Zoysia japonica cultivar “Zenith” seeds, purchased from the Patten Seed Company (Lakeland, GA, United States), were cultivated in the TS1 nutrient medium (Klasmann-Deilmann, Germany). The seedlings were kept in an RXZ-380D-LED growth chamber (Ningbo Jiang Nan Instrument Factory, Ningbo, China), which is set at 28/25 ^oC (day/night), 65% humidity, and approximately 400 μmol m^–2s^–1photosynthetic active radiation (PAR). Arabidopsis thaliana “Columbia” (wild-type background) and pph mutant (SALK_000095) were kept in a growth chamber set at 24/22 ^oC (day/night) with 65% humidity and 400 μmol m^–2s^–1 PAR. The photoperiod of the growth chambers was set to 16 h per 24 h. Nicotiana benthamiana was cultivated under identical growth conditions as Z. japonica. The plants were irrigated by 1/2 strength Hoagland’s solution weekly (Hoagland and Arnon, 1950).

Total RNA was extracted from “Zenith” leaves using Trizol and then it was used to generate cDNA using PrimeScript RT reagent kit (TaKaRa, Dalian, China). Genomic DNA was extracted using the CTAB methods from fresh leaves of Zenith. BLAST analysis was carried out using homologous proteins as queries in the Zoysia grass genome sequence to design primers used for cloning ZjPPH (Table 1). PCR amplification was performed to amplify the ZjPPH gene and its promoter sequence, respectively. The PCR products were then purified and inserted into the pMD-19T cloning vector (TaKaRa, Dalian, China) with ampicillin resistance. After antibiotic selection and PCR verification, the plasmids containing the complete coding sequence (CDS) (pMD-ZjPPH) and the promoter sequence (pMD-ZjPPHpro) from the base 0 were generated, respectively. The correct plasmids examined by Sanger sequencing and alignment were selected and used in the subsequent experiments.

Total RNA was isolated from different organs of Z. japonica, including roots, stems, and leaves, separately. Total RNA samples were also extracted from leaves at different senescent stages as described previously (Teng et al., 2016). Phytohormone treatments, including 10 μmol ABA, 10 μmol MeJA, and 0.5 mM SA, and dark induction were applied to investigate the expression characteristics following a protocol described by Teng et al. (2017). qRT-PCR data calculation was performed using the ΔΔCt method (Livak and Schmittgen, 2001). The Z. japonica β-actin gene (GenBank accession No. GU290546) was used as the housekeeping gene (Teng et al., 2017).

The NCBI BLAST analysis was carried out to search for the homologs. Phylogenetic tree was built with the neighbor-joining method using MEGA v.5 software (Tamura et al., 2011). The Ks/Ka ratio was generated using DnaSP6 software (Rozas et al., 2017). Based on the PlantCARE online database, the cis-elements in the promoter sequence were identified (Lescot et al., 2002). The molecular weight (MW) and the theoretical isoelectric point (PI) were calculated using the compute pI/Mw tool¹. The subcellular localization pattern was predicted using the online program TargetP 1².

The ZjPPH CDS was inserted into the plant binary vector 3302Y3 (Jia et al., 2016) to generate the 35S:ZjPPH:YFP construct used for subcellular localization investigation. The ZjPPH CDS was infused into the pTA7002 vector used for generating transgenic Arabidopsis lines. The promoter sequence of ZjPPH was inserted into the pCambia1391Z vector to generate the ZjPPHpro:GUS construct used for the GUS staining determination.

For the subcellular localization observation, Agrobacterium tumefaciens EHA105 harboring 35S:ZjPPH:YFP was transformed into the lower epidermal cells of N. benthamiana leaves using an injection syringe according to the transient expression protocol (Sparkes et al., 2006). After 48-h incubation, the fluorescence in the tobacco cells were checked and photographed using a SP8 laser confocal scanning microscope (Leica, Mannheim, Germany).

The A. tumefaciens GV3101 containing the pTA7002-ZjPPH construct, and pTA7002 empty vector were used to generate transgenic and control A. thaliana lines using the floral dip method, respectively (Clough and Bent, 1998). The seeds harvested were screened on MS medium supplemented with 30 mg L^–1 hygromycin. After PCR verification, T₃ lines with 100% hygromycin resistance were selected for further morphological assessment. Three-week-old healthy lines growing in MS medium were transplanted to filter paper immersed in 30 μM dexamethasone (DEX) (Sigma-Aldrich, Munich, Germany) and 0.01% Tween-20 solution diluted with ddH₂O. After 4 days of induction, the plants were photographed in a photostudio using an EOS 80D digital camera (Canon, Tokyo, Japan).

The chlorophyll content was extracted using 95% ethanol, and total chlorophyll contents were then quantified based on the absorbance recorded (Teng et al., 2016). ABA content was measured using an ELISA kit (H251, Jiancheng Bioengineering Institute, Nanjing, China) based on the protocol provided by the manufacturer. Soluble sugar content was determined using the 3,5-dinitrosalicylic acid method (Sun et al., 2020). Leaf fresh weight was adopted in assessing these indicators above.

A Handy PEA analyzer (Hansatech, Kings Lynn, United Kingdom) was used for the determination of chlorophyll fluorescence parameters. For dark incubation, fresh leaves were covered by a leaf clip at the middle of the leaf blades for 30 min. Then, the clips were removed and chlorophyll fluorescence was immediately measured by exposing the leaves to a two-second saturating light pulse of 3,500 μmol photons m^–2 s^–1. Photosynthetic parameters were recorded and calculated according to the protocol provided by the manufacture. Ten replicates were utilized for each of the CK, line-3, and line-7. The parameters were processed using Handy PEA v.1.3 software. Origin Pro v.2019b (OriginLab Corporation, Northampton, MA, United States) was used for data presentation.

Transmission electron microscopy observation was performed based on the protocol of Park et al. (2007) with moderate modifications. After DEX incubation, leaf tissues were soaked in the fixative buffer with 2% paraformaldehyde, 2% glutaraldehyde, and 50 mM sodium cacodylate (pH 7.2). Next, the samples were washed with 50 mM sodium cacodylate buffer (pH 7.2) three times at 4^°C. Then, the samples were treated in the 50 mM sodium cacodylate buffer containing 1% osmium tetroxide (pH 7.2) at 4^°C for 2 h. After being washed three times with distilled water at 25^°C, the samples were incubated in 0.5% uranyl acetate for at least 30 min at 4^°C. A gradient series of propylene oxide and ethanol was used for sample dehydration. Finally, the samples were embedded in resin for generating ultrathin sections using an ultra-microtome. The samples were then placed on copper grids after being polymerized at 70^°C for 24 h. Finally, they were treated with 2% uranyl acetate for 5 min and with Reynolds’ lead citrate for 2 min at 25^°C. The ultrastructure of chloroplasts was then observed and photographed using a HT7700 transmission electron microscope (Hitachi, Tokyo, Japan). We examined 10 cells in each sample and photographed the representative chloroplast structure in well-cut sections.

Data were analyzed through Student’s t-test or one-way ANOVA using SPSS version 15.0 (IBM, Chicago, IL, United States). *p < 0.05 were recognized statistically significant. All data are presented as the mean ± SD (n = 3 at least).

The ZjPPH (GenBank accession No.:MW882937) was successfully cloned by RT-PCR based on the genome sequence database of Z. japonica. The coding sequence of ZjPPH is 1479 bp in length, encoding 492 amino acids. Its theoretical pI and MW are 8.19 and 54.67 kD, respectively. A phylogenetic tree was constructed and indicated that ZjPPH was most closely related to the PPH orthologs in Zea mays and Sorghum bicolor (Figure 1A). Conserved domain analysis revealed that ZjPPH contained a conserved domain of α/β hydrolases and a PPH motif (PVYIVGNSLGG) containing a catalytic Ser residue (Figure 1B). The synonymous and non-synonymous substitution rates (Ka/Ks) were calculated to classify the evolutionary selection types. It showed that the Ka/Ks values of ZjPPH were less than one, suggesting that ZjPPH is under purifying selection (Figure 1C).

A 1500-bp nucleotide promoter sequence was obtained by PCR amplification. Cis-elements prediction results demonstrated that the ZjPPH promoter contained two ABRE elements involved in ABA responsiveness, one CGTCA motif and one TGACG motif involved in MeJA-responsiveness, two TCA elements participant in salicylic acid responsiveness, and one ERE elements correlated with ethylene-responsive element (Figure 2A).

Using transgenic Arabidopsis seedlings, GUS histochemical staining was performed to further characterize the promoter. It displayed that the promoter sequence could drive GUS expression in transgenic plants. Specifically, blue pattern was abundantly found in the leaves and root (Figure 2B). The GUS expression level showed that the 48 h of dark and ABA treatment could induce the expression of GUS gene, whereas the MeJA and SA treatment caused no significant changes, indicating the role of ABA and dark treatment in regulating the promoter activity (Figure 2C).

The expression of ZjPPH exhibited a tissue-specific character with a higher level in the leaf than in the root and stem (Figure 3A). The expression level of ZjPPH was found more abundantly in senescent leaves than in mature and young leaves (Figure 3B). Hormone inducement analysis revealed that the transcriptional activity of ZjPPH could be activated by ABA, MeJA, SA, and dark treatment within 24 h of treatment (Figures 3C–F). Specifically, ABA and dark treatments upregulated the expression level of ZjPPH from 6 h and maintained a higher level at 24 h. MeJA treatment induced the expression of ZjPPH from 3 h and continuously kept the highest level from 6 h to the end of the experiment. SA increased the expression of ZjPPH from 1 h and reached the highest level at 6 h, and then it decreased gradually from 6 to 24 h.

The TargetP 1.1 online program demonstrated that the ZjPPH was a chloroplast-specific localized protein. To further verify the subcellular localization pattern of ZjPPH, transient expression analysis was employed. As shown in Figure 4, the tobacco lower epidermis cells transformed with 35S:YFP were full of YFP signal. In contrast, the YFP signal was only detected in the chloroplast of the tobacco cells transformed with 35S:ZjPPH:YFP construct. The results proved that ZjPPH was a chloroplast-specific localized protein.

Transgenic Arabidopsis lines overexpressing ZjPPH were generated to explore the function of ZjPPH. The ZjPPH-overexpressing Arabidopsis lines showed an obvious yellowing phenotype compared with control (Figure 5A). The qRT-PCR analysis proved that the ZjPPH was efficiently expressed in the transgenic lines (Figure 5B). Correspondingly, chlorophyll content determination results showed that the chlorophyll contents was higher in the control than in the transgenics (Figure 5C). Moreover, ZjPPH also rescued the stay-green phenotype of the pph mutant (Figure 5A). ABA content determination showed that overexpression of ZjPPH increased the ABA content in transgenic lines compared with control (Figure 5D). The soluble sugar content was also found to be higher in the transgenic lines (Figure 5E).

Two senescence marker genes, SAG12 and SAG14, were selected to assess the impact of ZjPPH in senescence processes. It turned out that the expression levels of the two senescence marker genes were upregulated significantly (Figures 6A,B). It indicated that the overexpression of ZjPPH promoted senescence. In addition, the expression of four photosynthetic efficiency marker genes was monitored to evaluate the photosynthetic rate. The results revealed that the overexpression of ZjPPH significantly downregulated the expression of CAB1 (Chl a/b binding protein 1), PsaF (PSI component), RbcL (rubisco large subunit), and RCA (rubisco activase) (Figures 6C–F). It is rational to propose that ZjPPH could decrease the photosynthetic efficiency, which should be further evidenced by the determination of ultrastructure characteristics of chloroplasts and the detailed photosynthetic parameters.

Transmission electron microscopic analysis was carried out to reveal the ultrastructure changes of chloroplasts caused by the overexpression of ZjPPH. The results revealed that the overexpression of ZjPPH caused drastic changes to the ultrastructure of chloroplasts. In detail, the chloroplasts in the control maintained a high stability status with normal grana stacks, starch particles, and thylakoid membrane (Figure 7A). In contrast, the chloroplasts showed a decomposition tendency in the ZjPPH-overexpressing plants with fewer grana stacks and looser thylakoid membrane (Figure 7B).

The JIP test revealed that F_O/F_V and F_V/F_M were higher in the control than in the transgenic lines, indicating less of active photosynthesizing structures and photosynthetic membranes in the transgenic lines (Supplementary Table 1). Comparison of normalized OJIP curves provides detailed information of structural and functional differences of the photosystem (Hu et al., 2016). As shown in Figure 8A, the O-J step values were higher in the transgenic lines than in the control. It demonstrated that the expression of ZjPPH retarded the electron flow in the acceptor side of PSII.

The L-band is an indicator of the PSII status in the thylakoid grana membrane. A positive L-band was observed, suggesting the ungrouping of the antenna complexes and attenuated connectivity between LHCII and PSII reaction centers (Figure 8B and Supplementary Figure 1A). The positive K-band revealed the inactivation of OEC, resulting in the slower electron transport toward PSII reaction center (Supplementary Figure 1B). In this study, no obvious H-band was identified, indicating that the expression of ZjPPH caused no significant changes of the PQ pool volume and electron transduction between the two PS (Supplementary Figure 1C). The negative G-band implied that the PSI acceptors pool is bigger (Supplementary Figure 1D). Because the intersystem electron transduction did not show significant difference, it is rational to conclude that the PSI was suppressed in the transgenic lines. The lower RE₀/CS_m of the transgenic lines also supported this assumption (Supplementary Table 1).

Phenomenological energy pipeline models help to visualize the influence caused by the overexpression of ZjPPH. It showed that membrane model parameters, including ABS/RC, TR₀/RC, and DI₀/RC, were increased in the transgenic lines (Figure 8C). However, the ET₀/RC was showed no significant changes. It indicated the decreased efficiency of PSII reaction center in electron transport (Janeeshma et al., 2021). Cross-section (CS)-level indicators, including ABS/CS_m, TR₀/CS_m, and ET₀/CS_m, were lower in the transgenic lines, while the DI₀/CS_m increased (Figure 8C). In addition, less active reaction centers were identified in the ZjPPH-overexpressing lines. The decrease of ABS/CS_m, TR₀/CS_m, and ET₀/CS_m was related to the increased inactive reaction centers and suppressed PSII efficiency (Janeeshma et al., 2021). The increase of the DI₀/CS_m reflected the inefficiency in the utilization of light energy.