Transpiration from shoots triggers diurnal changes in root aquaporin expression

Authors

  • JUNKO SAKURAI-ISHIKAWA,

    Corresponding author
    1. Climate Change Research Team, National Agricultural Research Center for Tohoku Region, 4 Akahira Shimo-kuriyagawa, Morioka 020-0198, Japan
      J. Sakurai-Ishikawa. E-mail: junkoi@affrc.go.jp
    Search for more papers by this author
  • MARI MURAI-HATANO,

    1. Climate Change Research Team, National Agricultural Research Center for Tohoku Region, 4 Akahira Shimo-kuriyagawa, Morioka 020-0198, Japan
    Search for more papers by this author
  • HIDEHIRO HAYASHI,

    1. Climate Change Research Team, National Agricultural Research Center for Tohoku Region, 4 Akahira Shimo-kuriyagawa, Morioka 020-0198, Japan
    Search for more papers by this author
  • ARIFA AHAMED,

    1. Climate Change Research Team, National Agricultural Research Center for Tohoku Region, 4 Akahira Shimo-kuriyagawa, Morioka 020-0198, Japan
    Search for more papers by this author
  • KEIKO FUKUSHI,

    1. Climate Change Research Team, National Agricultural Research Center for Tohoku Region, 4 Akahira Shimo-kuriyagawa, Morioka 020-0198, Japan
    Search for more papers by this author
  • TADASHI MATSUMOTO,

    1. Graduate School of Bioresource Sciences, Akita Prefectural University, Shimoshinjo, Akita 010-0195, Japan
    Search for more papers by this author
  • YOSHICHIKA KITAGAWA

    1. Graduate School of Bioresource Sciences, Akita Prefectural University, Shimoshinjo, Akita 010-0195, Japan
    Search for more papers by this author

  • Financial source: This work was supported, in part, by Grant-in-Aid for Scientific Research (KAKENHI) to J.S.-I. (no. 21780235) and from the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN) to M.M.-H.

J. Sakurai-Ishikawa. E-mail: junkoi@affrc.go.jp

ABSTRACT

Root hydraulic conductivity (Lpr) and aquaporin amounts change diurnally. Previously, these changes were considered to be spontaneously driven by a circadian rhythm. Here, we evaluated the new hypothesis that diurnal changes could be triggered and enhanced by transpirational demand from shoots. When rice plants were grown under a 12 h light/12 h dark regime, Lpr was low in the dark and high in the light period. Root aquaporin mRNA levels also changed diurnally, but the amplitudes differed among aquaporin isoforms. Aquaporins, such as OsPIP2;1, showed moderate changes, whereas root-specific aquaporins, such as OsPIP2;5, showed temporal and dramatic induction around 2 h after light initiation. When darkness was extended for 12 h after the usual dark period, no such induction was observed. Furthermore, plants under 100% relative humidity (RH) showed no induction even in the presence of light. These results suggest that transpirational demand triggers a dramatic increase in gene expressions such as OsPIP2;5. Immunocytochemistry showed that OsPIP2;5 accumulated on the proximal end of the endodermis and of the cell surface around xylem. The strong induction by transpirational demand and the polar localization suggest that OsPIP2;5 contributes to fine adjustment of radial water transport in roots to sustain high Lpr during the day.

INTRODUCTION

Transpiration is the dominant process that controls and influences plant water relations because of the large volume of water involved (Kramer & Boyer 1995). Under high transpirational demand during the day, the main pathway of root water transport is generally considered to be apoplastic (Steudle & Peterson 1998). However, in radial water movement from the root surface to the central cylinder, the apoplastic pathway is restricted at the exodermis and endodermis by apoplastic barriers such as Casparian strips and suberin lamellae in the cell wall of these cells. Therefore, the cell-to-cell pathway plays a crucial role in these cells. Water channels called aquaporins are distributed in diverse plant tissues and regulate water permeability of cell membranes (Maurel et al. 2008). In roots, aquaporins accumulate at high levels in both endodermal and exodermal cells (Hachez et al. 2006; Sakurai et al. 2008). Aquaporins are responsible for the cell-to-cell water movement in these cells, and also determine root hydraulic conductivity (Lpr) (Maurel et al. 2008).

Plant aquaporins constitute a large protein family. Arabidopsis, maize and rice have 35, 33 and 33 aquaporins, respectively (Chaumont et al. 2001; Johanson et al. 2001; Sakurai et al. 2005). Plant aquaporins are usually divided into four subfamilies: plasma membrane-localized aquaporins (PIPs), tonoplast-localized aquaporins (TIPs), nodulin 26-like aquaporins (NIPs), and small and basic intrinsic proteins (SIPs). Recently, another subfamily, X intrinsic proteins (XIPs), was found in several plants, although monocots do not have XIPs (Danielson & Johanson 2008). PIPs are further divided into two groups (PIP1- and PIP2-type aquaporins) based on their sequence similarities. Among the plant aquaporins, PIP2-type aquaporins and several TIPs have been shown to have high water transport activity (Chaumont et al. 2000; Fetter et al. 2004; Suga & Maeshima 2004; Matsumoto et al. 2009). Although PIP1-type aquaporins do not show clear water transport activity when they are heterologously expressed in the egg of a Xenopus lavis oocyte or yeast cell (Fetter et al. 2004; Suga & Maeshima 2004; Matsumoto et al. 2009), they have a significant water transport activity in planta (Siefritz et al. 2002; Postaire et al. 2010). Furthermore, some researchers have proposed a mechanism in which PIP1 and PIP2 aquaporins interact to form heterotetramers with increased water transport activity when their genes are co-expressed in the same egg of an oocyte (Fetter et al. 2004; Zelazny et al. 2007). In contrast, most NIPs and SIPs have low water transport activity. Some NIPs transport molecules such as glycerol, urea, silicon, boron and arsenic (Ma et al. 2006; Takano et al. 2006; Wallace, Choi & Roberts 2006; Kamiya et al. 2009).

Root Lpr is higher during the day than at night (Parsons & Kramer 1974). In general, the amounts of PIPs in roots show similar diurnal changes (Henzler et al. 1999; Lopez et al. 2003; Beaudette et al. 2007). In Lotus japonicus, the expression of PIP genes homologous to Arabidopsis AtPIP1;1 reached their maximum value 4–6 h after light initiation and reached their minimum value shortly after the beginning of the dark period (Henzler et al. 1999). In a similar manner, maize PIP transcripts in roots reached a maximum within 2–4 h after light initiation and remained at low levels during the second half of the light period and the beginning of the dark period (Lopez et al. 2003). Similar fluctuation was observed at the protein level only in PIP2-type aquaporins, but not in PIP1-type aquaporins (Lopez et al. 2003). The expression of PsPIP2;1 in the lateral roots of pea plants showed peaks at 3 and 10 h after light initiation, and reached its minimum level at midnight (Beaudette et al. 2007). These results have led researchers to assume that diurnal changes in root aquaporin levels may contribute to regulating water movement through roots in accordance with the plant's need for water during the light–dark cycle.

However, it is not clear how root aquaporin levels are regulated diurnally. The expression of maize aquaporin genes continued to fluctuate diurnally even when plants were maintained in darkness for a day (Lopez et al. 2003). This suggests that the observed diurnal changes are regulated spontaneously by a circadian rhythm. Here, we have used the term ‘circadian rhythm’ to represent an endogenously generated pattern that can be entrained by external stimuli such as the light–dark cycle. Such circadian rhythms usually continue for several days even when plants are isolated from the light–dark cycle. In this report, we distinguish circadian rhythm from diurnal changes that comprise time-dependent changes that occur during a single day. In contrast with the previous hypothesis that the diurnal changes in aquaporin levels are regulated by a circadian rhythm, it is possible that they are strongly regulated and enhanced by other factors.

One possible mechanism may be transpirational demand from shoots, because light induces stomatal opening and thus results in increased transpiration. A few studies have shown that transpirational demand affects Lpr. For example, Lpr in wheat decreased by a factor of two under reduced transpiration (Carvajal, Cooke & Clarkson 1996). On the other hand, in Lotus japonicus, reduced transpirational demand slightly enhanced root Lpr (Henzler et al. 1999). Despite these contradictory results, both experiments showed an effect of transpirational demand on Lpr. A recent report showed that low relative humidity (RH) induced the expressions of aquaporin genes in Arabidopsis roots; AtTIP2;2, AtTIP2;3 and AtNIP2;1 were strongly induced by 2–5 h of low RH (RH = 17%) treatment, but PIP genes were less induced (Levin et al. 2009). Thus, transpirational demand appears to have an effect on Lpr and on aquaporin gene expressions, although there is no consensus on the mechanism.

In the present study, we tested the hypothesis that the induction of aquaporin gene expression during the day could be triggered by transpirational demand from shoots after light initiation. To do so, we compared changes in aquaporin gene expressions under 100% RH to those under 60% RH. To understand comprehensively the mechanism responsible for the observed diurnal changes, it is important to focus on all aquaporin members that could potentially play crucial roles in roots. Such a perspective has been lacking in previous works on the diurnal expression of root aquaporin genes in several plants. We used rice plants (Oryza sativa L.) in our research, because we have previously identified all 33 rice aquaporin genes and clarified their organ-specific expression patterns (Sakurai et al. 2005). We have also previously described water transport activity and tissue localization of some rice PIPs and TIPs that are abundant in roots (Sakurai et al. 2005, 2008). Based on these results, we focused on nine PIPs and three TIPs as candidates that may contribute greatly to root water transport. The effects of transpirational demand on diurnal changes in expression levels of these aquaporin genes were examined in detail. Furthermore, the differences in the diurnal patterns, and in the organ and tissue localization of these aquaporins will help to reveal the possible roles of each aquaporin in root water transport.

MATERIALS AND METHODS

Plant material and growth conditions

Rice (cv. Akitakomachi) seeds were germinated in the dark for 3 d at room temperature, and then grown in a growth chamber under a 12 h light/12 h dark photoperiod (with light at a photosynthetic photon flux density of 370 µmol s−1 m−2), and with day and night temperatures set at 25 and 20 °C, respectively. An RH of 75% was maintained for all experiments, except the high-humidity experiment described later in the Materials and methods. Plants were grown in hydroponic solution using a previously reported culture solution (Murai-Hatano et al. 2008). The solution was changed daily until the plants were used for our experiments. The volume of the container was 500 mL. Eight to 16 seedlings were grown in each container. We used 14- to 16-day-old plants for analysis of root osmotic hydraulic conductivity [Lpr(os)], transpiration rate and aquaporin gene expression. We used 16- and 38-day-old plants for the immunocytochemistry analysis described later in Materials and methods.

Measurement of Lpr(os)

We removed the shoots of 15-day-old plants with a razor blade at the root base. The xylem sap was collected in pre-weighed cotton swabs. The volume was calculated based on the change in the weight of the cotton. Simultaneously with the volume measurements, we collected xylem sap from other seedlings using a micropipette. We determined the osmotic potentials of the hydroponic medium and exuded xylem sap using a VAPRO 5520 osmometer (Wescor Inc., Logan, UT, USA). We measured the osmotic potential of the xylem sap for the sum from 4 to 13 plants, because it is not always possible to obtain enough volume (more than 50 µL for measurement of osmotic potential) of xylem sap from only one plant. Before calculating the Lpr(os), we examined the changes in the volume and the osmotic potential of the xylem sap at 5 min interval after the shoot removal. Although both the volume and the osmotic potential were a little unstable during the first 5 min, they were almost constant during the next 25 min. Therefore, we discarded the xylem sap during the first 5 min, and then collected the sap of the following 20 min for calculating the Lpr(os).We calculated Lpr(os) (m s−1 MPa−1), using the following equation:

image(1)

where Jv is the volumetric xylem sap flow per unit root surface area (m3 m−2 s−1), σ is the reflection coefficient for nutrient salts in the xylem and ΔΨs is the difference in osmotic potential (MPa) between the exuded xylem sap and the hydroponic solution. Miyamoto et al. (2001) estimated a σ value of 0.4 for rice roots.

Murai-Hatano et al. (2008) note that Lpr is affected by the temperature of the root medium (i.e. of the hydroponic solution). Therefore, all measurements were conducted at 25 °C that was maintained using a circulating water bath. The constant root temperature was initiated at 3 d before the Lpr(os) measurements. After measuring the volumetric xylem sap flow, we scanned the total root system using a scanner at a resolution of 500 dpi and calculated the surface area using version 2007d of the WinRHIZO software (Regent Instruments, Inc., Quebec City, QC, Canada).

Measurement of transpiration rate

Transpiration was measured by the decrease in pot weight during the 30 min measurement period. Simultaneously, the change in the pot weight without plants was measured. Transpiration was calculated from the subtraction of the decrease in pot weight without plants as a result of evaporation from that with plants. The transpiration rate (g cm−2 min−1) was calculated as the transpiration per unit leaf area. We determined the leaf area after the final measurement by collecting all leaf blades and scanning them using the same system used for the determination of root surface area.

Shoot-cut and high-humidity experiments

In the shoot-cut experiment, we removed all shoots using scissors at the root collar 10 min before turning on the lights. In the high-humidity experiment, plants were maintained in a humidified cage in a growth chamber. The light and temperature of the growth chamber were maintained at the same condition when the plants were grown (see the first section of Materials and methods). The cage was covered with transparent film used for greenhouses. The RH in the cage was maintained at almost 100% using a humidifier D-505 (PS MFG Co. Ltd., Tokyo, Japan). The plants were moved into the humidified cage 10 min before turning on the lights. For the non-treated plants, the RH of the growth chamber was maintained at 60%, instead of 75% that was established under normal growth condition, in order to highlight the difference of gene expressions between control plants and those under high humidity.

Total RNA extraction and quantitative real-time PCR

We collected all of the root systems except for 1 cm below the root base. Leaf blades were collected from the upper half of the shoots. These samples were immediately frozen in liquid nitrogen and ground in a mortar and pestle. Total RNA was extracted from the frozen powdered tissue using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) and the RNase-Free DNase Set (Qiagen) according to the manufacturer's instructions. The first-strand cDNA was synthesized from 1.5 µg of total RNA using the High Capacity cDNA Reverse Transcription Kit [Applied Biosystems (ABI), Foster City, CA, USA] according to the manufacturer's instructions. After dilution of the synthesized cDNA solution to 2% of the original concentration, we performed quantitative real-time PCR using the StepOne Real-Time PCR System (ABI). PCR was conducted using the Fast SYBR Green system (ABI) with the primers listed in Supporting Information Table S1. The PCR conditions were one cycle of 95 °C for 20 s followed by 40 cycles of 95 °C for 3 s and 60 °C for 30 s. The melt-curve reaction was conducted using the method recommended by the manufacturer for the instrument and the reagents.

To measure the mRNA copy numbers for the expressed aquaporin genes, we constructed standard plasmids for each aquaporin gene. We amplified each gene using the primers listed in Supporting Information Table S1 and cloned into the p-GEM-T easy vector (Promega, Madison, WI, USA). Plasmids were extracted using the QIAprep Spin Miniprep Kit (Qiagen). After digestion with SalI or SphI restriction enzyme (TOYOBO, Tokyo, Japan), we purified the linearized plasmids using phenol–chloroform and ethanol precipitation. Copy numbers in the plasmid solutions were determined by their concentration and molecular weight. We constructed standard curves using a dilution series of the plasmid solutions.

Preparation of crude membrane fractions and immunoblotting

We prepared crude membrane fractions and performed immunoblotting using a previously described method (Murai-Hatano et al. 2008), with isoform-specific peptide antibodies. All antibodies except for anti-OsPIP2;4 antibody were raised using previously described methods (Murai-Hatano et al. 2008; Sakurai et al. 2008). Anti-OsPIP2;4 antibody was raised against the peptide MGKEVDVSTLEAGGAR in the N-terminal region. The intensity of the band signal after immunoblotting was visualized using the Light Capture system (ATTO, Tokyo, Japan) and measured using version 3.0 of the CS Analyser software (ATTO).

Immunocytochemistry

We excised root tissue samples at 0–1 cm from the root tip of main roots of seminal and crown roots. All procedures of fixation, embedding in paraffin, section preparation and immunodetection were conducted as described previously (Sakurai et al. 2008), with slight modifications. After reaction with the aquaporin antibodies, sections were visualized with a fluorescent second antibody, Alexa Fluor 555 goat anti-rabbit IgG (H + L) (Invitrogen, Carlsbad, CA, USA). Fluorescence was observed under an IX81 fluorescence microscope (Olympus, Tokyo, Japan). The Alexa Fluor 555 was excited using a 75-W U-LH75XEAPO xenon lamp (Olympus) with a 530 to 550 nm filter, and the fluorescent emission was measured through a 575IF-nm filter using a U-MWIG3 mirror unit (Olympus). The fluorescence images were captured using a cooled CCD camera (Rolera-XR-C, Q-Imaging, Surrey, BC, Canada).

RESULTS

Diurnal changes in transpiration rate and Lpr(os) of rice seedlings

We investigated the change in transpiration rate and Lpr(os) during a 12 h light (25 °C)/12 h dark (20 °C) cycle. Because the temperature of the root medium affects the transpiration rate and Lpr(os), we conducted the measurements using a circulating water bath to maintain a constant root temperature of 25 °C. The transpiration rate was extremely low during the dark period (Fig. 1a). It increased rapidly after light initiation, and reached to 30.3 mg h−1 cm−2, then continued to increase more slowly to around 40.4 mg h−1 cm−2. After dark initiation, the transpiration rate decreased rapidly to reach a low level similar to that before light initiation.

Figure 1.

Diurnal changes in transpiration rate and osmotic root hydraulic conductivity [Lpr(os)] in rice. Transpiration rate (a, n = 3) was calculated by the loss of pot weight per unit leaf area. For analysis of Lpr(os), shoots were cut off at the root base. The volumetric xylem sap flow per unit root surface area, Jv (b, n = 4 or 5), and the difference in osmotic potential between the xylem sap and the hydroponic solution, ΔΨs (c), were measured at various times of the light–dark cycle. ΔΨs at each time was measured for the sum of xylem sap from 4 to 13 plants. Lpr(os) (d, n = 4 or 5) was calculated from Jv and ΔΨs using an equation indicated in Materials and methods. The root medium was maintained at 25 °C throughout the whole experiments. Gray and white bars below the graph represent dark and light period of the chamber, respectively. Error bars represent the SD.

Lpr(os) was calculated from the volumetric xylem sap flow per unit of root surface area (Jv) and the difference in osmotic potential between the xylem sap and the hydroponic medium (ΔΨs) using an equation indicated in Materials and methods. Jv was low during the dark period, but gradually increased to around 3.0 × 10−9 m s−1 towards the end of the dark cycle (Fig. 1b). After light initiation, Jv increased rapidly to a constant high level of around 4.4 × 10−9 m s−1 between 1 and 4 h after light initiation, then decreased steadily to a plateau at around 4.0 × 10−9 m s−1. After dark initiation, Jv decreased rapidly. ΔΨs remained high during the 8 h before the end of the dark period, and then decreased rapidly after light initiation, reaching a minimum level of 5.03 × 10−2 MPa at 1–2 h after light initiation(Fig. 1c). ΔΨs then increased gradually, reaching a relatively constant level of around 7.0 × 10−2 MPa until the end of the light period. After dark initiation, ΔΨs increased rapidly to reach a maximum of 12.4 × 10−2 MPa at 2 h after dark initiation, and then decreased slightly. Lpr(os) was low during the dark period, reaching a minimum of 5.89 m s−1 MPa−1 at 0700 h, but gradually increased towards the end of the dark period, reaching a value of 8.59 × 10−8 m s−1 MPa−1(Fig. 1d). After light initiation, Lpr(os) increased rapidly, reaching a maximum of 22.3 × 10−8 to 22.6 × 10−8 m s−1 MPa−1 at 1–2 h after light initiation. Lpr(os) then decreased gradually, reaching a plateau at around 13.4 × 10−8 m s−1 MPa−1 during the last half of the light period. Just before dark initiation, Lpr(os) decreased slightly to 11.8 × 10−8 m s−1 MPa−1. After dark initiation, Lpr(os) continued to decrease, reaching a minimum of 5.30 × 10−8 m s−1 MPa−1 at 2 h after dark initiation.

Comparison of expression levels of various aquaporin genes in roots and leaf blades

To clarify the absolute amounts of each aquaporin mRNA, we conducted quantitative real-time PCR using cloned plasmids of each aquaporin as standard templates. Rice has 11 PIPs and 10 TIPs (Sakurai et al. 2005). In this experiment, we focused on nine PIPs and four TIPs listed in Table 1. We chose these members of the family because they were shown to be expressed at a higher level in leaf blades and roots than other PIPs and TIPs based on a previous analysis using semi-quantitative RT-PCR (Sakurai et al. 2005). We did not analyse any NIPs or SIPs, because they have been reported to have lower water transport activity (Ishikawa et al. 2005; Wallace et al. 2006). In the roots, OsPIP2;1 was expressed at the highest level (7.14 × 106 copies µg−1 of total RNA; Fig. 2). Except for OsTIP1;2, the aquaporin genes were expressed in roots at levels more than 4 × 105 copies µg−1 of total RNA. On the other hand, all aquaporin genes in the leaf blades, except for OsTIP1;2, were expressed at lower levels than in the roots. The highest level of expression in the leaf blades was for OsPIP2;2, at 1.36 × 106 copies µg−1 of total RNA. OsTIP1;2 was expressed almost predominantly in the leaf blades, whereas OsPIP2;3, OsPIP2;4, OsPIP2;5 and OsTIP2;1 were barely expressed in the leaf blades, and were instead expressed almost predominantly in the roots. OsPIP1;3 was also expressed more in the roots than in the leaf blades.

Table 1.  Summary of properties of major rice plasma membrane-localized aquaporins (PIPs) and tonoplast-localized aquaporins (TIPs)
Aquaporin nameExpressions (organ specificity)Amplitude of diurnal changes in the roots (mRNA)Localization in the roots (immunocytochemistry)
Leaf bladesRoots
  1. Gene expression levels were categorized for each aquaporin according to the results in Fig. 2. Double circle, single circle, triangle and cross mean that the aquaporin genes were expressed more than 106, 105, 104 and less than 104 copies µg−1 of total RNA, respectively. The amplitude of the diurnal change in root aquaporin expression level was calculated from the ratio of the maximum to the minimum, according to the result of Fig. 3 (data from blue dot were used for calculation). Localization in the roots was summarized by the results of the previous report (italic, Sakurai et al. 2008) and this work (bold, Fig. 8).

OsPIP1;1inline image2.92Endodermis
OsPIP1;2inline image4.82Endodermis
OsPIP1;3inline image17.6Endodermis
OsPIP2;1inline image2.75Endodermis (proximal end), cortex (proximal end), outer part of roots, central cylinder
OsPIP2;2inline imageinline image3.75Not examined
OsPIP2;3×inline image16.5Endodermis, outer part of roots, cortex, central cylinder
OsPIP2;4×28.2Not examined
OsPIP2;5×145Endodermis (proximal end), xylem, outer part of roots, cortex, central cylinder
OsPIP2;6Little changedNot examined
OsTIP1;1Little changedOuter part of roots (rhizodermis, exodermis)
OsTIP1;2inline imageNot expressedNot localized
OsTIP2;1×inline image5.24Endodermis, central cylinder
OsTIP2;2inline imageinline imageLittle changedCentral cylinder
Figure 2.

Comparison of copy numbers of the expressed aquaporin genes in the leaf blades and the roots. Leaf blades (white columns) and roots (black ones) of 16-day-old rice plants were harvested 6 h after light initiation. Copy numbers were determined by quantitative real-time PCR method using standard curve for each aquaporin gene. Details for calculation were described in Materials and methods. Expression levels were corrected by the ratio of 18S rRNA amounts. Data represent means with the SD calculated for three independent experiments.

Diurnal changes in root aquaporin mRNA levels

We previously demonstrated using semi-quantitative RT-PCR method that rice aquaporin mRNA levels varied diurnally (Sakurai et al. 2005), that is, when seedlings grown under 12 h light/12 h dark conditions were analysed at 3 h intervals, most PIP members showed the highest expression 3 h after light initiation, and the lowest expression 3 h after dark initiation. In the present study, using quantitative real-time PCR method, we further investigated the changes in expression of each aquaporin gene at 1–5 h intervals. We focused on 12 aquaporin genes (OsPIP1;1, OsPIP1;2, OsPIP1;3, OsPIP2;1, OsPIP2;2, OsPIP2;3, OsPIP2;4, OsPIP2;5, OsPIP2;6, OsTIP1;1, OsTIP2;1 and OsTIP2;2) that were expressed strongly in the roots (Fig. 2). They were divided into three groups with different types of diurnal changes. The first is a group consisting of OsPIP2;6, OsTIP1;1 and OsTIP2;2 that showed no clear pattern or rhythm in the three independent experiments (Fig. 3a). The second and the third group showed diurnal variations. The second group, which comprised of OsPIP1;3, OsPIP2;3, OsPIP2;4 and OsPIP2;5 showed a dramatic peak induction around 2 h after light initiation, reaching values of more than 15 times the minimum level (Table 1). For OsPIP2;5, the peak mRNA level (2 h after light initiation) was 145 times the minimum level (4 h after dark initiation). The third group consisting of OsPIP1;1, OsPIP1;2, OsPIP2;1, OsPIP2;2 and OsTIP2;1 showed a moderate level of diurnal variation, which is less than six times the minimum level (Table 1).

Figure 3.

Diurnal changes in aquaporin gene expression levels in the roots. Rice plants were grown under 12 h light/12 h dark condition for 16 d. Expression levels were analysed at various times of the light–dark cycle of the 16 d (a). Blue, green and purple dot mean results from three independent experiments using different plant samples. At time of 1200 h, samples were taken just before light initiation. Relative proportion of each PIP gene expression levels to the total PIP expressions was calculated at each time of the light–dark cycle (b), according to the result from (a) (data from blue dot was used for calculation).

Figure 3b shows the relative proportions of each PIP to the total PIP expression level. OsPIP2;1 ranked first among all PIPs throughout the light–dark cycle, but the relative proportions of each PIP changed diurnally. Shortly after light initiation, the proportions of OsPIP1;3, OsPIP2;3, OsPIP2;4 and OsPIP2;5 increased. Their sum at 2 h after light initiation (38.5%) was higher than that of OsPIP2;1 (23.0%), suggesting significant roles of these root-specific aquaporins under the light condition.

Effect of disturbance of the light–dark cycle on the diurnal changes in root PIP levels

We investigated whether diurnal changes in root PIP expressions are caused by a circadian rhythm or by a response to light. Here, we focused on OsPIP2;1 and OsPIP2;2, which showed only moderate diurnal variation, and OsPIP2;4 and OsPIP2;5, which showed a high degree of diurnal variation. After rice seedlings were grown under the 12 h light/12 h dark cycle for 13 d, they were maintained in the dark for an additional 12 h. For control plants, we continued the normal light–dark cycle. The OsPIP2;4 and OsPIP2;5 levels in the control plants increased rapidly, but temporarily shortly after light initiation, whereas the dark-extended plants showed little increase (Fig. 4a,b). OsPIP2;1 and OsPIP2;2 also showed lower levels in the dark-extended plants than in the control plants. Similar down-regulation of aquaporin genes by the dark extension treatment was observed in other PIPs and TIPs, except for OsPIP2;6, OsTIP1;1 and OsTIP2;2, which showed a tendency to increase in expression level during 8 h of dark extension (Supporting Information Fig. S1).

Figure 4.

Effect of disturbance of the light–dark cycle on the diurnal changes in root PIP expression levels. After the rice seedlings were grown under the 12 h light/12 h dark until the time of 1200 h of day 13, they were maintained in the dark for an additional 12 h (closed circle) (a). For control plants, normal light–dark cycle was continued (open circle). Another set of seedlings were exposed to light after an 18 h of continuous dark (open triangle). Gray and white bars below the graph represent the light condition of each treatment (upper, normal light–dark cycle; middle, 6 h dark extension followed by 6 h light; lower, dark extension). Aquaporin expression levels between the three treatments were compared at indicated time of the 13th day (b). Black, white and grey bars represent dark-extended, normal light/dark cycle and 3 h of illumination after 18 h of continuous dark, respectively. Values are means ± SD from three different plant sample sets. Asterisks indicate statistically significant differences (P < 0.05) compared with plants under continuous dark at each time.

We also confirmed the effect of light on the dark extended plants. After an 18 h of continuous dark treatment, the plants were suddenly exposed to light. OsPIP2;1, OsPIP2;4 and OsPIP2;5 increased rapidly after illumination (Fig. 4a). Expression reached a peak at 3 h after illumination, and then decreased. However, the expression level at each peak was lower than that seen 2 h after light initiation in the plants under a normal light–dark cycle. For example, OsPIP2;5 transcription at 2 h after light initiation under normal light–dark conditions reached 3.62 × 106 copies µg−1 of total RNA, whereas peak transcription at 3 h after illumination in the dark-extended plants reached only 1.38 × 106 copies µg−1 of total RNA (Fig. 4b). The induction of OsPIP2;2 by light after the extended dark treatment was low (Fig. 4a,b). Similar induction by sudden illumination was observed for other PIPs and TIPs, except for OsPIP2;6, OsTIP1;1 and OsTIP2;2, which showed a tendency to decrease expression (Supporting Information Fig. S1).

Effect of shoot cut on the increase in root PIP levels after light initiation

We also examined whether the increase of PIP levels in roots occurs without shoots. Shoots were cut 10 min prior to light initiation, and these plants showed no induction of PIP expression at 2 h after light initiation, in contrast with clear induction in the intact plants (Fig. 5). The shoot-cut plants showed aquaporin expression lower than those of dark-extended plants.

Figure 5.

Effect of shoot cut on the increase in root PIP expression levels. Shoots of 15-day-old rice seedlings were removed at the root collar 10 min before light initiation. The aquaporin gene expression level of shoot-cut plants 2 h after light initiation (grey bars) was compared with those of intact plants under normal light/dark cycle (white bars) and dark-extended plants (black bars). Values are means ± SD from three different plant sample sets. Asterisks indicate statistically significant differences (P < 0.05) compared with plants under the normal light/dark cycle at each time.

Effect of inhibition of transpiration on the increase in root PIP levels after light initiation

We examined whether transpirational demand from the shoots after light initiation was involved in the observed increase in PIP expression in the roots. To confirm this hypothesis, we investigated changes in PIP levels under high-humidity conditions in the presence of light. In this experiment, we transferred seedlings into a humidified cage 10 min before light initiation. We used a humidifier to maintain nearly 100% RH. Under these conditions, the leaf blades were covered by small water droplets and their transpirational rate was considered to be extremely low. Although the control plants under 60% RH showed increased OsPIP2;1, OsPIP2;2, OsPIP2;4 and OsPIP2;5 expression after light initiation, the plants in the high-humidity condition did not (Fig. 6). Changes in these PIP expression levels in the high-humidity treatment were similar to or slightly higher than those in the dark-extended plants. These results indicated that the temporary and rapid induction of aquaporin gene expression may occur because of transpirational demand from shoots. Similar results were observed in other PIPs and TIPs, except for OsPIP2;6, OsTIP1;1 and OsTIP2;2 (data not shown).

Figure 6.

Effect of high humidity on the diurnal changes in root PIP expression levels. Rice seedlings (14 days old) were transferred to a humidified cage [relative humidity (RH) = 100%] 10 min before light initiation (open triangle). Changes in aquaporin gene expression levels were compared with those of control plants under 60% of RH (open circle) and dark-extended plants (closed circle).

Diurnal changes in root aquaporin protein levels

We examined diurnal changes in aquaporin protein levels by means of immunoblotting using anti-aquaporin antibodies. We previously raised isoform-specific antibodies for eight aquaporins (Murai-Hatano et al. 2008; Sakurai et al. 2008). In this report, we added an anti-OsPIP2;4 antibody raised against a peptide corresponding to 16 amino acid residues at the N-terminal region. The antibody recognized recombinant OsPIP2;4 protein well, and did not react with other aquaporin isoforms (data not shown). We also used anti-RsTIP1;1 antibody raised for radish TIP1;1, but this antibody also recognized OsTIP1;1 well (Sakurai et al. 2008). OsPIP1;3, OsPIP2;3, OsPIP2;4 and OsPIP2;5 showed clear diurnal changes (Fig. 7a). On the other hand, OsPIP1s (OsPIP1;1, OsPIP1;2 and OsPIP1;3 which are all recognized by the same anti-PIP1 antibody), OsPIP2;1 and OsTIP2;1 showed only moderate diurnal variation. The maximum protein levels were 1.91 and 4.78 times the minimum level for OsPIP2;1 and OsPIP2;5, respectively (Fig. 7b). No clear diurnal rhythms were observed for OsTIP1;1 and OsTIP2;2. Although the isoform-specific diurnal changes in protein levels agreed well with the mRNA results, there was a time lag between the two diurnal patterns. For example, OsPIP2;5 mRNA reached its maximum level 2 h after light initiation (Fig. 3a), whereas the maximum level of OsPIP2;5 occurred 6 h after light initiation (Fig. 7b).

Figure 7.

Diurnal changes in aquaporin protein levels in roots. Crude membrane fractions were prepared from the root samples and subjected to immunoblotting using anti-aquaporin antibodies (a). The band intensities were expressed as the ratio to the minimum ones (b). Protein amounts subjected to immunoblotting were 1.1, 11, 1.1, 10, 15, 12, 0.53, 0.90, 4.5 µg for anti-OsPIP1s, anti-OsPIP1;3, anti-OsPIP2;1, anti-OsPIP2;3, anti-OsPIP2;4, anti-OsPIP2;5, anti-RsTIP1;1, anti-OsTIP2;1 and anti-OsTIP2;2 antibodies, respectively. At time of 1200 h, samples were taken just before light initiation.

Localization of OsPIP2;1 and OsPIP2;5 in root tissues

To characterize the roles of OsPIP2;1 and OsPIP2;5 in roots, we determined their localization in root tissues. Tissue samples were fixed 3 h after light initiation. By using anti-aquaporin antibodies in combination with a peroxidase-linked secondary antibody and 3,3′-diaminobenzidine tetrahydrochloride, we previously clarified that both aquaporins were localized in various tissues in roots (Table 1), but they mainly accumulated in the endodermis (Sakurai et al. 2008). More precise localization was obtained in the present study using the fluorescent secondary antibody. OsPIP2;1 accumulated abundantly in the endodermis and in cortical cells one or two layers outside of the endodermis (Fig. 8b). In these cells, OsPIP2;1 accumulated in the proximal end (close to the central cylinder) of the cells. Accumulation of OsPIP2;5 was also observed in the proximal end of the endodermal cells(Fig. 8c). Interestingly, we also found OsPIP2;5 in the cell surfaces facing the xylem (Fig. 8d).

Figure 8.

Localization of OsPIP2;1 and OsPIP2;5 in root tissues. Samples were fixed from tap roots of 16- (a, d) or 38-day-old (b, c) rice plants at 3 h after light initiation. Root sections at around 4 mm from the root tip were subjected to immunocytochemistry using pre-immune (a), anti-OsPIP2;1 (b) and anti-OsPIP2;5 antibodies (c, d). Bars represent 50 µm.

DISCUSSION

Strong enhancement of the diurnal changes in root aquaporin gene expressions by transpirational demand after light initiation

Previously, Lopez et al. (2003) reported that diurnal changes in maize root aquaporin gene expression levels depend on a circadian rhythm. In this report, we examined other factors that may strongly regulate and enhance the amplitudes of diurnal changes in rice roots. Because dark extension did not cause rapid induction of PIP expression after light initiation (Fig. 4a,b), we considered that light, or a factor relating to light, is an important factor that governs PIP expression. The shoot-cut experiment demonstrated that increase in PIP expressions under light does not occur without shoots (Fig. 5). In the high-humidity experiment, we clarified that PIP expression in the roots did not increase even in the presence of light (Fig. 6). From this evidence, we concluded that the temporary and rapid increase in PIP expression may be caused by the signal of transpirational demand from the shoots after light initiation. Because PIP expression levels in the light under high humidity were a little higher than the corresponding levels under continuous dark (Fig. 6), other factors related to light might have affected the increase in PIP expression, but based on the low magnitude of the response, transpirational demand seems to be the most important factor.

Levin et al. (2009) recently reported that expression of several TIPs and NIPs in the roots of Arabidopsis was increased by low RH, but expression of PIPs was less increased. The difference between their work and our results may result from differences between the two plant species and in the experimental conditions. Among the aquaporins, PIPs are considered to play the most crucial roles in cell-to-cell water transport in the roots, because they are localized in the plasma membrane. Our result is the first report showing that transpirational demand can affect the expression of PIPs. Furthermore, we recently found that expression levels of OsPIP2;4 and OsPIP2;5 in the roots were significantly increased by exposure of the shoots to low air humidity (45% RH) from high air humidity (90% RH) (Murai-Hatano et al., unpublished data). From these results, we concluded that PIP induction is triggered by transpirational demand after light initiation.

Rice plants often suffer from excess water loss because of transpiration; under a high radiation level, leaf water shortages are observed even in paddy fields with enough water in the wet soil (Ishihara & Saito 1987; Jiang, Hirasawa & Ishihara 1988; Hirasawa, Tsuchida & Ishihara 1992). This is because water uptake by roots cannot compensate for the high transpirational water loss from shoots (Hirasawa et al. 1992; Miyamoto et al. 2001). The Lpr of rice roots is low compared with that in other herbaceous species (e.g. only 20% of that in maize) because of the existence of highly developed apoplastic barriers in rice (Miyamoto et al. 2001). In other words, the function of aquaporins in the cell-to-cell pathway may be more important in rice than in other plants, and may respond more sensitively and rapidly to changes in transpirational demand. This may be the reason of the lower contribution of circadian rhythm in root aquaporin expression levels in rice than in maize (Lopez et al. 2003). Our finding of rapid induction of root aquaporin gene expression by transpirational demand may be related to a strategy of rice plants to increase cell-to-cell water movement in order to sustain a high transpiration rate during the light period (Fig. 1a).

Although the diurnal changes in root aquaporin gene expression were strongly enhanced by transpirational demand, they were also moderately driven by a circadian rhythm. This is because mRNA and protein levels increased several hours before light initiation (Figs 3a & 7). Similarly, the decrease in aquaporin expression several hours before dark initiation indicated the involvement of a circadian rhythm. The effect of this circadian rhythm was also observed in the response to sudden illumination after 18 h of continuous dark (Fig. 4a,b). The treatment did not induce a high level of PIP transcripts as the level observed 2 h after light initiation under the normal light–dark cycle. This means that diurnal changes in aquaporin gene expressions in rice roots are moderately regulated by circadian rhythm, and the patterns are further strongly enhanced by transpirational demand.

Involvement of aquaporins at protein levels in diurnal changes in Lpr

Immunoblotting showed that protein levels of root-specific aquaporins, such as OsPIP2;4 and OsPIP2;5, peaked 6 h after light initiation (Fig. 7), whereas mRNA levels peaked 2 h after light initiation (Fig. 3a). This means that there may be a substantial time lag (about 4 h) between the increase in mRNA and accumulation of proteins after translation. Interestingly, the peak in protein levels for aquaporins, such as OsPIP2;4 and OsPIP2;5 (6 h after light initiation; Fig. 7), was inconsistent with the peak in Lpr(os) (1–2 h after light initiation; Fig. 1d). This inconsistency does not necessarily contradict the contribution of aquaporins to Lpr(os). Temporary peak in Lpr(os) 1–2 h after light initiation may depend on the activation of the aquaporin proteins, possibly because of an abrupt increase in transpiration after light initiation (Fig. 1a). Lopez et al. (2003) also suggested that the activity of some aquaporins in maize could be induced by light. Several mechanisms have been proposed for the activation of aquaporin proteins (Maurel et al. 2008). For example, aquaporin gates can be reversibly opened and closed by post-translational modifications such as phosphorylation (Johansson et al. 1998; Törnroth-Horsefield et al. 2006) and protonation (Tournaire-Roux et al. 2003). Recently, the assembly of PIP heterotetramers and their trafficking to plasma membranes have been suggested as a possible activation mechanism (Fetter et al. 2004; Temmei et al. 2005; Zelazny et al. 2007). Interestingly, PIP2;1 and PIP2;5 showed polar localization in the proximal end of the endodermis as discussed later. This finding opened the possibility that the activity of these aquaporins may be regulated by trafficking to a specific side of plasma membrane as well described in mammalian aquaporin-2 in renal collecting duct (Sasaki & Noda 2007). These mechanisms and other unknown activation systems might be involved in the temporary increase in Lpr(os) 1–2 h after light initiation.

In the present study, we measured the Lpr(os) in order to focus on the function of aquaporins, because Lpr(os) relates to the water movement along the cell-to-cell pathway in the roots. On the other hand, hydrostatic Lpr[Lpr(hy)] relates to the water movement along with both the apoplastic and the cell-to-cell pathways (Steudle 1993, 2000). Although a large difference in the values between Lpr(os) and Lpr(hy) has been reported in some species, such as maize and Phaseolus vulgaris (Steudle 2000), similar values are reported in rice roots (Miyamoto et al. 2001). Regarding the calculation of Lpr(os), it should be noted that reflection coefficient (σ) was assumed to be constant (σ = 0.4) in the present work as mentioned in Materials and methods. However, there is a possibility that σ changes to some extent in day/night cycle, depending on the possible diurnal changes in the transport activity for nutrient solutes in the root cells. Further studies are required for evaluating the diurnal changes in Lpr(os) considering the diurnal changes in σ.

Possible function of each aquaporin based on its organ specificity, diurnal expression pattern and tissue localization

Twelve aquaporins (nine PIPs and three TIPs) were expressed abundantly in roots (Fig. 2). Here, we will discuss possible function of each aquaporin according to the differences in their organ specificity, diurnal expression patterns and tissue localization.

In this study, we clearly demonstrated that the patterns of diurnal changes in expression levels differed among the aquaporins. In maize, ZmPIP1;1 and ZmPIP2;1 showed moderate diurnal changes, whereas ZmPIP1;5 and ZmPIP2;5 showed marked changes (Lopez et al. 2003). In rice, we found three aquaporin groups with different diurnal expression patterns. The first group (OsPIP2;6, OsTIP1;1 and OsTIP2;2) did not show clear diurnal changes in mRNA or in protein levels (Figs 3a & 7). Interestingly, these aquaporins showed increased mRNA levels in our dark-extended experiment (Supporting Information Fig. S1). These aquaporins may play distinct roles other than in radial water movement through roots, such as roles in root elongation and cell differentiation. In case of deep-water rice plants, gene and protein expression levels of OsTIP1;1 and OsTIP2;2 in internodes were induced by submersion, suggesting possible function of these aquaporins for cell enlargement in the growing internodes during the submersion (Muto et al. 2011). The second group, which included aquaporins such as OsPIP2;4 and OsPIP2;5, showed temporary, but dramatic induction after light initiation, whereas the third group, which included aquaporins such as OsPIP2;1 and OsPIP2;2, showed moderate diurnal changes (Figs 3a & 7).

The tissue localizations of OsPIP2;1 and OsPIP2;5 highlight the possible differences in their function. OsPIP2;1 accumulated in the root endodermis and in a few layers of cortical cells adjacent to the endodermis (Fig. 8b). On the other hand, OsPIP2;5 accumulated mainly in the endodermis (Fig. 8c). OsPIP2;1 therefore seems to play a constitutive role both in the endodermis and in cortical cells during water transport at any time of the light–dark cycle, because the pattern of diurnal change in OsPIP2;1 was moderate (Figs 3a & 7b). On the other hand, OsPIP2;5 may be involved in fine-scale regulation mainly at the endodermis, which is main resistance in radial water transport (Miyamoto et al. 2001; Ranathunge, Steudle & Lafitte 2003). Localization of OsPIP2;5 in the cell surfaces around the xylem (Fig. 8d) also suggests a water transport function from the xylem parenchyma cells towards the xylem vessels. Furthermore, the rapid and temporary increase in OsPIP2;5 expression levels in response to transpirational demand (Fig. 3a) supports our hypothesis that fine adjustments by this protein provide key roles in root water transport under rapidly changing environmental conditions.

Interestingly, both OsPIP2;1 and OsPIP2;5 were localized in the proximal end (close to the central cylinder) of the endodermis (Fig. 8b,c). Polar localization of aquaporins has been reported for a few other plants. For example, maize aquaporin ZmPIP2;5 accumulated in the external sides of epidermal cells (Hachez et al. 2006). Rice silicon transporter Lsi1 (OsNIP2;1) was also localized at the distal parts of both the exodermis and the endodermis (Yamaji & Ma 2007). Another rice silicon transporter, Lsi6 (OsNIP2;2), in xylem parenchyma cells of the leaf sheath and leaf blades was localized at the side facing towards the vessels (Yamaji, Mitani & Ma 2008). Our report is the first one that demonstrated aquaporin localization in the proximal end of root cells. OsPIP2;1 and OsPIP2;5 on the proximal end of the endodermis may play crucial roles in promoting the delivery of water from the endodermal cells towards the central cylinder. Rice roots may also have other PIPs that are expressed in the distal end of the endodermis and that may have a function complementary to that of OsPIP2;1 and OsPIP2;5. Further studies are needed in order to clarify whether such polar localization might be caused by polar trafficking of OsPIP2;1 and OsPIP2;5 as a kind of activation mechanism for these proteins because of high transpirational demand after light initiation.

Based on our present results and previous reports (Sakurai et al. 2005, 2008), Table 1 summarizes the properties of the major rice PIPs and TIPs. Root-specific aquaporins, such as OsPIP2;3, OsPIP2;4 and OsPIP2;5, which showed large diurnal changes in their mRNA and protein levels, and strong induction by transpirational demand, could possibly play crucial roles in fine adjustments of radial water transport in roots, especially under the high transpirational demand after light initiation. On the other hand, aquaporins, such as OsPIP2;1 and OsPIP2;2, which are localized both in leaf blades and in roots, may play a constitutive role in water transport in both organs, irrespective of the diurnal cycle and stress conditions.

In this report, we uncovered an aquaporin regulation mechanism in which transpirational demand from the shoots increases aquaporin expressions at mRNA and protein levels, and possibly increases aquaporin activity. This mechanism is an important way for plants to regulate root water uptake in response to shoot water requirements.

ACKNOWLEDGMENTS

We are grateful to Prof. Masayoshi Maeshima, Prof. Masumi Okada and Dr. Tsuneo Kuwagata for their valuable discussions throughout this work. We also thank Prof. Matsuo Uemura and Dr. Yukio Kawamura for technical support of the measurement of osmotic potentials. We greatly appreciate Mrs. Katsuko Takasugi for providing experimental assistance.

Ancillary