Zhenhui Guo,Wenyun Run,Qingyu Wu, Yuhu Lyu, Keke Yi,*

aKey Laboratory of Plant Nutrition and Fertilizer, Ministry of Agriculture, Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences,Beijing 100081,China

bXinyang Academy of Agricultural Sciences,Xinyang 464000,Henan,China

Keywords:Astragalus sinicus Green manure Pi accumulation Vacuolar Pi transporter

ABSTRACT Astragalus sinicus is a commonly used legume green manure that fixes atmospheric N2 and accumulates mineral nutrients and organic substances that are beneficial to soils and subsequent crops. However, little is known about genotypic variation in, and molecular mechanisms of, Pi (phosphate) uptake and storage in A. sinicus. We recorded the morphological responses of six A. sinicus cultivars from four regions of China to external Pi application and measured their Pi accumulation. We identified full-length transcripts of Pi-signaling and Pi-homeostasis regulators by sequencing and measured the expression level of these genes by qRT-PCR.The major components in Pi signaling and Pi homeostasis were largely conserved between A. sinicus and the model species rice and Arabidopsis.Different A.sinicus varieties responded differently to low-phosphorus(P)stress,and their Pi accumulation was positively correlated with the expression of vacuolar Pi influx gene(SYG1/PHO81/XPR1-MAJOR FACILITATOR SUPERFAMILY (SPX-MFS)-TYPE PROTEIN) AsSPXMFS2 and negatively correlated with the expression of the vacuolar Pi efflux gene(VACUOLAR Pi EFFLUX TRANSPORTER) AsVPE1. We identified key Pi-signaling and Pihomeostasis regulators in A. sinicus. The expression of vacuolar Pi transporter genes could be used as an index to select A. sinicus accessions with high Pi accumulation.

1.Introduction

Phosphorus (P) is essential for nucleic acid, ATP, and biomembranes and functions in numerous biochemical pathways [1]. It is absorbed by plants from the soil in the form of phosphate (Pi). Owing to its low availability in soil, Pi is a major limiting factor for crop production [2-4]. To achieve a higher yield, farmers often apply excessive Pi fertilizers,resulting not only in depletion of phosphate rock but in eutrophication[5-7].

Green manure is an alternative for mineral fertilizers including Pi fertilizer and offers the advantage of maintaining a sustainable agricultural system. Astragalus sinicus L. (Chinese milk vetch), a winter-growing legume green manure, is widely grown in China, Japan, and other southeast Asian countries.It fixes atmospheric N2and absorbs nutrients from the soil. After being incorporated into the soil, it increases subsequent rice yield by increasing soil organic matter, total nitrogen content, and available P content [8]. However, low P availability in soil impairs the growth of A. sinicus and ultimately reduces its yield. It is thus desirable to identify accessions with tolerance to low-P stress and to identify the underlying Pi signaling mechanism.

Although the mechanisms by which A. sinicus uses Pi remain largely unknown, studies in model species such as Arabidopsis and rice shed some light. Plants absorb Pi mainly from the soil via roots[9,10],and PHOSPHATE TRANSPORTER 1(PHT1) family genes play an important role in this process[11,12].Thirteen and nine PHT1 genes have been identified in rice and Arabidopsis,respectively[13,14],and most of them are induced by Pi deficiency in roots and participate in Pi absorption and transport [15,16]. The expression of Pistarvation-induced PHT1 genes is dependent on PHOSPHATE STARVATION RESPONSE REGULATORs (PHRs), which are central regulators of Pi signaling [17-19]. The transcriptional activity of PHRs is negatively regulated by a class of SYG1/PHO81/XPR1 (SPX) domain-containing proteins (SPXs), which are also commonly used as marker genes for Pi deficiency[20,21]. Under low-P conditions, plants maintain Pi homeostasis by quickly releasing Pi from vacuole to cytoplasm[9,22].Two protein families, SPX and the MAJOR FACILITATOR SUPERFAMILY (MFS) domains contained family members(SPX-MFSs), and VACUOLAR Pi EFFLUX TRANSPORTERS(VPEs, a subfamily of glycerol-3-phosphate transporters(GLPTs)), play key roles in regulating respectively Pi influx and efflux across the plant tonoplast [23-26].

In the present study, we investigated variation in growth and Pi accumulation among six A. sinicus cultivars from several regions of China.

2.Materials and methods

2.1. Plant materials and growth conditions

Six A. sinicus cultivars from four regions of China were used:Xin Bai-1 (XB) and Xin Zi-1 (XZ) from Henan province, Yu Jiangdaye(YJ)from Jiangsu province,Yijiang Zi(YZ)and Wan Zi-1 (WZ) from Anhui province, and Xiang Fei-2 (XF) from Hunan province. For hydroponic culture, the seeds were immersed in concentrated sulfuric acid for 8 min [27]and then germinated in distilled water at 30 °C for 12-24 h. The germinated seeds were transferred to solutions described by Yoshida [28]with three concentrations of NaH2PO4(high P(HP), 200 μmol L?1Pi; moderate P (MP), 20 μmol L?1Pi; low P(LP), 2 μmol L?1Pi) and grown for 12 days. For qRT-PCR,germinated seeds were transferred to c omplete Yoshida's rice nutrient solution (+P, 200 μmol L?1Pi) for 7 days, and then placed under Pi deprivation (-P, 0 μmol L?1Pi) for 12 days. All plants were grown in a greenhouse with a longday photoperiod of 12 h light/12 h dark(～200 μmol m?2s?1)at 29 °C/22 °C with ～60% humidity. For soil culture, XZ and YJ were grown in paddy soil with two Pi levels: Pi-sufficient(Olsen P content of 31.5 mg kg?1and total P content of 325.13 mg kg?1) and Pi-deficient (Olsen P content of 10.1 mg kg?1and total P content of 280.21 mg kg?1) for three weeks in the greenhouse.

2.2. Determination of inorganic Pi contents

Fresh shoot and root samples were harvested and ground into powder in liquid nitrogen.H2SO4(500 μL,5 mol L?1)was added to a 2-mL Eppendorf tube containing 0.1 g ground tissue. The Pi contents were measured as described[29].

2.3. Anthocyanin measurement

Twelve-day-old seedlings grown in nutrient solution with three concentrations of NaH2PO4, (HP, 200 μmol L?1Pi; MP,20 μmol L?1Pi; LP, 2 μmol L?1Pi) were harvested for anthocyanin measurement. Anthocyanin contents were determined as described previously, with modifications [30].Briefly, frozen homogenized seedlings (80-100 mg) were extracted for 1 h at 37 °C in extraction buffer (methanol: HCl:water [18:1:81]), and purified with chloroform. After centrifugation, the supernatant was collected for measurement of absorbance at 535 nm. Relative anthocyanin amount was calculated as A535FW?1(fresh weight,mg).

2.4. Isolation of Pi-signaling and homeostasis-associated genes and construction of phylogenetic trees

For full-length RNA-seq analysis,RNA was extracted from 12-day-old seedlings of XZ,and the RNA was quality-checked on a BioAnalyzer (Agilent Technologies, Waldbronn, Germany).First, an Isoform Sequencing (Iso-Seq) library was prepared according to the Iso-Seq protocol using the Clontech SMARTer PCR cDNA Synthesis Kit (#634926; Clontech, http://www.clontech.com). For Single Molecule Real-Time (SMRT) sequencing, the PacBio Sequel platform (Pacific Biosciences,Menlo Park,CA,USA)was used.Next,the sequence data were processed using SMRTlink 5.1 software,and the redundancies in corrected consensus reads were removed with CD-HIT software [31]to leave 54,172 transcripts for the subsequent analysis. Finally, protein sequences of Cicer arietinum and other closely related species to run the ANGEL[32]pipeline to identify protein coding sequences from cDNAs of A. sinicus.The mRNA or protein sequences of IPS1s(Induced by Phosphate Starvation 1), SPXs, PHRs, PHT1s, SPX-MFSs, and GLPTs from several species were retrieved from NCBI (https://www.ncbi.nlm.nih.gov/) or PLAZA (http://bioinformatics.psb.ugent.be/plaza/). Phylogenetic analysis was performed with MEGA 7.0 software using neighbor joining with 1000 bootstrap replicates.

2.5. RNA isolation and qRT-PCR analysis

Total RNA was isolated using RNeasy Mini kits (TIANGEN BIOTECH,Beijing,China),and cDNA was synthesized with MMLV Reverse Transcriptase (Promega Corporation, Madison,WI, USA) according to the manufacturer's instructions.Quantitative real-time PCR (qRT-PCR) was performed with a SYBR Premix kit(Roche,Basel,Switzerland)on a Quant Studio 6 Flex machine (Life Technologies, New York, NY, USA). The As18sRNA gene was used as an internal control[33].The genespecific primers are listed in Table S1. Three biological replicates per gene were performed. Three technical replicates per gene within an experiment were performed.

NCBI accession numbers of genes used in qRT-PCR analysis are listed in Table S2.

2.6. Complementation test in yeast

To generate vectors for yeast complementation, the coding sequences (CDS) of AsVPE1, AsSPX-MFS1, AsSPX-MFS2,PHO84, and PHO91 were cloned into PRS426-ADH1 vector between BamHI and SacI sites. The primers used are listed in Table S1. These constructs and the empty vector were transformed into the yeast mutant strain YP100 (Δpho84 Δpho87 Δpho89 Δpho90 Δpho91 Δgit1)[38],which is defective in Pi transport and does not grow on YNB medium (lacking galactose). Transformants were selected on synthetic complete medium containing 2% galactose, 0.67% yeast nitrogen base without amino acids,0.2%appropriate amino acids,and 2% agar at pH 5.5 without uracil (SC-uracil) for 3-4 days at 28 °C. Three colonies were selected from each transformation and grown in liquid SC-uracil medium.Mid-exponential phase cells were harvested, washed twice with water, and resuspended to OD600= 1. Equal volumes of five-fold serial dilutions were spotted on Yeast Nitrogen Base (MP Biomedicals,Santa Ana,CA,USA)medium(without phosphate)with 2% glucose, and indicated Pi concentrations. Plates were incubated at 28 °C for 5 days.

2.7. Protoplast transient expression

To generate GFP-AsVPE1, GFP-AsGLPT1 and GFP-AsPT7, the CDS of these genes were amplified, and cloned into the pCAMBIA 1300-35s-GFP-N vectors [26]. Primer sequences are listed in Table S1.Rice seedlings were grown in basal nutrient solution for 13 days, and rice protoplasts were isolated and transfected as described[34].For confocal microscopy,images were taken with a Zeiss LSM 880 microscope (Carl Zeiss Microscopy GmbH, Jena, Germany) using 488 nm laser excitation and 500-550 nm emission for detection of the GFP(green fluorescent protein), and 516 nm laser excitation and 600-650 emission for detection of FM4-64.

3.Results

3.1. Astragalus sinicus cultivars responded differently to external Pi application

To select the cultivars that are more tolerant to low-P stress and study the underlying mechanisms, we analyzed physiological phenotypes of six A. sinicus cultivars under three Pi concentration conditions. Shoot and root biomass of YJ, YZ,WZ and XF were approximately two-fold greater than those of XB or XZ under HP or MP conditions (Fig. 1-a, b, c). The shoot biomass and root/shoot ratios of XZ were least affected by low-P stress among the cultivars (Fig. 1-d, e). In contrast, the root/shoot ratios of the other cultivars, especially YJ, were dramatically increased with the decrease in Pi supplementation (Fig. 1-e). Consistently, XZ shoots accumulated more Pi than others, whereas the Pi content of YJ was the lowest under the HP and MP conditions (Fig. 1-f, g). These results suggested that XZ was more tolerant than YJ to low-P stress.

The anthocyanin content, an indicator of Pi starvation in plants [35], of XZ was not dramatically affected by LP treatment compared with HP,whereas that of YJ was induced up to 21-fold by LP treatment (Figs. 2-a, b; S1). This result further confirmed that XZ was more tolerant than YJ to low-P stress in the hydroponic system.

When YJ and XZ were grown in paddy soil with two Pi treatments, Pi-sufficient (31.5 mg kg?1Pi) and Pi-deficient(10.1 mg kg?1Pi). In agreement with the hydroponic results,low-P stress significantly reduced the biomass of YJ, the sensitive cultivar,but not XZ,the tolerant cultivar(Fig.3-a,b).Similarly, the Pi content of XZ was significantly higher than that of YJ(Fig.3-c).Thus,XZ was more tolerant to low-P stress and accumulated more Pi than YJ under both hydroponic and soil culture conditions.

3.2.Identification of Pi-signaling and homeostasis genes of A.sinicus

Given that different A. sinicus cultivars, especially YJ and XZ,responded differently to external Pi, we next asked whether the variation was caused by their differences in Pi-starvation signaling components and Pi-homeostasis regulators. Given that the genome of A.sinicus has not been fully sequenced,we obtained full-length transcripts of XZ seedlings by thirdgeneration full-length transcriptome sequencing. Based on the homologous sequences of rice and Arabidopsis, we identified Pi-starvation signaling and Pi-homeostasis regulator genes of A. sinicus, named AsIPSs, AsSPXs, AsPHRs, AsPTs(PHT1 gene families),and AsSPX-MFSs(PHT5 gene families).By sequence alignment analysis,we identified two Pi-starvationinduced marker genes AsIPS1/2, two Myb domain-containing genes AsPHR1/2,five SPX domain-containing genes AsSPX1-5,and six Pi transporter genes AsPT2/4/5/6/7/8 (Figs.S2 and S3).AsPT2/4/5/6 corresponded to the previously reported AsPT1-6 that was isolated by RACE-PCR [36]. We also identified two vacuolar Pi influx regulators,AsSPX-MFS1/2(Fig.4-a).

Given that our previous study had shown that some of the glycerol-3-phosphate transporters (GLPTs) are bona fide vacuolar Pi efflux transporters (VPEs) [26], we tried to isolate VPEs of A. sinicus from the GLPT families. We first identified two candidate AsGLPTs: AsGLPT1 and AsGLPT2 (Fig. 4-b), but phylogenetic and protein sequence alignment results showed that only AsGLPT2 harbored the dileucine motif (PLL), a tonoplast localization signal [37](Fig. 4-b). We designate AsGLPT2 as AsVPE1 hereafter. We further studied the subcellular localization of AsVPE1 by fusing a GFP at its N-terminus and transiently expressed in rice protoplasts. Confocal microscopy showed that AsVPE1 was localized on the tonoplast,as expected; whereas AsGLPT1 was localized on the plasma membrane like AsPT7 (Fig. 5), indicating that AsVPE1 is a vacuolar Pi efflux transporter.

Fig.1- Astragalus sinicus cultivars responded differently to three Pi treatments.(a) Representative images of six A.sinicus cultivars grown hydroponically under three Pi concentrations(HP,200 μmol L?1 Pi;MP,20 μmol L?1 Pi;LP,2 μmol L?1 Pi)for 12 days.Scale bar,5 cm.(b,c)Shoot and root biomass of the cultivars.(d)Shoot relative biomass(shoot biomass under MP,LP vs.shoot biomass under HP).(e)Root-to-shoot ratios(root vs.shoot biomass)of the plants.(f,g)Pi concentrations in shoot and root.Experiments were performed in biological triplicate,and three plants were measured in each replicate.Values are mean±SE.ANOVA and LSD tests were used to test mean differences between cultivars.Different letters in (b-g) indicate significant differences at P <0.05 within each group.FW,fresh weight.

Fig.2-XZ is more tolerant than YJ to low-Pi stress.(a)Representative images of the leaves of YJ and XZ grown in a hydroponic system under three Pi treatments(HP,200 μmol L?1 Pi;MP,20 μmol L?1 Pi;LP,2 μmol L?1 Pi)for 12 days.Scale bars,2 mm.(b)Anthocyanin contents of the two cultivars under the three Pi treatments.Values are means±SE of three replicates.Asterisks and ns in(b)indicate significant difference or no significant difference compared with the HP condition(#)within each group by Student's t-test:*,**, and***indicate significant differences at P ≤ 0.05,P ≤ 0.01,and P ≤ 0.001,respectively.FW,fresh weight.

Thus, key Pi-signaling and homeostasis components are largely conserved between A. sinicus and the model species rice such as and Arabidopsis.

3.3. Differential expression of Pi influx/efflux transport genes results in different responses to low-P stress

To determine whether differential expression of Pi-signaling and Pi-homeostasis regulator genes accounted for the different Pi responses for YJ and XZ,we measured the expression of these genes in the two contrasting cultivars under three Pi concentration conditions.We classified these genes into three groups: The Pi-starved signaling group (AsIPS1, AsSPX2,AsSPX4, AsSPX5, and AsPHR1/2), the plasma membrane Pi transporter group (AsPT2/5/6/7), and the vacuole Pi transporter group (AsSPX-MFS1/2 and AsVPE1). The expression of neither the Pi-starved signaling group nor the plasma membrane Pi transporter group genes was correlated with the Pi accumulation of the two cultivars under the -P condition (Fig. S4). Although the expression of Pi influx vacuolar transporter genes AsSPX-MFS1/2 was positively correlated with Pi accumulation in YJ and XZ, the expression levels of the Pi efflux vacuolar transporter gene AsVPE1 were negatively correlated between the two accessions(Fig.6).The correlation was further confirmed in the other four cultivars(XB, YZ, WZ, and XF) (Figs. S5 and S6). These results suggest that the accumulation in XZ of higher Pi is due partially to its high vacuolar Pi influx but low efflux activity.

We further asked whether these predicted vacuolar transporters from A. sinicus have Pi-transport capability like their rice and Arabidopsis orthologs. The Pi-transport capability of these predicted vacuolar transporters from A.sinicus was tested by heterologous expression in the yeast mutant strain YP100 [38]. Both AsVPE1 and AsSPX-MFS1/2 rescued the hypersensitivity of YP100 to Pi treatment, confirming their Pi-transport activities (Fig. 7). Together, our results showed that different Pi influx and efflux activities of A. sinicus cultivars resulted in different Pi accumulations.

4. Discussion

4.1. Different varieties responded differently to low P stress

Phosphate deficiency is one of the major limiting factors in plant growth and development. To adapt to low-P stress,plants have evolved multiple mechanisms, which differ even within the same species. Most Pi nutrition differences represent genetic variation [39]. For example, variation in phytochrome B was a dominating factor responsible for the difference of Pi uptake among various Arabidopsis genotypes[40].The difference in P absorption efficiency between the two rice cultivars Kasalath and Nipponbare corresponds to the presence or absence of the kinase Phosphorus Starvation Tolerance 1 (PSTOL1) [41]. Similarly, we found that A. sinicus cultivars responded differently to Pi application (Fig. 1). XZ was the cultivar most tolerant to low-P stress, while YJ was the most sensitive.However,the underlying mechanism in A.sinicus might differ from that of Arabidopsis and rice.

Fig.3- XZ was more tolerant than YJ to Pi-deficiency when grown in soil. (a)Representative images of 3-week-old YJ and XZ seedlings grown in soil with two Pi treatments(Pi-sufficient, Olsen P: 31.5 mg kg?1;Pi-deficient,Olsen P:10.1 mg kg?1).Scale bar,5 cm.Shoot biomass(b)and Pi concentration(c)were determined.Values are means±SE of three replicates.ANOVA and LSD tests were used to test differences between XZ and YJ.Different letters in(b)indicate significant difference at P<0.05.FW,fresh weight.

Upon low-P stress,plants reshape their root architecture to absorb more P [42]. We found that some root traits, such as root length and root biomass, differed between YJ and XZ,with the root length of XZ greater than that of YJ under low-P stress(Figs.1,3,S7),whereas YJ had greater root biomass than that of XZ. The biomass and root/shoot ratios of XZ were not significantly affected by external Pi levels as in YJ.This result is consistent with one in Arabidopsis where accessions such as Calver attaining relatively high Pi content had long roots with longer root hairs than Tenela,which showed relatively low Pi content [43]. Accordingly, we suggest that changes in root traits can be used as indicators of plant sensitivity to low-P stress.

Fig. 4 - Phylogenetic analyses of SPX-MFSs and GLPTs protein families. (a) A phylogenetic tree of SPX-MFSs proteins from Astragalus sinicus (As), Oryza sativa (Os), and Arabidopsis thaliana (AT). (b) A phylogenetic tree of GLPTs proteins from 9 plant species. Letters in the codes represent species names as follows: As, A. sinicus; AT, A. thaliana; MT, Medicago truncatula; CA,Cicer arietinum; GM, Glycine max; ME, Manihot esculenta; Os, O. sativa; PT, Populus trichocarpa; ZM, Zea mays. The genes cloned in this study are indicated by points. These neighbor-joining trees were constructed with MEGA 7.0 software. The percentages of replicate trees in which the associated taxa clustered together in the bootstrap test (with 1000 replicates) are shown above the branches.

Fig.5-AsVPE1 was localized on the tonoplast while being expressed in rice protoplasts.AsVPE1,AsPT7,and AsGLPT1 fused with GFP at their N-terminus were transiently expressed in rice protoplasts.Green color indicates GFP and red color indicates FM4-64,a plasma membrane marker.BF,bright field.Scale bars,25 μm.

4.2. Key Pi-signaling and Pi-homeostasis regulators are conserved in multiple species

Key Pi-signaling, Pi-homeostasis regulators and their Pi regulation networks are conserved in the model plants rice and Arabidopsis [44-46]. In our study, most of the regulators PHRs,SPXs, PTs,SPXs-MFSs,VPEs,and IPSs could be found in A. sinicus based on sequence similarities with rice and Arabidopsis (Figs. 4, S2, S3). Our qRT-PCR results showed that the induction patterns of PTs under low-P stress were also largely similar in A.sinicus,rice,and Arabidopsis.For example,the expression of AsPT5/6 was significantly induced in both shoot and root by -P treatment (Fig. S4-c, d), as was that of AtPHT1;4, suggesting that they play important roles in Pi absorption and transfer from root to shoot. The significant induction of AsPT7 in root by -P treatment was similar to those of AtPHT1;8, AtPHT1;9, and OsPT9/10 (Fig. S4-c, d)[11,12,47-49]. The expression of AsVPE1 was induced by low-P stress (Fig. 6-c), consistent with the expression of OsVEP1/2 in rice [26]. Similarly, the repression of SPX-MFSs by low-P stress was conserved in A.sinicus,rice,and Arabidopsis[23,24](Fig. 6-a, b). These results suggest that the trans-regulatory elements of these genes may also be conserved between these species.However,we noticed some exceptions,such as PHRs.For example, the AtPHR1, AtPHL1 (PHOSPHATE STARVATION RESPONSE 1 LIKE PROTEIN),OsPHR1,OsPHR2 were not dramatically induced by low-P stress [17,18,50]; however, both AsPHR1 and AsPHR2 were significantly induced upon -P treatment (Fig. S4-a, b). These results suggest that the key components of Pi regulation networks are largely conserved,though with exceptions.

4.3. The activity of vacuolar Pi transporters is correlated with Pi accumulation in A.sinicus

Differentially expressed vacuolar Pi transporter genes resulted in different Pi accumulation among these accessions.Given that we confirmed their Pi-transport activities in yeast,it is reasonable to speculate that their expression levels could reflex transport activities.There may be other factors we have not found acting to redistribute Pi, resulting in differing expression of vacuolar Pi transporter genes, in the presence of similar Pi content between YJ and XZ under the LP condition. In summary, lower expression of AsVPE1, a vacuolar Pi efflux gene, and higher expression of a Pi influx gene (AsSPX-MFS2) in XZ resulted in more Pi accumulation in vacuoles.Thus,these genes could be developed as markers for selecting A. sinicus accessions with high Pi accumulation. For example,high expression of AsSPX-MFS2 or low expression of AsVPE1 in an A. sinicus accession may indicate high Pi accumulation. Correspondingly, prediction of Pi content for A.sinicus by measuring the expression levels of these genes is promising.

Fig. 6 - Relative expression of vacuolar Pi transporter genes in YJ and XZ under three Pi treatments. Relative expression of AsSPX-MFS1 (a), AsSPX-MFS2, (b) and AsVPE1 (c) in YJ and XZ under three Pi treatments. Plant shoots from 12-day-old seedlings grown under HP (200 μmol L?1 Pi), MP (20 μmol L?1 Pi), or LP (2 μmol L?1 Pi) conditions were used for qRT-PCR. Data are means ±SE of three biological replicates. The expression level of YJ under HP condition was defined as 1. * in (a)-(c) indicate significantly different at P < 0.05, and ns indicates not significantly different between YJ and XZ, with Student's t-test.

5. Conclusions

Astragalus sinicus cultivars responded differently to external Pi application,with XZ the cultivar most tolerant to low-P stress.Pi-signaling and Pi-homeostasis regulators were largely conserved among A.sinicus and other species,and the expression of vacuolar Pi transporter genes correlated well with the Pi accumulation in the cultivars. They could be developed as markers for selection of A.sinicus accessions with tolerance to low-P stress.

Fig.7- AsVPE1 and AsSPX-MFS1/2 showed Pi transporter activity in yeast.Heterologous expression of AsVPE1 and AsSPXMFS1/2 in yeast mutant YP100(Δpho84 Δpho87 Δpho89 Δpho90 Δpho91 Δgit1)complemented its hypersensitivity to external Pi.Equal volumes of five-fold serial dilutions were spotted on Yeast Nitrogen Base without phosphate(YNB-P,pH 5.5)medium with different Pi concentrations and incubated at 28°C for 5 days.

Declaration of competing interest

The authors declare no conflicts of interest.

Acknowledgments

This work was supported by the China Agriculture Research System - Green Manure (CARS-22) and the Innovation Program of Chinese Academy of Agricultural Sciences.

Author contributions

Zhenhui Guo performed experiments; Zhenhui Guo,Wenyuan Ruan,Qingyu Wu,Yuhu Lyu,and Keke Yi analyzed the data;Zhenhui Guo,Wenyuan Ruan,Qingyu Wu,and Keke Yi wrote the manuscript.

Appendix A. Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.05.005.