Introduction

Aquaporins, which belong to the highly conserved major intrinsic protein (MIP) family, play an essential role in plant water transport 1, 2, 3. Various molecular and cellular mechanisms underlying aquaporin regulation in a challenging environment has been well documented. These mechanisms include three basic pathways: changes in aquaporin gene transcription rate and protein abundance, the subcellular localization of aquaporin and the gating of water channels 4. The expression of aquaporins, at mRNA level, has been the focus of many published studies.

One subgroup of MIPs, the plasma membrane intrinsic proteins (PIPs), which are localized in cellular plasma membrane, has been extensively investigated at the transcriptional level under water deficit/drought treatment. Different PIP genes exhibit different expression patterns. Water deficit was shown to enhance the expression of rice OsPIP1a and OsPIP2a 5, tomato TRAMP 6, Nicotiana excelsior NeMip2 and NeMip3 7 and radish RsPIP2 8, but reduced Nicotiana glauca NgMIP4 9 and rice RWC1 10. In a recent study, both up- and down-regulation of PIP expression were observed under water deficit induced by mannitol in Arabidopsis 11. The expression of AtPIP1;5, AtPIP2;2, AtPIP2;3 and AtPIP2;6 was down-regulated, whereas AtPIP1;3, AtPIP1;4, AtPIP2;1 and AtPIP2;5 was up-regulated in both the roots and aerial parts of the plants. The expression level of AtPIP1;1 increased in the aerial parts during the 12 h-drought treatment, but decreased in roots. In another study, similar expression regulation in AtPIPs was observed in response to drought 12. However, due to different experimental conditions, some PIPs were different regulated. For example, AtPIP1;3 and AtPIP2;1 were down-regulated by drought. The above results illustrate that the PIP expression pattern under water deficit is a complicated response and the mechanisms underlying the regulation of PIP expression under water deficit are still elusive.

Abscisic acid (ABA) accumulation in stressed plants was found to be protective against drought damage, causing stomata closure that reduces water loss via transpiration 13, 14, and increasing hydraulic conductance, promoting water movement from roots to leaves 15, 16. In addition, ABA plays an important role in mediating the gene expression in response to water deficit 17, 18. It was suggested that water deficit triggered the up-regulation of PIP genes in ABA-dependent and ABA-independent manner 11, 19, 20. Verification of a similar mechanism in rice would contribute to a greater understanding of the regulation of PIPs.

In our previous study, upland rice and lowland rice showed different mechanisms of drought resistance and diverse expression patterns of one PIP gene, RWC3 (OsPIP1;3) under water deficit 21. In the present study, upland rice and lowland rice were initially observed to have different PIP expression patterns at both mRNA and protein levels and had different ABA contents as a result of 20% polyethylene glycol (PEG). Based on the comparative analysis of the expression of the whole PIP subgroup at the mRNA level between upland rice and lowland rice treated with PEG or exogenous ABA, the potential role of ABA in the regulation of PIP expression is proposed.

Materials and Methods

Plant material and growth condition

Physiological experiments were performed using lowland rice (Oryza sativa L. spp. japonica cv. Xiushui 63) and upland rice (Oryza sativa L. spp. indica cv. Zhonghan 3). After germinating in the dark for 3 d at 28 °C, all seedlings were grown in nutrient solution 22 at a photon flux density of 300–350 mol m−2 s−1, 60–80% relative humidity, 12 h/12 h day-night cycle at 28 °C in a phytotron. Four-week-old plants were used for the water deficit treatments, the nutrient solution contained 20% PEG 6000 giving −0.82 MPa of stress determined by a Vapor Pressure Osmometer 5520 (Wescor, USA). For the ABA treatment, 60 μmol L−1 ABA solution was prepared by dissolving the crystals in 100 μL of ethanol and then adding it to the nutrient solution. The final concentration of ethanol in the nutrient was 100 μL/L. The entire experiments were repeated three times.

Leaf water status and osmotic potential

Leaf water status was assayed by measuring the fresh weight (FW) and dry weight (DW) of the leaves. DW was obtained after drying the leaves at 100 °C for 5 h. Leaf water content (WC) was calculated using the following formula:

WC (%)=(FW–DW)×100/FW

The osmotic potential was measured with a Vapor Pressure Osmometer 5520 (Wescor, USA). Leaves of PEG-treated and control seedlings were harvested and put in a syringe and frozen in a freezer overnight. The leaf juice was squeezed and then the osmotic potential was measured.

Preparation of plasma membranes and immunoblot analysis

Plasma membrane fractions were prepared by aqueous two-phase partitioning as described by Ohshima et al. 23. Protein concentration was measured according to Bradford 24. Western blots were carried out using anti-PIP antibody 23 following the standard method according to Sambrook et al. 25. The PIP antibody was raised against N-terminal common sequence “KDYNEPPPAPLFEPGELSSWS”, among six radish PIPs including three PIP1s and three PIP2s. The antibody has broader specificity and was able to detect the PIPs in graptopetalum and radish 23. We compared all the rice PIP proteins and found high conservation in this region. Standard procedures for SDS-PAGE in 12% polyacrylamide gels were used and proteins were transferred to nitrocellulose. The anti-PIP antibodies 23 were 1:500 diluted for immunoblotting. The primary antibodies were detected with goat anti-rabbit IgG-HRP (1:1000), and the signal was detected by chemiluminescence method (West Pico West Blotting, Pierce, IL, USA).

Extraction of RNA and first-strand cDNA synthesis

Total RNA was extracted from roots and leaves using RNArose reagent (Watson, China). The residual genomic DNA was removed by DNA-freeTM Kit (Ambion, USA). The concentration of RNA was accurately quantified by spectrophotometric measurements. The cDNA was synthesized using oligo-(dT)18 primer and ReverTra Ace M-MLV RTase (Toyobo, Osaka, Japan) according to the manufacture's recommendation. The cDNA mixtures were diluted to 60 μL with sterile ddH2O and 1.8 μL were used as template for real-time PCR in a total volume of 20 μL.

Primers and TaqMan-MGB probes

The primers and TaqMan-MGB probes were designed using Primer Express 2.0 software (Applied Biosystems-Perkin-Elmer, USA) according to the sequences of rice PIP genes from Sakurai et al. 26. Ubiquitin (NCBI accession number: D12629) was used as internal control. The sequence of PCR primers and TaqMan-MGB probes are shown in Table 1. All the primers and probes were synthesized by Shanghai GeneCore Company (Shanghai, China). The probes were labeled with the fluorescent reporter dye 6-carboxy-fluoroscein on the 5′-end, and with the fluorescent quencher dye 6-carboxytetramethylrhodamine (TAMRA) and MGB on the 3′-end. They formed extremely stable duplexes with single-stranded DNA targets, allowing shorter probes to be designed. All PCR primers were designed to produce the PCR products of about 70 bp in length.

Table 1 Sequence of the primers and TaqMan-MGB probes for real-time RT-PCR
Full size table

Real-time PCR

TaqMan-MGB real-time PCR was performed in PRISM 7700 (ABI, USA) using Real-time PCR Master Mix (Toyobo, Japan). The reaction mixture (20 μL) contained 1.8 μL of diluted cDNA, 4 pmol of each primer, 4 pmol TaqMan-MGB probe and appropriate amounts of other components as recommended by the manufacturer (Toyobo, Japan). The real-time PCR was programmed for 5 min at 95 °C, 50 cycles of 20 s at 94 °C and 30 s at 60 °C. Each sample was quantified in triplicate. For control reactions, either no sample was added or RNA alone was added without reverse transcription to test if the RNA sample contained genomic DNA contamination. The housekeeping gene ubiquitin was used as internal standard to normalize target genes 27. To generate standard curves for the PIPs and UBI, serial dilutions of the conventional RT-PCR products were made in steps of 1:10.

ABA determination

For extraction of ABA, leaves or roots were homogenized with a mortar and pestle in an extraction solution of 80% methanol. To remove plant pigment and other non-polar compounds, which could interfere in the immunoassay, extracts were first passed through a polyvinylpyrrolidone column and C 18 cartridges. The elutants were concentrated to dryness by vacuum-evaporation and re-suspended in tris-buffered saline before enzyme-linked immunosorbent assay (ELISA). ABA content was quantified by ELISA 28. The ABA immunoassay detection kit was purchased from the Nanjing Agriculture University.

Results

Water content and osmotic potential of leaves in upland rice and lowland rice during water deficit

Application of 20% PEG to the root systems of both upland rice and lowland rice resulted in water deficit. The response of seedling of both upland rice and lowland rice towards water deficit was compared by analyzing the water content and osmotic potential of leaves.

Treatment with 20% PEG rapidly and significantly decreased the water content and caused about 3% water loss in leaves of upland rice (Figure 1A). In contrast, the water content of leaves of lowland rice reduces slightly and slowly, and only caused 1% water loss (Figure 1A). During water deficit, the osmotic potential of leaves reduced from −1 MPa to −1.34 MPa and from −1.17 MPa to −1.45 MPa in upland rice and lowland rice, respectively (Figure 1B). Within 2 h of PEG treatment, the osmotic potential of the leaves was strongly reduced in upland rice, but not significantly decreased in lowland rice. After the 10 h of PEG treatment, the osmotic potential of leaves reduced 34% and 24% in upland rice and lowland rice, respectively. These results suggested that lowland rice had better water status than upland rice when treated with 20% PEG.

Figure 1
figure 1

Effect of water deficit induced by 20% PEG 6000 on the water content (A) and osmotic potential (B) of leaves in upland rice and lowland rice. Each point is the mean of measurements (n=3), and the bars represent the standard error of the mean.

Full size image

PIP protein abundance in upland and lowland rice during water deficit

The variations of PIP protein accumulation in upland rice and lowland rice under water deficit were investigated using Western blots. The antibodies used were raised against the common N-terminal peptide of PIPs. Indeed, only one band was detected at 30 kDa in upland rice and lowland rice. The antibodies did not cross-react with other proteins. As shown in Figure 2, upland rice had higher PIP protein abundance than lowland rice both in non-stressed conditions and after 10 h of 20% PEG treatment. The PEG treatment markedly enhanced the PIP protein abundance in the roots of both rice types and in leaves of upland rice. The PIP protein abundance remained approximately constant in leaves of lowland rice after 10 h of the water deficit treatment.

Figure 2
figure 2

Immunoblot analysis of plasma membrane protein in upland rice (U) and lowland rice (L) under the water deficit (+ PEG) for 10 h or non-stressed condition (− PEG). Plasma membrane proteins (10 μg) were loaded in each lane.

Full size image

Abundance of the PIP transcripts in upland rice and lowland rice under non-stressed condition

The expression of all 11 PIP, under non-stressed conditions, was determined with real-time RT-PCR in upland rice and lowland rice. Relative transcript levels of the PIP genes were normalized by comparing them to the expression of a housekeeper gene, ubiquitin. As shown in Figure 3, PIP genes were generally transcribed at similar levels in upland rice and lowland rice, but with higher expression levels of OsPIP2;5 and OsPIP2;6 in upland rice roots. For most of the PIP genes, roots and leaves had characteristic and distinct expression levels. Many, such as OsPIP1;2, OsPIP1;3, OsPIP2;1, OsPIP2;3, OsPIP2;4, OsPIP2;5 and OsPIP2;6, were expressed abundantly in roots, yet OSPIP2;7 was more highly expressed in leaves. Furthermore, the expression of OsPIP2;7 in roots and OsPIP2;8 were extremely low, which are 10−4 and 10−3 to that of OsPIP2;7 in leaves, respectively. However, very weak expression of OsPIP2;8 could be detected by conventional RT-PCR 26. Expression of OsPIP2;7 in roots and OsPIP2;8 upon PEG and ABA treatment was not analyzed in the next experiments.

Figure 3
figure 3

The transcript levels of the 11 PIPs in roots and leaves of upland rice and lowland rice. Total RNA was extracted from the roots and leaves under non-stressed condition. The transcript levels of each PIP were estimated by real-time RT-PCR and normalized to that of ubiquitin (see Materials and Methods). The values are means ± S.D. of three replicates.

Full size image

Expression profile of the PIP genes in upland rice and lowland rice during water deficit

Figure 4 shows the expression patterns of PIP genes in upland rice and lowland rice when subjected to water deficit induced by 20% PEG treatment. The expression patterns of PIP genes changed in different ways in response to water deficit. The transcript level of OsPIP1;2 was up-regulated more than 10-fold, and that of OsPIP1;3, OsPIP2;1 and OsPIP2;5 increased over six-fold in roots of upland rice. In contrast, no obvious enhancement in expression of these genes was observed in lowland rice under water deficit. Other genes, such as OsPIP1;1, and OsPIP2;6 in roots and OsPIP1;1, OsPIP2;1, OsPIP2;2, OsPIP2;3, OsPIP2;6 and OsPIP2;7 in leaves, were down-regulated in both upland rice and lowland rice in response to the PEG treatment.

Figure 4
figure 4

Relative expression of the PIP genes in upland rice (black column) and lowland rice (white column) under water deficit induced by 20% PEG 6000. The transcript levels of each PIP in the roots (A) and leaves (B) of treated plants were plotted as the relative expression (fold) of the non-stressed control plants at 2, 4 and 6 h. Values are means ± S.D. of three replicates.

Full size image

ABA accumulation in upland rice and lowland rice during water deficit

ABA acts in response to various abiotic stresses, and serves as a crucial component in the drought signaling pathway. As different responses to water stress were observed, e.g. water content and osmotic potential of leaves, PIP expression and protein levels in upland rice and lowland rice, it was of interest to determine endogenous ABA content in parallel to these changes. As shown in Figure 5, upland rice had a higher ABA level than lowland rice, especially in roots, during the 20% PEG treatment. The data clearly show that ABA increased to a maximum after 2 h and 4 h of water deficit in roots and leaves of upland rice with increases of 16.7- and 11.5-fold, respectively. After 10 h of water deficit, the root ABA content decreased to approximately the initial value, whereas it remained high (6.7-fold) in leaves of upland rice. In contrast, lowland rice ABA content did not change throughout the duration of the water deficit. Thus, ABA levels changed in parallel to the changes in expression of some of the PIP genes, such as OsPIP1;2, OsPIP1;3, OsPIP2;1 and OsPIP2;5 in roots and OsPIP1;2, OsPIP1;3 in leaves of upland rice in response to water deficit.

Figure 5
figure 5

Change in ABA content in roots (A) and leaves (B) of upland rice and lowland rice under water deficit induced by 20% PEG 6000. Each point is the mean of measurements (n=3), and the bars represent the standard error of the mean.

Full size image

Effect of exogenous ABA on the expression of PIP genes in upland rice and lowland rice

The fact that water deficit enhanced the ABA level and the transcription of some PIP genes in upland rice (Figures 4 and 5) supports the interesting possibility that PIP expression is regulated by ABA. In both upland rice and lowland rice, no visible difference in growth phenotype was observed between untreated and 60 μmol L−1 ABA-pretreated seedlings. The transcripts of OsPIP1;2, OsPIP2;5 and OsPIP2;6 in roots and OsPIP1;2, OsPIP2;4 and OsPIP2;6 in leaves of upland rice were up-regulated, whereas only OsPIP1;2 and OsPIP2;6 were enhanced in the roots of lowland rice (Figure 6). The up-regulation of OsPIP1;2, and OsPIP2;5 in roots as well as OsPIP1;2, in leaves of upland rice was reminiscent of the response to PEG treatment (Figures 4 and 6). An additional observation was that OsPIP2;3 and OsPIP2;4 in roots and OsPIP2;2, OsPIP2;3 and OsPIP2;7 in leaves were down-regulated in both upland rice and lowland rice during ABA treatment.

Figure 6
figure 6

Effect of exogenously applied ABA on the expression of the PIP genes in upland rice (black column) and lowland rice (white column). The transcript levels of each PIP in roots (A) and leaves (B) of treated plants were plotted as the relative expression (fold) of the non-stressed control plants at 2, 4 and 6 h. Values are means ± S.D. of three replicates.

Full size image

Discussions

Different physiological response to water stress in upland rice and lowland rice

Different physiological responses to water stress were observed in upland rice and lowland rice. Young leaf rolling was observed in the upland rice, but not in the leaves of lowland rice 21. We also found the leaf water content and osmotic potential decreased more upland rice than lowland rice (Figure 1A and 1B). Furthermore, long-term (7–8 d) and mild (−0.45 MPa) water stress also more strongly decreased the water content and osmotic potential in young upland rice (7–8 d after germination) than young lowland rice (data not shown). To reduce the efflux of water and regain turgor, lower osmotic potential in leaves is beneficial for upland rice. It was known that synthesis of compatible solute is involved in the osmoregulation 29. Our experiment showed that PEG strongly induced the accumulation of proline in leaves of upland rice but to a lesser degree in lowland rice (data not shown). These results suggested that lowland rice had better water status under PEG treatment.

PIP expression in upland rice and lowland rice under normal condition

At the protein level, there was higher PIP accumulation in leaves than roots (Figure 2); however, this was not reflected in the mRNA level measured using RT-PCR (Figure 3). Seven PIP genes, i.e. OsPIP1;2, OsPIP1;3, OsPIP2;1, OsPIP2;3, OsPIP2;4, OsPIP2;5 and OsPIP2;6, were transcribed at higher rates in roots than leaves and only one PIP gene, OsPIP2;7, followed the pattern of mRNA accumulation of higher levels in leaves than roots. Inconsistency between aquaporin expression at the mRNA level and protein accumulation was also noted in a previous study 8, 30. These authors postulated that post-transcriptional mechanisms resulted in this, e.g. the translation rate and degradation of PIP proteins could differentially influence PIP protein levels independently from transcription levels. Although the anti-PIP antibodies were raised against the common sequence of PIP1 and PIP2 subfamily, it is possible that some plasma membrane aquaporins were not detected, resulting in inconsistency of aquaporin expression between mRNA and protein level, due to low sequence identity in the antibody-recognizing site. Better antibodies that could discriminate the PIP1s and PIP2s in rice would be helpful in the future studies.

Drought-inducible PIP genes were ABA-dependent or ABA-independent

Tremendous efforts have concentrated on the expression of PIP gene in responses to drought/water deficit, and some of those were up-regulated under the stressed condition 31. In this research, based on the integrated expression profile of PIP gene family in upland rice and lowland rice during water deficit, some PIP genes, i.e. OsPIP1;2, OsPIP1;3, OsPIP2;1 and OsPIP2;5, were also induced in upland rice (Figure 4). ABA is known to serve as a crucial component in the drought signaling pathway in plants, and induces expression of many PIPs in a number of species 1, 11, 19. In these experiments, we also observed the induction of OsPIP1;2 and OsPIP2;5 by ABA (Figure 6). Thus, ABA signaling was presumed to mediate the regulation of expression of these PIPs in upland rice during water deficit. The hypothesis was supported by the fact that an increase in ABA accumulated was parallel to some PIPs up-regulation under water deficit (Figure 5).

To date, two ABA-dependent signaling pathways that function in the activation of drought-inducible genes under water deficit have been identified 18, 32. One is the, bZIP/ABRE pathway and the other is mediated by the MYC/MYB transcription factors. Coincidently, when we analyzed the promoter sequence (from −1000 bp to initiation codon) of rice PIP genes using the software Plant Care (http://intra.psb.ugent.be:8080/PlantCARE/), the ABRE (PyACGTGGC) motif was found in OsPIP1;2 and the potential MYB recognition sequence (AACCAAA) was found in OsPIP2;5. Therefore, it was proposed that ABA regulation of OsPIP1;2 was via the bZIP/ABRE pathway, and that of OsPIP2;5 might be mediated by the MYC/MYB transcription factors.

However, the OsPIP1;3 and OsPIP2;1 genes in roots and OsPIP1;3 in leaves of upland rice were up-regulated when treated with PEG, but did not respond to ABA treatment, indicating drought-inducible expression of OsPIP1;3 and OsPIP2;1 were ABA-independent. These data are consistent with the response of some PIP genes in other species, such as Cp-PIPa6, Cp-PIPa7 and Cp-PIPc in Craterostigma plantagineum 19, PIP1;3 and PIP2;5 in Arabidopsis 11 and TRAMP in tomato 6. It has been shown that some of the drought-responsive genes are induced in an ABA-independent pathway, and the dehydration-responsive-element/C-repeat (DRE/CRT) cis-element was found in the promoter region of these genes 32. Analysis of the promoter sequences of OsPIP1;3 and OsPIP2;1 (http://intra.psb.ugent.be:8080/PlantCARE/) resulted in the identification of two putative DRE/CRT boxes (TACCGACAT) in the OsPIP1;3 promoter at −239 bp and −721 bp, respectively. No DRE/CRT was found in the OsPIP2;1 promoter region from −2000 bp to the initiation codon. Thus, the up-regulation of OsPIP1;3 under water deficit was probably via the DRE/CRT-dependent pathway.

Roles of PIPs in rice drought avoidance

Although important progress has been made in understanding the role of aquaporins in plant adaptations to dehydration condition, opinions are still in conflict. One point of view is that an increase in aquaporin level results in higher membrane water permeability and facilitates water transport 6, 7, 33, 34. Overexpression of BnPIP1 in transgenic tobacco plants also increased tolerance to water stress at the whole plant level, and partial silencing of BnPIP1 resulted in reduced water uptake and decreased tolerance to water stress 34. A second point of view is that plants decrease their membrane water permeability to avoid excessive loss of water by down-regulation of some aquaporins under water deficit 9, 35, 36. For instance, Aharon et al. 36 constitutively over-expressed the Arabidopsis plasma membrane aquaporin PIP1b in tobacco, and these transgenic plants wilted more quickly during water stress.

In a previous study, upland rice and lowland rice were characterized as drought avoidance and drought tolerance, respectively 21. So the comparison of upland rice and lowland rice appears to be a paradigm for studying the role of aquaporins in drought resistance. The fact that the upland rice had deeper, wider roots (data not shown) as well as higher levels of PIP protein than lowland rice (Figure 2) might be beneficial to absorb more, deeper underground water in order to supply adequate water for aerial parts when drought occurs. At the mRNA level, we observed that upland rice up-regulated OsPIP1;2, OsPIP1;3, OsPIP2;1 and OsPIP2;5 in roots and OsPIP1;2 and OsPIP1;3 in leaves (Figure 4), and that this matched the changes in the amount of total PIP protein. On the other hand, the reduced transcript levels of OsPIP1;3 and OsPIP2;5 in roots of lowland rice (Figure 3) imply that a totally different drought tolerance mechanism may operate in this rice type. The reduction of aquaporins in lowland rice may inhibit membrane water permeability, thereby avoiding excessive loss of cellular water under water deficit. It appears that these seemingly opposite aquaporin gene regulation responses in upland rice and lowland rice, under water deficit, reflect different adaptive mechanisms for dealing with water deficit.

Subsequently, a question arose as to why the water deficit with 20% PEG treatment reduced the expression of these PIP genes (for instance, OsPIP1;3 and OsPIP2;5 in roots) in lowland rice, whereas enhanced it in upland rice. This probably could be explained by the different drought-sensitive between upland rice and lowland, although other mechanisms could not be ruled out. Upland rice was drought-sensitive and lowland rice was drought-tolerant crop under PEG treatment 21. Namely, 20% PEG treatment was a severe water stress for upland rice, whereas only a mild stress for lowland rice. It was proposed that mild water stress caused the down-regulation of PIP expression, whereas more severe drought caused PIP up-regulation 7.

Some PIP genes in our experiments (OsPIP1;1, OsPIP2;6 in roots and OsPIP2;3, OsPIP2;7 in leaves) were down-regulated in upland rice (Figure 4). We also observed that the transcript level of OsPIP2;4 in roots as well as some PIP genes in leaves of lowland rice remained unchanged under water deficit (Figure 3). These observations are not explained by the above-mentioned points, but in Arabidopsis, up- or down-regulation of PIP genes under water deficit was also described 11, 12. This diversity of PIP expression patterns in plants implies that the role of aquaporins in drought resistance is influenced by many factors. Owing to the complexity of the response of PIPs, a comprehensive understanding of these proteins in response to drought resistance remains a challenge to be explained by future research.