Identification of two glycerophosphodiester phosphodiesterase genes in maize leaf phosphorus remobilization

Jingxin Wng,Weno Pn,Alexiy Nikiforov,Willim King,Wnting Hong,Weiwei Li,Yng Hn, Jn Ptton-Vogt, Jino Shen, Lingyun Cheng,*

aKey Laboratory of Plant-Soil Interactions, Ministry of Education/Key Laboratory of Plant Nutrition, Ministry of Agriculture, Department of Plant Nutrition,China Agricultural University,Beijing 100193,China

bState Key Laboratory of Plant Genomics,Institute of Genetics and Developmental Biology,Chinese Academy of Sciences,Beijing 100101,China

cDepartment of Biological Sciences,Duquesne University,Pittsburgh,PA 15282,USA

Keywords:Glycerophosphodiester phosphodiesterases Maize Phosphate deficiency Phosphorus remobilization Phospholipid

ABSTRACT Phosphate deficiency is one of the leading causes of crop productivity loss. Phospholipid degradation liberates phosphate to cope with phosphate deficiency. Glycerophosphodiester phosphodiesterases (GPX-PDEs) hydrolyse the intermediate products of phospholipid catabolism glycerophosphodiesters into glycerol-3-phosphate, a precursor of phosphate.However,the function of GPX-PDEs in phosphate remobilization in maize remains unclear.In the present study, we characterized two phosphate deficiency-inducible GPX-PDE genes,ZmGPX-PDE1 and ZmGPX-PDE5, in maize leaves. ZmGPX-PDE1 and ZmGPX-PDE5 were transcriptionally regulated by ZmPHR1, a well-described phosphate starvation-responsive transcription factor of the MYB family.Complementation of the yeast GPX-PDE mutant gde1Δ indicated that ZmGPX-PDE1 and ZmGPX-PDE5 functioned as GPX-PDEs,suggesting their roles in phosphate recycling from glycerophosphodiesters. In vitro enzyme assays showed that ZmGPX-PDE1 and ZmGPX-PDE5 catalysed glycerophosphodiester degradation with different substrate preferences for glycerophosphoinositol and glycerophosphocholine, respectively.ZmGPX-PDE1 was upregulated during leaf senescence, and more remarkably,loss of ZmGPXPDE1 in maize compromised the remobilization of phosphorus from senescing leaves to young leaves,resulting in a stay-green phenotype under phosphate starvation.These results suggest that ZmGPX-PDE1 catalyses the degradation of glycerophosphodiesters in maize, promoting phosphate recycling from senescing leaves to new leaves. This mechanism is crucial for improving phosphorus utilization efficiency in crops.

1.Introduction

Phosphorus (P) is an essential macronutrient required for plant growth and development that is necessary for crucial cellular functions such as cellular constituent formation,energy production, metabolism and signal transduction [1,2].Due to the chemical properties of P in soils, nutritional inorganic phosphate (Pi) is a limiting factor for crop productivity worldwide [2,3]. Hence, improvement of P efficiency is pivotal for sustainable food production.

In order to achieve enhanced P efficiency, plants have developed highly integrated systems to improve acquisition of Pi from the environment and recycling of P within plants[4-7].Simply increasing P uptake would lead to P export from the environment, so enhancement of the efficiency of P remobilization from senescing organs to young and developing organs is a promising area for improvement of crop P efficiency[7,8].Phospholipid degradation is a major driver of P remobilization in many plant species[9,10].

Phospholipases implicated in recycling Pi are divided into three classes,the phospholipase A1(PLA1),phospholipase C(PLC),and phospholipase D(PLD)groups,according to cleavage sites in membrane glycerophospholipids[7].Glycerophospholipid hydrolysis mediated by PLA1 produces lysophospholipids and fatty acids.The expression of a carnation PLA1 gene is upregulated at the onset of flower senescence,and antisense suppression of the corresponding gene in Arabidopsis results in delayed leaf senescence [11,12]. Nonspecific phospholipase C (NPC) degrades glycerophospholipids into diacylglycerol(DAG)and Pi-containing head groups. NPC4 and NPC5 are responsive to Pi limitation and are involved in digalactosyldiacylglycerol (DGDG) synthesis[13,14]. PLD hydrolyses glycerophospholipids into phosphatidic acid and a free head group and is implicated in lipid degradation during abscisic acid (ABA)/ethylene-promoted leaf senescence and the response to Pi limitation[15,16].

Glycerophosphodiesters, which are produced by phospholipase A2 (PLA2) and lysophospholipase from membrane glycerophospholipids, are hydrolysed by glycerophosphodiester phosphodiesterases (GPX-PDEs), also known as GDPDs. In this route, GPX-PDEs decompose glycerophosphodiesters into glycerol-3-phosphate (G3P) and corresponding alcohols. GPXPDEs are highly conserved enzymes among species ranging from bacteria to plants and mammals, and have been linked to numerous biological functions [17-19]. The GPX-PDEs GlpQ and UgpQ,which were initially characterized in Escherichia coli,possess glycerophosphodiesterase activity with broad substrate specificity and differ in cellular location [20-22]. In Saccharomyces cerevisiae,the GPX-PDE gene Gde1p(YPL110c)has been shown to be required for yeast to use glycerophosphocholine(GPC)as the sole P source[23].In plants,GPX-PDEs were first characterized in the cell walls and vacuoles of carrot cells [24,25]. In Arabidopsis and rice, GPXPDEs are induced by Pi deprivation;AtGPX-PDE1 and OsGDPD2 can improve the resistance to Pi limitation [26-28]. In white lupin(Lupinus albus L.), two GPX-PDEs have been characterized to play roles in root hair development in response to Pi limitation[29].

Phosphorus deficiency-inducible GPX-PDEs are involved in adaptation to P deficiency.However,little is known about the roles of GPX-PDEs in Pi scavenging in maize leaves for adaptation to Pi limitation. In this study, we characterized the roles of GPX-PDE genes in P remobilization in maize leaves and demonstrated that ZmGPX-PDE1 probably participates the recycling of P from senescing to young leaves to enable adaptation to Pi limitation.

2. Materials and methods

2.1. Plant materials and growth conditions

Maize seeds were sterilized with 10% (v/v) H2O2for 20 min and soaked in saturated CaSO4solution overnight. The seeds were then germinated between filter paper sheets in the dark and wrapped in paper. The complete nutrient solution contained 2 mmol L?1Ca(NO3)2, 250 μmol L?1KH2PO4, 750 μmol L?1K2SO4,650 μmol L?1MgSO4, 1 μmol L?1H3BO3, 1 μmol L?1MnSO4,0.1 μmol L?1CuSO4, 1 μmol L?1ZnSO4, 0.005 μmol L?1(NH4)6MoO7, and 100 μmol L?1FeNa-EDTA (pH 6.0). For Pi deprivation treatment, KH2PO4was replaced with 250 μmol L?1KCl. The nutrient solution was replaced every three days. Four seedlings at the two-leaf stage were cultivated in hydroponic solution in a 2.2 L container under the following conditions: a light intensity of 230 μmol m?2s?1, a photoperiod of 16 h day/8 h night, a day/night temperature of 28/22 °C, and a relative humidity of 60%.

2.2. RNA preparation and quantitative real-time PCR (qRTPCR)

Total RNA was extracted from different samples using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA), and a reverse transcription kit (TaKaRa, Dalian, China) was used to synthesize cDNA. The qRT-PCR was performed using a SYBR Green PCR Master Mix kit (TaKaRa, Dalian, China). Relative expression levels were calculated with the comparative cycle threshold (CT) method. ZmTubulin and ZmUBCP were used as the internal controls. The primers are listed in Supplementary Table S1. Three biological replicates were used for each treatment. The PCR program was as follows: 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C, 30 s at 60 °C and 30 s at 72 °C.

2.3. Subcellular localization analysis of ZmGPX-PDE1 and ZmGPX-PDE5 in maize protoplasts

The full-length CDSs of ZmGPX-PDE1 and ZmGPX-PDE5 were inserted into pZES-NL vectors to generate GFP-fused proteins.The 35S:ZmGPX-PDE5-GFP and 35S:HDEL-RFP/35S:ZmGPX-PDE1-GFP plasmids were transformed into maize protoplasts with a polyethylene glycol-mediated method [30,31]. The pEZS-NL vector was used as the control. The GFP and RFP signals were observed under a Zeiss confocal microscope after 18 h of incubation at 22 °C.

2.4. Complementation of yeast mutants and enzyme assay

The full-length CDSs of ZmGPX-PDE1 and ZmGPX-PDE5 were amplified by PCR using primers (Table S1) and were subsequently cloned into the yeast vector p416 [32].

Wild-type (WT) and gde1Δ deletion mutant strains were transformed with uracil prototrophy plasmids using a standard lithium acetate-based technique. The transformants were grown in low-Pi medium lacking uracil overnight.Overnight cultures with A600values of 0.01 were used to inoculate fresh culture medium lacking phosphate or containing 200 μmol L?1KH2PO4or 200 μmol L?1GPC. The optical density at 600 nm was determined after five days of growth at 30 °C on a rotating wheel.

The enzymatic activity of recombinant ZmGPX-PDE1 and ZmGPX-PDE5 was analysed following the methods of Larson et al. [20]. The total assay mixture (0.5 mL) contained 0.5 mol L?1hydrazine-Gly buffer (pH 9.0), 1 mmol L?1NAD,5 units of G3P dehydrogenase, either 10 mmol L?1CaCl2or 10 mmol L?1MgCl2, 0.5 mmol L?1substrate (GPC or GPI), 10 μg of ZmGPX-PDE5 protein and 20 μg of ZmGPX-PDE1 protein.The concentration of NADH was observed from the absorbance change at 340 nm. All activity measurements were performed thrice, and the average values were used for calculations.

2.5. Yeast two-hybrid (Y2H) assays

Y2H assays were performed using the Gal4 vector system.The CDSs of ZmGPX-PDE1 and ZmGPX-PDE5 (without signal peptide sequences) were cloned into pGADT7 (AD) and pGBKT7 (BD), and the construct pairs (ZmGPX-PDE1-AD plus ZmGPX-PDE1-BD and ZmGPX-PDE5-AD plus ZmGPXPDE5-BD) were used to co-transform the yeast strain AH109.BD-ZmGPX-PDE plus AD, AD-ZmGPX-PDE plus BD, and BD plus AD were used as negative controls. The transformed cells were grown on SD/-Leu-Trp and SD/-Trp-Leu-His plates for 3-7 days at 30 °C.

2.6. Electrophoretic mobility shift assay (EMSA) and transient expression regulation assays in Nicotiana benthamiana leaves

The CDS of ZmPHR1 was amplified and inserted into the pGEX-4T-1 vector. The resulting plasmid was introduced into E. coli strain BL21(DE3)pLySS and was induced with 0.5 mmol L-?1isopropyl-β-D-thiogalactopyranoside (IPTG) at 18 °C for 16 h. The ZmPHR1-GST fusion protein was purified by Glutathione-Sepharose. The biotin-labelled and unlabelled probes are listed in Table S1. The details of the EMSA experiments were described by Chen et al. [33].

A transient expression assay was performed as described previously [34]. Promoter fragments (～2 kb) of the genes ZmGPX-PDE1 and ZmGPX-PDE5 were amplified by PCR and cloned into pGWB35 vectors, generating pZmGPX-PDE1:LUC and pZmGPX-PDE5:LUC reporter plasmids, respectively.ZmPHR1 was cloned into the pCanG-Flag vector to form the fused construct ZmPHR1-Flag.The recombinant plasmid pairs were introduced into Agrobacterium tumefaciens strain GV3101 and used to co-transform N. benthamiana leaves via an electroporation method. pZmGPX-PDE1:LUC/pCanG-Flag or pZmGPX-PDE5:LUC/pCanG-Flag pairs were used as internal controls. All the constructs were expressed for 3-5 days in tobacco leaves at 22 °C before charge-coupled device (CCD)imaging.The luminescence was imaged by a low-light cooled CCD imaging apparatus (NightOWL IILB983 with Indigo software). The leaves were sprayed with 100 mmol L?1luciferin and were placed in darkness for 5 min before luminescence detection.

2.7. Total P concentration assessment

Maize leaves were dried at 70 °C for 72 h and cut into small pieces with scissors. Approximately 0.3 g samples were digested in 5 mL of H2SO4-H2O2. The P concentrations of leaves were determined using the vanado-molybdate method.

2.8. Generation of a zmgpx-pde1 mutant by CRISPR/Cas9

The zmgpx-pde1 mutants used in this study were generated using the CRISPR/Cas9 genome editing system [35]. A 20 bp target sequence located at base pairs 2322-2341 was selected,and the primers are listed in Table S1. The corresponding fragments were cloned into the CPB vector using the Hind III restriction site by fusion using an In-Fusion PCR Cloning Kit(Transgene, Beijing, China). Transformation with the resulting plasmid was completed at Weimi Biotechnology Co. Ltd.(Jiangsu, China). PCR was performed with a pair of primers flanking the target sites, and the PCR product was sequenced to identify mutations in ZmGPX-PDE1.

2.9. Statistical analysis

The data were analysed using IBM Statistics SPSS 21 (SPSS Inc., Chicago, IL, USA). Following one-way analysis of variance(ANOVA), Tukey's HSD post hoc test was used to analyse differences between means at the 0.05 probability level.

3. Results

3.1. Identification and characterization of ZmGPX-PDEs in maize

To identify GPX-PDE family members in maize, we searched for the conserved GPX-PDE domain sequence of E. coli GlpQ(PF03009,Pfam)in the Maize Genetics and Genomics Database(MaizeGDB,https://www.maizegdb.org)and found 14 ZmGPXPDEs homologous to GlpQ.These fourteen genes were named sequentially as ZmGPX-PDE1-ZmGPX-PDE14. The phylogenetic relationships of ZmGPX-PDEs with homologous proteins from Arabidopsis,rice,E. coli(GlpQ and UgpQ),and L. albus are shown (Fig. 1A). ZmGPX-PDE1, ZmGPX-PDE2, ZmGPX-PDE3 and ZmGPX-PDE4 were grouped with UgpQ, which is a cytoplasm Mg2+-dependent enzyme, while ZmGPX-PDE5,ZmGPX-PDE6 and ZmGPX-PDE7 were similar to GlpQ, which is a periplasm Ca2+-dependent protein. Analysis of protein domain organization with the Simple Modular Architecture Research Tool (SMART, http://smart.embl-heidelberg.de) revealed that ZmGPX-PDE4-ZmGPX-PDE7 had signal peptides or transmembrane helices (Fig. 1B). Only ZmGPX-PDE13 and ZmGPX-PDE14 contained two GPX-PDE domains, and ZmGPX-PDE7-ZmGPX-PDE-12, which were grouped with AtGPX-PDE-like subfamilies, each contained only one GPXPDE domain, one transmembrane helix or signal peptide(Fig.1B). These results indicate that the ZmGPX-PDEs that belong to different clusters may have different functions or characteristics.

Fig.1- Phylogenetic relationships among GPX-PDEs of various organisms and the domain organization of ZmGPX-PDE proteins.(A)Phylogenetic tree of ZmGPX-PDE proteins and GPX-PDE proteins in other species.The tree was created in MEGA6.The organisms include Arabidopsis thaliana(At),Escherichia coli(GlpQ and UgpQ), Lupinus albus(La)and Oryza sativa(Os). (B)Domain organization of ZmGPX-PDE proteins.The scale bar represents 100 amino acids (aa).

Fig.2-Expression of ZmGPX-PDEs under different phosphorus conditions.The relative expression levels of ZmGPX-PDEs in the leaves of maize after 10 days of growth with phosphorus(+P)or without phosphorus(?P)are shown.ZmTubulin was used as an internal control for normalization.The values are presented as the mean±SD(n=3).*indicate significant differences between the mean values of the +P and ?P groups as determined by Student's t-test at P <0.05.

3.2. GPX-PDE activity and ZmGPX-PDE gene expression in maize leaves are induced by Pi limitation

The response of total GPX-PDE activity in maize leaves to Pi limitation was estimated by using GPC as a substrate. The GPX-PDE activity was significantly higher in Pi-starved leaves than in Pi-sufficient leaves (Fig. S1A). Quantitative RT-PCR was performed to study the responses of the ZmGPX-PDE genes to Pi limitation in maize leaves.Pi limitation enhanced the expression of all ZmGPX-PDE genes except ZmGPX-PDE10 and ZmGPX-PDE11 (Fig. 2). The expression levels of ZmGPXPDE1-ZmGPX-PDE3, which were grouped with E. coli UgpQ,under Pi limitation were markedly higher than those of ZmGPX-PDE4-ZmGPX-PDE6, which were grouped with E. coli GlpQ. In contrast, the rest of the ZmGPX-PDEs were slightly induced by Pi limitation.

3.3.ZmGPX-PDE1 and ZmGPX-PDE5 can complement a yeast GPX-PDE mutant and display in vitro GPX-PDE activity

To examine the potential roles of ZmGPX-PDEs in response to Pi limitation, we further characterized two of the most promising maize GPX-PDEs, ZmGPX-PDE1 and ZmGPX-PDE5,which were grouped with E. coli UgpQ and GlpQ, respectively.We expressed ZmGPX-PDE1 and ZmGPX-PDE5 in the GPX-PDE-deficient S. cerevisiae strain gde1Δ (a GPX-PDE mutant). As shown in Table 1, gde1Δ mutants expressing either ZmGPXPDE1 or ZmGPX-PDE5 were able to grow in medium containing GPC as the phosphate source, whereas yeast mutants containing only the empty vector had impaired cell growth.Thus, ZmGPX-PDE1 or ZmGPX-PDE5 from Zea mays can counteract the growth defects in yeast GPX-PDE mutants when GPC is the sole source of P.

?

To further explore the enzymatic functions of ZmGPXPDE1 and ZmGPX-PDE5, we purified recombinant ZmGPXPDE1 and ZmGPX-PDE5 from E. coli and subsequently measured their enzymatic activity using GPC and GPI as substrates.The results showed that ZmGPX-PDE1 exhibited a preference for GPI as a substrate and used Mg2+as the divalent cation (Fig. 3A). In contrast, ZmGPX-PDE5 had higher enzymatic activity with GPC than with GPI as the substrate in the presence of either Mg2+or Ca2+(Fig.3B).Free GST protein had no phosphodiesterase activity for GPC or GPI. Therefore,ZmGPX-PDE1 and ZmGPX-PDE5 both have phosphodiesterase activity towards GPC and GPI but show different substrate preferences and dependences on divalent cations.

3.4.ZmGPX-PDE1 and ZmGPX-PDE5 homodimerize,as determined by Y2H assay

To analyse the possible homodimerization states of the ZmGPX-PDE1 and ZmGPX-PDE5 proteins, we performed Y2H assays. The ZmGPX-PDE1 and ZmGPX-PDE5 open reading frames were separately cloned into prey pGADT7 (AD) and bait pGBKT7 (BD) plasmids. Four different bait/prey combinations were tested: AD/BD, AD-ZmGPX-PDE/BD, BDZmGPX-PDE/AD,and AD-ZmGPX-PDE/BD-ZmGPX-PDE(Fig.4).The results showed that all transformants grew well on SD/-Leu-Trp medium.The AD-ZmGPX-PDE1 and BD-ZmGPX-PDE1 co-transformed yeast strains grew well on SD/-Leu-Trp-His medium. However, the cells transformed with the negative control combination sets of BD-ZmGPX-PDE1/AD,AD-ZmGPXPDE1/BD, and AD/BD could not grow well on SD/-Leu-Trp-His medium. Similar results were obtained for ZmGPX-PDE5.These results show that ZmGPX-PDE1 and ZmGPX-PDE5 are able to homodimerize in yeast.

3.5. Subcellular localization of ZmGPX-PDE1 and ZmGPXPDE5 in maize protoplasts

The TargetP and SignalP servers predicted that the GPX-PDE5 protein contains an endoplasmic reticulum (ER)-targeting signal peptide with 21 N-terminal amino acids, while the ZmGPX-PDE1 protein has no signal peptide and is expected to be located in the cytoplasm. To confirm the subcellular localization of ZmGPX-PDE1 and ZmGPX-PDE5,the respective GFP-fused genes were transiently expressed together with the 35S:HDEL-RFP vector, which was used as a marker for the ER,in maize protoplasts. The results showed that the ZmGPXPDE5 signal co-localized with the ER,confirming that ZmGPXPDE5 localizes to this organelle (Fig. 5G-J). In contrast, the ZmGPX-PDE1-GFP signal was observed in the cytoplasm (Fig.5D-F).

3.6. ZmGPX-PDE1 and ZmGPX-PDE5 are transcriptionally regulated by ZmPHR1

PHR1 is a key transcription factor that regulates the expression of Pi starvation response genes in plants. Analyses of ZmGPX-PDE1 and ZmGPX-PDE5 promoter sequences identified two and one conserved PHR1 binding site (P1BS), respectively(Fig. 6A). Binding of ZmPHR1 to the promoter sequences of ZmGPX-PDE1 and ZmGPX-PDE5 was verified by EMSA. Fulllength ZmPHR1 was expressed in E. coli and affinity purified.As shown in Fig. 6B, the PHR1 fusion proteins were able to bind DNA probes containing the P1BSs of the ZmGPX-PDE1 and ZmGPX-PDE5 promoters. In contrast, additional unlabeled DNA probes competed for binding. These results reveal that ZmPHR1 activates ZmGPX-PDE1 and ZmGPX-PDE5 expression through direct association with their promoters.

Fig.3- GPX-PDE activity of ZmGPX-PDE1 and ZmGPX-PDE5 in vitro.The Mg2+and Ca2+ dependence of ZmGPX-PDE1(A)and ZmGPX-PDE5(B) GPX-PDE activity towards GPC and GPI was examined. The values are presented as the mean±SD(n =3).Different letters indicate significant differences among treatments,P<0.05.

Fig.4-Homodimerization of ZmGPX-PDE proteins as revealed by Y2H assay.Cells were transformed with BD-ZmGPX-PDE plus AD or AD-ZmGPX-PDE plus BD.Cells transformed with BD plus AD were used as negative controls.The transformed strains were plated on SD/-Leu/-Trp or SD/-Leu/-Trp/-His selection medium. AD,target construct vector;BD,bait construct vector.

Using a well-established transient expression assay with N. benthamiana leaves, we verified the promoting effect of ZmPHR1 on the expression of separate reporters containing the ZmGPX-PDE1 and ZmGPX-PDE5 promoters fused with the firefly luciferase gene (LUC). When pZmGPX-PDE1:LUC or pZmGPX-PDE5:LUC and p35S:ZmPHR1 were used to cotransform N. benthamiana, substantial amounts of luminescence could be detected; in contrast, no or much weaker LUC activity was observed in the negative controls (co-transformed with flag and pZmGPX-PDE1:LUC or pZmGPX-PDE5:LUC) (Fig. 6C). Together, the transient assays of N. benthamiana leaves confirm that ZmPHR1 promotes expression of ZmGPXPDE1 and ZmGPX-PDE5 in vivo.

3.7. ZmGPX-PDE1 likely impacts P remobilization in leaves

Since GPX-PDE proteins can scavenge Pi from organic P sources, ZmGPX-PDEs may be involved in P remobilization from senescing leaves to new leaves. We first detected the expression levels of ZmGPX-PDE1 and ZmGPX-PDE5 in senescing lower leaves, middle leaves and new upper leaves in maize seedlings grown in Pi-sufficient conditions for 15 days. The results showed that ZmGPX-PDE1 transcript abundance in the lower leaves was significantly higher than that in the upper leaves (Fig. 7A). The expression levels of ZmGPX-PDE5 were almost the same between the lower leaves and upper leaves (Fig. 7B). Then,we measured the total GPX-PDE enzyme activity in different leaves, and the results showed that the total GPX-PDE activity in lower leaves was higher than that in upper leaves (Fig. S1B).

To examine whether the expression of ZmGPX-PDE1 was correlated with P remobilization, we measured ZmGPX-PDE1 expression in different maize leaves under Pi-deprived conditions(Fig.7C-E).In lower leaves,1 day of Pi deprivation resulted in increases in ZmGPX-PDE1 transcript abundance.After 3 and 5 days of Pi deprivation,the transcript abundance of ZmGPX-PDE1 remained high (Fig. 7C). In middle leaves,ZmGPX-PDE1 transcript abundance was induced after 3 days of Pi deprivation (Fig. 7D). In upper leaves, ZmGPX-PDE1 transcript abundance was induced after 5 days of Pi deprivation (Fig. 7E). These results indicate that ZmGPXPDE1 in senescing lower leaves responds quickly to Pi deprivation.

Fig.5- Subcellular localization of ZmGPX-PDE1 and ZmGPX-PDE5 in maize protoplasts.35S:ZmGPX-PDE5-GFP and ER marker 35S:HDEL-RFP plasmids were used to co-transform maize protoplasts.35S:ZmGPX-PDE1-GFP plasmids were used to transform maize protoplasts.ZmGPX-PDE5-GFP fluorescence co-localized with the ER marker HDEL-RFP when the images were merged(G-J).ZmGPX-PDE1-GFP fluorescence was located in the cytoplasm(D-F).The GFP signal from the control 35S:GFP construct pEZS-NL was distributed throughout the maize protoplasts (A-C).Scale bars,10 μm.

To further evaluate the role of ZmGPX-PDE1 in P remobilization, we performed targeted mutagenesis of ZmGPX-PDE1 via the CRISPR/Cas9 system. We acquired a translational frameshift mutant of the target line zmgpx-pde1(Fig. 8A). The activity of GPX-PDE was significantly lower in zmgpx-pde1 plants than in WT plants under Pi-deprived conditions (Fig.8B). Maize seedlings were grown hydroponically in Pisufficient solution for 7 days and subsequently transferred to Pi-deprived solution. After 2 weeks, WT plants displayed chlorosis in the lower leaves(Fig. 8C).In contrast,zmgpx-pde1 plants showed delayed leaf senescence, as evidenced by the greater greenness in zmgpx-pde1 leaves than in WT leaves(Fig.8D). The P concentrations of lower leaves were higher in zmgpx-pde1 plants than in WT plants(Fig.8E).P remobilization was estimated by the ratio of the P concentration in the upper leaves to that in the lower leaves.The results showed that the ratio in zmgpx-pde1 plants was significantly lower than that in WT plants (Fig. 8F), suggesting that ZmGPX-PDE1 is probably involved in P remobilization from senescing to new leaves in maize.

4.Discussion

Internal P scavenging in plants via liberation of Pi from different molecular components helps to improve crop P utilization efficiency, thereby reducing the rate of depletion of nonrenewable rock P reserves [7,8]. Disruption of plasma membrane integrity and degradation of phospholipids are key events during senescence or phosphate deprivation. In this study, we characterized the Pi deprivation-induced glycerophosphodiester phosphodiesterase genes GPX-PDEs in maize leaves. We uncovered the biochemical functions of ZmGPX-PDE1 and ZmGPX-PDE5 based on yeast complementation experiments, in vitro enzyme assays, and subcellular localization and transcriptional regulation experiments.The high expression level of ZmGPX-PDE1 in senescing leaves at the early Pi deprivation stage and the decrease in P remobilization from senescing to new leaves in zmgpx-pde1 mutants further indicate the role of ZmGPX-PDE1 in P recycling in maize.

4.1. Identification and characterization of ZmGPX-PDEs

Fourteen GPX-PDEs were identified from maize leaves that showed sequence similarities to related genes in other organisms, such as the human MIR16 [36]and GPX-PDEs in plants[25-29],bacteria[20,22],and yeast[18,23];the findings indicated that these maize genes belong to a large, evolutionarily conserved family of GPX-PDEs. ZmGPX-PDE1 and ZmGPX-PDE5, two distinct P deficiency-inducible GPX-PDEs in maize leaves, both exhibited phosphodiesterase activity and formed dimers in the native state(Figs.3 and 4;Table 1);however, they showed different substrate preferences and were located in the cytoplasm and ER, respectively (Fig. 5).Similar findings have been observed for two divergent GPXPDEs in white lupin [29]. It is noteworthy that ZmGPX-PDE1 showed substrate preference towards GPI, indicating that it provides inositol for consequent metabolism. Inositol polyphosphates (InsPs), which are highly phosphorylated molecules containing inositol rings that can be sequentially phosphorylated,are characterized as ligands that enable SPX proteins to interact with PHR1/2 in P sensing and signaling[37,38]. Whether inositol produced by the hydrolysis of GPI via ZmGPX-PDE1 contributes to the roles of InsPs in P signaling deserves to be further investigated.

Fig.6- ZmGPX-PDE1 and ZmGPX-PDE5 were transcriptionally regulated by ZmPHR1.(A)Diagram of the ZmGPX-PDE1 and ZmGPX-PDE5 promoter regions showing the relative positions of the P1BS cis-elements(?231,?274 and ?1102).(B) ZmPHR1 bound to the P1BS cis-elements in the ZmGPX-PDE1 and ZmGPX-PDE5 promoters in vitro.A biotin-labelled ZmGPX-PDE probe was incubated with GST-ZmPHR1 or 2-fold GST-ZmPHR1. A competition assay was performed by adding a 50-fold excess of unlabelled probe(cold probe).The arrow indicates the position of the bound probe.(C)ZmPHR1 promoted the expression of ZmGPX-PDE1 and ZmGPX-PDE5 in vivo.Tobacco leaves were transformed with pZmGPX-PDE:LUC plus ZmPHR1 or pZmGPX-PDE:LUC plus a Flag vector as the negative control.The right panel indicates the infiltrated construct pair.

Since the ZmGPX-PDE5 protein contains a signal peptide at the N-terminus, ZmGPX-PDE5 exhibited ER subcellular localization (Fig. 5). Consistent with this finding, our previous study showed that the orthologue of ZmGPX-PDE5 in white lupin, LaGPX-PDE1, also localizes to the ER [29]. It is well known that the ER plays an important role in phospholipid metabolism and that many gene products involved in phospholipid metabolism localize to this organelle. Hydrolysis of phospholipids mediated by several PLA2enzymes occurs at the ER [39,40]. Phosphatidic Acid Phosphohydrolase 1 and Phosphatidic Acid Phosphohydrolase 2, which regulate phospholipid synthesis, localize to the ER in Arabidopsis [41].Overall, it appears that many phospholipid-related gene products are associated with the ER and play roles in phospholipid degradation in response to Pi limitation.

Fig.7- Effects of different phosphorus conditions on the expression patterns of ZmGPX-PDE1 in different leaves.The relative expression of ZmGPX-PDE1(A)and ZmGPX-PDE5(B)in the upper leaves,middle leaves and lower leaves of maize grown for 15 days under Pi-sufficient conditions was examined.The values are presented as the mean±SD(n =3). Different letters indicate significant differences among treatments,P< 0.05.The relative expression of ZmGPX-PDE1 in the lower leaves(C),middle leaves(D)and upper leaves(E)of plants after different phosphorus treatments on day 1(D1),day 3(D3)and day 5(D5)was examined.Ten-day-old seedlings(approximately six leaf stages)that were precultured in phosphorus-sufficient solution were transferred to nutrient solution with or without phosphorus for 1, 3, or 5 days.ZmUBCP was used to normalize the expression levels.The values are presented as the mean± SD(n= 3).* indicates significant differences in the mean values between treatments as determined by Student's t-tests at P< 0.05.

4.2. ZmGPX-PDE1 and ZmGPX-PDE5 expression is regulated by ZmPHR1

ZmGPX-PDEs were induced by Pi deprivation,and ZmGPX-PDE1 and ZmGPX-PDE5 expression was greatly increased (Fig. 2).PHR1, the known transcription factor regulating responses to P starvation, binds to P1BS elements in the promoters of a large number of P starvation-responsive genes[42,43].ZmGPXPDE1 and ZmGPX-PDE5 were found to contain P1BS motifs.EMSA and transient expression assays in N. benthamiana demonstrated that ZmGPX-PDE1 and ZmGPX-PDE5 are regulated by ZmPHR1(Fig.6).PHR1 has been shown to regulate the process of phospholipid remodeling in plants to enable adaptation to Pi limitation. Glycerolipid composition and the expression of most lipid-remodeling gene transcripts have been found to be altered in phr1 mutants under Pi limitation[44]. The expression levels of AtGPX-PDE6, the orthologue of ZmGPX-PDE5 in Arabidopsis, have also been found to be significantly reduced in phr1 mutants during Pi deprivation[14]. Our data further indicate a linkage between PHR1 and phospholipid metabolism pathways upon Pi limitation.

Fig.8-Characterization of zmgpx-pde1 via CRISPR-Cas9.(A)Targeted mutagenesis of ZmGPX-PDE1 was performed via CRISPRCas9.The protospacer adjacent motif(PAM) sequences are labelled in red.(B) GPX-PDE activity in WT and zmgpx-pde1 plants under phosphorus-deficient(?P)conditions.(C,D)Growth condition of leaves in WT and zmgpx-pde1 plants under ?P conditions. Scale bars,2 cm.(E)The total P concentrations of the upper and lower leaves were determined in WT and zmgpxpde1 plants under ?P conditions.(F) Phosphorus remobilization(estimated as the ratio of the phosphorus concentrations in upper leaves to that in lower leaves)in wild type and zmgpx-pde1 plants under ?P conditions was determined.The values are presented as the mean±SD(n=3).*indicate significant differences in the mean values between treatments as determined by Student's t-test at P <0.05.

4.3. ZmGPX-PDE1 may be involved in P recycling from senescing leaves to new leaves

Nucleic acids,phospholipids and phosphorylated metabolites are the most abundant sources of P in leaves, and Pi can be liberated from organic phosphate molecules [8]. P-containing molecules are major contributing factors to phosphate remobilization.AtPAP26 and OsPAP26,which can hydrolyse a broad range of organophosphates, are involved in P remobilization during leaf senescence [45,46]. The exonuclease DPD1 participates in the relocation of P to upper tissues and the efficient use of phosphate [47]. In addition, RNases show increased expression at the gene and/or protein level in response to P deficiency and leaf senescence[48-50].

Phospholipids are another macromolecule P reserve from which P can be liberated to meet phosphate demands during plant development. Phospholipid degradation and replacement by sulfolipids and glycolipids is an important strategy for plant adaptation to phosphate stress[7,51].The phospholipase genes PLA, PLC and PLD, which conduct the initial step of phospholipid degradation, are induced by senescence as well as by Pi starvation in Arabidopsis [12-16]. In addition,upregulation of GPX-PDE genes can improve phosphate deficiency tolerance in Arabidopsis and rice [26,28]. Interestingly, PLC and PLD family members are not induced by phosphate starvation in maize, and the transcription of six GPX-PDE (GPDE) genes has been found to increase during phosphate deficiency[52].Therefore,the phospholipid hydrolysis route mediated by GPX-PDEs may play a more important role than other pathways in resistance to phosphate deficiency in maize. Consistent with this possibility, we found that ZmGPX-PDE1 was highly upregulated in senescent leaves and P-deficient leaves. Compared with WT maize, transgenic maize with attenuated ZmGPX-PDE1 activity showed decreased ratio of phosphorus concentration in upper leaves to that in lower leaves (Fig. 8B, E, F), which is similar to dpd1 mutant in Arabidopsis [47]. An RNA-seq study previously revealed that an OsGPX-PDE in rice orthologous to ZmGPXPDE1 is highly induced in senescing flag leaves during grain filling and indicated that this gene is involved in the remobilization of P from flag leaves to developing grains [53].Taken together, these results suggest that ZmGPX-PDE1 contributes to the efficient recycling of P from old to young leaves. The role of ZmGPX-PDE1 in phosphate remobilization during maize grain filling deserves further study.

There is considerable overlap among differentially expressed genes during Pi deficiency and leaf senescence[43,54], highlighting the relationship between phosphate deficiency-related P remobilization and senescence-related P remobilization.Identification of these molecular components together with the finding that ZmGPX-PDE1 is involved in P remobilization during leaf senescence will help to unravel the complexity of P limitation and senescence signaling pathways.The regulation of ZmGPX-PDE1 expression in senescent leaves is still unknown. The transcription factors PHR1 and WRKY regulate the expression of P starvation genes; it is possible that these factors also control genes associated with P remobilization. In Arabidopsis, NAC family transcription factors such as NAP and ORE1 have been implicated in the early senescence process [55,56]. Whether ZmGPX-PDE1 is regulated by the homologous proteins ZmNAP and ZmORE1 needs to be explored.

5. Conclusions

In summary, we characterized two P deprivation-inducible GPX-PDE genes (ZmGPX-PDE1 and ZmGPX-PDE5) in maize. In vitro enzyme assays and yeast complementation experiments showed that ZmGPX-PDE1 and ZmGPX-PDE5 hydrolyse glycerophosphodiesters into G3P and corresponding alcohols with different substrates, indicating their functions in P recycling from glycerophosphodiesters. In particular, we found that ZmGPX-PDE1 is upregulated during leaf senescence. The compromised remobilization of P from senescing leaves to young leaves in the maize transgenic line with attenuated ZmGPX-PDE1 expression further indicates the role of ZmGPX-PDE1 in P recycling in plants. The results of the current study regarding the mechanism of P recycling via ZmGPX-PDE-dependent enzymatic degradation suggest new approaches with which to increase P utilization efficiency and optimize nutrient management for sustainable P use.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.05.004.

Declaration of competing interest

The authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by the National Key Research and Development Program of China(2017YFD0200204),the National Natural Science Foundation of China (31972496, 31572190),the Deutsche Forschungsgemeinschaft (328017493/GRK2366) and the National Institutes of Health Grant(R15 GM 104876)to Jana Patton-Vogt.We thank Prof.Chuanxiao Xie(Chinese Academy of Agricultural Sciences) for the supply of CPB-CRISPR/Cas9 construct.

Author contributions

Jingxin Wang and Lingyun Cheng designed the study. Jingxin Wang, Wenbo Pan, Alexiy Nikiforov, William King, Wanting Hong, Weiwei Li, and Yang Han performed experiments.Jingxin Wang, Jana Patton-Vogt, Jianbo Shen, and Lingyun Cheng analysed data. Jingxin Wang, Jana Patton-Vogt, and Lingyun Cheng wrote the manuscript.