Centipedegrass seedlings of the acid-soil-resistant strain ‘E041’ were used for RACE (Rapid Amplification of cDNA Ends) and Real-time PCR analysis. The seedlings were collected from the Turfgrass Germplasm Resource nursery at the Institute of Botany, Chinese Academy of Sciences, Jiangsu Province. Potted seedlings of Arabidopsis thaliana ‘Col-0’ were used for protoplast transfection and plant transformation.

Genomic DNA was extracted from the newly expanded leaves using the DNAsecure Plant Kit (Tiangen Biotech, Beijing, China). Total RNA was extracted from various tissues or organs with the RNAprep Pure Plant Kit (Tiangen) with on-column DNaseI digestion. The nucleic acid concentration was quantified using an ND-1000 spectrophotometer (Nanodrop Technologies Inc., Rockland, DE, USA).

Based on the transcriptome sequencing data of centipedegrass, nested primers of 3′-RACE and 5′-RACE were designed for the EoPHR2 gene (Supplementary Table). The 3′-Full RACE Core Set Kit and 5′-Full RACE Kit (Takara, Dalian, China) were used to amplify the full-length sequences. The PCR products were separated on 1% agarose gels, cloned into the pMD19-T vector (Takara), and finally transformed into competent cells of Escherichia coli strain Top10. Colonies were checked by PCR, and inserts from positive colonies were sequenced. The full-length cDNA sequence of the EoPHR2 gene was obtained by aligning and comparing 3′-RACE and 5′-RACE products. The predicted ORFs were subsequently amplified by PCR and verified by sequencing.

Online BLAST software of The National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov) was used to analyze the DNA and protein sequences. The theoretical isoelectric point (pI) and mass value for the proteins were predicted and calculated using Expasy Protparam (http://web.expasy.org/protparam/). The functional domains of the protein were predicted using the SMART program (http://smart.embl-heidelberg.de/). Secondary structure of amino acid sequences was predicted by the SOPMA program (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html). The tertiary structure of deduced amino acid sequences was predicted using the SWISS MODEL (https://www.swissmodel.expasy.org/). The protein affinity and hydrophobicity were predicted using the ProtScale (https://web.expasy.org/protsca/). The protein signal peptide was predicted using the SignalP 4.1(http://www.cbs.dtu.dk/services/SignalP-4.1/). Phosphorylation site was predicted using the NetPhos (http://www.cbs.dtu.dk/services/NetPhos/). Protein transmembrane structure was predicted using the TMHMM 2.0. Multiple-sequence alignments of EoPHR2 proteins were carried out using the ClustalX 2.1 software [15]. The phylogenetic tree was constructed using MEGA6 software with the Neighbor-Joining method and 1000 bootstraps.

In this study, plasmids were constructed using the Gateway technology (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s protocol. The EoPHR2 coding region was cloned into the entry vector, pCR8/GW-TOPO (Invitrogen), via a simple T-A cloning reaction. An LR clonase enzyme mix (Invitrogen) was used to transfer the insert from the entry vector to its destination vector, p2GWF7.0 for C-terminal GFP fusion. The generated GFP fusion vector (35S:: EoPHR2-GFP) was a high copy vector driven by a double 35S cauliflower mosaic virus (CaMV) promoter with ampicillin as the bacterial selection marker. The plant protoplast preparation and transformation kit (Real-Times Biotechnology, Beijing, China) was used to isolate and transform Arabidopsis protoplasts. The protoplasts were observed using a BX53 microscope (Olympus, Tokyo, Japan).

Based on the centipedegrass hydroponic culture system [16], the hydroponic seedlings were pre-cultured for 2 weeks to the best growth state, and then treated using the “long-term Al-P alternate treatment” method [17]. Four treatment conditions, defined as ‘Control’ (−Al/+Pi), ‘P-deficiency’ (−Al/−Pi), ‘Al-toxicity’ (+Al/+Pi) and ‘Al-toxicity & P-deficiency’ (+Al/−Pi), were set up as follows:

Control (−Al/+Pi): A normal solution (1/2 Hoagland, in which the concentration of Pi was 500 μM, pH 4.0) and a Mock solution (0.5 mM CaCl₂ solution, pH 4.0) were set up as alternating treatments.

P-deficiency (−Al/−Pi): A P-deficient solution (1/2 Hoagland, the concentration of Pi was adjusted to 10 μM, pH 4.0) and Mock solution (0.5 mM CaCl₂ solution, pH 4.0) were set up as alternating treatments.

Al-toxicity (+Al/+Pi): A normal solution (1/2 Hoagland, in which the concentration of Pi was 500 μM, pH 4.0) and an Al-toxicity solution (0.5 mM CaCl₂ solution with 1.5 mM AlCl₃, pH 4.0) were set up as alternating treatments.

Al-toxicity & P-deficiency (+Al/−Pi): A P-deficient solution (1/2 Hoagland substrate, the concentration of Pi was adjusted to 10 μM, pH 4.0) and an Al-toxicity solution (0.5 mM CaCl₂ solution with 1.5 mM AlCl₃, pH 4.0) were set up as alternating treatments.

After 2 weeks of growing in the Normal nutrient solution, the seedlings were exposed to a 0.5 mM CaCl₂ solution either with (+Al) or without (−Al) AlCl₃ for 1 day. Thereafter, they were grown in the Normal (+Pi) or P-deficiency solution (-Pi) on alternate days. After 3 weeks of alternating treatments, the root, stem and leaf tissues were harvested (3 biological replicates). The fresh samples were first cleaned and then immediately frozen in liquid nitrogen.

Total RNA extraction from various tissues was performed as described above. The cDNA was synthesized from the total RNA using Oligo(dT)₁₈ with the Promega ImPromII Reverse Transcription System (Promega, Madison, WI, USA). For semi-quantitative RT-PCR, specific primers were designed for the EoPHR2 gene (Supplementary Table). Real-time RT-PCR was carried out in a 20 μl reaction mixture using the FastStart Universal SYBR Green Master Mix (Roche Applied Science, Indianapolis, IN, USA). Real-time PCR was performed on an ABI ViiA7 Real-time PCR System (ABI, Foster City, CA, USA). The cycling conditions were: 50°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min. All reactions were performed in triplicate. The relative quantitative method [18] was used for data analysis.

Plasmids used in this study were constructed by Gateway technology (Invitrogen). The insert was transferred from the entry vector to its destination vector, pBI121-3HA, with a C-terminal HA-tag. The generated GFP fusion vector (35S::EoPHR2-HA) was a high copy vector driven by overexpression elements [35S promoter from cauliflower mosaic virus (CaMV)]. Potted seedlings of Arabidopsis thaliana ‘Col-0’ were used for the transgenic experiments. The seedlings were transformed by the Agrobacterium-mediated inflorescence infiltration method. Then, the homozygous transgenic seedlings of the T3 generation were screened, and tested for their function with the methods as follows:

Potted seedlings culture and treatment: After germinated on the solid MS medium, the wild-type/transgenic seedlings were transplanted to pots. The culture substrate (peat: vermiculite: perlite = 1:1:1) pre-mixed with a base fertilizer of 1:10 diluted MS nutrient solution was no longer fertilized. Potted seedlings were grown under a cycle of 16 h light/8 h dark at 22°C/18°C (day/night) and irrigated with pure water for 3 weeks. Thereafter, the phenotypes were observed.

Aseptic seedlings culture and treatment: Based on the solid MS medium (pH 5.8; added sucrose 20 g/L, agar 9 g/L), the phosphorus concentration was controlled by adjusting the addition amount of KH₂PO₄ (KCl was added to supplement K to the standard level). Four treatment conditions were set up as follows: 0 mM (Pi-absent), 0.0625 mM (low-Pi), 0.625 mM (half-Pi), 1.25 mM (full-Pi). The seeds of the wild-type and 2 transgenic lines were first sterilized and then seeded on the medium containing any of the four treatments.. The phenotypes were observed 3 weeks later: (1) The growth phenotype were scanned by a scanner (WinRHIZO Pro 2017 system); (2) A total of 25 seedlings were weighed to evaluate the fresh weight from each of the wild-type or transgenic lines exposed to the diverse treatment conditions. Meanwhile, the seedlings were frozen in liquid nitrogen and stored at −80°C for RNA extraction and real-time RT-PCR analysis.

The full-length cDNA sequence of the EoPHR2 gene was cloned from E. ophiuroides ‘E041’ by RACE (Rapid amplification of cDNA ends) procedures. It was 1991 bp in length, containing an open reading frame (ORF) of 1287 bp, flanked by 246 bp of 5´-untranslated region (UTR) and 458 bp of 3´-UTR. The encoded polypeptide consisted of 429 amino acids, with a corresponding molecular weight (MW) of 46.93 kDa (Fig. 1).

The physical and chemical properties of the protein encoded by EoPHR2 were predicted by the ProtParam software. It was shown that the theoretical isoelectric point, fatty acid coefficient, average hydrophilicity and instability parameter were 4.99, 70.44, −0.634 and 55.58, respectively. The protein domain predicted by SMART showed a low complexity domain (222-243), a Myb_DNA-binding domain (251-302), and a Myb-CC_LHEQLE domain (334-380) with the conserved LHEQLE motif (Fig. 2A). The secondary structure analysis of the EoPHR2 protein, identified using the SOPMA program, revealed the following proportions of each component: 26.57% alpha helices, 6.06% extended strands, 2.33% beta turns and 65.03% random coils (Fig. 2B). Furthermore, the three-dimensional model of EoPHR2 constructed by the SWISS MODEL was consistent with the prediction of a secondary structure (Fig. 2C). In addition, phosphorylation site analysis of the EoPHR2 protein, predicted using the DISPHOS program, showed that 59.615% of serine sites, 36.364% of threonine sites and 50.000% of tyrosine sites were phosphorylated (Fig. 2D).

According to the protein BLAST software, the predicted EoPHR2 protein belonged to the Myb-CC family with the conserved LHEQLE motif. The various alignments of the amino acid sequences were performed with the homologous proteins from different plant species using the ClustalX 2.1 software (GeneBank: Eremochloa ophiuroides, EoPHR2; Sorghum bicolor, XP_021309376.1; Zea mays, XP_008651544.1; Panicum miliaceum, RLN35937.1; Panicum hallii var. hallii, PUZ69446.1; Setariaitalica, XP_004956084.1; Dichanthe- liumoligosanthes, OEL30463.1; Eragrostiscurvula, TVU39510.1; Aegilops tauschiisubsp.Tauschii, XP_020191223.1; Triticum turgidum subsp. Durum, VAN57702.1; Hordeum vulgare, KAE8802583.1; Brachypodium distachyon, XP_003563187.1; Oryza meyeriana var . Granulata, KAF0909536.1; Oryza sativa Japonica Group, XP_015647735.1, OsPHR2; Oryza brachyantha, XP_006657651.1). The multiple alignments of the homologous proteins ranged, in percentage identity to EoPHR2 (Eremochloa ophiuroides), from 92.09% (Sorghum bicolor) to 71.89% (Oryza brachyantha), and revealed a conserved Myb-CC_LHEQLE domain in the C-terminus (Fig. 3).

To understand the evolutionary relationships of the EoPHR2 proteins from different species, the amino acid sequences of 15 EoPHR2 proteins were aligned, and the phylogenetic tree of the homologous proteins was constructed by the Neighbor-Joining method using the MEGA6 software with bootstrap testing on 1000 replicates. The results showed the evolutionary relationships among the 15 proteins from different species (Fig. 4). The shortest evolutionary distance from E. ophiuroides was Sorghum bicolor, another member of the Panicoideae. The longest evolutionary distance from E. ophiuroides was Oryza brachyantha in the Oryzoideae.

Subcellular localization information is one of the key features of protein function research. The amino acid sequence of the EoPHR2 protein was predicted and analyzed by SignalP 4.1. The results showed that the average S-score of the EoPHR2 protein was 0.103 and no signal peptide sequence was found. This indicated that the protein was a non-secretory protein. Further, the cellular localization of the EoPHR2 protein was probed using the Arabidopsis protoplast transient expression system. Using fluorescence microscopy, the EoPHR2-GFP fusion protein was detected in the nucleus in protoplasts, while the fluorescence signal of 35::GFP fusion protein had no clear cellular localization as the positive control (Fig. 5). The observed subcellular localization of the EoPHR2 protein was consistent with its functional localization as a transcription factor.

To verify the expression pattern of EoPHR2 in E. ophiuroides, the expression levels of the EoPHR2 gene were measured in diverse tissues under different Al-P conditions by real-time PCR. It was found that under the ‘Control’ condition, EoPHR2 mainly expressed in root (Fig. 6A). Under the P-deficiency stress, the expression of EoPHR2 appeared no significantly differ in root, stem or leaf tissue, relative to the Control condition (Fig. 6B). These results showed that the EoPHR2 gene was mainly expressed in the root and would not be significantly induced by the P-deficiency stress, which confirmed the conclusions drawn from the homologous genes of other plant species [3,5,10].

On the other hand, interestingly, the expression level of EoPHR2 could be significantly up-regulated by Al-toxicity stress or Al-toxicity & P-deficiency multiple-stress in root, stem and leaf tissues, among which the gene expression was induced strongly in root. In addition, in root and leaf tissues, the EoPHR2 gene showed a stronger up-regulation under the multiple-stress of Al-toxicity & P-deficiency than the single stress of Al-toxicity (Fig. 6B). The results suggest a potential role of EoPHR2 in the Al-toxicity stress response of centipedegrass.

Transgenic Arabidopsis plants overexpressing the EoPHR2 gene were generated to analyze the function of EoPHR2 in vivo. The full-length coding sequence (CDS) region of the EoPHR2 gene was amplified from cDNAs, subcloned into a modified pBI121-HA overexpression vector, and transferred into Arabidopsis using an Agrobacterium-mediated transformation method. After a period of iteratively screen, two homozygous transgenic lines were confirmed for the further experiments, termed EoPHR2-OV1 and EoPHR2-OV2 (Supplementary Figure).

In the condition of potted culture and treatment, growth of the 2 transgenic Arabidopsis lines was observed better than that of the wild-type (Fig. 7A). Then, the seeds of wild-type (WT) and 2 transgenic (EoPHR2-OV1, EoPHR2-OV2) lines were sterilized and sown on the solid medium with inorganic phosphorus concentration of 0 mM (Pi-absent), 0.0625 mM (low-Pi), 0.625 mM (half-Pi) or 1.25 mM (full-Pi), respectively. Three weeks later, the phenotype was observed and the fresh-weight was determined. In the extremely Pi-absent condition (0 mM Pi), the wild-type and transgenic lines showed obvious growth inhibition as a result of absence of the essential macronutrient. Then, in the condition of low-Pi (0.0625 mM Pi), the transgenic plants began to show advantages in vegetative growth compared with the wild-type; the fresh weight of transgenic plants was slightly higher than that of the wild-type. Moreover, the transgenic plants appeared to have more developed lateral roots. Further, in the condition with the half-Pi concentration of 0.625 mM, the transgenic plants showed obvious growth advantages compared with the wild-type; the two transgenic lines exhibited higher fresh weight with better developed root systems than the wild-type. In addition, under the full-Pi condition (1.25 mM Pi), there was no significant difference in plant fresh weight between the transgenic and the wild-type plants. However, the root systems of the transgenic lines were still observed superior to that in the wild-type (Figs. 7B and 7C).

Previous studies in the model plants have revealed that AtPHR1/OsPHR2 is a key regulator in the pathway responding to phosphorus signals [3,5,10]. Overexpression of EoPHR2 should be able to mimic Pi-starvation signalling and activate the transcription of the Pi-starvation-induced PHT1 genes in transgenetic Arabidopsis plants. The expression levels of PHT1 genes were measured by real-time RT-PCR to verify the mode of action of EoPHR2. According to previous studies [19], nine Arabidopsis PHT1 genes were found, and primers were designed to quantify the level of expression of each AtPHT1 gene. Under the full-Pi condition (1.25 mM Pi), the expression levels of most PHT1 genes were up-regulated in the transgenic plants without Pi-starvation stress, relative to the wild type. Overexpression of EoPHR2 could induce the expression of eight AtPHT1 genes apart from the AtPHT1-9 (Fig. 8, Fig. S2). As members of an important phosphate transporter famliy, PHT1 genes have frequently been used to study phosphate starvation responses in plants. The induction of PHT1 genes in transgenic plants suggests a potential in vivo role for EoPHR2 in the Pi-signaling pathway and low-phosphorus stress resistance.