The Distribution of Aquaporin Subtypes (PIP1, PIP2 and γ-TIP) is Tissue Dependent in Soybean (Glycine max) Root Nodules

Abstract

• Background and Aims The inner cortical cells (IC-cells) of legume root nodules have been previously shown to regulate the resistance to nodule O₂ diffusion by a rapid contraction/expansion mechanism, which controls the volume of intercellular spaces and their occlusion by a liquid phase. The expression of aquaporins in IC-cells was also found to be involved in this nodule O₂ diffusion mechanism. The aim of this study was to compare the expression of plasma membrane intrinsic proteins (PIP) aquaporin isoforms with tonoplast intrinsic protein (γ-TIP) in both IC-cells and adjacent cell types.

• Methods Using immunogold labelling in ultra-thin sections of Glycine max nodules, the expression of two PIP isoforms was observed and compared with the γ-TIP pattern.

• Key Results The plasma membrane aquaporins PIP1 and PIP2 were expressed more in IC-cells and endodermis than in pericycle and infected cells. The tonoplast aquaporin γ-TIP has shown a distribution pattern similar to that of the PIPs.

• Conclusions PIPs and γ-TIP aquaporins are highly expressed in both plasmalemma and tonoplast of nodule IC-cells. This distribution is consistent with the putative role of water fluxes associated with the regulation of nodule conductance to O₂ diffusion and the subsequent ATP-dependent nitrogenase activity. In the endodermis, these aquaporins might also be involved in nutrient transport between the infected zone and vascular traces.

INTRODUCTION

Water flux into cells is known to occur not only by diffusion through lipid bilayers but also through water channel aquaporins (Chrispeels and Maurel, 1994). The aquaporin-mediated water flux has been demonstrated for tonoplast aquaporins (TIP = tonoplast intrinsic proteins) and plasmalemma aquaporins (PIP = plasma membrane intrinsic proteins) of plant cells (Steudle and Henzler, 1995). The distribution of both types of aquaporins may be a key point in understanding water flows in plant tissues.

Recently, a series of experiments was carried out to analyse the distribution of plasma membrane and tonoplast aquaporin proteins among tissues and cells by a quantitative immunocytological method. Previously a great amount of γ-TIP aquaporin in the motor cells in Mimosa pudica pulvini was found (Fleurat-Lessard et al., 1997). It is proposed that this quantity is associated with the ability of motor cells to exhibit large and rapid turgor variations in response to stimuli affecting leaf posture. Moreover, in soybean nodules, it had been found previously that the expression of γ-TIP in the tonoplast of the inner cortex cells (IC-cells) was twice that found in other cell types. This expression might be related to the ability of IC-cells to change their shape and volume rapidly (Serraj et al., 1995) in response to stimuli affecting nodule permeability to O₂ diffusion (Serraj et al., 1998).

Understanding the molecular mechanisms of nodule O₂ diffusion in the Rhizobia–legume symbiosis may be essential to understand how nitrogen fixation activity is regulated. This could be important for agronomic improvement through better symbiotic N₂ fixation. It has been hypothesized that by a contraction/expansion mechanism the IC-cells may control the intercellular space volume (Serraj et al., 1995), where the high resistance to O₂ diffusion would be due to their occlusion by a liquid phase (Minchin, 1997). With this in mind, observations in the present work were focused on the expression of two PIP isoforms in four adjacent cell types of root nodules, i.e. cells in the infected zone at the centre of the nodule, IC-cells, endodermis and pericycle surrounding the vascular traces (Fig. 1A). In addition, the observation that there is γ-TIP aquaporin in the endodermis supplements previous reports of this protein being detected. Also, its distribution compared is with that of PIP1 and PIP2 in the same cell types.

Fig. 1.

(A) Cross-section of a soybean nodule. IZ, infected zone; IC, inner cortex; E, endodermis; P, pericycle around vascular bundle; X, xylem vessels; PH, phloem. Arrows indicate Casparian strips in the endodermis layer. Scale bar = 10 μm. (B) Bands detected in 10 μg total proteins of nodule plasmalemma: antibodies saturated with PIP1 (a) and PIP2 (b) peptides; 1/1000 diluted polyclonal anti-radish PIP1 (c) and PIP2 (d); 1/2000 diluted secondary antibody, GAR-HRP.

MATERIALS AND METHODS

Soybean Glycine max (L.) ‘Kingsoy’ seeds were inoculated with Bradyrhizobium japonicum strain PJ17 and grown in an aerated liquid nutrient solution (Serraj et al., 1995).

Purified fractions enriched in plasmalemma vesicles were extracted from root nodules using a protocol derived from that used for Beta vulgaris (Gallet, 1989). Thirty grams of root nodules were crunched in 275 mL of extraction medium containing 500 mm of sorbitol, 50 mm HEPES, 5 mm EDTA, 2H₂O, 2 mm DTT, 1 % (w/v) PVP, 1 % (w/v) PVPP, 0·01 caffeine, 0·1 % (w/v) ascorbic acid and 275 mm PMSF. Centrifugation of the liquid phase occurred at 10 000 g for 13 min and that of the supernatant at 50 000 g for 30 min. The purification of plasmalemma vesicles was obtained on the medium phase separation (6·5 % dextran T6500, 6·5 % PEG 3350).

Thin sections of chemically fixed nodules, included in London White Resin, were immunotreated as previously described (Fleurat-Lessard et al., 1997) and observed with a microscope (JEM 1010; JEOL, Tokyo, Japan). For this immunocytochemical analysis, the antibodies used were specific to aquaporin isoforms of radish tonoplast (γ-TIP; Maeshima, 1992) and plasma membrane [PAQ1 and PAQ2 (Suga et al., 2001) that are the equivalent of PIP1 and PIP2 in Arabidopsis thaliana]. The number of gold particles indicating immunoreaction on thin sections was calculated from three independent fixations and at least four immunodetections. Counts were made on six nodules, in five cells per nodule, i.e. a total of 30 cells for each cell type. Determinations were made on a 500-µm length of plasmalemma (or tonoplast). The means of antigenic sites recognized in the four cell types were compared statistically using Student's t-test. The distribution of each aquaporin was considered independently since the respective protein titre was unknown. Controls in each cell type indicated that the background noise had the same value.

RESULTS AND DISCUSSION

Knowledge of the expression patterns of aquaporins is essential, as it provides clues about their function at the whole plant level. The expression pattern may suggest in which cells rapid transmembrane water transport is especially important. The coexistence of members of the same PIP (and TIP) subfamily raises the questions of their specific functions and their possible collaboration under various physiological conditions.

The PIP antibodies used are directed against the conserved N-terminal peptides of radish PIP1 and PIP2 and, therefore, may cross-react with several PIP1 (PIP2) isoforms, including those of different species, but are specific for either PIP1 or PIP2. It is, however, relevant to mention here that the amino acid sequences of PIP1 and PIP2 proteins are 67 % identical and that identity between their subfamilies is >91 % (Suga et al., 2001). As shown in a Western blot a cross-reaction occurred in soybean root nodules (Fig. 1B) where the anti-PIP1 and anti-PIP2 antibodies gave, respectively, an immunostained band of 30 kDa in the plasmalemma fractions. The anti-γ-TIP recognized a band of 23 kDa in the microsomal fraction of soybean nodules as demonstrated by Serraj et al. (1998).

Using the anti-PIP1 antibody on ultra-thin sections of soybean nodules the density of gold particles, corresponding to aquaporin sites, was significantly higher in the plasma membrane of IC-cells (Fig. 2A) than in the control, where this antibody has been saturated with the PIP1 peptide (Fig. 2B). This isoform was as abundant in the endodermis surrounding the vascular traces as in the IC-cells. By contrast, little of this protein was expressed in the pericycle and in the cells of the infected zone (Table 1).

Fig. 2.

(A) Immunolabelling (arrows) of the PIP1 protein in plasmalemma (black arrowhead) of IC-cells, no labelling on tonoplast (white arrow). (B) Scarce and non-specific labelling in control (antibody saturated by the corresponding peptide). Legend as in (A). (C–E) High immunolabelling (arrows) of γ-TIP protein in the tonoplast of IC-cells (C and D) and endodermis cells (E), in particular in small vacuoles (*). There is no labelling in the plasmalemma. Legend as in (A). Sections incubated with 1/50 diluted anti-radish PIP1 (A) and 1/50 diluted anti-radish γ-TIP (C–E) followed by GAR-IgG-15 nm gold. CS, Casparian strip; V, main vacuole; W, wall. Scale bar = 1 μm.

Table 1.

Distribution of aquaporins in four adjacent cell types of root nodules

Antibody/cell type	Infected cells	IC-cells	Endodermis	Pericycle
PIP1	0·4 ± 0·3a	6·4 ± 0·8b	5·9 ± 0·3b	2·1 ± 1·6c
PIP2	0·5 ± 0·3a	3·3 ± 0·3b	2·4 ± 0·4b	1·4 ± 0·3c
γ-TIP	1·1 ± 0·3a	21·2 ± 1·8b	14·8 ± 3·2c	2·8 ± 0·4d
Controls*	0·3 ± 0·1a	0·3 ± 0·1a	0·4 ± 0·1a	0·3 ± 0·1a

Means ± standard deviation (n = 30) of antigenic sites (labelled by gold particles) per 10 µm of membrane. Means followed by the same letter, within the same line, do not differ significantly at P < 0·05 according to Student's t-test.

^*Antibodies saturated with corresponding peptides.

The PIP2 aquaporin was also immuno-detected in the sections of soybean root nodules. Its distribution in the four nodule tissues observed showed the same pattern as that of PIP1 (Table 1). In addition, in mature tap roots of radish, as in etiolated cotyledons, i.e. in non-photosynthetic organs a low PIP2 level has been reported (Suga et al., 2001). Similarly, it has been observed that the PIP2 density range is lower (minus 30 %) in the soybean nodules than in the leaves, the source of the photo-assimilates (result not shown). Therefore, the involvement of this PIP2 isoform in the water transport associated with the loading of photo-assimilates in leaf minor veins is suggested.

As previously documented (Serraj et al., 1998) the γ-TIP was abundant in the IC-cells (Fig. 2C) but scarce in the pericycle and the infected cells of soybean root nodules. In this study, it was found that γ -TIP was also abundant in the endodermis (Fig. 2E and Table 1). Also it was observed that high labelling occurred not only in the central vacuole but also in small vacuoles in both IC- (Fig. 2D) and endodermis cells (Fig. 2E). This labelling might be related to the regulation of aquaporin expression to recover IC-cell turgor after O₂-induced collapse. Vera-Estrella et al. (2004) recently suggested that the mechanism responsible for osmotic stress-induced changes in the ice plant involved the redistribution of a TIP1;2 aquaporin in endomembranes. The labelling of small vacuoles in the present work also point towards a putative role in metabolite routing between prevacuolar compartments and the central vacuole, similar to the one recently assigned to TIP1;1 aquaporin in Arabidopsis (Ma et al., 2004). In addition, as casparian bands may block the apoplastic pathway, the well-expressed aquaporins may provide an alternative path of low resistance in endodermis (Steudle and Henzler, 1995). Thus, in root nodules this layer may form a water reservoir able to play a buffering role against water constraints, as proposed in root endodermis of maize where aquaporins are also highly expressed (Barrieu et al., 1998).

In the inner cortex and endodermis of root nodules, the density range of PIP1, PIP2 and γ-TIP aquaporins is the same as that occurring in other mature cells involved in large nutrient exchanges, in particular motor cells in leaf pulvini (Fleurat-Lessard et al., 1997). Nodule IC-cells have structural features similar to those of motor cells, particularly a thin cytoplasm layer all around the large central vacuole and a distorted wall when these cells change shape (Serraj et al., 1995). In motor cells, the shape changes depend on turgor variations associated with ion compartmentation, in particular reversible fluxes of K⁺ and Cl⁻ (Satter and Galston, 1981) The rapid (<1 s) turgor variations that occur in several sensitive plants may involve water channels known to favour rapid water fluxes. Similar ion and water fluxes might be involved in the O₂-induced IC-cell collapse in nodules (Denison and Kinraide, 1995; Serraj et al., 1995). Indeed, the higher expression of aquaporins in IC-cells, compared with the adjacent cells, suggests that water fluxes may be rapid in the inner cortex. In addition, changes in water status across the plasma membrane may induce fluxes across the vacuolar membrane, since the vacuole constitutes a potential water reservoir in plant cells. Thus, the relatively high densities of γ-TIP and PIP in IC-cells are consistent with the hypothesis of osmoregulation of nodule conductance to O₂ diffusion (Serraj et al., 1998). Another important feature of IC-cells is the existence of intercellular spaces containing a liquid phase, which may favour water fluxes between cytoplasm and intercellular spaces (Purcell and Sinclair, 1993). It has been suggested that PIP proteins could increase the gas permeability of membranes in nodules (Prasad et al., 1998; Guenther and Roberts, 2000). Further research is needed to investigate this PIP role in IC-cells.

Although a direct link cannot be made between the presence of aquaporins in situ and their activity, provided the measurements reflect a real quantitative difference in the abundance of γ-TIP and PIP, the present results would be in accordance with water permeability of the tonoplast being greater than that of the plasmalemma in both IC- and endodermis nodule cells. In this context, VM 23 (γ-TIP) is known to account for about 30–40 % of the vacuole membrane proteins but PIP1 is <10 % of the plasma membrane proteins (Higuchi et al., 1998). However, until now differences between the osmotic water permeability of plasmalemma compared with that of tonoplast have not yet been resolved.

Root nodules display very original physiological features as they behave as sinks for photo-assimilates exported from the leaves and as sources that provide nitrogenous compounds to aerial parts. Their large water content may explain the high sensitivity of these organs to water deficiency (Serraj et al., 1999). Further work should focus on the possible rapid turnover of aquaporins in nodule tissues related to regulation of exchanges of assimilates between source and sink function in these organs.