Introduction

The processes of transmembrane water transfer in plant cells attract the increasing attention of researchers due to the discovery in cell membranes of the large family of integral proteins—aquaporins, which can regulate the passive transfer of water and some low molecular electroneutral compounds (Maurel 1997, 2007; Johansson et al. 2000; Verkman and Mitra 2000; Chaumont and Tyerman 2014). The wide range of selectivity profiles and regulation properties allows the aquaporins to participate in many processes of plant development and adaptation to variable environmental conditions (Maurel et al. 2008). Under stress, the contribution of the transmembrane pathway to the total water flow through the plant might become considerable (Steudle and Herzler 1995).

Various molecular mechanisms are involved into the regulation of plant aquaporin activity. These mechanisms have been studied poorly, though a number of aquaporin amino acid residues were identified to be responsible for sensitivity of water permeability of these integral proteins to phosphorylation–dephosphorylation, to their oligomerization in the membrane, and to the cytoplasmic level of pH and Ca2+ (Chaumont et al. 2005; Zhao et al. 2008; Bienert et al. 2012; Prado et al. 2013; Chaumont and Tyerman 2014). Some other intracellular and external factors, such as heavy metals, nutrients, temperature, cell turgor, solute gradients, and reactive oxygen species (ROS), were shown to be able to modify aquaporin activity (Preston et al. 1993; Daniels et al. 1996; Clarkson et al. 2000; Henzler and Steudle 2000; Niemietz and Tyerman 2002; Lee et al. 2004; Chaumont et al. 2005; Boursiac et al. 2008; Chaumont and Tyerman 2014).

Accumulation of the excess ROS in plant cells resulting from the disturbance of the normal equilibrium is one of the earliest responses to any biotic and abiotic stress (Minibayeva et al. 1998; Hernandes et al. 2001). Under stress conditions, the production of free oxygen radicals by mitochondria increases, and first of all, there is accumulated superoxide anion radical, then the product of its dismutation—hydrogen peroxide, and finally, the most toxic hydroxyl radical. The increase in ROS production in plants under stress can also result from the activation of the plasmalemma NADPH oxidase (Sagi and Fluhr 2006) and/or extracellular peroxidase (Minibayeva et al. 2001; Garrido et al. 2012).

Proteins are one of the most ROS-sensitive targets in cells (Stadtman 1992), which can bind from 50 to 75 % of oxygen radicals. The X-ray structural analysis of the glycerol transporter from bacteria (Fu et al. 2000) and of erythrocyte membrane water channel (Murata et al. 2000) showed the large degree of similarity of molecular structure of all aquaporins. The aquaporin topographic model based on these data includes six transmembrane domains which form a kind of a keg in the membrane due to inter-domain loops consisting of amino acid residues. Two submersed into the lipid bilayer inter-domain loops, containing highly conserved asparagine–proline–alanine (NPA) motifs, and participated directly in the formation of a water channel. These two loops are folded into the bilayer from the opposite sides of the membrane and create inside the membrane an hourglass-like structure. This transmembrane domain forms a water pore (Daniels et al. 1996; Maurel 2007). For most aquaporins, in particular tonoplast aquaporins, a conserved cysteine residue containing a thiol group is located immediately at the water pore zone. This residue is believed to be in charge of aquaporin activity and aquaporin sensitivity to mercury compounds (Preston et al. 1993; Hachez and Chaumont 2010). Highly conserved cysteine residues with thiol groups are also located in other structural sites of aquaporins, including hydrophilic loops on both sides of a membrane (Daniels et al. 1996; Kukulski et al. 2005; Zhao et al. 2008; Bienert et al. 2012). From analysis of the literature data, a complex picture of the relationship between cysteine residues and mercury response in plant aquaporins emerges (Daniels et al. 1996; Frick et al. 2013). Mercury binds to three out of four cysteine residues (Frick et al. 2013). Thus, there is admitted the existence of one more way of aquaporin functional state regulation, based on changes in thiol group redox status of these structural sites. It is likely that some questions concerning the effect of oxidative stress and ROS on aquaporin functional activity should be related to redox processes of thiol groups of these proteins. The high conservatism of cysteine residues points to the possibility of their participation in aquaporin activity regulation by affecting the level of aquaporin oligomerization in membranes and/or transitions between open–closed states of water channels via formation of inter- or intra-molecular disulfide bonds, correspondingly (Bienert et al. 2012).

Currently, the reversible dithiol-disulfide transitions in proteins are actively studied, since the redox regulation is assigned to have a leading role in endogenous regulation mechanisms of vital processes (Berczi and Moller 2000; Gelhaye et al. 2005; Minibayeva et al. 2009, 2012; Bienert et al. 2012). Earlier, the attempts were made in experiments in vitro to confirm the role of such transitions in transmembrane water transport via aquaporins (Ampilogova et al. 2006; Zhestkova et al. 2009). In order to simulate the effect of endogenous redox regulators, the authors of these papers studied the effect of exogenous oxidizing or reducing thiol group agents (diamide, dithiothreitol, tributylphosphine) on the process of osmotic shrinkage of plasma membrane vesicles. The data obtained from in vitro experiments (Ampilogova et al. 2006; Zhestkova et al. 2009) testify in favor of the possibility of aquaporin water permeability regulation due to changes in the status of their thiol groups by endogenous ROS.

To date, there are numerous studies of changes in root hydraulic conductivity under exogenous factors, which are able to cause ROS accumulation (e.g., Boursiac et al. 2008; Ehlert et al. 2009). Using a cell pressure probe, hydraulic properties (half-time of water exchange, inversely proportional to water permeability) of parenchyma cells in the midrib tissue of maize leaves have been measured (Kim and Steudle 2007). This methodical approach was also used to estimate the effect of exogenous ROS on cell hydraulic permeability. Results were discussed in terms of an oxidative gating of aquaporins by ROS (Kim and Steudle 2009). The contribution of aquaporin-mediated water transport in the experiments in vivo is estimated mostly using mercuric inhibitors (Javot and Maurel 2002; Bramley et al. 2009; Hachez and Chaumont 2010). To date, however, there are studies demanding caution while using mercury test to estimate the aquaporin role in water membrane permeability regulation (Frick et al. 2013; Chaumont and Tyerman 2014).

The NMR method with a pulsed magnetic field gradient proved to be successful in studies of water diffusional transport in biological samples, such as cells and tissues (Anisimov et al. 1998; Cho et al. 2003; Duval et al. 2005; Van As 2007). This method allows the non-invasive registration of water self-diffusion and separation of contributions of different water transport pathways in plant tissues (Ishida et al. 2000; Anisimov et al. 2004; Velikanov 2007; Ionenko et al. 2012). It was applied for plant roots to study mercury-induced changes in the permeability of cell membranes in intact tissue (Ionenko et al. 2003, 2012; Volobueva et al. 2004; Ionenko and Anisimov 2007).

In the present study, the changes in the total water permeability of cell membranes (plasmalemma and tonoplast) in maize seedling roots under two levels of oxidative stress were analyzed by the NMR method.

Materials and methods

Plant growth conditions and preparation of samples

Maize (Zea mays L., cv. Mashuk) seedlings were grown hydroponically in 0.25 mM CaCl2 (pH 6.3) at 22 °C for 5 days under the 16-h photoperiod (irradiance of 15 W m−2) and relative humidity of 60 %. Ten-millimeter-long root segments from the elongation zone, cut with a blade, were used in the experiments. The root segments (300 mg) were then incubated for 1 to 6 h with moderated shaking (gentle aeration) in (i) 0.25 mM CaCl2 (6 ml, pH 6.3), wound stress variant, and (ii) 0.25 mM CaCl2 plus 100 μM paraquat (6 ml, pH 6.3), wound–paraquat stress variant. In all experiments, the incubation of roots was started directly after excision. The excision of roots from seedlings (wound stress) and the subsequent incubation activated root cells to produce oxygen radicals (Minibayeva et al. 1998, 2001, 2009, 2012).

Diffusion measurements by pulsed field gradient NMR

After incubation, the root segments were gently wiped with a filter paper. Forty-five to 50 arranged in parallel segments were placed into a test tube for NMR measurements. Experiments were carried out at 25 °С on the time domain 1H NMR analyzer “Spin Track” (Resonance Systems Ltd., Yoshkar-Ola, Russia) operating at 19.1 MHz and equipped with an electromagnet (Bruker, Karlsruhe, Germany). Pulsed field gradient sequences were used to measure the translational diffusion coefficients (Stejskal and Tanner 1965). The pulsed magnetic field gradient was applied perpendicularly to the test tube. It allowed the observation of water self-diffusion in a cross direction of the root. It is in the radial direction that the main water transporting role of aquaporins occurs (Maurel et al. 2008).

During the experiments, we registered diffusional decays (DD) of spin echo signals as a function of parameters of the pulse sequence: the amplitude of magnetic field gradient pulses (g), pulse duration (δ), and interval between pulses (t d), conventionally called the diffusion time. DD of the echo is expressed as

$$ R= \exp \left[-{\gamma}^2{\delta}^2{g}^2\left({t}_d-\delta /3\right)D\right] $$
(1)

where R is the relative echo amplitude, which is equal to the ratio of echo amplitudes in the presence and absence of magnetic field gradient, A (g)/A (0); γ is the proton gyromagnetic ratio (the constant is equal to 2.67 × 108 T−1 s−1 for protons); and D is the self-diffusion coefficient.

Diffusional decays were obtained while changing the values of g up to 3 T m−1 with fixed values of δ of 350 μs and t d of 700 ms. An initial part of DD (at g → 0) was fitted with Eq. 1, resulting in an average over a sample value of water diffusion coefficient.

For quantitative description of the experimental results, we employed the formalism of effective coefficient of water diffusion (D ef) (Cooper et al. 1974; Anisimov et al. 1998). The formalism D ef is based on the fact that cell membranes significantly restrict the mobility of water molecules when the distances travelled by molecules become comparable with the cell size. As a result, the measured value of D ef is less than the self-diffusion coefficient of bulk water. The decrement of D ef directly correlates with the decrement of water permeability of cell membranes, which restrict the diffusional process (Cooper et al. 1974; Anisimov et al. 1998). The longer is the diffusion time, the stronger is the effect of the decrease in D ef.

The relaxation time of water molecule magnetic induction in root tissue is a factor that determines the maximum possible, in the NMR experiment, time t d. In plant tissues, the longitudinal (along the steady-state magnetic field) relaxation time (T 1) of magnetization is larger than the transverse one (T 2). We used the three-pulsed sequence (Tanner 1970) for diffusion measurements because in this case, the limitation of t d is determined by T 1. The heterogeneity of a plant cell on a microlevel is the reason for differences in relaxation times of water molecules which are located in different cell compartments. For mature cells of higher plants, these differences are most contrast between an apoplast and a large central vacuole. The fast component of the magnetic induction multiexponential decay was attributed to the apoplast water, and the slowest component with the maximum population was attributed to the cell vacuole water (Van Dusschoten et al. 1995; Van der Weerd et al. 2001, 2002; Velikanov 2007; Sibgatullin et al. 2007; Ionenko et al. 2012). Eighty to 85 % of a cross section area of root segments used in the experiments consists of large (with a diameter of about 40 μm) strongly vacuolated cortical cells (Fig. 1). At t d equal to 700 ms, the echo signal of the root is formed mostly by water molecules of the central vacuoles of cortical cells (Sibgatullin et al. 2007; Velikanov 2007; Ionenko et al. 2012), and the displacement of water molecules becomes comparable with the diameter of cortical cells. As a result, the measured D ef directly correlates with the overall water permeability of a barrier between vacuoles of adjacent cells, namely, the tonoplast, plasmalemma, and thin cytoplasm layer (Fig. 1b). At short t d (1–2 ms), the water self-diffusion in the cytoplasm does not differ from the bulk water self-diffusion, and at long fixed t d (700 ms), changes in D ef might be caused only by membrane barrier properties (Crick 1970). Additional evidence in favor of correctness of the applied method is presented in the “Results” section.

Fig. 1
figure 1

A transverse section of maize root cuts from the elongation zone. Most of the transverse section area is occupied by cortex cells (a). A typical view of cortex cells at a higher magnification obtained by electron microscopy. Diffusion of water molecules between adjacent cell tonoplasts is restricted by two membranes (tonoplast and plasmalemma) and a thin layer of cytoplasm between them (b). Key to the lettering in the figures: CC cortex cells, CW cell walls, V vacuole, C cytoplasm, IS intercellular space

Full size image

Calorimetric measurements

Metabolic heat production of root segments was registered with differential dark microcalorimeter LKB-2277 (Bio Activity Monitor, Sweden). The root segments (50 mg) were submerged into 1 cm3 of 0.25 mM CaCl2 solution (wound stress variant) or 1 cm3 of 0.25 mM CaCl2 plus 100 μM paraquat solution (wound–paraquat stress variant) in the calorimetric test tube with the volume of 3 сm3. The time of the sample temperature stabilization prior to the isothermal (25 °C) registration of heat production was 30 min.

Registration of oxygen consumption

Oxygen consumption was registered with manometric method in a Warburg apparatus (Myers and Matsen 1955). The excised root weights of 150 mg each were placed into Warburg vessels with 3 cm3 of 0.25 mM CaCl2 (wound stress variant) or with 3 cm3 of 0.25 mM CaCl2 plus 100 μM paraquat (wound–paraquat stress variant). After 10-min temperature stabilization, the oxygen consumption was registered every hour at the temperature of 25 °C.

Determination of lipid peroxidation level

The most general indicator of the development of oxidative stress caused by the accumulation of the excessive amount of ROS is the state of peroxidation of lipids (POL), which was judged by the amount of produced malondialdehyde (MDA). The contents of MDA were determined by a method (Heath and Packer 1968) based on the formation of a colored complex of MDA and thiobarbituric acid during heating. Root samples were fixed with liquid nitrogen. The frozen samples were stored at −84 °C before the measurements.

Electron microscopy

The conventional technique of chemical fixation of root segments described earlier was used (for details, see Velikanov et al. 2011). Preparations were examined in a JEM-1200EX electron microscope (JEOL Ltd., Japan).

Statistics

The NMR experiments were repeated for three to five samples. The time of registration of one DD in NMR measurements was determined by the number of accumulations of echo signals (four to five accumulations) and was about 10–15 min. Other experiments (determination of oxygen consumption, heat production, lipid peroxidation level) were performed in five recordings each. The statistical analysis was carried out using the OriginPro 7.0 (OriginLab Corp., Northampton, MA, USA) software.

Chemicals

β-Mercaptoethanol (β-ME), dithiothreitol (DTT), thiobarbituric acid, and paraquat were purchased from Sigma Chemicals, USA.

Results

The time dynamics of D ef changes in response to the excision of maize roots from seedlings and following incubation in 0.25 mM CaCl2 during 1, 2, 3, or 6 h is shown in Fig. 2 (wound stress variant). With the increase in incubation time, the value of D ef first increased from the initial value at the moment of excision to the maximum by the second to the third hour of incubation and then reduced a little by the sixth hour. In order to verify whether the observed increase in D ef is related to membrane aquaporin activity, we tested the aquaporin mercury sensitivity. For this purpose, we added 0.2 mM HgCl2 solution for 20 min after 2, 3, or 6 h of incubation (Fig. 2). Under the effect of mercuric chloride, the value of D ef at the mentioned time points reduced (solid arrows) to a constant level (a horizontal dashed line) corresponding to the mercuric chloride effect on D ef at the 0 time point, i.e., immediately after root excision. The inhibitory effect of mercury was mostly reversed by a 15-min exposure of HgCl2-treated roots in 5 mM β-mercaptoethanol (Fig. 3). After the 2-h incubation of excised roots (wound stress), we added the reductant of thiol groups—dithiothreitol (DTT)—into the incubation medium (concentration of 5 mM) for 15 min. This resulted in the removal of the D ef increase caused by 2-h wound stress (the value of D ef returned to the value of D ef immediately after excision) (Fig. 3).

Fig. 2
figure 2

Time course of effective water diffusion coefficient in response to wound stress (white square) and wound–paraquat stress (white circle). Stressed samples were treated with mercuric chloride (0.2 mM HgCl2, 20 min) to inhibit water transport through aquaporins (black square, black circle, respectively). In both stress conditions, the mercuric chloride treatment resulted in the decrease in D ef (solid arrows for wound stress and dashed arrows for wound–paraquat stress) to the same level (horizontal dashed line) independent of the duration of stress. Bars show SE (n = 5)

Full size image
Fig. 3
figure 3

Effect of HgCl2, β-mercaptoethanol (β-ME), and dithiothreitol (DTT) on the effective water diffusion coefficient (D ef ) in cells of excised maize roots preliminary subjected to 2-h incubation in 0.25 mM CaCl2 (wound stress). —no chemicals; —HgCl2 (0.2 mM, 20 min); —β-mercaptoethanol (5 mM, 15 min) following HgCl2 (0.2 mM, 20 min); —dithiothreitol (5 mM, 15 min)

Full size image

As far as wound stress produces some excess of ROS (Orozco-Cárdenas et al. 2001; Ross et al. 2006; Garrido et al. 2012), we wondered whether D ef changes if we artificially increase ROS production against the background of the initial response to wound stress. For this purpose, we used paraquat which is a well-known “generator” of superoxide radicals in plant cells (Radyukina et al. 2008; Lascano et al. 2012). Paraquat gradually eliminated D ef response to wounding, and by the sixth hour of incubation, the resulting value of D ef achieved a constant level, corresponding to the mercury effect in the wound stress variant (Fig. 2, wound–paraquat stress). Under the treatment with mercuric chloride (0.2 mM, 20 min) in the wound–paraquat variant, the value of D ef after 1, 2, 3, or 6 h of incubation also reduced (dashed arrows) to a constant level, corresponding to the effect of mercuric chloride on D ef at the 0 time point (Fig. 2).

The monitoring of cell membrane permeability using the NMR method is based on the theory of restricted diffusion (Tanner and Stejskal 1968). Water molecule diffusion between two compartments, contrast with respect to magnetization relaxation time (central vacuole and apoplast), in root cells is restricted by a thin layer of cytoplasm with two membranes—tonoplast and plasmalemma (Fig. 1). The degree of diffusion restriction (decrease in D ef compared to bulk water diffusion coefficient) depends on the total permeability of these two membranes and the average vacuole size (Crick 1970). However, D ef under the effect of mercuric chloride during both stresses (wound stress and wound–paraquat stress) reached plateau, corresponding to the effect of mercuric chloride on D ef at the 0 time point (Fig. 2). It means that there occurred no significant changes in the vacuole size in the course of both stresses. Thus, for the studied samples, the increment of the overall water permeability of two membranes—tonoplast and plasmalemma—participated in the increment of D ef under the applied effects. Therefore, in the following analysis of results, we shall directly relate the registered changes in D ef to changes in the overall permeability of these two membranes. Under the transient conditions during cell volume changes, or artificially maintained unidirectional hydraulic flow, the role of plant membrane aquaporins can be partially fulfilled by potassium channels (Wayne and Tazawa 1990). Therefore, the lack of vacuole volume changes in our studies (the steady with respect to transmembrane water exchange conditions) prevented the possibility of potassium channel influence.

The excess ROS accumulation in cells presents an early event, accompanying the effect of actually all abiotic and biotic stresses in plants (Hernandes et al. 2001). The most general indicator of the development of oxidative stress caused by accumulation of the excess ROS is the peroxidation of lipids (POL) level. The amount of malondialdehyde (MDA), as a criterion of POL intensity, is shown in Fig. 4 for both stresses. In the course of wound stress, the amount of MDA gradually reduced or did not change. Under wound–paraquat stress, the amount of MDA increased by the first hour and then remained at a constant level. In the latter case, the amount of MDA at all time points reliably exceeded the amount of MDA at wound stress alone.

Fig. 4
figure 4

MDA content as a measure of lipid peroxidation (POL) level during the development of wound stress (white square) and wound–paraquat stress (white circle). Bars show SE (n = 5)

Full size image

The disturbance of the balance between the ROS level and activity of antioxidant defense systems under stress is associated with the involvement of the complex of metabolic and physiological changes. So, we registered the oxygen consumption and metabolic heat production by root segments. The data on oxygen consumption are shown in Fig. 5. For wound–paraquat stress, there was observed a surplus of the amount of oxygen consumed by root segments compared to the wound stress variant for all time points. The corresponding data on heat production are shown in Fig. 6. In the case of wound–paraquat stress, the total level of heat production during the first 3 h was twice as large as the one for wound stress. But in the wound stress variant, after 3 h, there began the significant increase in heat production. In the wound–paraquat variant, the increase in heat production during this time period was not that drastic.

Fig. 5
figure 5

The rate of oxygen uptake for excised roots during wound stress (white square) and wound–paraquat stress (white circle). Bars show SE (n = 5)

Full size image
Fig. 6
figure 6

Heat production of excised roots during wound stress (white square) and wound–paraquat stress (white circle). The level of heat production in the wound–paraquat variant during the first 3 h is twice larger compared to wound stress alone. Beginning with the third hour, a considerable increase in heat production is observed for wound stress. SE was less than 7 % (n = 5)

Full size image

Discussion

Maize roots are widely used to study cell water self-diffusion using the NMR method (Anisimov et al. 1998; Ionenko et al. 2012). In our experiments, the maize roots were stressed by excising them from seedlings and subsequent incubating in growth medium for several hours. Earlier, this method of wound stress was used for wheat (Minibayeva et al. 1998) and pea (Karimova and Petrova 2007) seedling roots. It was shown that in this case, similarly to other kinds of stress factors, a complex of metabolic and physiological changes in cells was switched on (Minibayeva et al. 2009). In particular, wound stress activated root cells to produce oxygen radicals. The increase in production of superoxide anion radicals and a corresponding increase in peroxidase activity were shown (Minibayeva et al. 1998, 2001, 2009, 2012). It is well established that production of ROS is one of the universal responses of various plants to wounding (Orozco-Cárdenas et al. 2001; Ross et al. 2006). The mechanism of ROS production in excised roots was studied in detail by Garrido et al. (2012).

During maize root response to wound stress (Fig. 2), there was observed the increase in D ef from the initial value at the time point of excision to the maximum one by the second to the third hour of incubation. These changes in D ef result from corresponding changes in the effective total water permeability of two membranes (tonoplast and plasmalemma) in excised root cells (see “Materials and methods”). However, the water uptake capacity of plant roots (i.e., their hydraulic conductivity) is known to be determined largely by aquaporins of the plasma membrane intrinsic protein (PIP) subfamily (Boursiac et al. 2008). The increase in D ef mostly resulted from the increase in plasmalemma permeability.

The contribution of aquaporin-mediated water transport across the membrane is, as a rule, estimated using an inhibitor. Mercury is the most commonly applied inhibitor for aquaporins (Martre et al. 2001; Javot and Maurel 2002; Volkov et al. 2007; Bramley et al. 2009; Hachez and Chaumont 2010). Mercury can inhibit the channel activity of a hetero-oligomer composed of Z. mays PIP1;2 and PIP2;5 through its interaction with a cysteine residue located in the loop A of Z. mays PIP1;2, a residue involved in disulfide bond formation between PIP monomers (Bienert et al. 2012). The structure of the PIP2;1–mercury complex has been solved and reveals three binding cysteine residues for mercury, which could act on the channel gating (Frick et al. 2013). However, strangely, reconstitution of PIP2;1 in liposomes showed that mercury did not inhibit but increased their water channel activity in a cysteine-independent way, possibly through changes in the properties of the lipid bilayer (Frick et al. 2013). These authors considered it premature to apply the results, obtained with liposomes, to real plant membrane aquaporins and pointed to the need of special studies. Apparently, the mercury test for plant membranes should be applied with caution (Chaumont and Tyerman 2014).

In the case of wound stress, the addition of mercuric chloride after 2, 3, or 6 h of incubation resulted in the decrease in membrane water permeability to the constant level corresponding to the mercuric chloride effect at the 0 time point (Fig. 2, horizontal dashed line). Obviously, we failed to observe the abnormal response of the membrane water permeability to mercury effect, predicted by Frick et al. (2013). The inhibitory effect of mercury was mostly reversed after the addition of 5 mM β-mercaptoethanol in the incubation medium for 15 min (Fig. 3). This corresponds to the well-known response of aquaporins to the treatment with HgCl2 and β-mercaptoethanol (Preston et al. 1993). After the 2-h wound stress, DTT was added into the incubation medium for 15 min, resulting in the removal of the D ef increase caused by 2-h wound stress (Fig. 3). The increase in D ef under wound stress and its subsequent removal by DTT appear to show the processes of oxidation and reduction of SH– groups of water permeating paths in membranes, respectively. In this case, the role of lipid bilayer can be excluded since mercury reduced D ef to the constant level (Fig. 2) during the whole period of root response to the stress. One can suppose that the increase in membrane water permeability under wound stress is associated with aquaporin activation.

Wound stress was shown earlier to activate wheat root cells to produce oxygen radicals (Minibayeva et al. 1998, 2001, 2009, 2012; Ross et al. 2006). In addition, the superoxide production achieved maximum in 2 h after root excision and at the beginning of incubation (Minibayeva et al. 1998). One can suppose that aquaporin activation during wound stress results from redox processes of thiol groups of these proteins. In vitro studies (Ampilogova et al. 2006) gave the experimental evidence in favor of this concept. The isolated from pea roots plasmalemma responded adequately to the presence in the medium of agents that oxidize (diamide) or reduce (dithiothreitol, tributylphosphine) thiol groups by changing the balance in the number of SH– groups and S–S– bonds. Changes in this balance resulted in modification of plasmalemma osmotic water permeability in vitro: it increased with the increase in S–S– bonds and decreased with the increase in the number of SH– groups.

Bienert et al. (2012) tested whether the loop A cysteine residue involved in disulfide bond formation between PIP monomers had an effect on oxidative or reductive treatments that might affect PIP gating. This hypothesis was investigated by measuring the osmotic water permeability coefficient (P f) in oocytes expressing ZmPIP2;5 after treatment with oxidants (H2O2 or t-butylhydroperoxide) or reductants (2-mercaptoethanol or DTT). No significant differences were observed in the P f values for non-treated, or oxidant-treated, or reductant-treated oocytes, indicating that the disulfide bonds, as well as other potential modifications of the cysteine thiol group, do not affect the activity of ZmPIP2;5 (Bienert et al. 2012). The same researchers showed that the loop A cysteine residue of ZmPIP1;2, but not that of ZmPIP2;5, is involved in mercury sensitivity. The latter does not exclude the possibility that PIP isoforms, differing from ZmPIP2;5, or some other cysteine residue can be involved in ROS regulation. Indeed, the research (Ampilogova et al. 2006) clearly demonstrated that a change in the ratio of thiol to disulfide bonds in plant plasma membrane proteins correlated with a change in water permeability. These researchers supposed that it is not the loop A cysteine residue, but two cysteine residues from CLGAIC sequence in the transmembrane domain, associated with the loop C of PIPs, that might be involved into the realization of the revealed regulation. The short distance between these cysteine residues allows the coupling and redox transitions between thiol groups of this pair of cysteine residues (Ampilogova et al. 2006). Apparently, the possibility of direct influence of ROS on aquaporin activity on the molecular level is not yet studied in detail. However, our data (in vivo) on regulation of membrane water permeability under wound stress agree nicely with the abovementioned data on regulation of water permeability for the plasma membrane vesicles (Ampilogova et al. 2006).

In the wound–paraquat variant, the increment of D ef inherent to wound stress appeared to reduce after the first hour, and by the fifth to the sixth hour of incubation, D ef reached the value on the plateau (dashed line), which tallies with the blocked by mercury state of the membrane at the 0 time point. Under mercuric chloride effect, the value of D ef at all time points of exposition (1, 2, 3, and 6 h) achieved a constant level corresponding to that for wound stress. This fact additionally indicates that in our experiments, mercury does not influence water permeability of the membrane lipid phase. We did not observe the abnormal effect of mercuric chloride (see aforesaid, Frick et al. 2013). The mercury test in the experiment with wound–paraquat stress proves that, in this case as well, the time course of membrane water permeability decrease might be determined by the corresponding aquaporin reaction.

By using the measurements of MDA, oxygen consumption, and heat production, as conventional factors of general oxidative stress, we showed the presence of real differences between two variants of the experiments. The development of wound–paraquat stress was characterized by the exceeding amount of oxygen consumed by root segments during all time points of exposition compared to the wound stress experiment (Fig. 5). In addition, the overall level of heat production during the first 3 h of simultaneous effect of stressors was twice as large as for the wound stress alone (Fig. 6). During the disturbance of coordination of processes occurring in the mitochondrial matrix, caused by stress impact, there increases a leakage in the electron transfer chain. The direct interaction of leaking electrons with oxygen is the principal pathway of superoxide synthesis during the paraquat treatment (Lascano et al. 2012). Comparing the data from Figs. 5 and 6, one can see that the increase in oxygen consumption during wound–paraquat stress is coupled with a large gain in heat production. Presumably, a large part of the exceeding amount of the consumed oxygen is spent for production of ROS, which activate exothermal reactions and appear to be one of the factors of the overall heat production increase. Apparently, the observed changes in energy metabolism imply that paraquat indeed fulfills the assigned function of the inductor of ROS excess in our experiments as well.

It should be noted that after 3 h of wound stress, a significant increase in heat production began. With the restricted energy resource in the excised root, the oxidative stress can induce a catabolic process of intracellular degradation of some macromolecules and organelles—autophagy (Minibayeva et al. 2012). This process is characterized by production of one- and two-membrane vesicles (autophagosomes), containing the fragments of degradation. Presumably, this exothermal process causes the abovementioned increase in heat production during wound stress. This kind of exothermal process can be considered as an independent proof of the presence of ROS excess (compared to the norm) during the initial hours of development of wound stress in our experiments.

The double excess of heat production during the first 3 h of the wound–paraquat stress provides a convincing evidence of the general increase in oxidative reaction intensity compared to that for wound stress alone. It is also indicated by POL intensification that is directly confirmed by a higher level of MDA in the wound–paraquat variant. Therefore, two variants of our experiments differed essentially in levels of general oxidative stress. At the same time, characteristics of the general oxidative stress allowed us to judge indirectly about both ROS production and the state of the membrane lipid phase under the effect of two stress factors.

The results of our experiments point to the counter-directed response of membrane water permeability under wound and wound–paraquat stresses (increase and decrease, respectively). There apparently exists an optimal level of oxidative stress (excess of ROS), against the background of antioxidant cell activity, to maximize water flow across root cells. In our experiments, wound stress, apparently, is associated with the weak oxidative stress and wound–paraquat stress—with the considerably increased ROS impact on the cells.

This explanation is supported by literature data. Thus, the hydraulic permeability of maize roots increases twice during the exogenous 1-h treatment with 100 μM hydrogen peroxide (Aroca et al. 2005). On the contrary, 1-h treatment with 2 mM hydrogen peroxide (the higher dose) resulted in inhibition of hydraulic permeability in the maize roots (Boursiac et al. 2008; Ehlert et al. 2009). Other authors measured hydraulic permeability (L p) of parenchyma cells in the midrib tissue of maize leaves using a cell pressure probe (Kim and Steudle 2009). In response to low light intensity of 200 μmol m−2 s−1, the cell L p was increased. In contrast, high light intensities of 800 and 1800 μmol m−2 s−1 decreased the cell L p. Together with it, the treatment of the tissue with oxidants (H2O2 and hydroxyl radical produced in the Fenton reaction) had an effect similar to high light, and the presence of the antioxidant (glutathione) tended to prevent the inhibition by high light.

The state of the membrane lipid phase during the wound–paraquat stress differed from that during the wound stress (Fig. 4). Therefore, changes in the lipid phase related to the intensification of lipid peroxidation can be supposed to affect directly the conformation and transport function of integral proteins—aquaporins, resulting in their downregulation. Under wound–paraquat stress, the proteins participating in the system of intracellular phosphorylation, for example, protein phosphatase of the tyrosine type, can experience destruction and inactivation by ROS (Karimova and Petrova 2007). Aquaporin traffic from the plasma membrane to the intracellular vesicular membrane structures can also provide the important mechanism of downregulation under this stress (Boursiac et al. 2008). However, for all three processes (peroxidation of lipids, cell protein destruction, and relocalization of aquaporins), it is possible that ROS do not gate aquaporins through a direct oxidative mechanism (Boursiac et al. 2008).

In this paper, we used the NMR method of water self-diffusion measurement in maize root cells, allowing for express monitoring of cell membrane water permeability. Changes in membrane water permeability was studied in response to (i) root excision from seedling and the following 6-h incubation in the growth medium (wound stress) and (ii) the superposition of wound stress plus paraquat. The noticeable difference in the general level of the oxidative stress was shown for these variants of the experiments. Wound stress is associated with the weak oxidative stress, and wound–paraquat stress with the considerably increased ROS impact on the cells. Results of our experiments point to the counter-directed response of membrane water permeability under wound and wound–paraquat stresses (increase and decrease, respectively). There apparently exists an optimal level of oxidative stress (excess of ROS), against the background of antioxidant cell activity, to maximize water flow across root cells. We showed the reversible by dithiothreitol increase in cell membrane water permeability in maize roots under wound stress. The obtained data in our experiments for root cell membranes in vivo under this stress (weak oxidative stress) agree nicely with the data presented in literature about upregulation of water permeability of isolated plasma membrane vesicles, occurring under the conditions of changes in equilibrium between the number of SH– groups and S–S– bonds towards the latter. The possibility of direct influence of ROS on aquaporin activity on the molecular level is not yet studied in detail. In our experiments, the applicability of mercury test to aquaporin activity was verified. The results of wound stress effect, obtained using this test, are discussed in terms of oxidative upregulation of aquaporin activity by ROS. It is supposed that under wound–paraquat stress, ROS affected the aquaporins not directly, but via such processes as peroxidation of lipids, intracellular protein destruction, and relocalization of aquaporins in the cell. Results of our research might be important to cope with environmental factors.