Constitutive overexpression of soybean plasma membrane intrinsic protein GmPIP1;6 confers salt tolerance

Background Under saline conditions, plant growth is depressed via osmotic stress and salt can accumulate in leaves leading to further depression of growth due to reduced photosynthesis and gas exchange. Aquaporins are proposed to have a major role in growth of plants via their impact on root water uptake and leaf gas exchange. In this study, soybean plasma membrane intrinsic protein 1;6 (GmPIP1;6) was constitutively overexpressed to evaluate the function of GmPIP1;6 in growth regulation and salt tolerance in soybean. Results GmPIP1;6 is highly expressed in roots as well as reproductive tissues and the protein targeted to the plasma membrane in onion epidermis. Treatment with 100 mM NaCl resulted in reduced expression initially, then after 3 days the expression was increased in root and leaves. The effects of constitutive overexpression of GmPIP1;6 in soybean was examined under normal and salt stress conditions. Overexpression in 2 independent lines resulted in enhanced leaf gas exchange, but not growth under normal conditions compared to wild type (WT). With 100 mM NaCl, net assimilation was much higher in the GmPIP1;6-Oe and growth was enhanced relative to WT. GmPIP1;6-Oe plants did not have higher root hydraulic conductance (Lo) under normal conditions, but were able to maintain Lo under saline conditions compared to WT which decreased Lo. GmPIP1;6-Oe lines grown in the field had increased yield resulting mainly from increased seed size. Conclusions The general impact of overexpression of GmPIP1;6 suggests that it may be a multifunctional aquaporin involved in root water transport, photosynthesis and seed loading. GmPIP1;6 is a valuable gene for genetic engineering to improve soybean yield and salt tolerance.

Background A significant proportion of cultivated land is salt affected representing about 2% of dry-land and 20% of irrigated agriculture (FAO Land and Plant Nutrition Management service, http://www.fao.org/nr/aboutnr/nrl/en/). Soil salinity arises from natural or human-induced processes that inhibits plant growth via osmotically induced water deficit and/or ion toxicity if excessive sodium (Na + ) and chloride (Cl − ) accumulate in the shoot via transpiration [1]. Osmotic stress reduces the ability of the plant to extract water from the soil and growth will reduce rapidly and significantly as salt concentration around the roots increases past a threshold level. Ion toxicity occurs when salt (Na + and Cl − ) gains entry via the transpiration stream and accumulates in the shoot to toxic concentrations resulting in injury to cells and causing further reductions in growth [1,2]. Salt tolerance/sensitivity is indicated by the relative degree of biomass reduction in saline soil compared to plants in a non-saline soil, over an extended period of time [3]. Clearly water flow is linked to both types of stresses induced by salinity, yet the role of water transport in plant salt tolerance is not yet clearly defined.
Plants have evolved three distinct mechanisms of salinity tolerance including osmotic adjustment to allow turgor to be maintained, Na + and Cl − exclusion from leaf blades, and compartmentalization of Na + and Cl − at cellular or intracellular sites [1]. Numerous transporters have been identified as likely to be involved in Na + and Cl − exclusion and compartmentation [1,2,[4][5][6][7], but the proteins that transport water across membranes, the aquaporins, are not considered to be directly involved in these processes, though indirect effects could occur through their impact on osmotically driven water flow and pathways for water and solute flow in roots and leaves [8].
The radial flow of water from soil solution toward the root xylem encounters a relatively high resistance compared to subsequent axial flow in the xylem to the shoot. The radial flow pathway in the root consists of the apoplastic pathway along the intracellular spaces and the cell-to-cell pathway, in which water moves through plasmodesmata or across membranes [9]. Apoplastic water flow can be blocked by Casparian bands and suberin lamellae at key cellular barriers such as the endo and exodermis [10,11] where water transport across membranes occurs. Depending on the plant species and conditions, as well as the position along the root, there are variable contributions of the apoplast pathway compared to the cell-to-cell pathway [8]. The conductance of the cell-to-cell pathway can be largely determined by the activity of aquaporins (AQPs) [12]. AQPs are suggested to play a key role in plant water balance and water use efficiency [8,[13][14][15][16][17].
Aquaporins are members of the major intrinsic protein (MIP) family, which in plants are divided into five subfamilies that include the plasma membrane intrinsic proteins (PIPs). These are considered as the main water transport pathway across plasma membranes in root and leaf tissues that play important roles in plant water relations [8,[16][17][18][19]. According to the N terminal length of the proteins, the PIPs are further divided into two subclasses (PIP1 and PIP2). PIP1s require co-expression of PIP2s to show high water permeability in Xenopus laevis oocytes [20][21][22][23][24][25][26][27]. PIP1s and PIP2s interact affecting targeting to the plasma membrane [20,21] and forming hetero-tetramers of variable stoichiometry that appears to affect their transport efficiency [27]. Plant genomes have variable numbers of aquaporin genes, ranging from 35 in Arabidopsis thaliana [28], 33 in Oryza sativa [29] and 66 in soybean, including 22 PIPs [30]. Compared with other species, little is known about the function of AQP genes in soybean.
Aquaporins are clearly involved in water transport in roots and leaves [8] and have been linked to water uptake required for cell expansion [18,26,[31][32][33][34][35]. Water is the carrier of Na + and Cl − in the transpiration stream contributing to shoot ion toxicity, and in salinity-induced osmotic stress, free energy gradients need to be developed to drive water diffusion to the sites of cell expansion. In this context aquaporins could affect the root's ion selectivity by determining the proportion of water that flows via membrane pathways relative to the apoplast, while in osmotic stress, they could allow continued water supply under diminished osmotic and pressure gradients by increasing membrane hydraulic conductivity.
Abiotic stresses such as salt, drought and cold influence the water balance of plants and the expression of AQP genes [36]. Overexpression of several AQP genes in plants confers abiotic stress resistance. Overexpressing NtAQP1 in tobacco increased photosynthetic rate, water use efficiency and yield under salt stress [17]. Overexpression of several wheat AQPs, including TaNIP, TaAQP8 and TaAQP7 genes in Arabidopsis or tobacco also increased salt tolerance or drought tolerance of the transgenic plants [37][38][39]. Recently, overexpression a MusaPIP1;2 in banana displayed high tolerance to multiple abiotic stresses including salt, cold and drought [40].
Soybean is a major source of protein and oil for humans and animals, yet relatively mild salt stress significantly reduces soybean growth, nodulation, seed quality and yield [41]. Recently it was found that the expression of GmPIP1;6 in roots correlated with rapid and longer term changes in root L o in response to shoot treatments and appeared to be more concentrated in stellar tissue [42]. These results indicated that GmPIP1;6 may be important in the control of root water transport particularly in response to shoot signals. In this study, GmPIP1;6 was cloned and functionally characterized. Overexpression of GmPIP1;6 significantly increased salt tolerance of soybean by improving root L o and Na + exclusion.

Subcellular localization of GmPIP1;6
In soybean, GmPIP1;6 was proposed to be one of the major water transporter genes in roots [42]. The full-length cDNA of GmPIP1;6 (Phytozome No. Gm08g01860.1) was amplified from soybean roots cultivar Williams 82 by PCR. The cDNA of GmPIP1;6 is comprised of 1128 bp with an 870 bp open reading frame. The GmPIP1;6 belongs to the PIP1 subgroup and has an orthologous gene, GmPIP1;5, in the soybean genome (Additional file 1: Figure S1). GmPIP1;6 protein contains the characteristic motifs of PIPs and is predicted to be localized on the plasma membrane. To verify the subcellular localization of GmPIP1;6, the GmPIP1;6 was fused with green fluorescent protein (GFP) and driven by a constitutive Cauliflower mosaic virus 35S promoter (CaMV 35S). The final construct GmPIP1;6::GFP was transiently co-expressed in onion epidermal cells and compared with the mCherry plasma membrane marker ( Figure 1). GFP fluorescence of GmPIP1;6 was confined to the plasma membrane and co-localized with the RFP fluorescence of mCherry. When the control construct with GFP alone was transiently expressed in onion epidermal cells, the GFP fluorescence was observed in the nucleus and cytoplasm ( Figure 1).

Expression patterns of GmPIP1;6
To analyze GmPIP1;6 gene expression in different soybean tissues, we measured the expression of GmPIP1;6 in root, stem, unifoliolate leaf, trifoliolate leaf, flower and pod by quantitative RT-PCR. This showed that GmPIP1;6 was highly expressed in root, stem, flower and pod whereas it was lowly expressed in leaves (Figure 2A). To investigate the response of GmPIP1;6 to salt stress, we determined expression of GmPIP1;6 in root and leaf after 100 mM NaCl treatment for 6 hours, 12 hours, 1 day, 3 days and 5 days. The expression of GmPIP1;6 was suppressed by NaCl treatment in 6 and 12 hours in both roots and leaves. However, expression was induced in the roots after 1 day and further increased at 3 days and 5 days of NaCl treatment ( Figure 2B). A similar response was observed in the leaves although the absolute expression of GmPIP1;6 was much lower than that of the roots (Additional file 1: Figure S2). To characterize the role of GmPIP1;6 in salt stress, the cDNA of GmPIP1;6 driven by a modified CaMV 35S promoter was introduced into soybean via soybean cotyledonary node transformation system ( Figure 3A). Positive transgenic lines were selected by spraying the herbicide Liberty (Additional file 1: Figure S3). A total of 11 independent lines which overexpressed GmPIP1;6 were generated and confirmed by semi-qRT PCR ( Figure 3B). Two transgenic lines were selected and measured by qRT-PCR. These two lines, which showed more than 100-fold higher expression levels of GmPIP1;6 than WT control in leaves ( Figure 3C), were selected for further experiments.

Overexpression of GmPIP1;6 enhances salt tolerance in soybean
The growth of WT and GmPIP1;6-Oe transgenic soybean plants were similar when grown in aerated hydroponic solution ( Figure 4A, Table 1). For salt tolerance analysis, 10-day-old WT and GmPIP1;6-Oe transgenic seedlings were treated with 100 mM NaCl for 7 days. Salt treatment suppressed the growth of WT and GmPIP1;6-Oe transgenic soybean plants, all of which exhibited a decreased plant length and fresh weight in both leaves and roots ( Figure 4A, Table 1). However, the leaves of WT plants turned yellow after treatment for 7 days while the leaves of transgenic plants were still green ( Figure 4A). The relative measure of leaf greenness was carried out with a portable chlorophyll meter. Soil-plant analyser development (SPAD) values of unifoliolate leaf in GmPIP1;6-Oe were significantly higher than WT under salt stressed condition ( Figure 4B). Moreover, the shoot length and fresh weight of GmPIP1;6-Oe transgenic soybean were significantly higher than that of WT plants under salt stressed conditions (Table 1). These results indicated that GmPIP1;6-Oe plants were more tolerant to salt stress than WT plants.
Overexpression of GmPIP1;6 increased photosynthesis and root water conductance in soybean under salt stress conditions The impact of GmPIP1;6 overexpression on net assimilation (A N ) under saturating light, stomata conductance (g s ) and transpiration rate (T r ) were measured using an infrared gas analyser (LI-6400) under normal and salt treatment conditions. Diurnal photosynthesis of soybean was measured every 2 hours in a light period from 8:00 AM to 4:00 PM. As expected, the A N , g s and T r showed diurnal changes and peaked at about 2:00 PM in both WT and GmPIP1;6-Oe plants ( Figure 5A-C). Under normal growth conditions, GmPIP1;6-Oe plants showed significantly higher A N , g s and T r than that of WT at all the time points measured ( Figure 6A-C). We then took the values of A N , g s and T r at 2:00 PM to compare WT and GmPIP1;6-Oe plants. Under normal growth conditions, A N , g s and T r was significantly increased in GmPIP1;6-Oe plants than that of WT plants ( Table 2). Interestingly, the A N , g s and T r was more than 2-fold higher in GmPIP1;6-Oe1 plants compared with that of WT under salt treatment. In the other transgenic line, GmPIP1;6-Oe2, the A N , g s and T r were 1.71, 1.75 and 2.1-fold higher than that of WT (Table 2). These results indicated that overexpression of GmPIP1;6 increased photosynthetic activity and stomatal conductance, especially under saline conditions. and root (R). Soybean seedlings were grown in nutrient solution. Total RNA was extracted from different tissues for qRT-PCR. (B) Ten-day-old soybean seedlings were treated with or without 100 mM NaCl in nutrient solution. RNA was extracted from the leaves of these seedlings at 6 hours, 12 hours, 1 day, 3 days, 5 days after treatment. All data are means of four biological replicates with error bars indicating SD. Expression level of treated plants was relative to control plants at each time point.
Under normal conditions, the substomatal concentration of CO 2 (C i ) of GmPIP1;6-Oe was lower than that of WT though no significant difference was observed. In contrast, the C i of GmPIP1;6-Oe was significantly lower than that of WT under salt treatment (Table 2, Figure 6D). This is in accordance with the higher rate of net photosynthesis of GmPIP1;6-Oe compared to WT plants under saline conditions. Instantaneous water use efficiency (IWUE = A/T) was significantly increased in GmPIP1;6-Oe plants under both normal and salt stress conditions compared with WT (Table 2). Changing stomata density and/or pore area will influence the g s and T r . Examination of the abaxial leaf surface revealed a significantly wider stomatal aperture in GmPIP1;6-Oe plants under both normal and salt stress conditions while the stomata density was not changed ( Table 2, Additional file 1: Figure S4A). As a result, the water loss rate was increased in the transgenic plants compared with WT plants (Additional file 1: Figure S4B).
We also measured root hydraulic conductance (L o ), normalized to root dry weight, in GmPIP1;6-Oe and WT plants. Interestingly the L o was similar between GmPIP1;6-Oe and WT plants irrigated with nutrient solution. However when irrigated with nutrient solution containing 50 mM NaCl, L o of WT plants decreased almost 50% while L o of GmPIP1;6-Oe plants remained unchanged ( Figure 5).

Overexpression of GmPIP1;6 affects Na uptake and exclusion of transgenic plants under salt stress
The sodium concentration of plants was analyzed under normal and salt stress conditions. Sodium (Na + ) concentration was similar between WT and GmPIP1;6-Oe plants in roots and leaves under normal conditions ( Figure 7A, B). Salt treatment increased Na + concentration in the roots and leaves of both WT and GmPIP1; 6-Oe plants. However, the Na + concentration was significantly lower in the leaves of GmPIP1;6-Oe plants than WT plants under salt stress ( Figure 7B, Additional file 1: Figure S5). We examined relative Na + exclusion of WT and GmPIP1;6-Oe plants after salt treatment revealing that the relative exclusion of Na + from the shoot of GmPIP1;6-Oe plants was higher relative to WT ( Figure 7C). Moreover, salt treatment induced the expression of GmNHX1 in the leaves and roots of WT but not in GmPIP1;6-Oe plants ( Figure 8).
Overexpression of GmPIP1;6 increased yields of soybean in the field Four independent GmPIP1;6-Oe lines, where two of these lines were the same as the ones used in the physiology experiments, were grown in field conditions for an entire growing season in two continuous years. Each year, the transgenic plants were compared with WT and null transgenic plants, which were segregated from heterozygous transgenic plants. Interestingly, GmPIP1;6-Oe plants have a significantly higher seed weight per plant and per 100 seeds than WT (Table 3). Other yield parameters were similar between WT and GmPIP1;6-Oe plants ( Table 3, Additional file 1: Figure S6A). A detailed analysis showed that the increased seed weight of GmPIP1;6-Oe plants was caused by large seed size (Additional file 1: Figure S6B, Table 3).

Discussion
Recently, 66 AQP genes were identified in soybean by a genome wide analysis [30]. The GmPIP subfamily contained 8 PIP1 genes and 14 PIP2 genes, all of which were predicted to localize on the plasma membrane. It is found that PIP2 aquaporins when expressed in Xenopus oocytes have high water permeability while PIP1 aquaporins do not. However, PIP1 aquaporins can work cooperatively with PIP2s in targeting to the plasma membrane and in water permeation as heterotetramers [20][21][22][23][24][25][26][27]. This is accordance with the fact that GmPIP1;6 protein fused with GFP localized on the plasma membrane ( Figure 1). GmPIP1;6 is the ortholog of AtPIP1;2, NtAQP1, HvPIP1;6/1;1 and TaAQP8 in Arabidopsis, tobacco, barley and wheat (Additional file 1: Figure S1). AtPIP1;2 and NtAQP1 play a key role in regulating root hydraulic conductance (L o ) in Arabidopsis and tobacco [17,43,44], respectively. In situ PCR showed that GmPIP1;6 was highly expressed in the stellar region of the root [42], which is similar as NtAQP1. Shoot topping rapidly decreased root hydraulic conductance (L o ) by 50% to 60%, which is correlated with the reduced expression of GmPIP1;6 in roots of soybean. Therefore, GmPIP1;6 was suggested to control the L o as AtPIP1;2 and NtAQP1.
Water stress caused by drought, salt or cold has a complex effect on the expression of AQP genes [36]. In summary, the expression of AQP genes could be divided into two stages. In the early stress response, the plant usually suppresses the expression of PIP genes, which is hypothesised to avoid water flow from the root to the soil when the soil water potential decreases [45,46]. After a few days of acclimation, the expression of PIP genes recovers or even increases and is correlated with increased hydraulic conductance [47][48][49]. The expression of GmPIP1;6 in both roots and leaves showed this two stage response under salt stress ( Figure 2B, Additional file 1: Figure S2), indicating GmPIP1;6 may be involved in the salt stress acclimation of soybean.
Overexpression of several PIP1 genes increased the hydraulic conductance and salt tolerance of the transgenic plants, such as NtAQP1, OsPIP1;1, TaAQP8 and Musa-PIP1;2 [17,37,40,50]. Here we show that GmPIP1;6 conferred salt tolerance, but also under normal conditions the overexpression resulted in higher growth and greater yield under field conditions compared to WT plants ( Figure 4A, Table 3). However, the mechanism of how these PIP1 genes can improve plant growth and salt tolerance is largely unknown, though a high K + /Na + ratio was mentioned with overexpression of TaAQP8 [37].
It is highly unlikely that GmPIP1;6 can transport Na + , therefore salt tolerance of transgenic GmPIP1;6 plants is more likely to occur through indirect mechanisms: First, improvement in water uptake by roots and leaf cell hydration, could improve energy capture and conversion by leaves. Greater energy availability in turn could improve Na + exclusion by roots and improve tissue Na + compartmentalization [1]. We compared the root L o of WT and GmPIP1;6-Oe plants under normal and salt stressed conditions (Figure 7). As expected, NaCl treatment decreased L o by 50% in WT plants. In contrast, GmPIP1;6-Oe plants maintained L o under salt stress conditions. Therefore, GmPIP1;6-Oe plants may have a higher water uptake activity than WT plants under  saline conditions. Secondly, Na + is transported to shoots in the transpiration stream through the xylem, but it can return to root via the phloem [43,[51][52][53]. Export of Na + from leaves in the phloem could conceivably help to maintain low salt concentration in the leaves and may be enhanced by greater water permeability in phloem cells. Also we show that net assimilation and gas exchange are enhanced in the GmPIP1;6-Oe plants, and especially so under saline conditions compared to WT. This would potentially translate to a higher capacity to exclude Na + via energy demanding salt exclusion mechanisms in the roots and the leaves. Thirdly, we measured the expression of GmNHX1 [54] to analyze the effect of Na + compartmentalization in the vacuole (Figure 8). Salt treatment induced the expression of GmNHX1 in the leaves and roots of WT but not in GmPIP1;6-Oe plants. This is accordance with the lower Na + concentration of GmPIP1;6-Oe plants and indicated that vacuole compartmentalization of GmPIP1;6-Oe plants was not necessarily enhanced. Another possibility that may account for reduced Na + transport to the shoot in the over expressing plants could be that more water flow occurs radially across roots via the cell-to-cell (membrane) pathway, as opposed to the apoplast  pathway. This would occur because of the higher activity of GmPIP1;6 in root membranes under salinity stress demonstrated by the higher root L o compared to WT. A higher proportion of water flow via the membrane pathway in roots would confer a greater degree of ion selectivity relative to flow in the apoplast pathway. Altogether, we clarified that overexpression of GmPIP1;6 increased soybean salt tolerance by maintaining water uptake ability and Na + exclusion.
In addition to function as a water channel, AtPIP1;2 and NtAQP1 may function to facilitate CO 2 transport and enhance photosynthesis by increasing the mesophyll conductance to CO 2 diffusion [55][56][57][58][59]. Overexpression of NtAQP1 in tobacco and tomato increased the A N , which resulted in increased WUE. The overexpression of NtAQP1 produced higher dry biomass and yield under normal irrigation and salt stressed conditions [17]. GmPIP1;6-Oe plants also exhibited higher A N , g s and IWUE than WT under both normal and saline conditions ( Figure 6, Table 2). However, the growth of GmPIP1;6-Oe plants was only enhanced under saline conditions compared to WT plants ( Figure 4A, Table 1). Whether GmPIP1;6 has a similar function as NtAQP1 to facilitate CO 2 diffusion across leaf cell membranes requires further research.
Importantly, GmPIP1;6-Oe plants showed higher yield in the field than WT because the seed weight and size of GmPIP1;6-Oe were increased (Table 3, Additional file 1: Figure S6). This may be reflecting the higher net assimilation, but also may indicate sink limitation of seed loading that could be enhanced by greater water permeability in the seed loading process [60]. In addition to being highly expressed in roots and stems, the transcripts of GmPIP1;6-Oe were abundant in flower and pod, which supports a role of GmPIP1;6-Oe in seed loading of assimilates via enhanced water permeability.

Conclusions
In this study, the function of GmPIP1;6 was analyzed by constitutive expressing in the soybean plants. The expression of GmPIP1;6 was influenced by salt stress. Overexpression of GmPIP1;6 improved salt tolerance of transgenic plants by increasing water transport, photosynthesis and Na + exclusion. Moreover, the yield of GmPIP1;6 overexpression plants was improved in the field indicating the potential of GmPIP1;6 in genetic engineering of soybean.

Plant materials, growth conditions and treatments
Soybean cultivar Williams 82 was used for all physiological experiments and soybean transformation. Seeds were germinated in nursery pots with sand. Five days after germination, the seedlings grown uniformly were transferred into pots with nutrient solution or soil. 1/2 Hoagland solution was used for hydroponic culture containing 2.5 mM KNO 3  Ten-day-old seedlings were transferred into nutrient solution with or without 100 mM NaCl. The nutrient solution was changed every two days. In the soil experiments, plants were irrigated nutrient solution every three days.

Subcellular localization of GmPIP1;6
Full length cDNA of GmPIP1;6 without stop code was amplified via PCR using the primers in Supplementary Additional file 2: Table S1. The PCR product was cloned into vector pCAMBIA1302 under the control of the CaMV 35S promoter. The resulting construct (pCAM-BIA1302:GmPIP1;6) placed GmPIP1;6 in-frame, upstream of the sGFP. Plasmids DNA of pCAMBIA1302: GmPIP1;6 and CD3-1007 (AtPIP2A::mCherry fusion) was mix with 50 μl gold particles and bombarded into onion inner epidermal cells using the Biolistic PDS-1000/ He particle delivery system (BIO-RAD). Fluorescence was observed by confocal laser scanning microscopy (LSM700; Carl Zeiss) after incubation at 25°C for 16-18 h on MS medium in dark.

Construction of transgenic plants
Full-length cDNA of GmPIP1;6 was amplified by PCR with cDNA of Williams 82 and ligated into pMD-18 T vector (Takara). After sequencing, the correct GmPIP1;6 was digested from pMD-18 T vector using BamHI and XbaI restriction enzymes. GmPIP1;6 was then cloned into binary plasmid pTF101-35S which was modified by introducing CaMV 35S promoter and nos terminator into pTF101. The vector was transformed into Williams 82 via Agrobacterium tumefaciens media soybean cotyledon node transformation system as described [61].

RNA extraction
Total RNA was isolated from tissues of soybean cultivar Williams 82 using TRIzol reagent (Invitrogen, Carlsbad, CA) according the manufacturer's instruction. 50 mg soybean tissues with three bilogical replicate were quickly harvested, frozen in liquid nitrogen and stored at -80°C.
Contaminating DNA was removed with DNaseI treatment for 20 min at 25°C (Takara), and RNA was stored at -80°C. Total RNA was quantified with nanodrop.

Semi-quantitative RT-PCR and quantitative real-time PCR
First-strand cDNAs were synthesized from total RNA using SuperScript II reverse transcriptase (Invitrogen). Semi-quantitative RT-PCR was performed using a pair of gene-specific primers. The housekeeping gene GmACTIN was used as an internal control. Quantitative real-time PCR was performed using a SYBR Green I on a Light Cycler 480 II machine (Roche Diagnostics), according to the manufacturer's instructions. The amplification program for SYBR Green I was performed at 94°C for 10 sec, 58°C for 10 sec and 72°C for 10 sec. Triplicate quantitative assays were performed on each cDNA sample. The relative level of expression was calculated using the formula 2 -△(△cp) . All primers used for RT-PCR are given in Supplementary Additional file 2: Table S1.

Gas-exchange measurements
Homozygous lines were selected from the T 2 generation of transgenic GmPIP1;6 overexpression plants and used for the physiology experiment. A N , g s , T r and C i were recorded in GmPIP1;6 overexpression and control plants in green house on fully expanded leaves, using an Li-6400 portable gas-exchange system (LI-COR). All measurements were conducted between 8:00 AM and 4:00 PM. Photosynthesis The content of Na + in roots and leaves. Ten-day-old WT and GmPIP1;6-Oe transgenic soybean plants were treated with or without 100 mM NaCl for 7 days. The roots or leaves of these seedlings were sampled for measurement. (C) Relative Na exclusion. Data are means of four biological replicates with error bars indicating SD. Asterisks indicate a significant difference between the WT and the transgenic lines (**P < 0.01). DW, dry weight.
was induced in saturating light (1000 μmol m −2 s −1 ) with 400 μmol mol −1 CO 2 surrounding the leaf. The leaf-to-air VPD was kept at around 2 to 4 kPa and leaf temperature was approximately 30°C (ambient temperature) during all measurements. For each treatment, there were four biological replicates.

Stomata aperture and density
Epidermis of soybean abaxial leaf was separation by forceps. All samples were collected around 2:00 PM (at peak transpiration). Counting and photographing were performed with a bright-field microscope (80i; Nikon) mounted with a camera. Stomata images were later  analyzed to count the number per 0.1 mm 2 area and determine aperture using the microscope software (NIS elements) measurement tool. A microscopic ruler was used for the size calibration.

Determining Na + concentration
Leaves and root from 17-day old transgenic lines and WT were sampled and dried at 80°C for 3 days. 50 mg of the material was weighed and dissolved with 3 ml of nitric acid and 2 ml of H 2 O 2 (30%). The digested samples were diluted to a total volume of 50 ml with ultrapure water and transferred into new tubes before analysis by using an inductively coupled plasma-mass spectrometer (ICP-MS, ELAN DRC-e).
To analyze the relative Na + exclusion, Ten-day-old WT and GmPIP1;6-Oe transgenic soybean plants in hydroponics were treated with 100 mM NaCl for 7 days. Soybean plants after treatment were transferred into narrow neck flask individually, which filled with same volume of normal nutrient solution, and cultured for 24 hours. Na + concentration of solution was measured by ICP-MS described above. The relative Na + exclusion was calculated by the formula: relative Na + exclusion = (Na + concentration of solution × Volume of solution)/(Na + concentration of shoot × DW of shoot).

Root hydraulic conductance
Root hydraulic conductance were measured with a hydraulic conductance flow meter (HCFM) (Dynamax, Houston, TX, USA) as described in Vandeleur [23]. 5week-old potted soybean plants grown in greenhouse. 1 day before root hydraulic conductance were measured, control plants were irrigated with normal nutrient solution and treatment plants were irrigated with nutrient solution containing 50 mM NaCl. Measurements were made between 10:00 AM to 12:00 AM. Hydraulic conductance (L o ) was normalized by dividing total root dry weight. The soil was washed from the roots, and baked at 80°C for 3 days.

Additional files
Additional file 1: Figure S1. Phylogenetic analysis of GmPIP1s and other AQPs by MEGA 5.04. Figure S2. Expression pattern of GmPIP1;6 under NaCl treatment in leaves relative to control.. Ten-day-old soybean seedlings were treated with or without 100 mM NaCl in nutrient solution. RNA was extracted from the leaves of these seedlings at 6 hours, 12 hours, 1 day, 3 days, 5 days after treatment. All data are means of four biological replicates with error bars indicating SD. Expression level of treated plants was relative to control plants at each time point. Figure S3. Detection of transgenic soybean with herbicide Liberty. One half of the leaf was painted with 135 mg/L Liberty W , the bar-containing positive transgenic soybean leaves were green and the negative ones were yellow and wilted. Treated leaves were labeled with marker pen which can be seen in the images. Figure S4. Measurement of stomata aperture water loss rate. Ten-day-old WT and GmPIP1;6 overexpression plants in nutrient solution were treated with or without 100 mM NaCl for 3 days. Leaves were sampled at 2:00 PM to observe the abaxial leaf surface with microscope and measured stomata aperture. Bar = 100 nm. Figure S5. Distribution of intracellular Na + in WT and transgenic soybean plants. Ten-day-old WT and GmPIP1;6 overexpression plants in nutrient solution were treated with or without 100 mM NaCl for 2 days. Samples of leaves were sliced and stained with CoroNa-Green at 2:00 PM and observed with a confocal microscope. Bar = 100 μm. Figure S6. Phenotypic characterization of GmPIP1;6 overexpressing soybean seeds. Mature dried seeds from WT and GmPIP1;6-Oe transgenic soybean plants were recorded. Bar = 3 cm.
Additional file 2: Table S1. Primers used in this study.