The soybean cultivar “Williams 82” and “Hedou 12” were obtained from the Chinese Academy of Agricultural Sciences (Cheng et al., 2016). The Gmlmm2-1 mutant was isolated from an M₂ population induced by ethyl methane sulfonate (EMS) (Feng et al., 2019). For further analysis, the Gmlmm2-1 mutant was backcrossed to “Williams 82” four times to purify the GmLMM2 mutation. All the plants were grown in Changchun, China. The Columbia (Col-0) ecotype of A. thaliana plants were grown in a growth chamber (Percival, IA, United States) under 150 μmol m^–2s^–1 16 h light/8 h dark cycles at 25°C.

The first compound leaf of the 15-day-old “Williams 82” and the Gmlmm2-1 mutant were used in a half-leaf shading experiment. Small pieces of aluminum foil were used to cover half of the Gmlmm2-1 mutant and “Williams 82” leaves to prevent light exposure; 5 days later the Gmlmm2-1 mutant and “Williams 82” leaves were photographed and staining with different reagents. Leaf trypan blue (TB) staining was performed as the method described by Yin et al. (2000). H₂O₂ accumulation was detected by 3, 3’-diaminobenzidine (DAB) staining (Thordal-Christensen et al., 1997).

To determine pigment content, leaves of 3-week-old Gmlmm2-1 mutant and “Williams 82” were collected and measured as previously described (Lichtenthaler and Wellburn, 1983). Net photosynthetic rate (Pn) and transpiration rate (Tr) of leaves were measured using Li-6400 photosynthesis equipment (Li-Cor, Lincoln, NE, United States) (Yamori et al., 2011). F₀ (initial fluorescence) and Fm (maximal fluorescence) values were measured using a chlorophyll fluorometer OS-30p (Opti-Sciences, Hudson, NY, United States). Fv/Fm (maximum quantum efficiency of photosystem II) and Fv/F₀ (maximum primary yield of photochemistry of PSII) were calculated as previously described (Genty et al., 1989). All measurements were made using three plants from 10:00 am to 12:00 noon during the R1 period with three biological replicates.

Malonyldialdehyde (MDA) levels in leaves were measured as described by Hodges et al. (1999). H₂O₂ content was quantified using an H₂O₂ detection kit according to the manufacturer’s instructions (Comin Biotechnology, Suzhou, China). The absorbance of yellow compound was measured at 415 nm using a NanoPhotometer (Implen).

Coprogen III content was measured as described by Pratibha et al. (2017). The absorbance of Coprogen III was measured at a wavelength of 402 nm. Uroporphyrinogen III (Urogen III) content was determined as described by Bogorad (1962). The absorbance of Urogen III was measured at a wavelength of 405 nm. The Proto IX, Mg-protoporphyrin IX (Mg-proto IX) and protochlorophyllide (Pchlide) contents were measured as described by Rebeiz et al. (1975). All above measurements were obtained from three biological replicates.

True leaves were cut into smaller sections of approximately 1 × 1 mm, and placed in 1.5 mL fixation solution (2.5% glutaraldehyde with phosphate, pH 7.2), and vacuumed until completely immersed in the solution. The samples were subsequently fixed with 1% osmium tetroxide at 4°C. Samples were then treated according to previously described methods (Kowalewska et al., 2016). The samples were sliced to 70 nm thickness with a MT-X (RMC, Tucson, AZ, United States) ultramicrotome and stained with uranyl acetate and lead citrate. The samples were observed using a Hitachi H-7650 electron microscope (Tokyo, Japan).

To determine the subcellular location of GmLMM2, a full-length cDNA fragment of GmLMM2 was amplified using the primer pairs listed in Supplementary Table S1 and inserted into the modified expression vector pUC19-GFP (Zheng et al., 2016). The pFL1004 construct (Supplementary Figure S8) was introduced into Arabidopsis (Col-0) mesophyll protoplasts using 20% polyethylene glycol (Meyer et al., 2006). Fluorescence was monitored using a confocal laser scanning microscope Leica SP8 (Leica, Solms, Germany). We detected the fluorescence of GFP wavelengths at 488 nm (excitation) and 495–540 nm (emission) and chlorophyll auto fluorescence wavelengths at 488 nm (excitation) and 680–700 nm (emission).

The INDEL markers for preliminary mapping were same as described previously (Song et al., 2015). New molecular markers for fine mapping are listed in Supplementary Table S1. The Gmlmm2-1 mutant was re-sequenced with a depth of approximately 30× using illumina Hiseq2000 (Song et al., 2015). The Genome Analysis Toolkit (GATK, version 3.8) was used to detect single nucleotide polymorphisms (SNPs; McKenna et al., 2010). Sequences were deposited at the National Center for Biotechnology Information (NCBI) under the accession number SRP149750.

The amino acid sequences of CPOs/HemF were downloaded from Phytozome¹, Phytozome V12.0. CPOs/HemF sequences were aligned with DNAMAN 8.0. MEGA7² and used to build a phylogenetic tree via the maximum likelyhood method with 1000 bootstrap replications. CPOs/HemF motifs were analyzed using MEME³ and TBtools software (Bailey et al., 2015). Synteny was analyzed using MCScanX (Wang et al., 2012).

For mutation complementation, the 5,374 bp genomic DNA fragment of GmLMM2 (including 2,008 bp promoter region and 3,366 bp gene region) was inserted into the pCAMBIA3301 vector between Sac I and Bam HI sites. The resultant pFL1002 plasmid (Supplementary Figure S6) was then transformed into the cotyledonary explants of the Gmlmm2 mutant via Agrobacterium-mediated transformation (Zhao et al., 2016).

To knockout the GmLMM2 gene in the ‘DongNong50 (DN50)’ soybean cultivar, the modified pSC1-Cas9 was used as described by Du et al. (2016). A 20-nt single guide RNA (sgRNA) was identified using the web-based tools CRISPR-P⁴ and CRISPR-PLANT⁵, which targeted 195 to 214 position in the first exon of GMLMM2. The oligo pairs corresponding the sgRNA were annealed and ligated into a plasmid pSC1-Cas9 digesting with BspQ I. The sgRNA expression was driven under GmU6-16g-1 promoter in the recombinant plasmid pFL1003. The recombinant plasmid pFL1003 (Supplementary Figure S7) was transformed into DN50 as above (Zhao et al., 2016). The primers used for the construct are listed in Supplementary Table S1.

Total RNA was isolated from different soybean tissues using TRIzol (Invitrogen, Carlsbad, CA, United States), and genomic DNA was digested with DNase I (Takara Biotechnology, Dalian, China). cDNA was synthesized using a Fast Quant RT Kit (TIANGEN, Beijing, China). RT-qPCR was performed in a Gene Amp 5700 sequence Detection System with SYBR^® premix Ex Taq^TM regent (Takara). GmActin11 was used as the reference gene (Zeng et al., 2017), the data were analyzed using the 2^–ΔΔCT method (Livak and Schmittgen, 2001). Three biological replicates, each with three technical replicates, were performed. The primers used in RT-qPCR are listed in Supplementary Table S1.

Phytophthora sojae strain P7076 (P. sojae P7076) was cultivated at 25°C on 10% (v/v) V8 juice agar in a polystyrene dish (Förster et al., 1994). P. sojae P7076 infection assay was carried out as previously described with some modifications (Jia et al., 2013; Hwu et al., 2017). Briefly, 20 μL 5‰ Tween 20 was dripped on middle of 10-day-old detached leaves, and then transferred 5-day-old fresh mycelial disks approximately 5.5 mm diameter with a 200 μL Eppendorf pipette tip on the middle of detached leaves. Immediately after inoculation, the leaves of soybean were plated in trays covered with plastic wrap to maintain humidity and infected for 48 and 60 h at 25°C in the dark. Infection dead cell was determined by trypan blue staining as described above. Disease resistance was evaluated by measuring the lesion area using Image J software⁶.

A lesion mimic mutant was isolated from the EMS-induced “Williams 82” mutant population and named Gmlmm2-1 (Glycine max lesion mimic mutant 2-1). In this mutant, small chlorotic spots were initially visible along the veins of fully expanded leaves, before spreading gradually over the entire leaf. Older leaves of mutant with irregular brown spots began to shrink and senesce rapidly. The adult leaf of mutant was yellow compared with that of “Williams 82” (Figures 1A,B). In addition to necrotic spots, the number of nodules, hundred-grain weight, number of grains per plant, and plant height were reduced in the Gmlmm2-1 mutant (Supplementary Figure S1).

To examine cell membrane damage situation of leaves, we performed a leaf trypan blue (TB) staining assay of both wild type and mutant at V2 stage (Figure 1C). After trypan blue staining, numerous dark blue spots appeared at the lesion sites on the Gmlmm2-1 mutant leaves; whereas the surrounding cells in the Gmlmm2-1 mutant and “Williams 82” cells were healthy (Figure 1C). The content of MDA in the Gmlmm2-1 mutant was 44.75 ± 6.42 μmol/g, which was 1.70-fold higher than that of “Williams 82” (Figure 1D). The increased level of MDA suggested that lipid peroxidation occurred in the Gmlmm2-1 mutant. The results of these experiments suggested that PCD occurred during lesion formation.

We subsequently performed DAB staining and observed red-brown polymer deposition in necrotic areas, indicating the generation of excess H₂O₂ (Figure 1E). The H₂O₂ level in the Gmlmm2-1 mutant leaves was higher than that in “Williams 82” leaves (Figure 1F), which was consistent with the DAB staining results. These findings indicated that ROS accumulation in cells was responsible for cell death and the visible necrosis in leaves.

To map the location of GmLMM2, we crossed Gmlmm2-1 mutant with “Hedou 12,” and the F₁ plants exhibited the wild-type phenotype. Genetic analysis of the F₂ population revealed that the necrotic phenotype of the Gmlmm2-1 mutant was controlled by a single recessive nuclear gene (WT: mutant = 618:195). The F₂ population was used to identify the GmLMM2 locus. A total of 107 INDEL markers covering all 20 chromosomes were used for preliminary mapping, and the mapping result showed that GmLMM2 was restricted to the top of Chromosome 14 (Figure 2A). To fine-map the GmLMM2 locus, we developed four INDEL markers, MOL3470, MOL3552, MOL3560, and MOL4565; the GmLMM2 locus was further narrowed down to a 223 kb region between 203,085 and 426,404 bp on chromosome 14 (Chr 14) containing 32 annotated genes (Figures 2A,B). To identify the causative mutation, the Gmlmm2-1 mutant was re-sequenced with a depth of approximately 30 × using illumine Hiseq2000. We identified a A_–667 to G_–667 transition at the second exon of Glyma.14G003200 (Figure 2C), which caused a nonsynonymous substitution of Tyr_–192 to Cys_–192 in the predicted protein. No other mutations were discovered among the 32 genes in the candidate GmLMM2 genomic region (Figure 2B and Supplementary Table S2).

A complementary vector (pFL1002, Supplementary Figure S6) was transformed into the Gmlmm2-1 mutant. Basta-resistance progenies of the transformed the Gmlmm2-1 mutant had a wild type phenotype in the T₂ COM1001-02, COM1001-09, and COM1001-12 populations (Figure 2D, Supplementary Figure S3). Furthermore, we constructed a CRISPR/Cas9 expression vector (pFL1003, Supplementary Figure S7), and the construct was transformed into soybean DN50. Five heterozygous mutants were found in the T₀ plants, and two independent homozygous knock-out lines, named Gmlmm2-2 and Gmlmm2-3, were found in T₁ population (Figure 2D). The Gmlmm2-2 mutant contains a 16 bp deletion corresponding to CDS region of GmLMM2 gene from 200 bp to 215 bp, and a 1 bp insertion in above region. The Gmlmm2-2 mutant caused five amino acids deletion from 67th to 71th and one amino acid substitute (Glu to Arg) at 72th of GmLMM2 (Figure 2E). The Gmlmm2-3 mutant has a 1 bp deletion at 210 bp of CDS and caused a frameshift mutation to a truncated protein with 85 aa (Figure 2E). The Gmlmm2-2 and Gmlmm2-3 mutant variedly phenocopied cell death on leaf compared with the Gmlmm2-1 mutant. The Gmlmm2-2 and Gmlmm2-3 mutant also caused a lesion mimic phenotype in the stem, which was not observed in the Gmlmm2-1 mutant (Supplementary Figure S4). In addition, the Gmlmm2-2 and Gmlmm2-3 mutant exhibited pleiotropic phenotypes, including slow growth, dwarfing, and delayed development during vegetative stages, and the Gmlmm2-3 mutant was a seedling lethal mutant (Figure 2D). These data suggested the Gmlmm2-2 and Gmlmm2-3 mutants are strong alleles, whereas the Gmlmm2-1 mutant is a weak allele, and that complete loss of GmLMM2 function strongly affects soybean growth and development, and regulates cell death.

The open-reading frame (ORF) of GmLMM2 is 1158 bp in length, and the deduced protein contains 385 amino acids (Supplementary Figure S5). Phylogenetic analysis showed that GmLMM2 is more closely related to Phaseolus vulgaris and Vigna radiate homologs, which constitute an isolated branch in the phylogenetic tree. The identities of GmLMM2 and the P. vulgaris homolog are up to 91%. The results of MEME analysis showed that CPOs contain six conserved motifs (Figure 3A).

To find the homologs of GmLMM2 gene, we identified the 41.6 kb syntenic block on chromosome 2 from 48,264,660 bp to 48,306,267 bp, which is homologous to a 48.9 kp region to GmLMM2 located from 300,925 bp to 349,856 bp on chromosome 14 (Figure 3B, Supplementary Table S3). They shared 8 holomgous genes between these two regions. However, the putative paralog of GmLMM2 was absent from the syntenic block on chromosome 2 between Glyma.02G309600 and Glyma.02G309700 (Figure 3B). Thus, GmLMM2 is more like a single copy gene in the soybean genome.

We examined GmLMM2 expression in the root, stem, leaf, flower, nodule, SAM, and pod via quantitative real-time PCR. GmLMM2 was expressed in all organs and tissues tested (Figure 3C). The highest expression levels were detected in leaf and nodule, which may explain why the GmLMM2 mutation severely affected the growth of those tissues.

A chloroplast transit peptide in the first 40 amino acids of GmLMM2 was predicted by TargetP and ChloroP software; and GmLMM2 may be a chloroplast protein. To confirm its subcellular location, a GmLMM2-GFP fusion protein was transiently expressed in Arabidopsis protoplasts. GFP fluorescence only overlapped with the chlorophyll auto fluorescence signal (Figure 3D), which indicated that the GmLMM2 protein is localized to the chloroplast.

To confirm whether GmLMM2 is associated with chloroplast development, we examined the ultrastructure of plastids in 20-day-old seedlings in “Williams 82” and the Gmlmm2-1 mutant. The Gmlmm2-1 mutant contained fewer granal thylakoid membranes compared with the “Williams 82” leaves (Figure 3E). The abnormal chloroplast development in the Gmlmm2-1 mutant may related to the decreased pigment contents. The contents of Chl a, Chl b, and Car in the Gmlmm2-1 mutant were approximately 51, 50, and 40% of those in “Williams 82,” respectively (Figures 3F–H). The decreased chlorophyll contents also affected photosynthesis. We found that the net photosynthetic rate of “Williams 82” was 5.93 ± 0.46 μmol CO₂ m^–2s^–1, which was 2.72-fold higher than that of the Gmlmm2-1 mutant (Figure 3I). The transpiration rate of the Gmlmm2-1 mutant was 3.75 ± 0.12 mmol H₂O m^–2s^–1, which was 1.49-fold higher than that of “Williams 82” (Figure 3J). However, maximum chlorophyll fluorescence (Fv/Fm) and initial fluorescence ratio (Fv/F₀) did not vary between the Gmlmm2-1 mutant and “Williams 82” (Supplementary Figure S2).

To further understand the impact of the GmLMM2 mutation on the tetrapyrrole biosynthesis pathway, the contents of some metabolic intermediates of tetrapyrrole biosynthesis were examined. No significant difference in Uroporphyrinogen III (Urogen III) levels was observed between wild type and mutant (Figure 4A). Whereas, the Coproporphyrinogen III (Coprogen III) level in the Gmlmm2-1 mutant was two-fold higher than that in wild type (Figure 4B). The levels of Protoporphyrinogen-IX (Proto IX), Mg-protoporphyrinogen-IX (Mg-proto IX), and Pchlide in the Gmlmm2-1 mutant were significantly lower than those in the leaves of “Williams 82” (Figures 4C–E). Therefore, light-dependent cell death in the Gmlmm2-1 mutant may be attributed to the accumulation of Coprogen III.

We further analyzed the expressions of some enzymes related above intermediates, including uroporphyrinogen III decarboxylase (GmUROD, Glyma.18G021500); coproporphyrinogen III oxidase (GmCPO, Glyma.14G003200), protoporphyrinogen oxidase (GmPPO, Glyma.10G138600); Mg-chelatase subunit H (GmCHLH, Glyma.10G097800) and NADPH: Pchlide oxidoreductase (GmPOR, Glyma.12G222200). The transcript levels of GmUROD, GmCPO, and GmPPO were slightly elevated, but the transcript levels of GmCHLH and GmPOR remained substantially unchanged (Figures 4F–J). The expression raise of GmUROD and GmCPO might response to their substrates increasing in mutant plant.

To confirm whether the formation of lesions in the Gmlmm2-1 mutant leaves was light-dependent, half of the leaflets before lesion emergence were covered by aluminum foil, while the other half of the leaflets were exposed to light. Five-days later, no lesion had formed on the non-shaded or shaded leaflets of “Williams 82.” However, irregular brown lesions appeared on the light-exposed leaflets of the Gmlmm2-1 mutant, but not on the shaded part leaflets of the Gmlmm2-1 mutant (Figures 5A,B).

NADPH oxidase (NOX) is related to ROS homeostasis, and catalase (CAT) is a ROS-scavenging gene (Lee et al., 2005, 2018). To determine the relationship of ROS and light condition, we compared expression of GmNOX and GmCAT of half leaflet with light covering. The transcript level of GmNOX in the Gmlmm2-1 mutant was 1.35-fold higher than that in the “Williams 82” under light. The expression of GmCAT in the Gmlmm2-1 mutant was 1.65-fold higher than that in the “Williams 82” under light. There was no significant difference in the expression of GmCAT between the Gmlmm2-1 mutant and “Williams 82” in the aluminum foil-covered leaves (Figures 5C,D). The results demonstrated that the GmLMM2 defect disrupts ROS homeostasis in the Gmlmm2-1 mutant, and that these processes are light-dependent.

The effect of light intensity on the lesion induction was also observed under different light condition. We planted “Williams 82” and Gmlmm2-1 mutant under high (20,000 Lux) and low (7,500 Lux) intensity light, respectively. We found that the higher intensity light rapidly and broadly induced severe necrosis and development, and chlorophyll deficiency compared with low intensity light (Figures 5E,F). These results suggested that light is a critical causal factor for lesion induction in the Gmlmm2-1 mutant and light intensity is correlated with the degree of cell death.

The previously study of lesion mimic mutant suggested that the Gmlmm2-1 mutant may enhanced resistance to pathogen infection. To confirm this hypothesis, the fully expanded 10-day-old true leaves of “Williams 82” and the Gmlmm2-1 mutant were inoculated with P. sojae P7076. Visible lesions were observed on the leaves of “Williams 82” and the Gmlmm2-1 mutant at 48 h post-inoculation (48 hpi) (Figure 6A). We measured the infectious area by TB staining, the infectious area in Gmlmm2-1 mutant reduced to 34.5 and 43.7%, compared with “Williams 82” at 48 and 60 hpi separately (Figures 6B–D). These results suggested that mutation of GmLMM2 inhibits the invasive of P. sojae P7076 and the Gmlmm2-1 mutant enhanced resistance to P. sojae P7076.

To test whether spontaneous lesion formation in mutant correlated with the expression of resistance-related genes, we monitored the transcript levels of these genes in “Williams 82” and the Gmlmm2-1 mutant. SA-signaling genes (GmNPR1, GmPR1, GmPR10, and GmSnRK1.1) were more highly expressed in the Gmlmm2-1 mutant compared with those in “Williams 82” (Figure 6E). The transcript level of GmSnRK in the Gmlmm2-1 mutant was 5.5-fold higher than that in “Williams 82.” The expression of GmPR10 in the Gmlmm2-1 mutant was 1373.3-fold higher than that in “Williams 82.” As shown in Figure 6F, these JA-signaling genes (GmAOS, GmLOX, GmCOI1, and GmMYB014) were activated at different levels in the Gmlmm2-1 mutant. The expression of GmCOI1 in the Gmlmm2-1 mutant was 4.17-fold higher than that in “Williams 82.” These results indicated that the expressions of defense-related genes are up-regulated during lesion development.

Coproporphyrinogen III Oxidase (CPO) catalyzes the coproporphyrinogen-III to yield protoporphyrinogen-IX in tetrapyrrole biosynthesis. Defects in CPO have been reported in Nicotiana tabacum, Arabidopsis, and rice (Kruse et al., 1995; Ishikawa et al., 2001; Sun et al., 2011; Wang et al., 2015). This study is the first report to demonstration the function of CPO in soybean. Many lmms mutants exhibit light-dependent cell death, accompanied by the upregulation of ROS-related genes (Sathe et al., 2019). In our study, the GmLMM2 defect in the Gmlmm2-1 mutant also induced a light-dependent lesion mimic phenotype, and higher light intensity led to severe necrosis and chlorophyll deficiency (Figures 5A,B,E,F). Trypan blue and DAB staining assays showed that cell death in the Gmlmm2-1 mutant was caused by abnormal H₂O₂ accumulation (Figure 1). The up-regulation of GmNOX led to ROS accumulation in the illuminated region of the Gmlmm2-1 mutant, and not in the dark region (Figure 5C). As a large amount of H₂O₂ accumulated in the illuminated region of the Gmlmm2-1 mutant, expression of the genes encoding H₂O₂-scavenging enzymes, such as GmCAT, was significantly upregulated (Figure 5D). ROS accumulation can trigger the expression of resistance-related genes, such as PR genes; therefore, many lmms have enhanced resistance to pathogens (Dangl and Jones, 2001; Guo et al., 2013; Wang et al., 2015; Lv et al., 2019).

Resistance-related genes may be activated in lmms and contribute to enhanced pathogen resistance (Guo et al., 2013; Wang et al., 2015). In the present study, the expressions of important genes involved in JA and SA signaling pathways were upregulated in the Gmlmm2-1 mutant and the Gmlmm2-1 mutant displayed enhanced resistance to P. sojae P7076. The SA pathway has been implicated in conferring resistance to biotrophic pathogens, such as Pseudomonas syringae and Hyaloperonospora arabidopsidis, and the JA pathway in resistance to necrotrophic pathogens, such as Alternaria brassicicola and Botrytis cinerea (Penninckx et al., 1998; Pieterse and Van Loon, 1999; Glazebrook, 2005; Spoel et al., 2007). The induction of marker genes involved in the two defense signaling pathways suggested that the Gmlmm2-1 mutant may have acquired a broad spectrum of resistance to multiple pathogens.

In summary, the Gmlmm2-1 mutant exhibits light-induced lesions. The identification of GmLMM2 improves our understanding of the molecular mechanism underlying cell death and the defense response in soybean. In addition, it helps us to further understand the complex regulation of the tetrapyrrole biosynthesis pathway in soybean.