. Data are means ± SD (n = 5). Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05. (C) GFP signals of the Fd-GOGAT-GFP fusion protein located in the chloroplasts of rice protoplasts. (D) GFP signals were dispersed in rice protoplasts. Bars = 10 μm.

To further investigate subcellular localization of Fd-GOGAT protein in rice, we use ChloroP27 and TargetP28 to search subcellular localization signals. We found that N-terminus of Fd-GOGAT protein comprised a putative chloroplast-targeting signal. Thus, we constructed a fusion protein with Fd-GOGAT and green fluorescent protein (GFP), and transformed it into rice protoplasts. Confocal microscopy was used to observe the fluorescent signals in 16–18 h after transformation. We observed that green fluorescent signal from Fd-GOGAT-GFP was colocalized with the autofluorescent signals of chlorophylls in the chloroplasts (Fig. 5C,D), suggesting Fd-GOGAT protein is located in the chloroplasts.

The reassimilating of ammonia is inhibited in the spl32 mutant

In photosynthetic green leaves and shoots of Arabidopsis, Fd-GOGAT accounts for more than 96% of total enzyme activity and the remaining is derived from NADH-GOGAT29. Here we measured GOGAT activity in 9311 and the spl32 mutant (Fig. 6A). The result revealed that GOGAT activities in mutant leaves were decreased by 13.2% in non-spotted stage, 12.7% in spot-appearing stage and 13.2% in serious-spotted stage, respectively. We also evaluated effect of the decreased GOGAT activity on GS expression by ELISA Kit (Fig. 6B). It showed that GS activity in the mutant had no difference from 9311 in the all three stages.

Figure 6: Analysis of GOGAT and GS activities, determination of photorespiration rate, contents of ammonia (NH4+) and glutathione (GSH).
figure 6

(A and B) Enzyme activity of GOGAT and GS in wild-type 9311 and spl32 mutant, non-spotted, spot-appearing and serious-spotted leaves of spl32 mutant and the corresponding sections of leaves in the wild-type 9311 at tillering stage. (C) Determination of NH4+ content in the wild-type 9311 and spl32 mutant at tillering stage. (D) Determination of net photosynthesis rate in atmospheric condition and hypoxia conditions using portable photosynthesis system LI-6400XTOPEN6.1. At tillering stage, new-fully expanded leaves of wild-type 9311 and spl32 mutant were chosen and net photosynthesis rate of mid-blade was measured before 12 o’clock. (E) Determination of GSH content in the wild-type 9311 and spl32 mutant at tillering stage. Data represent means ± SD of three independent measurements. Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05. FW, fresh weight.

Besides, GOGAT and GS were known as key enzymes involved in inorganic nitrogen assimilation14,15,16. To investigate amino acid changes in the spl32 mutant, we measured free amino acid contents in 9311 and the spl32 mutant at tillering stage (Table 2). Many amino acids such as Glu, Ser, Asp, Gly, Ala, Val, Met, Ile, Leu, Tyr, Phe, Lys, and His were significantly decreased, while Gln, Cys and Arg were dramatically increased in the spl32 mutant. Since Glu and Gln are produced from ammonia though GS/GOGAT cycle and both amino acids are significantly altered in the spl32 mutant, so we determined ammonia content at the tillering stage and found ammonium was noticeably increased in the mutant (Fig. 6C).

Table 2 Free amino acid content in leaves of wild-type 9311 and spl32 mutant.

Previous study indicated Fd-GOGAT plays a major role in photorespiration by reassimilating ammonia20. To evaluate effect of the increased ammonia content on photorespiration, we measured photorespiration rates by supplying hypixic gas (<2%) to wild-type and mutant leaves, as described previously30. The result showed the spl32 mutant had much lower net photosynthetic rate than its wild type under hypoxic conditions, but no difference under atmospheric conditions (Fig. 6D).

Generally photorespiration can affect synthesis of glutathione (GSH), which is produced from condensation reaction of Gly and γ-glutamylcysteine31. Consistent with decreased Gly (Table 2), GSH content was reduced in the spl32 mutant (Fig. 6E).

Antioxidant enzyme activities and active oxygen burst in the spl32 mutant

GSH is an important active substance in plants and participates in formation of disulphide, sulfur ether and thioester. Meanwhile, it can remove free radicals to scavenge their poison in plants32. Hence we determined reactive oxygen species (ROS) content in plant cells by trypan blue dyeing. Leaves with no spots and serious spots in spl32 mutant and the corresponding parts of leaves in the wild type were stained at tillering stage (Fig. 7A). The result showed that no dyeing was observed on non-spotted leaves from either the mutant or wild type. In contrast, leaves with serious spots in spl32 mutant were dyed blue. This result indicated that severe cell necrosis occurred in spl32 mutant when spots emerged.

Figure 7: Dying experiments and the determination of enzyme activities involved in scavenging and generating mechanism of ROS.
figure 7

(A) non-spotted leaves and serious spotted leaves of spl32 mutant and the corresponding sections of leaves in the wild-type 9311 at tillering stage were stained by trypan blue and 3, 3-diaminobenzidine (DAB) respectively. Controls were not stained and decolorized directly by alcohol. (B) The contents of hydrogen peroxide (H2O2) and malonaldehyde (MDA) in the wild-type 9311 and spl32 mutant at tillering stage. FW, fresh weight. (C) The enzyme activities of catalase (CAT), superoxide dismutase (SOD), ascorbate peroxidase (POD), and peroxidase (APX) in wild-type 9311 and spl32 mutant at tillering stage. FW, fresh weight. (D) The enzyme activities of NADPH oxidase and polyamine oxidase (PAO) in wild-type 9311 and spl32 mutant at tillering stage. FW, fresh weight. Data represent means ± SD of three independent measurements. Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05. FW, fresh weight.

To determine accumulation level of hydrogen peroxide in the spl32 mutant, similar leaves in the above treatment were stained with 3,3-diaminobenzidine (DAB). No staining was detected in the non-spotted leaves in both the spl32 mutant and 9311; nevertheless, leaves with serious spots in the spl32 mutant were dyed bronzing (Fig. 7A). Moreover, we determined hydrogen peroxide contents at tillering stage, and found that spl32 mutant leaves accumulated much more hydrogen peroxide than the wild type (Fig. 7B).

Malonaldehyde (MDA) is one of the most important products of membrane lipid peroxidation, and its accumulation can aggravate membrane damage, indirectly reflecting the degree of cellular damage33. Hence, we measured MDA content at tillering stage (Fig. 7B). As expected, the spl32 mutant showed significantly increased MDA in leaves, compared to the wild type. Thus the spl32 mutant tissue could be susceptible to oxidative damage and started senescence more quickly than the wild type.

A previous study showed that enzymes involved in antioxidant systems were expressed in lesion mimic mutants34. Here, we detected activities of four oxidative stress related enzymes, including catalase (CAT), superoxide dismutase (SOD), ascorbate peroxidase (APX), and peroxidase (POD), respectively. As shown in Fig. 7C, activities of CAT and SOD were significantly increased while POD and APX activities were remarkably decreased in the mutant. Meanwhile, two other enzymes that catalyze ROS generation, viz. NADPH oxidase and polyamine oxidase (PAO), were also determined (Fig. 7D). It revealed that activity of PAO was increased while NADPH oxidase showed no change in the spl32 plants.

The spl32 mutant has enhanced disease resistance and up-regulated PR marker genes

The appearance of necrotic spots in the spl32 mutant resembles hypersensitive response (HR), which constitutes an important resistance mechanism in plants35. To evaluate response of the spl32 mutant to pathogen, bacterial blight pathogen Xoo strain Zhe173 was used for inoculation. As shown in Fig. 8A, the spl32 plants demonstrated significantly enhanced resistance compared to its wild type.

Figure 8: Detection of bacterial blight pathogen resistance and expression of resistance related genes.
figure 8

(A) Lesion lengths were determined after plant leaves inoculated by bacterial blight pathogen zhe173. Bars represent ± SD of six replicates. (B) The expression of pathogenesis-related (PR) marker genes at tillering stage. Bars represent ± SD of three measurements. The gene numbers of PR genes are as follows: PR1a, AJ278436; PR1b, B109D03; PR5, X68197; PR10, D38170. Asterisks indicate the statistical significance levels according to Student’s t test: **P < 0.01 and *P < 0.05.

Previous studies indicated that oxygen stress not only activated protection mechanism but also acted as a signal factor in undamaged tissues to enable systemic acquired resistance (SAR)36,37. So we detected expressions of four pathogenesis-related (PR) marker genes (PR1a, PR1b, PR5 and PR10) associated with defense response. The results revealed that all PR genes were consistently up-regulated in the spl32 mutant (Fig. 8B).

Discussion

Early leaf senescence occurred in spl32 mutant, confirmed by decreased pigment content and increased MDA content and SOD activity as physiological indicators, chloroplast degradation as cellular indicators, and both upregulation of senescence transcription factors (OsWRKY23 and OsWRKY72) and senescence-associated genes (Osl43 and Osl85) and downregulation of photosynthesis-related genes as molecular evidence. ROS plays an important role in early leaf senescence38. The main sources of ROS are from NADPH oxidase and PAO36. In the spl32 mutant, activity of NADPH oxidase showed no change but PAO activity was increased compared to the wild type. NADPH oxidase converts molecular oxygen into superoxide anion, and PAO interconverts polyamines, producing hydrogen peroxide. Overexpression of AtPAO3 resulted in an increased production of both hydrogen peroxide and superoxide anion39. Here, a large amount of hydrogen peroxide were accumulated, which thereafter induced CAT activity; while superoxide anion did not accumulate in the spl32 mutant by nitrotetrazolium blue chloride test (Supplemental Fig. S2), which might be that the increase of superoxide anion was promptly eliminated by SOD. Moreover, photorespiration provides Gly for GSH synthesis31 and plays a major role in the readjustment of redox homeostasis40. We found photorespiration was dramatically inhibited by over-accumulation of ammonium in the spl32 mutant. Meanwhile, contents of Gly and GSH were both decreased, accompanied by extremely decreased activities of POD and APX in the spl32 mutant. Therefore, the ROS accumulated in the mutant can probably be attributed to overproduction of hydrogen peroxide and damaged scavenging pathway.

OsWRKY23 played an important role in senescence and resistance, and overexpression of OsWRKY23 enhanced expression of PR related genes and increased resistance to the bacterial pathogen Pseudomanas syringae in Arabidopsis41. In the spl32 mutant, disease resistance to Xoo was significantly enhanced, consistent with a recently published study showing Fd-GOGAT plays an important role in broad spectrum bacterial blight resistance42. Moreover, we found four PR marker genes were up-regulated in the spl32 mutant. There are two major signaling pathways in plant disease resistance, viz. salicylic acid (SA) and jasmonic acid (JA)43. Since expressions of PR1a and PR1b were both enhanced in the mutant, suggesting the SPL32 might involve in SA and JA signal transduction.

Through coordinated action of GS and GOGAT, inorganic nitrogen is transformed to organic nitrogen20. GS catalyzes ammonium into Glu, thereby producing Gln. Subsequently, GOGAT transfers amide amino group of Gln to 2-oxoglutrate, forming two molecules of Glu again22. In this study, as contents of ammonium and Gln were remarkably increased, while Glu was decreased in the spl32 mutant, we concluded that Glu could be effectively catalyzed into Gln by functionally normal GS in the mutant. In contrast, decrease of GOGAT activity due to the spl32 mutation caused the over-accumulation of Gln. In addition to changed contents of Glu and Gln, we also found many other free amino acids, except Cys and Arg, were decreased in spl32 mutant, thus it was intriguing to investigate whether and how Fd-GOGAT involves in a complex cross-pathway regulatory mechanism in amino acid metabolism.

It is noteworthy that photorespiration plays an important role in protecting photosynthetic organs from damage through excessive absorption of light energy44. In the spl32 mutant, the lesion was induced by light and photorespiration rate was reduced, therefore light might be a trigger to accelerate oxidative damage in leaves, compared to the wild type.

In Arabidopsis, Fd-GOGAT contributes to major GOGAT enzyme activity, and the remaining is derived from NADH-GOGAT20. There are two NADH-GOGAT genes in rice19. OsNADH-GOGAT1 is mainly expressed in growing tissues, such as immature leaves, the early development of spikelet and roots, while OsNADH-GOGAT2 mRNA is only expressed in leaves and leaf sheathes. These isoforms have distinct functions: OsNADH-GOGAT1 is essential in assimilation of ammonia in roots45; OsNADH-GOGAT2 plays an important role in remobilization of nitrogen during senescence of rice leaves46. In this study, we found transcript levels of the Fd-GOGAT gene were dramatically reduced at 23 DAS and nearly abolished at 31 DAS in the spl32 mutant plants (Fig. 5C), whereas the GOGAT enzyme activity was only reduced constantly by 13% (Fig. 6A) during the period. The discrepancy between transcript levels and enzyme activities is possible due to following reasons: (1) potential post-transcriptional and post-translational regulation of Fd-GOGAT, such as alternative splicing, RNA stability, protein stability and modification, etc.; (2) compensation of another rice Fd-GOGAT, viz. osNADH-GOGAT. In the spl32 mutant, we indeed observed when expression level of Fd-GOGAT fell down to the lowest level at the most serious period of spots, NADH-GOGAT2 had a dramatic increase at the same stage (Supplemental Fig. S3). Therefore we deduce NADH-GOGAT2 might partially but not completely, compensate the function of Fd-GOGAT.

In conclusion, we report herein that SPL32 encodes an Fd-GOGAT that is highly expressed in young leaves. In the spl32 mutant, reassimilation of Gln and ammonia is inhibited due to Fd-GOGAT knockdown. Photorespiration rate and Gly content are decreased as well as glutathione content and some antioxidant enzyme activities, resulting in uncontrolled oxidative stress in the mutant. Moreover, defense response genes are activated and disease resistance is enhanced in the spl32 mutant. This provides a prospect in developing new rice resistant lines by manipulating SPL32 in gene engineering.

Materials and Methods

Plant materials and growth conditions

The rice lesion mimic mutant spl32 was selected from an M2 population of 9311 induced by 60Co radiation. M2 seeds were grown in a paddy field under natural growing conditions in Rice Station, Nanjing Agricultural University, Nanjing, China. The mutant was stabilized through backcrossing with 9311 and selection for three times.

SPL32 gene cloning

For genetic analysis, F2 populations were generated from two crosses of 9311/spl32 and spl32/9311. In both F2 populations, segregation of green-leaf and spotted-leaf individuals fitted a 3:1 ratio. To map the SPL32 gene, we constructed an F2 mapping population generated from a cross spl32 × 02428 (a japonica rice variety). Totally 330 SSR markers evenly spanning 12 rice chromosomes were used to rough map SPL32. The SPL32 was initially mapped to the markers between RM118 and RM172 on chromosome seven using 588 F2 mutant plants. Afterwards, the SPL32 locus was further narrowed down between markers zzy-21 and zzy-23 using 801 F2:3 mutant plants. The mapping primers were listed in Supplemental Table S2. The target region was analyzed with RiceGAAS (http://ricegaas.dna.affrc.go.jp) and eight ORFs were predicted. Then coding sequences of eight ORFs in both 9311 and spl32 mutant were amplified for sequencing.

Complementation of the spl32 mutant

To create complementation construct pFd-GOGAT, cDNA of Fd-GOGAT was amplified by RT-PCR from 9311 with the primer pairs listed in Supplemental Table S3. The resulting fragments were inserted into the vector 1300–221-FLAG in which the Fd-GOGAT gene was under the control of 35S promoter. Since transformation events with 9311 as receptor was difficult, then the construct was introduced into some F2 mutant plants selected from a cross between Nipponbare and spl32 mutant, by Agrobacteriun tumefaciens-mediated method as described previously47.

Determine of pigment contents

Pigment contents were measured according to the method described previously48. Fresh leaves of spl32 mutant plants and 9311 at three different developmental stages (15, 23, and 31 DAS) were collected. Absorption values were measured using UV/Vis spectrophotometer (Beckman DU800 USA).

Transmission electron microscopy

Samples of 9311 and spl32 mutant leaves were prepared for electron microscopy from the fourth leaves of seedling after growing 15, 23, 31 days under standard field conditions. The collected leaves were cut into small pieces and fixed in 2.5% glutaraldehyde in a phosphate buffer at 4 °C for 4 h, rinsed and incubated overnight in 1% OsO4 at 4 °C, dehydrated by an ethanol series, and infiltrated with a series of epoxy resin, and then embedded in Spurr’s medium prior to thin sectioning. Sections were stained again using uranyl acetate and measured with a JEM-1230 electron microscope.

RT-PCR and Real-Time RT-PCR

Total RNA of rice leaves was extracted with a RNA Prep Pure Plant kit (Tiangen) following the manufacturer’s instructions. Each RNA sample (2 μg) was reverse transcribed using the QuantiTect reverse transcription kit (Qiagen). Primer pairs were designed using Primer Express (Applied Biosystems) and listed in Supplemental Table S3. Additionally, the primer information of some genes including STAY GREEN (SGR), two senescence-associated transcription factors, senescence-associated genes and photosynthesis-associated genes were referred from a previous study49. Real-time PCR analysis was performed using ABI7300HT fast real-time PCR system with the SYBR Premix Ex Taq (TaKaRa; catalog no. RR041A). For each sample, we performed real-time PCR with three technical replicates on three biological replicates. ACTIN gene in rice was used as an internal control.

Amino acid sequence alignment and phylogenetic relationship

Searching and downloading the homologous proteins of Fd-GOGAT in different species were carried out in the NCBI website, and then homologous proteins of Fd-GOGAT were analyzed using biological software BioEdit7.0. Multiple sequence alignment was conducted using ClustalW program50, and phylogenetic analysis method described previously51.

Subcellular location of GFP protein

To investigate subcellular location of Fd-GOGAT, coding sequences of Fd-GOGAT were amplified by PCR using primer pairs listed in Supplemental Table S3. PCR products were cloned into pAN580 vector. The vector was transformed into rice protoplasts and incubated in the dark at 24 °C for 14–16 h according to the method described previously52. Finally, GFP fluorescence was observed using a confocal laser scanning microscopy (Zeiss LSM780).

Shading experiment and photorespiration determination

At heading stage, flag leaves (leaves were all green and no spots) in the spl32 mutant and 9311 were wrapped in tinfoil, then the plants were grown in the paddy filed for 10 days under natural light. Five flag leaves (each from mutant and wild type) were randomly selected for treatment. To investigate photorespiration in the spl32 mutant, we measured net photosynthesis rate in atmospheric and hypoxia condition using portable photosynthesis system LI-6400XTOPEN6.1. At tillering stage, we chose new-fully expanded leaves of the spl32 mutant and its wild type, and measured net photosynthesis rate of mid-blade before 12 o’clock. In both mutant and wild type, 10 plants were randomly selected. The gas access of instrument for measuring photosynthesis was input with 98% nitrogen and 2% oxygen to produce hypoxia conditions. Net photosynthesis rates were examined according to the manufacturer’s recommendation.

Determination of various antioxidant indexes

Commercial kits for assaying contents of GSH, hydrogen peroxide, and MDA, as well as enzyme activities for CAT, SOD, and POD were all purchased from Biological Engineering Institute of Nanjing JianchengTM. The enzyme activities of APX, PAO and NADPH oxidase were determined by corresponding Enzyme-Linked Immunosorbnent Assay (ELISA) Kit purchased from Biological Engineering Institute of Jiangsu lyuyeTM. Determination experiments were conducted at tillering stage. All determination protocols referred to manufacture’s instruction. In presence of peroxidases, DAB and hydrogen peroxide could react quickly and form brown polymer deposition. Integrity of cell membrane was examined by trypan blue. Dyeing experiments with trypan blue and DAB were performed according to Wang et al.53 using leaves with non-spotted and serious spotted leaves of spl32 mutant respectively, and their corresponding sections of normal leaves from 9311 at tillering stage. Finally, the leaves were photographed after the decolorization.

Assays of GOGAT and GS activities

GOGAT were determined by GOGAT ELISA Kit purchased from the company of Shanghai YanshengTM and GS activities by GS ELISA Kit from Biological Engineering Institute of Jiangsu lyuyeTM. We selected leaves from different sections of 9311 and spl32 plants (carrying no spots, newly-appearing spots, and serious spots) at tillering stage. Each sample was repeated three times according to the manufacturer’s instructions.

Inoculation with bacterial blight pathogen

Xoo strain Zhe173 (race IV) was provided by Jiangsu Academy of Agricultural Science. New-fully expanded leaves of six independent spl32 plants and 9311 at the tillering stage were inoculated with Zhe173 suspensions (optical density of 0.5 at 600 nm) using clipping leaf method54. Lesion lengths on inoculated plants were measured 21 days after inoculation.

Free amino acids and ammonium content analyses using automatic amino acid analyzer

Free amino acids and ammonium extraction method is as follows: At tillering stage, we chose new-fully expanded leaves of the spl32 mutant and 9311. Samples corresponding to ~300 mg fresh weight (FW) were collected and ground to powders in liquid nitrogen. We also set a blank control without samples. Aliquots of 100 mg FW and blank control were added into 10 mL of 0.1 mol/L HCl, then sealed the lid and properly soaked with shaking. The mixture was extracted by an ultrasonic equipment at temperature 40–50 °C. After 1–1.5 h, the samples were cooled and centrifuged (10 min at 8000 rpm). The proper amount of supernatant was added into equal volume of 10% sulfosalicylic acid. After cooling for an hour at 4 °C, the samples were centrifuged for 30 min at 15,000 rpm, then pH of the supernatant was adjusted with 8 mol/L NaOH to 2.0. Finally, the samples were filtered by 0.45 μm membrane filters and then determined by High Speed Amino Acid Analyzer (Hitachi L-8800). For ammonium content determination, absorbance was measured at 570 nm when retention time was 72.68 min.

Additional Information

How to cite this article: Sun, L. et al. Isolation and characterization of a spotted leaf 32 mutant with early leaf senescence and enhanced defense response in rice. Sci. Rep. 7, 41846; doi: 10.1038/srep41846 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.