Enhanced Boron Tolerance in Plants Mediated by Bidirectional Transport Through Plasma Membrane Intrinsic Proteins

High boron (B) concentration is toxic to plants that limit plant productivity. Recent studies have shown the involvement of the members of major intrinsic protein (MIP) family in controlling B transport. Here, we have provided experimental evidences showing the bidirectional transport activity of rice OsPIP1;3 and OsPIP2;6. Boron transport ability of OsPIP1;3 and OsPIP2;6 were displayed in yeast HD9 mutant strain (∆fps1∆acr3∆ycf1) as a result of increased B sensitivity, influx and accumulation by OsPIP1;3, and rapid efflux activity by OsPIP2;6. RT-PCR analysis showed strong upregulation of OsPIP1;3 and OsPIP2;6 transcripts in roots by B toxicity. Transgenic Arabidopsis lines overexpressing OsPIP1;3 and OsPIP2;6 exhibited enhanced tolerance to B toxicity. Furthermore, B concentration was significantly increased after 2 and 3 hours of tracer boron (10B) treatment. Interestingly, a rapid efflux of 10B from the roots of the transgenic plants was observed within 1 h of 10B treatment. Boron tolerance in OsPIP1;3 and OsPIP2;6 lines was inhibited by aquaporin inhibitors, silver nitrate and sodium azide. Our data proved that OsPIP1;3 and OsPIP2;6 are indeed involved in both influx and efflux of boron transport. Manipulation of these PIPs could be highly useful in improving B tolerance in crops grown in high B containing soils.

Boron (B) is an essential and immobile micronutrient required for all plant nutrition. The main functions of B relate to cell wall strength, nucleic acid synthesis, hormone responses, membrane function and cell cycle regulation 1,2 . Higher concentration of B is toxic to plants and leads to nutritional disorder that eventually limits plant production in arid and semi-arid environment 3 . High concentrations of B may occur naturally in soil or in groundwater, or added to the soil from mining, fertilizers, or irrigation water 4 . Recent studies have implicated the involvement of efflux type B transporters and members of major intrinsic protein (MIP) family in controlling B toxicity tolerance 5 .
Major Intrinsic Protein (MIP) superfamily is highly conserved with members ranging in size from 23 to 31 kDa 6 . In higher plants, MIPs are divided into five main subfamilies based on their sequence similarities and localization: Plasma membrane Intrinsic Proteins (PIPs), Tonoplast membrane Intrinsic Proteins (TIPs), Nodulin 26-like Intrinsic membrane Proteins (NIPs) and the Small basic Intrinsic Proteins (SIPs) 7,8 . Recently, uncharacterized X Intrinsic Proteins (XIPs) were identified in some plant and moss species 9,10 . Among the MIP subfamilies, members of the PIP subfamily are the most studied. PIP subfamily is further divided into two groups: PIP1s and PIP2s. PIP1s isoforms have very low water channel activity 11 , whereas, PIP2s isoforms have been shown to posses high water channel activity 12,13 . In rice, MIP family is comprised of 11 PIPs, 10 TIPs, 10 NIPs and two SIPs members 14 .
Several researchers have identified a number of B efflux transporters in plants. The first efflux-type B transporter identified was AtBOR1 from Arabidopsis thaliana 15 . Overexpression of AtBOR1 conferred tolerance

OsPIP1;3 Expression Enhances B Sensitivity and Increases B Content in Yeast.
We tested the ability of OsPIP1;3 and OsPIP2;6, as a representative members of PIP1 and PIP2 subgroups, to functionally complement the function of aquaglyceroporin Fps1 in S. cerevisae. Both OsPIP1;3, and OsPIP2;6 genes were expressed in the HD9 yeast strain (∆fps1∆acr3∆ycf1) lacking the expression of the bidirectional aquaglyceroporin pump Fps1, arsenite export pump ACR3, and the vacuolar transporter YCF1 30 , for complementation of the B sensitivity and mobilization. The preliminary results with full-length coding region of OsPIPs failed to show any B transport activity in yeast HD9 strain. There was no difference in the growth of HD9 strain (∆fps1∆acr3∆ycf1) expressing OsPIP genes and the empty vector controls (Fig. 1A). It has been reported that some of the plant aquaporins required a truncation of the N-terminal hydrophobic region in order to complement the function in yeast Fps1 mutant 31 . Therefore, we attempted to complement the B transport activity of OsPIP1;3, and OsPIP2;6 after truncating their N-terminal hydrophobic region. Yeast growth was strongly impaired in cells expressing truncated OsPIP1;3 compared to yeast transformed with the pYES3 empty vector control on plates containing medium supplemented with 10 mM and 20 mM of boric acid (Fig. 1B). However, there was no significant difference between pYES3 empty vector and yeast expressing OsPIP2; 6. Total B content in the yeast cells expressing the truncated OsPIP1;3 and OsPIP2;6 was analyzed. The yeast cells expressing OsPIP1;3 showed significantly higher B accumulation as compared with pYES3 vector control (Fig. 1C), whereas, yeast cells expressing OsPIP2;6 had no significant difference in the B accumulation level compared to yeast cells expressing pYES3 empty vector. This B accumulation data is in accordance with the B sensitivity phenotype in HD9 strain complemented with OsPIP1;3.

Short Duration Influx and Effux of B in HD9 Yeast Strain Expressing OsPIP1;3 and OsPIP2;6.
The role of OsPIP1;3 and OsPIP2;6 in transporting B was evaluated by measuring short duration influx assay using tracer 10 B. HD9 strain expressing the empty vector pYES3, OsPIP1;3, and OsPIP2;6 were grown to an OD 600 of 1. For influx, cells were exposed to tracer boron 10 B enriched boric acid for 0, 15, 30 and 60 min and 10 B contents were measured by ICP-MS. Expression of OsPIP1;3 led to a significant increase in 10 B uptake after 15, 30 and 60 min of exposure. While, there was no significant difference between yeast cells expressing the empty vector pYES3 and OsPIP2;6 (Supplementary Fig S1A). This time dependent influx results corresponded well with the levels of B sensitivity observed when growing the respective transformants on B-containing medium as shown in Fig. 1A.
To test whether OsPIP1;3 and OsPIP2;6 have B efflux activity, HD9 strains expressing the empty vector pYES3 and truncated versions of OsPIP1;3 and OsPIP2;6 were allowed to accumulate 10 B till the end of log phase, harvested, washed, and then cells were re-suspended in B free media. 10 B contents were analyzed by the method previously described after 15, 30 and 60 minutes. Efflux from yeast cells expressing OsPIP2;6 was faster than yeast cells expressing the empty vector pYES3 and OsPIP1;3 after 15 minutes of resuspension in B free media. However, there was no difference between the empty vector pYES3 and OsPIP2;6 after 30 and 60 minutes of resuspension in B free media. Whereas, at all time points, yeast cells expressing OsPIP1;3 showed less efflux activity than the empty vector pYES3 and OsPIP2;6 (Supplementary Fig. 1B we performed a quantitative RT-PCR (qRT-PCR) analysis. In shoots, transcript level of OsPIP1;3 were downregulated after 6, 12, and 24 hrs of B treatment, whereas, OsPIP2;6 transcript levels showed contrasting pattern as it was upregulated after 6, 12 and 24 hrs of B toxicity ( Fig. 2A). In roots, our results showed that the transcript levels of OsPIP1;3 and OsPIP2;6 were strongly upregulated up to 24 hrs of B treatment compared to the untreated controls (Fig. 2B). The transcript level for OsPIP1;3 showed 15-fold increase at 24 hrs of B exposure, whereas, OsPIP2;6 were induced by 30-fold at 24 hrs of B toxicity (Fig. 2B).

Enhanced Boron Toxicity Tolerance in Arabidopsis Plants Overexpressing OsPIP1;3 and
OsPIP2;6. To investigate the function of OsPIP1;3 and OsPIP2;6 in boron permeability in plants, OsPIP1;3 was overexpressed in A. thaliana under actin2 promoter-terminator cassette (ACT2pt) (Supplemental Fig. S2A). Following transformation, several independent transgenic lines were screened for kanamycin resistance, and then three T 2 homozygous Arabidopsis lines overexpressing OsPIP1;3 (43, 51, and 60) were selected for further analysis. In addition, we used the three T 2 homozygous Arabidopsis lines overexpressing OsPIP2;6 (27, 33, and 40) that we generated in our previous study 32 for further analysis. The overexpression of each transgene in the transgenic T 2 Arabidopsis plants was confirmed by semi-quantitative RT-PCR analysis (supplementary Fig. S2B).
To evaluate the phenotypic effect of B on the Arabidopsis transgenic lines overexpressing OsPIP1;3 and OsPIP2;6, seeds of transgenic lines and wild-type Arabidopsis were germinated on an 1/2x MS agar medium containing 0 or 2.5 mM of boric acid. On the control media without B, there were no differences in the phenotypes between the transgenic lines overexpressing OsPIP1;3 and OsPIP2;6, and wild-type control plants. However, on media supplemented with 2.5 mM of boric acid, the transgenic lines exhibited strong tolerance to B toxicity as compared with wild type plants (Figs 3 and 4). The growth of both shoots and roots of the transgenic lines were more vigorous than that of wild type plants on media containing B. The OsPIP1;3 and OsPIP2;6 transgenic plants had well developed green leaves compared to the wild types that had smaller and pale yellow leaves. Further, Boron Accumulation in Transgenic Arabidopsis Overexpressing Rice PIP1;3 and PIP2;6. To test whether OsPIP1;3 and OsPIP2;6 are involved in B transport and uptake, wild type and transgenic Arabidopsis lines overexpressing OsPIP1;3 and OsPIP2;6 were grown hydroponically for three weeks and then treated with 2.5 mM boric acid for four days. Total B from shoots and roots of wild type and transgenic Arabidopsis were measured separately by ICP-MS. In shoots, overexpression of OsPIP1;3 and OsPIP2;6 showed no significant difference in B accumulation compared to the wild type ( Fig. 5A,B). Similarly, there were no significant differences in B accumulation levels between the transgenic lines overexpressing OsPIP1;3 and OsPIP2;6 root tissues compared to wild type plant roots (Fig. 5C,D).
To confirm these results, another long-term uptake assays were conducted using a stable isotope of B ( 10 B). Regular Boric acid composition is 11 B : 10 B = 81.0 : 19.0. Wild type and transgenic Arabidopsis lines overexpressing OsPIP1;3 and OsPIP2;6 were grown on 1/2x MS for three weeks and then treated with 2.5 mM 10 B enriched boric acid for four days. Accumulation of 10 B in shoots and roots of wild type and transgenic plants were measured separately by ICP-MS. Similar to previous results for boric acid, there was no significant difference in 10 B accumulation levels between the Arabidopsis wild type plants and the transgenic lines overexpressing OsPIP1;3 and OsPIP2;6 in both shoot and root tissues (Supplemental Fig. S3).

OsPIP1;3 and OsPIP2;6 Exhibited Bidirectional Transport Activity in Transgenic Arabidopsis
Plants. In order to check if OsPIP1;3 and OsPIP2;6 are involved in the influx of B, a short-term B uptake (influx) using a stable isotope 10 B was performed. Wild type and transgenic Arabidopsis plants overexpressing OsPIP1;3 and OsPIP2;6 were first grown on 1/2x MS containing required amount of boric acid ( 11 B : 10 B = 81.0 : 19.0) for 3 weeks, and then plants were exposed to media containing 5 mM 10 B enriched boric acid and the amount of 10 B taken up by roots and shoots were determined at specified time points. In shoots, OsPIP1;3 and OsPIP2;6 overexpressed plants exhibited greater 10 B uptake than wild type plants after 1, 2, and 3 hrs (Fig. 6A). Similarly, a rapid influx of 10 B across the roots of OsPIP1;3 and OsPIP2;6 transgenic plants after 1, 2 and 3 hrs of exposure was observed (Fig. 6B).
To determine whether OsPIP1;3 and OsPIP2;6 are involved in the efflux of B, a short-term B efflux assay using a stable isotope 10 B was conducted. Arabidopsis wild type and transgenic plants overexpressing OsPIP1;3 and OsPIP2;6 were first exposed to 5 mM 10 B enriched boric acid in 1/2x MS liquid medium for 3 hrs and then Rice seedlings were exposed to 5 mM boric acid before harvesting shoots and roots tissues after 0, 6, 12 and 24 hrs of boron treatment. 18S rRNA was used for normalization of gene expression.
Scientific RepoRts | 6:21640 | DOI: 10.1038/srep21640 incubated in 1/2x MS liquid medium without any B up to 3 hrs. Concentration of 10 B in roots was measured at each time point (Fig. 6C). A rapid efflux of more than half of 10 B was observed within the first hour after transferring to 1/2x MS media without B in roots of the transgenic Arabidopsis expressing OsPIP1;3 or OsPIP2;6. However, after 2 and 3 hours period there was no significant difference in 10 B concentration in roots of the transgenic and wild types control plants (Fig. 6C).

Effect of Aquaporin Inhibitors on Boron Tolerance in OsPIP1;3 and OsPIP2;6 Transgenic
Plants. These above-described findings prompted further investigation into the effect of two different aquaporin channel inhibitor/blocker, sodium azide (NaN 3 ) and silver nitrate (AgNO 3 ), on the phenotype of the Arabidopsis plants overexpressing OsPIPs. Seeds of Arabidopsis wild type control and three independent transgenic lines were grown for three weeks on 1/2x MS medium containing 0, 3 mM B, 100 μ M NaN 3 , 50 μ M AgNO 3 , 3 mM B + 100 μ M NaN 3 , and 3 mM B + 50 μ M AgNO 3 . There was no difference between the wild type plants and transgenic lines overexpressing OsPIP1;3 and OsPIP2;6 grown on the control plates without B or NaN 3 or AgNO 3 (Figs 7 and 8). However, as expected, the transgenic lines of OsPIP1;3 and OsPIP2;6 grown on 3 mM B were more tolerant than the wild type plants. Interestingly, when plants were grown on media containing 3 mM B + 100 μ M NaN 3 , tolerance to B was abolished and there were no phenotypic differences between transgenic OsPIP1;3, OsPIP2;6 lines and wild-type control plants. Similarly, the addition of 50 μ M AgNO 3 to the 3 mM B media decreased the tolerance phenotype of the transgenic lines overexpressing OsPIP1;3 and OsPIP2;6 (Figs 7 and 8). The fresh shoot biomass of the OsPIP1;3 and OsPIP2;6 transgenic lines had 2 to 3-fold more the biomass  Fig. S4). Whereas, the biomass of the OsPIP1;3 and OsPIP2;6 transgenic lines and wild type plants was almost same when plants were grown on 3 mM B + 100 μ M NaN 3 or 3 mM B + 50 μ M AgNO 3 (Supplementary Fig. S4). This proves the specific action of the inhibition of PIPs by sodium azide and silver nitrate and the tolerance to B in the transgenic lines is as a result of the activity of the OsPIP1;3 and OsPIP2;6.

Discussion
Aquaporins are shown to play a pivotal role in metalloids transport in various organisms and our previous reports potentiate the role of members of rice PIPs, a subfamily of aquaporins in arsenite (AsIII) and B transport in plants 32,29 . In current study, we have provided experimental evidences showing that OsPIP1;3 and OsPIP2;6 are involved in both influx and efflux of boron transport. Regulation analysis of OsPIP1;3 and OsPIP2;6 transcripts showed that both genes exhibited a strong upregulation in roots under B toxicity. The reason for slight decreases in OsPIP1;3 transcript levels in shoots compared to OsPIP2;6 is not known. However, given to the very high levels of transcripts of both OsPIP1;3 and OsPIP2;6 in roots (15-to 30-fold higher), this difference (2-fold decrease) in shoot is relatively minor. Further, this difference in the regulation of both genes in shoots might be due their difference in tissue-specificity or functional variability, which need further studies to clarify. It has previously been reported that expression of HvPIP1;3 boron transporter was significantly increased upon exposure to 5 mM of B in barley 28 . Similarly, the expression of OsPIP2;4 and OsPIP2;7 were strongly induced in roots by toxic B treatment, that favors the current evidences 29 . The ability of OsPIP1;3 and OsPIP2;6 to transport B was examined using a yeast complementation assay with HD9 strain defective in Fps1, ACR3, and Ycf1 30 . Preliminary results with full-length genes failed to show any B transport activity in yeast strain HD9. It has been reported that some members of plant PIPs require partial truncation of the hydrophobic N-terminal domains in order to be functionally expressed in yeast 28,29 . Therefore, in the current study, we attempted to complement the B transport activity of OsPIP1;3 and OsPIP2;6 after truncating their N-terminal hydrophobic region. The B transport ability of OsPIP1;3 could be demonstrated as a result of increased B sensitivity and increased B influx and accumulation in yeast cells. Whereas, there was no difference in the growth of HD9 strain expressing OsPIP2;6 and the empty vector control. However, our observation should not be directly taken as the inability of OsPIP2;6 to transport boron as it showed strong efflux activity of B to external media within 15 minutes. This rapid efflux explains the B-insensitive phenotype and lack of B accumulation in yeast strain. Similarly, earlier studies also showed that expression of OsNIP1;1 did not affect the growth of Δ fps1Δ acr3Δ ycf1 mutant yeast on AsIII containing medium 31 . However, OsNIP1;1 was able to mediate AsIII transport when expressed in Xenopus oocytes 33 . In addition, heterologous expression of NIP1;2, NIP5;1, NIP6;1, NIP7;1 in yeast demonstrated their ability to transport AsIII 31,34 , although AsIII tolerance was not observed in the T-DNA lines of nip1;2, nip5;1, nip6;1, and nip7;1 35 . Therefore, these inconsistent results may rely on the respective expression assays and experimental conditions. Likewise, considering the fact that the failure of complementation of OsPIP2;6 in HD9 strain, our study also demonstrated the B transport ability of OsPIP2;6 in the overexpressed Arabidopsis transgenic plants. Further, we stressed that yeast cells are different than plant cells in terms of their ability to tolerate B toxicity. Optimization of yeast and Arabidopsis growth on various concentrations of boric acids showed clearly that yeast can tolerate much higher (up to 20 mM) concentration of B than Arabidopsis (up to 3 mM). When OsPIP1;3 and OsPIP2;6 were overexpressed in Arabidopsis to clarify their physiological function in planta, the transgenic plants exhibited enhanced tolerance to B toxicity. Similar results were previously observed for other aquaporins members. Transgenic Arabidopsis plants overexpressing AtTIP5;1 conferred tolerance to B toxicity 26 . Overexpression of OsPIP2;4 and OsPIP2;7 in Arabidopsis also resulted in an increased tolerance to toxic B levels 29 . It is also intriguing that total B concentration was not affected when the transgenic Arabidopsis plants overexpressing OsPIP1;3 and OsPIP2;6 treated with boric acid for four days while the B concentration was significantly increased after 2 and 3 hours of 10 B treatment. Interestingly, a rapid efflux of 10 B from the roots of the OsPIP1;3 and OsPIP2;6 overexpressed transgenic Arabidopsis plants was observed within 1 h of 10 B treatment. These results are in accordance with our recent report demonstrated the involvement OsPIP2;4 and OsPIP2;7 in both influx and efflux of B from roots and shoot tissues 29 , which is also consistent with the well-known fact of the metalloids bidirectional transport properties of plant aquaporins 31 . Therefore, tolerance to toxic B in the OsPIP1;3 and OsPIP2;6 overexpressed transgenic Arabidopsis plants presumed to be as a result of the rapid efflux of the excess B from root tissues to external medium.
Further, it is interesting to note that this B tolerance phenotype in OsPIP1;3 and OsPIP2;6 overexpressed lines caused by efflux mechanism was inhibited by two aquaporin inhibitors; silver nitrate and sodium azide. It has been reported that silver nitrate and sodium azide can block the aquaporins channels in plants 36,37 . Arsenite uptake in roots and fronds of P. Vittata was reduced by 64% and 58% in response to silver nitrate treatment as a possible result of inhibiting aquaporin transporters 38 . Boron transport via PIP aquaporins in barley roots was reduced by 40-50% in response to metabolic inhibition by sodium azide 28 . In this study, we demonstrated phenotypic evidence showing the inhibition of aquaporin gating by silver nitrate and sodium azide. Adding these two inhibitors to the media containing toxic B, abolished the tolerance phenotype of the transgenic Arabidopsis plants overexpressing OsPIP1;3 and OsPIP2;6 as a result of blocking the transport (both influx and efflux) mechanism of PIPs.
In conclusion, the findings presented here suggesting a bidirectional B transport activity of OsPIP1;3 and OsPIP2;6 and their differential regulation in roots and shoots exposed to high B. These results are in agreement with our previous findings for OsPIP2;4 and OsPIP2;7 29 , supporting our hypothesis that PIPs are involved in providing B tolerance by maintaining the B homeostasis in plant cells via their bidirectional transport activity under elevated B stress conditions. Similarly, we have previously shown the overexpression of OsPIP2;4, OsPIP2;6 and OsPIP2;7 showed strong tolerance to AsIII via similar bidirectional transport activity and were differentially regulated by AsIII 32 . Based on these findings, we drew a working model (Fig. 9) showing a proposed mechanism along with conclusion that for OsPIP1;3, OsPIP2;4, OsPIP2;6 and OsPIP2;7, contribute largely to the B and AsIII transport and tolerance mechanism in plants. OsPIP1;3, OsPIP2;4, OsPIP2;6, and OsPIP2;7 proteins localized at the plasma membrane are involved in the bidirectional transport of AsIII and B, and most likely other metalloids as well, which need to be tested in the future studies. More studies are needed to investigate the other possible mechanisms marked with '?' in Fig. 9. PIPs could transport metalloids through early endosome to vacuole or to the plasma membrane, supporting the hypothesis that it might be involved in the internal re-distribution of metalloids (Fig. 9). Further characterization of these and other PIPs is required to clarify the functional significance of each member in boron transport in plants. Manipulation of these PIPs in crops could prove highly useful in improving B tolerance in crops grown in high B containing soils.

Methods
Plant Material and Boron Treatment. Rice (Oryza sativa cv. Nipponbare) seeds were germinated for two weeks in vermiculite under the green house conditions of 16/8 hours light/dark cycle, respectively, at 28 °C. Two weeks old rice plants were transferred to hydroponic solution with Hoagland's nutrients and kept to acclimatize for one week before the treatment. Five mM boric acid were added to hydroponic solution before harvesting shoots and roots tissues at 0, 6, 12 and 24 hrs after B treatments. Harvested tissues were frozen into liquid nitrogen, and stored at − 80 °C until further use.

RNA Extraction and RT-PCR Analysis.
Total RNA from shoot and root tissues were isolated using RNeasy plant mini kit (Qiagen) following the manufacturer's instructions. Five μ g of total RNA was used for reverse transcription using the ThermoScript ™ RT-PCR System (Invitrogen) for first-strand cDNA synthesis following the manufacturer's instructions. Gene-specific primers used for the quantitative real time PCR (qRT-PCR) analysis were designed using the PrimerQuest (Intergrated DNA Technologies). The sequences of the qRT-PCR primers are as follows: OsPIP1;3 forward primer: CTGGTGATCGATGAAGCTAG; reverse primer: ACACAAGTACCATTTCTCACAC; and OsPIP2;6 forward primer: GCCAGGTGCATGATTTGTT; reverse primer: GCCGAAGCAGTTTGTATCTC. qRT-PCR was performed following the instruction for Mastercycler ® ep realplex (Eppendorf AG, Hamburg, Germany) with ABsolute Blue qPCR SYBR Green Mix (Thermo Fisher Scientific, Surrey, UK). Relative expression level was calculated using 2 −ΔΔCT method 39  Yeast Complementation Assays. Full-length or truncated OsPIP1;3 and OsPIP2;6 gene sequences were amplified by PCR. For N-terminal truncation of OsPIPs, first 30-40 amino acids, were deleted by designing forward primer lacking these N-terminal sequences as previously described 29 . Amplified products were cloned into the pGEM-T easy vector (Promega, UK) using manufacturer's instructions. The sequences were verified by DNA sequencing and then were cloned into EcoRI and XhoI sites of pYES3 vector for subsequent transformation into HD9 (∆fps1∆acr3∆ycf1) yeast (Saccharomyces cerevisae) strain 30 . The transformants were selected onto minimal SD medium minus tryptophan and confirmed by PCR. Yeast culture of HD9 expressing empty vector pYES3 or OsPIP1;3 and OsPIP2;6 were grown in liquid SD medium with either 2% glucose or 2% galactose and the appropriate supplements. Yeast phenotypes assays were performed as previously described by Kumar et al. (2013) 29 . Briefly, the cultures were diluted to an OD 600 of 1.0. Cells with 10-fold dilutions series were spotted onto SD + Gal plates and SD + glu plates (as a control) with various concentrations of boric acid as indicated. The indicated concentration of boric acids were selected based on optimization of yeast strain growth conditions. Yeast growth on the plates was observed after 3-4 days of incubation at 30 °C.
Boron Transport Assays in Yeast. HD9 strain transformed with OsPIP1;3 and OsPIP2;6 and empty vector pYES3 were grown at 30 °C in SD minus tryptophane medium supplemented with 2% galactose to an OD 600 of 1.0. For B uptake assay, cells were supplied with 10 mM boric acid and allowed to grow overnight. Yeast cells were washed with Tris-HCl buffer containing 0.5% NaCl and then quickly washed by deionized water and dried at 70 °C, which were then digested with concentrated HNO 3 and H 2 O 2 for total B was analysis by ICP-MS (Perkin Elmer). For B influx assay, cells were exposed to 10 mM 10 B enriched boric acid (99%, Cambridge Isotope Laboratories, Andover, MA, USA) for 0, 15, 30 and 60 min. Cells were harvested and washed once with chilled 25 mM phosphate buffer pH 6.0 containing 0.5% NaCl, then by 10 mM Tris HCl followed by deionized water and dried at 70 °C. Total B contents were estimated by ICP-MS (Perkin Elmer) after digestion with concentrated HNO 3 and H 2 O 2 . For B efflux assay, cells were allowed to accumulate B till the end of log phase in 10 mM 10 B enriched boric acid containing SD (-tryptophan) media. Cells were washed in chilled phosphate buffer as described above and re-suspended in the media without B. Aliquot of cells were withdrawn at 0, 20, 40 and 60 minute and the B content were analyzed by the method previously described. Total B contents at 20, 40 and 60 min were plotted as a fraction of B content at time 0, which is the maximum amount of B accumulated.
Overexpression of OsPIP1;3 and OsPIP2;6 in Arabidopsis and Characterization of Transgenic Plants. Full-length fragments of OsPIP1;3 gene was amplified by PCR using the following primers; Forward: TACGTCGAATTCTCCATGGAGGGGAAGGAGGA; Reverse: TAGCTGCTCGAGTTAGTCCCGGCTCTTGAA, and then cloned into the pGEMT easy vector (Promega, UK) using manufacturer's instructions. The NcoI/XhoI fragment of OsPIP1;3 was sub-cloned under the constitutively expressed actin2 gene promoter-terminator expression cassette, ACT2pt 40 , and cloned into the binary vector pBIN19 using the KpnI/SacI restriction sites. Plasmids were introduced into the Agrobacterium tumefaciens strain C58 using the heat shock method.
Arabidopsis plants were transformed by vacuum infiltration as described previously 41 . Transgenic plants were selected on kanamycin and the resulting T 0 seedlings of Arabidopsis expressing OsPIP1;3 were grown in soil and independent T 1 transgenic lines were obtained. Homozygous T 2 lines were generated by growing the T 1 transgenic lines showing 3:1 (resistant: sensitive) Mendelian segregation ratio. Transcript analysis was performed on the T 2 transgenic Arabidopsis lines by a semi-quantitative RT-PCR as described above. For studying the role of OsPIP2;6 in response to B treatment, we used the three T 2 homozygous Arabidopsis lines overexpressing OsPIP2;6 (27, 33, and 40) that were generated in our previous study 32 . For B phenotypic analysis, seeds of the wild type and transgenic overexpression lines were germinated and grown on 1/2x MS agar plates containing 0 or 2.5 mM boric acid for three weeks. Shoot biomass and root lengths were determined. For aquaporin inhibitors treatment, 100 μ M NaN 3 (sodium azide) and 50 μ M AgNO 3 (silver nitrate) were used with or without boric acid.
Boron Transport Assays in Arabidopsis. Wild type and transgenic Arabidopsis seeds were grown on a nylon mesh placed on 1/2x MS plates for two weeks with 16/8 hrs light/dark cycles at 22 °C. The nylon mesh along with seedlings was then transferred, placed on a support into the magenta boxes containing1/2x MS liquid media and plants were grown for one-week for acclimatization. For long-term B accumulation assay, the 1/2x MS liquid medium was replaced with new medium containing 2.5 mM boric acid and plants were grown for another four days. Shoot and root tissues were harvested, washed with deionized water, and dried at 70 °C for 48 hrs. Dried tissues were digested in concentrated nitric acid and hydrogen peroxide and then samples were analyzed by ICP-MS. For B influx assay, 5 mM 10 B-enriched boric acid (99%, Cambridge Isotope Laboratories, Andover, MA, USA) was used. Boric acid concentration was selected based on the optimization of Arabidopsis plant growths on wide range of concentrations. The shoot and root tissues were harvested separately after 1, 2, and 3 hrs of 10 B treatment. 10 B in roots and shoots was measured by ICP-MS. For B efflux assay, plants were treated with 5 mM 10 B-enriched boric acid for 3 hrs and washed quickly with deionized water and then transferred to 1/2x MS liquid medium without B. Roots and shoots were harvested after 0, 1, 2, and 3 hrs of transferring to B free media. 10 B was analyzed in roots, shoots, and efflux media by ICP-MS Statistical Analysis. One-way ANOVA followed by least significant difference (LSD) multiple comparison test (p < 0.01 and p < 0.05) was used to determine all differences of statistical significance among treatments. Standard error of the mean were calculated and represented in all figures. Elemental content is expressed on a dry weight basis.