A Panax notoginseng phosphate transporter, PnPht1;3, greatly contributes to phosphate and arsenate uptake

Keywords: arsenate (AsV) exposure, arsenic (As) acquisition and accumulation, heterologous expression, molecular mechanism, Panax notoginseng, phosphate (Pi) transporter, Pi deficiency, relative expression level.

Panax notoginseng (Burk.) F.H. Chen is a commonly used traditional Chinese herb, of which the rhizome is a historically and officially utilised medicinal part that has been used for hundreds of years in clinical treatment (Kim 2018; Xiong et al. 2019). The primary bioactive constituents are saponins, mainly consisting of ginsenosides Rb1 and Rg1 and notoginsenoside R1. All are classified as dammarane-type tetracyclic triterpenoids (Zu et al. 2018). Evidence showed that notoginsenoside has therapeutic effects on dissipating blood stasis and arresting bleeding, thereby promoting the subsidence of swelling and relieving pain (Ou et al. 2016). Unfortunately, arsenic (As) contamination increases the drug-use risk of P. notoginseng, mainly caused by high As background concentrations in the soil and wide-scale use of As-containing pesticides (Zhu et al. 2017). Investigations found that total As concentrations in organs of P. notoginseng and their preparation planted in Wenshan Autonomous Prefecture occasionally exceeded the threshold value of 2.0 mg/kg listed in the As standard of China Green Trade Standards of Importing and Exporting Medicinal Plants and Preparation, and the concentrations in excess of the maximum permissible level were as high as 24%, 81% 14%, 57%, 44% and 56% in rhizomes, fibrous roots, stems, leaves, seeds and preparation, respectively, causing a serious health risk (Yan et al. 2011; Lin et al. 2013). As a highly toxic heavy metal, arsenic is listed at the top of the world’s top 10 most toxic hazardous substances by the ‘Agency for Toxic Substances and Disease Registry’ (ATSDR) (Li et al. 2008; Jedynak et al. 2010), which could pose a serious threat to human health through the food chain. Thus, it would be a major health threat if P. notoginseng was utilised in clinical treatment.

Due to similar chemical characteristics and competition relationships between phosphate (Pi) and arsenate (AsV), supplementation with Pi suppresses the As uptake in plants (Woolson et al. 1973; Cao et al. 2003; Sun et al. 2020). This theory has been proved in Pteris vittata L. (Huang et al. 2007; Lei et al. 2012) and Pennisetum clandestinum Hochst (Panuccio et al. 2012), where phosphorus (P) addition reduced As translocation and accumulation. As a macronutrient, P is involved in some key biological functions in plant growth; e.g. structural cell components, energy transfer and photosynthesis (Knudson et al. 2003; Rufyikiri et al. 2006). Soil P in various forms and organic and inorganic weathering are taken up by Pi transporters (Pht) in the plants, which are usually driven by the generation of a proton gradient (Shen et al. 2011; Qin et al. 2012a; Doki et al. 2013; Li et al. 2019; Wang et al. 2019).

Pi transporters are usually categorised into high- and low-affinity types based on Km values. High-affinity Pi transporters with a Km value of micromoles are usually induced by Pi deficiency (López-Arredondo et al. 2014; Wang et al. 2019), while low-affinity Pi transporters are constitutive proteins with a millimolar Km value (López-Arredondo et al. 2014). Evidence suggested that plant Pi transporters consist of Pht1, Pht2, Pht3, Pht4 and PHO1 (PHOSPHATE 1), the first four of which were reported to localise in the plasma membrane, plastid inner membrane, mitochondrial inner membrane and Golgi compartment, respectively (Mimura 1999; Rausch et al. 2001; Liu et al. 2011). Numerous studies found that Pht-type proteins (Pht1, Pht2, Pht3 and Pht4) are mainly expressed in the epidermis and cortical tissue of the root hair zone under the stress of Pi deficiency and contribute to Pi uptake (Shin et al. 2004; Qin et al. 2012b). To date, many plant Pi transporters have been implicated in the uptake of Pi and AsV; e.g. PvPht1;3 from P. vittata (Ditusa et al. 2016), PHT1;3, PHT1;4 and PHT1;12 from Salix spp. (Puckett et al. 2012). Moreover, it is a common way of tolerance employed for As resistance by suppressing Pi/AsV uptake on the basis of the interplay between AsV uptake and Pi nutrition (Puckett et al. 2012).

Currently, the roles of P. notoginseng Pi transporters in the uptake of Pi and AsV are still unclear under the stress of Pi deficiency and As exposure. In this work, we identified a P. notoginseng Pi transporter-encoding gene, PnPht1;3, and uncovered the expression characteristics at the transcriptional level. In addition, an ideal approach to uncover the mechanism of Pi/AsV uptake is to employ mutant Saccharomyces cerevisiae MB192, which significantly reveals this uptake with Pi–AsV interplay. Our work helps to illustrate the role of P. notoginseng Pi transporters in the uptake of Pi and AsV; therefore, it will be useful for decreasing As accumulation in the rhizome.

One-year-old Panax notoginseng (Burk.) F.H. Chen seedlings with non-significant in biomass (fresh weight, leaf number and plant height) were selected as treatment materials, and grew well at a standard garden Wenshan Prefecture Yunnan Province. These P. notoginseng were authenticated by professor Rong-hua Zhao (Yunnan University of Chinese Medicine). Growing media in garden pots consisted of 10% silt, 20% light aggregate, 30% clay and 40% expanded vermiculite (Mandal et al. 2015), in which dissolved phosphorus content was cut down to an extremely low level by washing with 1% NaHCO₃. Then, dissolved phosphorus concentrations were adjusted to 0.07 mM (low P, lP), 0.7 mM (medium P, mP) and 1.4 mM (high P, hP) by adding KH₂PO₄ (Cao et al. 2020). Before planting, sodium arsenate (Na₃AsO₄) was thoroughly mixed into media with a 0.2 mM final concentration in dry weight for As treatment groups. The As concentration is in line with the background value of soil in Wenshan Prefecture Yunnan Province, where is the main producing area (Feng et al. 2005). In this experiment, there were six treatments in total, including lPnAs (low P, no As), lPhAs (low P, high As), hPnAs (high P, no As), hPhAs (high P, high As), mPhAs (med P, high As) and mPnAs (med P, no As; CK). Due to the absence of mineral nutrition caused by repeatedly media-washed, 50 mL 1/4 Hoagland solution without P was added to nourish P. notoginseng regularly. The growing environment of P. notoginseng was extremely strict in the greenhouse, keeping 25°C and 85% relative humidity, and protecting against direct sunlight and water-accumulated. After 5 months, the fresh fibrous roots of P. notoginseng were collected for RNA extraction. Eight biological replicates were included in each treatment, and four samples were in every replicate.

Open reading frame (ORF) sequence of PnPht1;3 was first obtained from a transcript of P. notoginseng fibrous roots treated as above described. The primer pair of PnPht1;3 for ORF amplification were listed in Table 1. The first-strand cDNA was used as a template, which was gained by reverse transcription from total RNA with Primescript II 1st strand cDNA synthesis kit (Takara, Japan). Total RNA was extracted from the fibrous roots according to the protocol of miniBEST plant RNA extraction kit (TaKaRa, Japan). The PCR programme for the PnPht1;3 consisted of initial denaturation step (94°C/5 min), followed by 35 cycles of 94°C/1 min, 58°C/30 s, 72°C/1 min, holding at 4°C.

Table 1.  Primer pairs used for ORF cloning, qPCR, and recombinant plasmid construction.

The relative expression levels of PnPht1;3 in different treatment groups were validated by quantitative real-time polymerase chain reaction (qPCR) with TB Green Premix Ex Taq (Tli RNaseH Plus), ROX plus (TaKaRa, Japan). Primer pairs of PnPht1;3 and reference gene (26S-2) (Wu et al. 2015) were also listed in the Table 1. The qPCR result of PnPht1;3 was calculated using the formula 2^−ΔΔCt (Livak and Schmittgen 2001).

The ORF of PnPht1;3 was authenticated via online software at https://www.ncbi.nlm.nih.gov/orffinder/. Biological information, including the location of hydrophobic, isoelectric point, protein molecular weight and putative transmembrane domains were predicted with the software package mounted at http://expasy.org/tools/protscale.html. The prediction of subcellular localisaiton was performed at http://wolfpsort.hgc.jp/. Multiple peptide alignments were analysed with DNAMAN v6.0. Phylogenetic tree was constructed using MEGA ver. 7.0 software with a parameter of 1000 bootstrap replications. The sequence of PnPht1;3 has been submitted to NCBI and assigned a GenBank accession number MT406774.

Mutant yeast, Saccharomyces cerevisiae MB192 (MATa pho3-1 pho84::HIS3 ade2 leu2-3, 112 his3-532, trp1-289 ura3-1, 2 can1) was used as a heterologous expression tool for uptake-functional verification, of which the high-affinity Pi transporter gene PHO84 is knocked-out, and inserted an HIS3 DNA fragment in situ (Bun-Ya et al. 1991; Qin et al. 2012a). The full-length of PnPht1;3 was amplified with primers containing restriction enzyme cutting sites are listed in Table 1. The amplicon was digested with BamHI/Kpn I, and then attached to the expression vector YEplac112 digested with corresponding enzymes using T4 DNA Ligase (NEB, USA). The recombinant plasmid and empty vector YEplac112 were transformed into S. cerevisiae MB192 by electro-transformation (Costaglioli et al. 1994). Two types recombinant yeasts were obtained, and named MB192-PnPht1;3 and MB192-YEplac112.

For the effect of Pi concentrations on the growth of yeasts, monoclonal cells were first cultured to the logarithmic phase (OD₆₀₀ = 0.6) in the YNB liquid medium. Then, 100 μL suspension liquid were diluted to 5 mL by adding YNB medium, and cultured at 200 rpm and 30°C. The medium were adjusted to different Pi concentrations, including 0.002, 0.02, 0.06, and 0.1 mM with the same initial pH of 6.8 (Liu et al. 2014b). Bromocresol purple was used as a pH indicator in the medium, which shift a colour change from light yellow to purplish-red within a pH range of 5.2–6.8. The change of pH was affected by the concentration of cells and acid phosphatase (ACP) activity (Jia et al. 2011). For pH-dependent Pi uptake experiments, identified cells were cultured in the YNB liquid medium containing 80 μM Pi for 24 h at 200 rpm and 30°C, and the initial pH values were adjusted to 4, 5, 6, 7 and 8, respectively. For the determination of growth curve and growth rate, OD₆₀₀ of yeast cells was determined every 3 or 5 h culturing in 5 mL SD-Trp^- medium at 200 rpm and 30°C, of that the medium contained 20 μM Pi and 2% glucose with a pH value of 6 and an initial OD₆₀₀ of 0.03 (Ditusa et al. 2016). Growth rate coefficients of logarithmic phase of yeast cells were calculated via exponential regression.

The 100 μL yeast suspension (OD₆₀₀ = 0.6) was diluted into 5 mL SD-Trp^- medium containing 2% glucose and 80 μM KH₂PO₄, and adjusted an initial pH to 6.0, with or without carbonyl cyanide m-chlorophenylhydrazone (CCCP) (10, 50 μM) or 2,4-dinitrophenol (2,4-DNP) (100, 200 μM) (Mitsukawa et al. 1997; Zhang et al. 2014). CCCP should be initially dissolved in ethanol, and then added into the medium with a final concentration of 0.01% (v/v) (Li et al. 2008). Growth density was measured after shake cultivation for 20 h at 200 rpm, 30°C.

For the growth rate and As tolerance of yeast cells, exponential yeast suspension of MB192-PnPht1;3 and MB192-YEplac112 were inoculated into 10 mL SD-Trp^- medium containing 50 μM Pi and 2% glucose, respectively, that made an initial OD₆₀₀ to 0.03. Then, AsV was mixed into the medium with a final concentration of 80 μM before culturing at 200 rpm and 30°C. The OD₆₀₀ of transformants was determined every 3 or 5 h for analysing growth rate coefficients and AsV tolerance during logarithmic phase (Ditusa et al. 2016). To investigate the effect of Pi concentrations on the As uptake of transformants, the OD₆₀₀ and total As concentrations of MB192-PnPht1;3, MB192-YEplac112, mutant MB192 and wild type (WT) yeast cells were determined. Firstly, 1 mL yeast suspension (OD₆₀₀ = 0.6) were transferred into 50 mL SD-Trp^- medium containing 2% glucose, 20/100 μM Pi and 80 μM AsV, and then adjusted their initial pH to 6.0. Determination of OD₆₀₀ of yeast suspension was carried out after shake cultivation at 200 rpm and 30°C for 30 h. Then, yeast cells were collected by centrifuging at 5000 rpm for 5 min, and the cell pellets were washed twice with 25 mL 10 mM EDTA for removing adhered As (Gravot et al. 2004). The total As concentrations in cells were determined via inductively coupled plasma mass spectrometry (ICP-MS) (Agilent 7500c, USA) as described by Wu et al. (2011) and Xu et al. (2017), in which yeast samples were ground to fine powder with liquid nitrogen and digested with HNO₃:H₂O₂ (85:15, v/v). Data used for analysing were performed with four biological replicates, and three technical replications in each biological replicate were conducted independently.

All data collected were processed and analysed statistically with SPSS 17.0 and Sigmaplot 12.0. The method of significant difference was the Turkey HSD tests of one-way analysis of variance (ANOVA) at the level of 0.05, that was conducted based on the assumptions of normality and homogeneity of variances, including qPCR result, ACP activity, OD₆₀₀, As concentration, growth rate coefficient and AsV tolerance. An independent-samples t-test at the 0.05 or 0.01 level was also employed to analyse the difference, including OD₆₀₀ between each treatment (CCCP or 2,4-DNP) and CK, OD₆₀₀ or As concentrations between 20 μM and 100 μM Pi. Data used in figures were expressed as the means ± standard deviation (s.d., n ≥ 3).

The ORF length of PnPht1;3 cDNA is 1605 bp; and the predicted translation product consists of 535 amino acids with a calculated molecular mass and isoelectric point of 59.04 kDa and 8.87, respectively. Transporter PnPht1;3 contains 10 transmembrane domains and a Pht1 signature sequence (GGDYPLSATIxSE) (Karandashov and Bucher 2005) in a red-line box, as shown in Fig. 1. Peptide sequence alignment of the PnPht1;3 and other Pi transporters revealed that PnPht1;3 shared 84.83%, 84.83%, and 70.24% identity with NtPht1 (GenBank accession number, AF156696.1), CmPht1 (QAU07451.1), and PvPht1;3 (KM192137.1), respectively (Fig. 1).

Fig. 1.  Alignment of PnPht1;3 and the peptide sequences of known Pi transporters present in Nicotiana tabacum L. (NtPht1, AF156696.1), Castanea mollissima BL. (CmPht1, QAU07451.1) and P. vittata (PvPht1;3, KM192137.1). Identical peptides are highlighted in black, and conservative substitutions are highlighted in pink. Predicted transmembrane domains of PnPht1;3 were marked by underlines. The signature sequence of Pht1 is shown in a red box.

To investigate the evolutionary relationship of PnPht1;3 with other homologues, a phylogenetic analysis was performed using the MEGA ver. 4.0 programme (Fig. 2). The results demonstrate that PnPht1;3, PvPht1;3 and CmPht1 belong to the Pht1 subfamily. In addition, subcellular localisation prediction showed that PnPht1;3 located in the plasma membrane. These results demonstrate that PnPht1;3 may be closely related to the uptake of Pi in P. notoginseng.

Fig. 2.  Phylogenetic relationship of PnPht1;3 and other plant Pi transporters. CmPht1 (QAU07451.1) from C. mollissima; PvPht1;3 (KM192137.1) from P. vittata; ZjPht2;1 (XP_015894666.1) from Ziziphus jujuba Mill.; VvPht2;1 (XP_002271876.1) from Vitis vinifera L.; TaPht2;1 (AY293827.1) from Triticum aestivum L.; NnPht2;1 (XM_010250335.2) from Nelumbo nucifera Gaertn.; AtPHO1;1 (NM_113246.5) from Arabidopsis thaliana L.; DcPHO1;2 (XM_017360779.1) from Daucus carota var. sativus Hoffm.; GhPHO1 (XM_016884382.1) from Gossypium hirsutum L.; DzPHO1 (XM_022904580.1) from Durio zibethinus Merr. The bootstrap value was calculated with 1000 replications.

In Fig. 3, it was found that PnPht1;3 significantly responded to Pi deficiency, and the fold change in expression became larger with AsV supplementation in the low-concentration Pi treatment groups; e.g. 24.8-fold increase with lPnAs and 81.7-fold increase with lPhAs. In addition, compared with low-concentration Pi (lP) treatment, the expression level of PnPht1;3 sharply decreased by supplementation with sufficient phosphate and presented a significant difference; e.g. a 24.8-fold increase with lPnAs, a 1-fold increase with mPnAs (CK), and a 0.9-fold decrease with hPnAs. The fold changes of mPhAs and hPhAs were 0.5 and 2.2, respectively, and were significantly less than 81.7 for lPhA, indicating that a high-concentration of Pi also weakened the stress of AsV. These results suggest that transporter PnPht1;3 may be involved in the uptake of Pi and AsV, and the supplementation with a high-concentration of Pi may decrease the stress caused by Pi deficiency or AsV exposure.

Fig. 3.  Relative expression levels of PnPht1;3 in fibrous roots of P. notoginseng under the stress of Pi deficiency and AsV exposure. One-year-old P. notoginseng plants in good condition were treated with different phosphate concentrations [(KH₂PO₄), 0.07 mM (lP), 0.7 mM (mP), and 1.4 mM (hP)] and supplemented with or without 0.20 mM AsV (Na₃AsO₄). The treatment of mPnAs (0.7 mM Pi and non-AsV) was set as a control group. Different lowercase letters represent difference of PnPht1;3 among different treatments, P ≤ 0.05. Error bars indicate mean values ± s.d. (n = 4).

Heterologous expression of PnPht1;3 in mutant yeast MB192 supported its survival at low concentrations of Pi (0.002 mM and 0.02 mM) by improving Pi uptake. The OD₆₀₀ of strain MB192-PnPht1;3 was remarkably higher than that of MB192 and MB192-YEplac112, particularly at low concentrations of Pi, close to that of the WT type. The logarithmic phase of MB192-PnPht1;3 was 10–25 h (Fig. 4a). The colour of medium-cultured MB192-PnPht1;3 and WT cells was yellow at 0.002 mM, 0.02 mM, and 0.06 mM Pi, while MB192 and MB192-YEplac112 cells were purple or faint yellow under low-Pi conditions (0.002 mM, 0.02 mM, and 0.06 mM) (Fig. 4b). This change in medium colour is in accordance with the pH value and ACP activity of yeast cells. As shown in Fig. 4c, the ACP activity of MB192-PnPht1;3 was higher than that of MB192-YEplac112 and MB192 and presented a significant difference (P < 0.05). The optimal pH value for MB192-PnPht1;3 and WT was 6 (Fig. 4d). In addition, the OD₆₀₀ of MB192-PnPht1;3 cells was significantly suppressed by supplementation with the respiratory inhibitors CCCP and 2,4-DNP. This phenomenon became more obvious with increasing respiratory inhibitor concentrations (Fig. 4e). These results confirmed that Pi transporter PnPht1;3 is a putative high-affinity H⁺/H₂PO₄⁻ symporter that mediates Pi uptake through a proton gradient generated on the membrane.

Fig. 4.  Complementation assays of MB192-PnPht1;3. (a) Growth curves of WT, MB192, MB192-YEplac112 and MB192-PnPht1;3 cultured for 40 h in the presence of low Pi (20 μM). (b) Medium colour changed with pH under different Pi concentrations. (c) ACP activity of of WT, MB192, MB192-YEplac112 and MB192-PnPht1;3 in the presence of low Pi (20 μM) with an initial pH of 6. Different lowercase letters represent the difference of ACP activity among cells, P ≤ 0.05. (d) The effect of different medium pH on the growth of WT, MB192, MB192-YEplac112 and MB192-PnPht1;3 supplemented with 100 μM Pi. Different lowercase letters represent the difference of OD₆₀₀ under the same pH among cells, P ≤ 0.05. (e) The growth of MB192-PnPht1;3 was suppressed by protonophores, carbonyl cyanide m-chlorophenylhydrazone (CCCP) and 2,4-dinitrophenol (2,4-DNP). **P ≤ 0.01; *P ≤ 0.05. Error bars indicate mean values ± s.d. (n = 4).

In addition, the growth rate coefficients of MB192-PnPht1;3 and MB192-YEplac112 were calculated via exponential regression, which was based on the logarithmic phase, as shown in Fig. 5a, b. A number of independently obtained growth rate coefficients (n = 4) for each transporter are shown in Fig. 5c. These results showed that under low-Pi conditions (20 μM), the growth rate coefficient of MB192-PnPht1;3 was significantly higher than that of MB192-YEplac112, suggesting Pi transporter PnPht1;3 performed optimally in complementing the yeast Pi-transport defect.

Fig. 5.  Growth rate of MB192-PnPht1;3 under a low-concentration Pi (20 μM). OD₆₀₀ of logarithmic growth phase were used to generate exponential trend lines (y(t) = a × e^kt), where k is the growth rate coefficient. (a) Growth rate of MB192-PnPht1;3. (b) Growth rate of MB192-YEplac112. (c) Growth rate coefficient. * represents a difference of growth rate coefficient between MB192-PnPht1;3 and MB192-YEplac112 at P ≤ 0.05. Error bars indicate mean values ± s.d. (n = 4).

As a common carrier, Pi transporters play important roles in the acquisition of Pi and AsV. Competitive inhibition revealed that high-concentration Pi could suppress AsV acquisition. As shown in Fig. 6, the growth curves of MB192-PnPht1;3 and MB192-YEplac112 seemed similar in 50 μM Pi medium without AsV. Correspondingly, the growth rate coefficients of MB192-PnPht1;3 and MB192-YEplac112 under 50 μM Pi were higher than the values determined under low-Pi (25 μM) described in the previous section. However, the growth of the two types of yeast cells was completed suppressed by supplementation with 80 μM AsV, of which the growth rate coefficients sharply decreased, e.g. 0.1588/0.0916 for MB192-PnPht1;3 and 0.1156/0.0562 for MB192-YEplac112 (Fig. 6a, b). In contrast, the growth curve of MB192-YEplac112 was gentler than that of MB192-PnPht1;3 under AsV stress. The As tolerance of MB192-PnPht1;3 was assessed by calculating the percentage of growth under As exposure relative to growth in the absence of As. The results revealed that MB192-PnPht1;3 had slightly stronger As tolerance than MB192-YEplac112. However, there was a non-significant difference between the two transgenic lines (Fig. 6c).

Fig. 6.  Growth rates and As tolerance of MB192-PnPht1;3 supplemented with 50 μM Pi and non-(−As) or 80 μM AsV. OD₆₀₀ of logarithmic growth phase were used to generate exponential trend lines (y(t) = a × e^kt), where k is the growth rate coefficient. (a) Growth rates of MB192-PnPht1;3 supplement with (+As) or without (−As) 80 μM AsV. (b) Growth rates of MB192-YEplac112 supplement with (+As) or without (−As) 80 μM AsV. (c) As tolerance (%) indicates the proportional growth rate of MB192-PnPht1;3 and MB192-YEplac112 in the presence of AsV compared with control growth. A non-significant difference of As tolerance was shown between MB192-PnPht1;3 and MB192-YEplac112, P > 0.05. Values are the mean ± s.d. (n = 4).

Under 80 μM AsV stress, the entire OD₆₀₀ of yeast cells, including WT, MB192, MB192-PnPht1;3 and MB192-YEplac112, increased with increasing Pi concentrations from 20 to 100 μM, particularly in WT and MB192-PnPht1;3 with a significant difference, suggesting that a high level of Pi can alleviate As stress (Fig. 7a). In addition, the OD₆₀₀ of WT and MB192-PnPht1;3 was higher than that of MB192-YEplac112 and mutant strain MB192 under the same Pi concentration (20 or 100 μM), but a non-significant difference existed between WT and MB192-PnPht1;3. This phenomenon demonstrates that Pi addition improved the probability that Pi transporters assimilated Pi under the competition of AsV (Cao et al. 2020). Another finding needs to be noted that PnPht1;3 preferred to combine Pi rather than AsV under a high concentration of Pi. These discoveries were reinforced by the result of As accumulation in yeast cells, of which the As contents of WT and MB192-PnPht1;3 significantly decreased by the addition of high-concentration Pi, while the As accumulation of MB192-YEplac112 and mutant strain MB192 changed very little (Fig. 7b). Under 20 μM of Pi and 80 μM AsV stress, MB192-PnPht1;3 had the second highest As concentration in cells and presented a significant difference from each of the four other yeast lines. When the Pi concentration reached 100 μM, the As content of MB192-PnPht1;3 sharply decreased to a value near the level of MB192-YEplac112 and mutant strain MB192 with a non-significant difference. Taken together, although overexpression of PnPht1;3 improved the survival of the mutant strain under Pi deficiency or AsV exposure, it could not completely complement the defect in the uptake of Pi and AsV, similar to WT. Generally, transporter PnPht1;3 has a stronger capacity for improving Pi and AsV acquisition in complementary mutants. In addition, a high Pi concentration played a positive role in alleviating AsV stress.

Fig. 7.  Growth (a) and As-accumulated concentration (b) of WT, MB192, MB192-YEplac112, and MB192-PnPht1;3 supplemented with low (20 μM)- or high (100 μM)-concentration Pi under the stress of 80 μM AsV. Different lowercase letters represent the difference of yeast cells under the same Pi concentration; * represents the difference of the same cell between 20 and 100 μM Pi at P ≤ 0.05. Error bars indicate mean values ± s.d. (n = 4).

P. notoginseng is a valued traditional Chinese medicinal herb. The rhizome is the main medicinal part, in which saponins are the primary bioactive components with clear pharmacological effects (Feng et al. 2018; Zhang et al. 2019; Wei et al. 2020). Unfortunately, evidence has indicated that As contamination commonly exists in notoginseng radix – up to 56% in 31 samples (Liu et al. 2014a). Arsenic accumulation in P. notoginseng was mainly caused by high As concentration background of the soil. As a result, almost half of the cultivated fields experienced a crisis of excessive As in the Wenshan Prefecture, of which 21 sample plots were analysed in total (Yan et al. 2012). Thus, the crisis of As contamination gradually entered people’s view. Consistent with reports that increasing phosphorus absorption would greatly inhibit As acquisition (Luan et al. 2018), our previous study found that the total As content in the notoginseng radix gradually increased with elevated AsV concentration in the soil but significantly decreased by supplementation with high-concentration Pi (Cao et al. 2020). These findings suggested a similar electrochemical characteristic of phosphate and AsV, whose absorption and transport were closely associated with Pi transporters (Tang et al. 2018). Nevertheless, the role of P. notoginseng Pi transporters in the uptake of Pi and As remains unclear.

Herein, the P. notoginseng gene PnPht1;3 was cloned and characterised based on a transcript of P. notoginseng fibrous roots under the stress of Pi deficiency and AsV exposure. According to the phylogenetic tree analysis, transporter PnPht1;3 belongs to subfamily Pht1 (Fig. 1), which plays important roles in the acquisition, transport and remobilisation of Pi in plants; e.g. transporters AtPht1, AtPht2, AtPht3 and AtPht4 are mainly involved in the uptake of Pi from soil to roots in Arabidopsis thaliana L. (Lapis-Gaza et al. 2014). The responsiveness effect of plant Pht1 genes to AsV varies with regard to the specific Pht1 gene and tissue type, as well as the concentration and duration of AsV exposure (Catarecha et al. 2007; Puckett et al. 2012; LeBlanc et al. 2013; Muehe et al. 2014). Many studies have shown that Pht1 could significantly respond to the induction of Pi deficiency or As exposure, e.g. downregulation of PvPht1;1 in P. vittata under As exposure, and upregulation of PvPht1;3 in P. vittata under Pi deficiency or As exposure (Ditusa et al. 2016). Our qPCR results showed that the relative expression level of PnPht1;3 significantly responded to both Pi deficiency and As exposure, particularly in the interaction of the two factors. The relative expression level of PnPht1;3 sharply increased under the interaction of low Pi and As exposure, but the stress response was weakened by supplying high Pi (Fig. 3). These results indicated that PnPht1;3 preferred to combine Pi iron rather than AsV. However, the roles of PnPht1;3 in Pi uptake and AsV accumulation remains unclear.

Hence, the characteristics of PnPht1;3 were revealed via complementation assays in S. cerevisiae MB192, which was defective in the high-affinity Pi transporter. Transporter PnPht1;3 played the role of PHO84 in the uptake of Pi in MB192-PnPht1;3 cells, which supports its growth under low Pi concentrations (2 and 20 μM Pi) (Fig. 4). Moreover, the results of pH-dependent ACP activity assays and proton pump inhibition experiments showed that PnPht1;3 was an H⁺-dependent Pi transporter (Fig. 4). Correspondingly, transformant MB192-PnPht1;3 was significantly inhibited by supplementation with CCCP or 2,4-DNP. In accordance with our results, Catharanthus roseus(L.) G. Don decreased cytoplasmic pH and increased pH in the extracellular medium (Sakano et al. 1992; Srivastava et al. 2018), as well as the acidification of cytoplasmic pH in the root hairs of Limmobium stoloniferum(G.F. Mey) Griseb. during Pi uptake, and ACP activity was improved (Sakano et al. 1992; Srivastava et al. 2018). Combined with the predicted subcellular localisation of the cytomembrane, the operation method of PnPht1;3 could be explained by transporter PnPht1;3, as an energised transport system, via H⁺/Pi co-transport to overcome the negative membrane potential entering plant cells. However, the process of H⁺/H₂PO₄⁻ symporting in the membrane has not been uncovered and is likely related to the mechanism of proton and glycerol-3-phosphate symporting in Escherichia coli (Abramson et al. 2003; Huang et al. 2003; Zhang et al. 2014).

The adaptation and tolerance of plants to As was initially uncovered at the levels of physiological and molecular mechanisms. AsV is usually reduced to AsIII in the cells of plants and further forms less toxic inorganic AsIII via methylation; e.g. dimethylarsenate (DMAsV) and trimethykarsenic oxide (TMAs) (Zhu et al. 2017; Alam et al. 2019; Roy et al. 2020). Arsenic only AsIII can bind to PCs and activate phytochelatin synthase, the reduction of AsV to AsIII is the first step in the detoxification of As, and As(V) reductase (AR) and its activity are therefore critical for As tolerance in plants (Li et al. 2016). Plant physiology changes against As stress, of which the activities of antioxidases; e.g. superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GPX) were improved, as well as enzymes involved in the ascorbate-glutathione (Asc-Glu) cycle and glyoxalase cycle, while the contents of antioxidants, such as proline and anthocyanin were also increased (Ahmad et al. 2020; Li et al. 2021). Many metabolic processes in plants, including availability of essential nutrients, photosynthesis, carbohydrate metabolism, lipid metabolism, protein metabolism and sulfur metabolism closely participate in As detoxification (Zhang et al. 2021). In addition, increasing evidence has shown that the expression levels of genes relating to transporters or phytochelate proteins are significantly upregulated; e.g. OsLsi2, OsLsi1 and OsABCC1 in rice (Abedi and Mojiri 2020; Pan et al. 2020). Notably, supplementation with some exogenous materies; e.g. sodium nitroprusside, melationin and Pi, would be an effective strategy for alleviating As toxicity (Ahmad et al. 2020; Li et al. 2021). Evidence has shown that high concentrations of external Pi competitively decrease the uptake and accumulation of As in plant, that could decrease membrane damage by lowering the activity of CAT, APX and lipid peroxidation (Gunes et al. 2009; Li et al. 2016). This explanation was confirmed by our previous study, in which high-concentration Pi lowered the oxidase activity of P. notoginseng, promoted the growth and improved the content of notoginsenoside, thereby decreasing the damage caused by AsV (Cao et al. 2020). Therefore, Pi application may be an important strategy for As detoxification or phytoremediation in plants and phytoplankton (Kertulis et al. 2005; Ye et al. 2011; Yan et al. 2012; Shaibur et al. 2013).

Despite the dependence on species and genotype, it could be concluded that arsenic tolerance in higher plants hinges on decreased As accumulation by suppression of the high-affinity Pi/As (V) uptake system in roots and decreased As transport to shoots (Meharg and MacNair 1992; Pigna et al. 2009), which is a common mechanism employed by arsenic-resistant plants (Ditusa et al. 2016). Thus, it is also a reasonable explanation for PnPht1;3 that transformants of MB192-PnPht1;3 decreased As accumulation in cells by suppression of the PnPht1;3 uptake system under high-concentration Pi (100 μM) (Figs 6 and 7). These reports suggest that plant high-affinity Pi transporters had comparable specificities for AsV uptake and play an important role in the AsV accumulation. For example, pht1;1-3 of A. thaliana showed a slow rate of AsV uptake that ultimately enabled the mutant to accumulate more As under appropriate Pi levels (Anawar et al. 2018), while AtPht1;5 or AtPht1;7 in A. thaliana preferred Pi over AsV (Ditusa et al. 2016). Taken together, PnPht1;3 responded to the stress of Pi deficiency and As exposure and was involved in the uptake of Pi and AsV. High-concentrations of Pi could also decrease the stress of AsV.

PnPht1;3 cloned from the roots of P. notoginseng encodes a putative high-affinity Pi/H+ symport transporter protein that enhances the uptake of Pi and AsV, which may be responsible for As accumulation. The role of the Pi transporter PnPht1;3 in the acquisition of Pi and As was revealed by a series of experiments. The qPCR results showed that the relative expression level of PnPht1;3 was significantly upregulated under Pi deficiency and AsV exposure. Heterologous expression in S. cerevisiae MB192 revealed that the expression of PnPht1;3 performed optimally in complementing the yeast Pi-transport defect. In addition, As accumulation in the yeast cells could be suppressed by supplementation with Pi. Taken together, we confirmed that PnPht1;3 encoded a functional plasma membrane-localised transporter protein that mediated putative high-affinity Pi/H⁺ symport activity and enhanced the uptake of Pi and AsV, which was likely responsible for the As accumulation of P. notoginseng. Our results provide insight into the acquisition mechanism of As in P. notoginseng and strategies for reducing As accumulation in taproots for enhanced pharmacal security.

Data and material not provided in this publication or previously published are available on request to the corresponding author.

The authors declare no conflicts of interest.

SH received stipends from the National Natural Science Foundation of China (81560612), the Yunnan Basic Research Projects (202001AT070109), the Yunnan Provincial Science and Technology Department-Applied Basic Research Joint Special Funds of Yunnan University of Chinese Medicine (202001AZ070001-010), and the Key Project at Central Government Level: The Ability Establishment of Sustainable Use for Valuable Chinese Medicine Resources (2060302-2101-24). GC received stipends from the Yunnan Basic Research Projects (2019FB122, 202101AT070608) and Yunnan Provincial Science and Technology Department-Applied Basic Research Joint Special Funds of Yunnan University of Chinese Medicine (202101AF070039). JY received a stipend from the Key Research and Development Projects of Yunnan Provincial Science and Technology Department (2019IB009).

GC, DL, XW, XZ, XL, FZ and WG conducted the experiments, completed the statistical and data analysis. GC, WX, SH and JY wrote most of the manuscript. XW, JY and SH contributed to the experimental setup and wrote part of the manuscript. GC, JY and SH were involved in planning the experiments and wrote parts of the manuscript. All authors read and approved the manuscript.