The rice lmpa mutant was obtained from an ethyl methane sulfonate (EMS)-treated population of japonica rice (Oryza sativa) variety, Zhonghua11 (ZH11). The lmpa mutant was back-crossed to ZH11 to produce M₂ plants in an isogenic background. In this study, ZH11 was used as the wild type (WT) in all the analyses. All rice plants were grown in the field under natural growth conditions in Hangzhou, Zhejiang Province, China. For the temperature treatment experiments, WT and lmpa seeds were grown in a chamber (12 h dark/12 h light cycle at 60% humidity) at 20 and 30°C from germination to 2 weeks. For the salt stress tolerance treatments, the germinated seeds of WT and lmpa were sown in a 96-well plate. Then the 7-day-old seedlings were grown in Yoshida's culture solution and treated with different concentrations of NaCl for 7 days. Then, the survival rate of the seedlings was determined after 7 days of recovery under normal conditions.

Plant materials were collected and fixed in 70% FAA (50% ethanol, 5% glacial acetic acid, and 5% formaldehyde) overnight. After fixing, the samples were dehydrated in a graded alcohol series, immersed in xylene, embedded in paraffin, and finally sliced into 8-μm thick sections using a rotary microtome (Leica, HistoCore AUTOCUT). After being dewaxed with xylene and rehydrated with decreasing ethanol concentrations, the prepared slices were stained with 1% safranin and 1% fast Green and observed under a microscope (Nikon, ECLIPSE 90i).

For TEM, the flag leaves of WT and lmpa mutant plants with the lesion phenotype were harvested at tillering stage and fixed in fixation buffer (2.5% glutaraldehyde in 100 mM phosphate buffer, pH 7.4). After fixing, the samples were washed three times with PBS and fixed with 1% OsO₄ for 1 h, then dehydrated with a graded alcohol series, and embedded with Suprr Kit (Sigma-Aldrich). The specimens were sliced into 70-nm thick sections with ultratome (Leica, EM UC7), stained with uranyl acetate and alkaline lead citrate, and finally observed under a Hitachi-7500 TEM (Tokyo, Japan).

The flag leaves of WT and lmpa mutant plants with the lesion phenotype were used for physiological measurements. The content of malonaldehyde (MDA) and hydrogen peroxide (H₂O₂) and the activities of superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), and ascorbate peroxidase (APX) were measured using an assay kit (Suzhou Keming Biotechnology Co, Ltd.) according to the manufacturer's instructions.

Nitroblue tetrazolium (NBT) and 3, 3′-diaminobenzidine staining (DAB) were performed to detect the accumulation of superoxide anion (O2-) and H₂O₂, respectively. In brief, fresh leaves were collected and incubated in 0.05% (w/v) NBT or 0.1% (w/v) DAB (pH 5.8) staining buffer overnight at 28°C, decolored with 75% ethanol until all the chlorophyll had been removed, and finally photographed.

For the GUS staining assay, different tissues of transgenic plants were collected and stained in GUS staining buffer for 12–16 h at 37°C. The samples were then cleared in 75% ethanol and photographed. The GUS activity in the root was analyzed by using the GUS gene quantitative detection kit (Coolaber, China, SL7161).

The flag leaves and apical spikelets of the WT and lmpa plants were collected and fixed in the FAA solution overnight. Briefly, the samples were embedded and sliced according to the method described in the section “Paraffin Sectioning.” After rehydrating with ethanol and being treated with proteinase K, the TUNEL assay was performed with a TUNEL Kit (Promega, G3250) according to the manufacturer's instructions. The nuclei were stained with DAPI, and the apoptotic cells were stained with fluorescein-12-dUTP. The blue fluorescence signal of DAPI and green fluorescence signal of fluorescein (TUNEL signal) were observed at 460 and 520 nm, respectively, under a confocal laser scanning microscope (Zeiss, LSM700).

For genetic analysis, the lmpa mutant was crossed with the indica variety Nanjing 6 (NJ6) to generate the F₁ plants, and the plants displaying the panicle abortion phenotype were selected as the mapping population from the F₂ mapping population. For fine gene mapping, new InDel molecular markers were developed by Primer Premier 5 software. LMPA was mapped to a 122.6-kb region on chromosome 4, and the genes predicted to be present within this region were amplified and sequenced, and compared between lmpa and WT plants. The primers used are listed in Supplementary Table 1.

Total RNA was isolated using AxyPrep™ total RNA Miniprep Kit (Axygen, AP-MN-MS-RNA) from different tissues. The first strand of complementary cDNA was synthesized with ReverTra Ace kit (Toyobo, FSK-101) from 1 μg of total RNA. The cDNA was diluted for qRT-PCR using the SYBR Green PCR Master Mix kit (Applied Biosystems, 4367659) in ABI7900 (Applied Biosystems). The OsActin (Os04g0177600) gene of rice was used as an internal reference to normalize the gene expression data. Three replicates were performed for all experiments. The cycle threshold (Ct) method was used to calculate the relative amounts of mRNA. The student's t-test was used to analyze the significance of the differences. The primers used are listed in Supplementary Table 1.

To create lmpa CRISPR lines, the gene-editing constructs were generated using CRISPR/Cas9 technology. The sgRNA targeting the 12th exon of LMPA was cloned into the pYLCRISPR-Cas9Pubi-H vector according to the method previously described (Xie et al., 2017). To compare the promoter activity, the 2,300 and 2,733-bp promoter fragments were amplified from WT and lmpa, respectively, and cloned into the pCAMBIA1305.1 vector to generate the plasmid pLMPA::GUS and plmpa::GUS. To generate the overexpression constructs, the coding sequence of LMPA was amplified and cloned into the pUbi-GFP-NosT vector to create the plasmid pubi::LMPA-GFP. All the vectors were introduced into the rice callus via Agrobacterium (EHA105)-mediated transformation, as previously described (Toki et al., 2006). The sequences of the PCR primers used for vector construction are listed in Supplementary Table 1.

To determine the subcellular localization of LMPA, the pUbi::LMPA-GFP vector was transformed into N. benthamiana leaves by initiating an Agrobacterium infection. The roots of 3-day-old seedlings of pUbi::LMPA-GFP transgenic plants were used to observe the subcellular localization of LMPA. GFP fluorescence was detected with LSM 700 confocal microscope (Zeiss). Plasma membranes were stained with NerveRedTM C2 (Coolaber, China, FM4-64).

To determine whether the promoter mutation affects the LMPA expression, the promoter fragments of WT and lmpa were amplified and cloned into pGreenII 0800-LUC vector. The LMPA-LUC and lmpa-LUC plasmids were transformed into rice protoplasts, respectively, according to a previously described method (Ruan et al., 2020). Firefly luciferase (fLUC) and Renilla luciferase (rLUC) were measured using the Dual-Luciferase Reporter Assay System (Promega, E1910). The relative luciferase activity was calculated as the ratio of fLUC/rLUC. The primers used are listed in Supplementary Table 1.

The lmpa mutant was identified from the EMS mutagenesis library of ZH11. At the seedling stage, the mutant first exhibited white patches on the leaves. Subsequently, lesions appeared and extended to the whole leaf blade with the growth of the plant (Figures 1A,B; Supplementary Figures 1A–D). Moreover, lmpa showed severe panicle degeneration at the apical portion of each panicle, and the development of spikelets in the apical part of the panicle stopped and became shriveled (Figures 1A,C–F). The panicle undergoes development in different stages, and the results showed that the whitish spikelets appeared after the panicle reaches a length of 5 cm (Supplementary Figures 2A–F). In addition, the iodine-potassium iodide (I₂-KI) staining assay displayed that no viable pollen grains were found in the apical spikelets, but the pollen grains of middle and basal spikelets were found to be normal (Supplementary Figures 3A–E). Due to the apical panicle abortion, the panicle length, the number of grains per panicle, and plant height were significantly reduced in lmpa mutants (Figures 1F–H). However, the 1000-grain weight of WT and lmpa has no significant difference (Figure 1I).

To reveal the abnormalities associated with lesion leaf and apical spikelets, a detailed cell structure was investigated by observing the paraffin sections. The results showed that the non-silicified cells were absent in lmpa glume (Figures 2A–F). Moreover, the morphological structures between the region of the lesion and normal regions also showed significant differences, including distorted bulliform cells, vascular bundle sheath cells, and mesophyll cells in lesion parts (Figures 2G–I). TEM analysis showed that the matrix thylakoids and stroma lamellae structures of chloroplast were disorderly arranged, and more osmiophilic granules were found in lmpa. In addition, the content of Chlorophyll a (chl a), Chlorophyll b (chl b), and carotenoids (car) were remarkably decreased in lmpa compared to that of WT (Supplementary Figures 4A–E).

In view of the formation of mimic leaf lesion and degeneration of apical spikelets in lmpa, we conducted TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling) assay to detect cell death. The results showed that TUNEL-positive signals were barely found in WT leaves, while strong signals were observed in lmpa (Figures 3A–F). In addition, the strong positive TUNEL signals were also detected in hulls and anthers of apical spikelets in lmpa, whereas almost no obvious TUNEL signals were observed in the WT tissues (Figures 3G–R). The excessive accumulation of ROS easily causes oxidative damage, resulting in cell death (Van Breusegem and Dat, 2006). NBT and DAB staining were performed to detect O2- and H₂O₂ accumulation, respectively. The results showed that the in lmpa, leaves with lesions were colored deep brown-red by DAB staining and exhibited denser deep-blue spots by NBT staining (Figures 4A,B). In addition, we carried out the measurements of related indicators and found that the levels of H₂O₂, MDA, POD, SOD, and APX were significantly higher in lmpa than the levels in WT, and the CAT activity was significantly decreased in the lmpa mutant (Figures 4C–H). We further examined the expression of PCD representative genes and ROS scavenging-related genes, such as VPE2, VPE3, CATA, CATB, and CATC, and the results revealed that the expression levels of these genes were significantly increased in lmpa (Figures 4I–M).

To locate the mutant gene, an F₂ population was constructed by crossing lmpa with indica cultivar NJ6. With the map-based strategy, the gene was initially mapped on chromosome 4 between the markers M1 and M7 and finally narrowed to a 122.6-kb region between the markers M4 and M5. Ten open reading frames (ORFs) are annotated within the region in the Rice Genome Annotation Project database (http://rice.plantbiology.msu.edu/cgi-bin/gbrowse/rice/) (Figure 5A; Supplementary Table 2). Sequence analysis showed that a 433-bp fragment was inserted into the promoter of LOC_Os04g56160 in lmpa (Figures 5B,C). Compared to the WT, the expression levels of the LOC_Os04g56160 gene were significantly decreased in lmpa (Figure 5E). To verify the gene function, we carried out gene editing by CRISPR/Cas9 technique, and a total of five homozygous knockout lines were obtained. As expected, the expression levels were significantly downregulated in the knockout plants and displayed the lesion mimic phenotype and apical panicle abortion (Figures 5D–H). Based on these findings, we conclude that LOC_Os04g56160 is LMPA gene. According to the Rice Genome Annotation Project, LMPA encodes a plasma membrane H⁺-ATPase (PM H⁺-ATPase) and is an allele of OSA7.

To confirm that the fragments inserted in the promoter region affect the expression of LMPA, we constructed the LMPA::GUS and lmpa::GUS reporter genes and introduced them into the WT, respectively. GUS staining indicated that LMPA was highly expressed in young roots and coleoptile of the LMPA::GUS transgenic lines, while only faint staining was observed in lmpa::GUS transgenic plants (Figures 6A,B). We further examined the GUS activity and found that it was significantly higher in LMPA::GUS than that of lmpa::GUS (Figure 6C). Moreover, the LUC reporter was employed to compare the promoter activity by transient expression in rice protoplasts. The results showed that the WT promoter exhibited higher transcriptional activity than lmpa promoter (Figures 6D,E).

To investigate the spatial expression pattern of LMPA, the qRT-PCR was employed to measure the expression in various organs, including root, stem, leaf, leaf sheath, and panicle. The results indicated that LMPA was expressed in all the investigated tissues, particularly higher in leaves and panicles (Figure 7A). Moreover, the GUS reporter gene driven by the native promoter was expressed in the WT background. Strong GUS staining was mainly detected in roots, stems, leaves, and panicles, which was consistent with the results obtained by qRT-PCR analysis (Figure 7B).

To determine the subcellular localization of LMPA protein, we constructed an LMPA-GFP fusion protein driven by the Ubiquitin1 promoter and expressed protein in rice protoplasts. The GFP fluorescence was observed at plasma membranes and colocalized with the FM4-64 marker (Figures 7C–F). In addition, the pubi::LMPA-GFP plasmid was introduced into N. benthamiana leaves. The result indicated that the GFP fluorescence was also localized in the plasma membranes (Figures 7G–J).

Extreme environmental conditions, such as high temperature, high humidity, strong light, and low nitrogen levels, often induce lesion mimic symptoms and lead to the panicle apical abortion during the rice growth (Yao et al., 2000; Zhang et al., 2017; Cui et al., 2021; Xia et al., 2022). We performed the temperature treatments and found that the lmpa mutant was sensitive to high temperatures. Under the 30°C conditions, the reddish-brown lesions were scattered on the surface of leaves and spread gradually with growth in the lmpa mutants. However, there were no obvious lesions on the leaf surface in lmpa when grown under low temperatures (20°C) (Figures 8A,B). Meanwhile, the NBT and DAB staining techniques were applied to detect the excessive accumulation of ROS. Under normal conditions, the lmpa leaves were stained with deep brown-red and deep blue spots by DAB and NBT staining, respectively, whereas the staining was minimal at low temperatures (Figures 8C–F). In addition, we detected the expression levels of LMPA when the plants were grown at different temperatures. The results showed that the expression levels of LMPA were significantly reduced in lmpa mutant compared to WT when grown at both 20 and 30°C, but the expression level in lmpa mutants grown at 20°C was higher than that observed at 30°C (Supplementary Figure 5). These results indicated that the lesion mimic phenotype of lmpa can be rescued by low temperature.

Plasma membrane H⁺-ATPases are the primary active transporters that pump protons out of the cell and mediate Na⁺ extrusion by providing proton motive force (Vitart et al., 2001). Therefore, PMAs play an important role in salt tolerance in a plant. In this study, we performed salt treatments and found that the disruption of LPMA impairs the salt stress tolerance in rice. When 7-day-old WT and lmpa seedlings were treated with different concentrations of NaCl (75 mM and 150 mM) for 7 days and recovered after 7 days, only a few WT leaves showed a wilted phenotype, while the majority of lmpa leaves exhibited a markedly wilted phenotype, especially under high concentrations of NaCl (Supplementary Figures 6A–F). In addition, the survival rate of WT was about 78.9% and 41.3% but that of the mutants was only 47.1% and 14.4%, respectively, after 7 days of recovery (Supplementary Figure 5G). These findings suggest that LPMA is important for salt stress tolerance in rice.