INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Water deficit is one of the most detrimental environmental stresses in plants; it leads to large physiological modifications, such as photosynthesis reduction, transcriptional and post-transcriptional regulation of various genes, protein turnover, and osmolyte biosynthesis (Bohnert et al., 1995). In maize (Zea mays) the increase of soluble and insoluble invertase activities during pollination and early kernel development was blocked by water stress conditions (Zinselmeier et al., 1995); this effect was correlated with a low level of reducing sugars, an increase in Suc concentration, starch depletion, and an inhibition of ovary growth leading to abortion. In contrast it was shown recently that a marked accumulation of hexoses was correlated to an induction of vacuolar invertase activity in mature maize leaves under water stress (Pelleschi et al., 1997). These results address the question of an organ- and tissue-specific response of invertase to water stress.

Suc is the major end product of leaf photosynthesis and also is the phloem-transported sugar in most higher plants. Suc cannot be used directly for metabolic processes, but must be cleaved into hexoses by invertase (-fructosidase, EC 3.2.1.26) or Suc synthase (UDP-D-Glc: D-Fru-2--D-glucosyl-transferase, EC 2.4.1.13; Susy) before entering into carbohydrate metabolism.

Invertase is present in most plant organs in multiple forms, which are classified by pH optima (acid and neutral/alkaline), solubility properties (soluble and insoluble), and cellular locations. Cell wall-bound and vacuolar forms are acidic, the first one being insoluble and the second one soluble; the cytosolic is both neutral/alkaline and soluble. The physiological role of invertase isozymes appears to be diverse and recent studies suggest that their function varies depending upon the organ/tissue or cells in which they are expressed. Invertases take part in Suc partitioning between source and sink organs. In young sink organs cell wall-bound invertases can enhance carbohydrate movement by maintaining a Suc concentration gradient between phloem and sink tissues (Sonnewald et al., 1991; Sturm et al., 1995; Cheng et al., 1996); in mature organs they can alter Suc and starch balance (Huber, 1983). More specifically, in source organs, invertase may limit carbohydrate export rate by hydrolyzing Suc, thus hindering Suc loading into phloem. The resulting hexose could also be involved in a feedback regulation of photosynthesis (Foyer, 1987).

To study the roles of invertases expressed in different cellular compartments, transgenic plants expressing yeast invertase have been produced in potato, tomato, tobacco, and Arabidopsis (Sonnewald et al., 1991, 1997; Heineke et al., 1994). Expression of a yeast-derived invertase in the cell wall or in the vacuole in potato plants leads to an accumulation of hexoses, Pro and starch, an inhibition of photosynthesis, and an increase in respiration (Heineke et al., 1994; Büssis et al., 1997). Furthermore in transgenic tobacco plants expressing yeast-derived cell wall or vacuolar invertase the reduction of photosynthesis-related gene transcripts and accumulation of pathogenesis-related gene transcripts were associated with hexose accumulation (Herbers et al., 1996).

The effect of invertase on the expression of transcripts may be better understood when considering sugars as physiological signals able to repress expression of photosynthesis-related genes (Sheen, 1990; Jang et al., 1997), or activate gene expression in Suc and starch synthesis and/or degradation (Chen et al., 1994; Koch, 1996), as well as specific stress-response genes in infection and wounding (Herbers et al., 1996; Ehness et al., 1997). Sugars can also regulate cell differentiation, elongation, and growth (Wu et al., 1993; Weber et al., 1996). In this regard some works deal with invertases as potential producers of signaling factors (Glc and Fru) via Suc hydrolysis (Koch et al., 1995; Herbers et al., 1996; Weber et al., 1996).

In maize two genes/cDNAs (Incw1 and Incw2) encoding two cell wall-bound invertases (Shanker et al., 1995; Cheng et al., 1996; Taliercio et al., 1999) and two partial cDNAs (Ivr1 and Ivr2) encoding two vacuolar invertases (Koch et al., 1995; Xu et al., 1995) have been cloned. We have recently cloned two new genes (Incw3 and Incw4) encoding cell wall invertases (Kim et al., 2000).

In the present paper the expression of three types of invertase activities (vacuolar, cell wall, and cytosolic) and of Susy activity was analyzed in well-watered and water-stressed vegetative organs (mature leaf, leaf sheath, young leaf, primary roots, and adventitious roots). The transcript levels for all the genes cloned so far and the invertase protein levels of two types (vacuolar and cell wall-bound) were investigated. It was observed that Ivr2, encoding a vacuolar invertase, was the only gene among all the known invertase genes to be specifically induced under water stress and in an organ/tissue specific manner. In addition a diurnal regulation of the induced invertase activity and of the Ivr2 mRNA content was observed in shoots. The putative role of specific temporal and spatial patterns of Ivr2 expression during water stress is discussed in relation to photoassimilate source-sink partitioning, osmotic adjustment, and sugar signaling.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Activity of Vacuolar Invertase Is Induced by Water Stress in Specific Organs

The activity of all Suc cleaving enzymes (vacuolar, cell wall, cytosolic invertases, and Susy) in well-watered and water-stressed maize source and sink organs was investigated for 11 d after deprivation (Fig. 1).

View larger version (41K): [in this window] [in a new window]   Figure 1.   Effect of drought on acid soluble (vacuolar), acid insoluble (cell wall), and neutral (cytosolic) invertase activity, and Susy activity in five vegetative organs from well-watered and water-stressed maize plants. Samplings were carried out at 9 AM (3 h after light on) from d 2 to 11 after water interruption. White symbols, Control plants; black symbols, water-stressed plants. Mean ± SE from four independent determinations. Control/stress Student's comparison: *, P < 0.05 significant; **, P < 0.01 very significant; ***, P < 0.001 highly significant.

In well-watered plants activities of all invertases were higher in sink organs (young leaf, primary, and adventitious roots) than in the source leaf with the exception of cell wall-bound invertase activity, which was undetectable in young leaf (Fig. 1). Water stress effect was either nil or marginal on cell wall-bound and cytosolic invertases and on Susy activities, whatever the organ tested (Fig. 1). In contrast a vacuolar invertase activity was markedly induced by water stress in mature leaf, leaf sheath, and primary roots (Fig. 1), but not in adventitious roots. Early effects were detectable at 2 d in primary roots and young leaves and a few days later in mature leaf laminae and sheath. Maximum effect was, in general, noted around 5 to 8 d.

Hexoses Accumulate in All Water-Stressed Organs

To determine if stress-induced vacuolar invertase activity alters carbohydrate pools, Suc and hexose (Fru and Glc) contents were measured (Fig. 2). In control plants Suc levels were higher than hexose levels in both mature leaf (source) and sheath (sink), whereas the opposite was observed in other sink organs (young leaf and both root types). Water stress produced a marked hexose accumulation in all organs, whereas Suc increased, to a lesser extent, in leaf sheath and in primary and adventitious roots. The daily hexose accumulation rate was higher in sinks (from 15-30 mg g¹ fresh weight d¹, depending the organ) than in source leaf (7 mg g¹ fresh weight d¹). This accumulation was not well correlated to invertase activity, but it was related to water loss: relative water content (RWC) decreased earlier and more rapidly in young leaf and primary root than in mature leaf (Fig. 3). This raised the question as to a concentration effect due to water loss. Nonetheless this question may discarded in mature leaf since vacuolar invertase activity and hexose content increased in parallel before any significant changes occurred in RWC (compared values at 5 and 6 d in Figs. 1-3). This retarded RWC response tends to dismiss leaf RWC as being the trigger for a response to water stress.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Effect of drought on Suc and hexoses (Glc and Fru) concentrations on a fresh weight basis in five vegetative organs from well-watered and water-stressed maize plants. Same samples as in Figure 1.

View larger version (22K): [in this window] [in a new window]   Figure 3.   Variation of RWC in mature leaf, young leaf, and root after water deprivation.

The Induced Activity Presents Diurnal Changes

It is interesting that the stress-induced vacuolar invertase activity showed diurnal modulations in stressed mature leaf, in sheath, in young leaf, and to a lesser extent in primary roots (Fig. 4, compare beginning and end of photoperiod). This was confirmed by measuring vacuolar invertase activity every 3 to 4 h for 3 d in water-stressed leaves under 16-h light-day cycles (not shown). The invertase activity increased during the dark period, reached a peak before the end of the dark period, and decreased rapidly after 3 h in the light period (Kim, 1998).

View larger version (29K): [in this window] [in a new window]   Figure 4.   Diurnal variation of acid soluble invertase (vacuolar) in five vegetative organs of watered (white bars) and water-stressed plants (gray bars). Samplings were carried out at the beginning, 9 AM (3 h after light on), and at the end, 9 PM, of the photoperiod (1 h before light off). Symbols for statistical comparisons between stressed and control plants as in Figure 1.

Drought Induced a Parallel Increase of Both Activity and Protein for Vacuolar Invertase

To understand the induction of vacuolar invertase activity it is necessary to evaluate the variation in invertase protein. Specificity of the anti-IVR2 antiserum was checked on two-dimensional isoelectric focusing (IEF)-SDS-gel blots by comparison with antiserums raised against a carrot vacuolar invertase and an Arabidopsis cell wall invertase peptide. The carrot vacuolar invertase antiserum displayed the same immunospot pattern as the IVR2 antiserum on two-dimensional IEF-SDS gel blots (data not shown). The IVR2 pattern (Fig. 5, A and B, spots a and b) was clearly distinct from that detected by Arabidopsis cell wall invertase antiserum (Fig. 5, C and D, spots e, f, and g) in mature leaf. Thus IVR2 antiserum does not recognize cell wall invertases and appears to be specific to vacuolar invertases. In mature leaf vacuolar invertase IVR2 antiserum showed at least two peptide groups with different M_rs (Fig. 5, A and B): group a (68 kD) composed of an apparent spot cluster showing a unique pI of 5.0, and group b (74 kD) with aligned spots between pI 4.7 and 7.0, but with two main spots between pI 5.8 and 6.0. Main spots in a and b groups from the morning samples (at 9 AM) were larger in water-stressed (Fig. 5B) than in control organs (Fig. 5A). The same samples probed with cell wall antibodies (Fig. 5, C and D) labeled three peptides groups: e (52 kD, pI 4.8-5.1), f (60 kD, pI 4.9-5.6), and g (44 kD, pI 7.8-8.2). The fragment g, being only 44 kD, could be a partial fragment, which was frequently detected by SDS-PAGE during invertase purification. This intrinsic protein lability is reported for many plant invertases (Unger et al., 1992). Comparison of band intensities between stress and control samples showed no significant differences in the cell wall invertase spots, except for the g band. As there is no difference in activity (Fig. 1) this would mean that the g peptide is inactive; this is consistent with a degradation product.

View larger version (108K): [in this window] [in a new window]   Figure 5.   Immunodetection from two-dimensional gels of vacuolar and cell wall invertase in mature leaf (A-D) and primary root (E and F) from well-watered (left column: A, C, and E) or water-stressed plants (right column: B, D, and F) for 7 d . Homologous groups of spots were designated a, b, c, and d for vacuolar invertase antibodies and e, f, and g for cell wall invertase antibodies. Antiserums raised against an IVR2 oligopeptide (A and B for mature leaf; E and F for primary root) and a cell wall invertase peptide (C and D for mature leaf) were used for invertase immunodetection from crude protein extracts (50 µg) in mature leaf and root. All sampling was done at 9 AM. Comparison among loaded protein quantities was carried out from Coomassie Blue gel staining (data not shown). To measure pI, four gels were cut into 15 fragments and four gel fragments were incubated together in 1 mL of distilled water overnight; pH was measured from these eluted solutions.

The immunoblot patterns for vacuolar invertase were rather complex in the primary roots. Two additional peptide groups were noted: c (68 kD, pI 5.3) and d (80 kD, with a broad pI range pI 5.5-6.3); both spots were slightly enhanced under stress (Fig. 3F). The b and d groups had the same spot pattern for pI, but the d group was slightly shifted toward a higher mass than b group.

Ivr2 Transcripts, Encoding a Vacuolar Invertase, Are Specifically Induced by Water Stress and Correlate with Vacuolar Invertase Activity

Since vacuolar invertase activity induction was accompanied by an increase in vacuolar invertase proteins, it is expected that the corresponding transcript levels would also be correlated to vacuolar invertase activity. Membrane blots of total RNAs from five organs of well-watered and water-stressed plants were hybridized with specific gene probes encoding two vacuolar invertases (Ivr1 and Ivr2), four cell wall invertases (Incw1, 2, 3, and 4), and two Susy (Sus1 and Sh1).

No significant changes in the expression pattern for genes encoding the cell wall invertases were detected in water-stressed organs, nor for the two genes encoding Susy (Kim, 1998); in contrast, the expression of one vacuolar gene (Ivr2) was affected.

Ivr1 transcripts were detected at very low levels in primary and adventitious roots and were undetectable in mature and young leaf, irrespective of the water treatment (data not shown). Only Ivr2 gene transcripts increased under water stress (figure 6) displays the time-course expression of Ivr2 gene transcripts in various organs under watered or water-stress conditions. Water stress clearly modified the Ivr2 gene expression pattern in mature leaf, leaf sheath, and roots. The Ivr2 expression pattern followed that of vacuolar invertase activity in all organs (Fig. 1), except in adventitious roots in which there was no significant induction of vacuolar invertase activity. Diurnal fluctuations of the Ivr2 transcripts were also apparent in stressed young leaf, mature leaf, and leaf sheath, but not in roots. Transcript levels were consistently the highest at the beginning and the lowest at the end of the light period. These fluctuations suggest a diurnal regulation of the gene.

Water Stress Induces Rab17 Expression and Reduces RbcS Expression

To examine if hexoses accumulated under water stress may trigger modifications in gene expressions, total RNAs from mature leaf were hybridized with probes of a maize RbcS gene encoding the small subunit of ribulose-bisphosphate carboxylase, known to be repressed by Glc (Sheen, 1990), and of a maize Rab17 gene, called Abscisic acid (ABA)- or water-stress responsive gene (Vilardell et al., 1990). The RbcS gene transcripts decreased rapidly after 5 to 6 d in stressed mature leaf compared with controls (Fig. 7). At this time point hexose content reached about 6 mg g¹ fresh weight (Fig. 7), which is the reported threshold for hexose regulation of gene expression in tobacco (Herbers et al., 1996). In mature leaf the induction of vacuolar invertase activity and the accumulation of hexoses preceded the reduction of RbcS gene transcripts. In contrast the Rab17 gene expression was induced by water stress from d 5; it is interesting that its temporal expression pattern suggested a diurnal regulation similar to that of Ivr2 transcripts (Figs. 6 and 7). The EF- gene transcript level used as a constitutive expression standard remained constant and unaffected by stress.

View larger version (25K): [in this window] [in a new window]   Figure 6.   Time course of Ivr2 transcript accumulation in water-stressed (black symbols) and watered plants (white symbols) in five vegetative organs. Total RNA samples (20 µg/lane) extracted from organs were hybridized with an Ivr2 cDNA probe, radioactivity was stripped, and the membrane was rehybridized with EF- cDNA. The hybridization signal intensity was measured with an image analyzer and the relative Ivr2 mRNA expression was normalized from the average between the intensity of EF- and of 18S rRNA staining.

View larger version (22K): [in this window] [in a new window]   Figure 7.   Time course of hexose accumulation in mature leaf in comparison with the expression of a hexose repressed gene (RbcS), an ABA-responsive gene (Rab17) and a constitutively expressed gene (EF-), under watered and water-stressed conditions. The membrane for mature leaf used in Figure 6 was rehybridized with 32P-labeled Rab17, RbcS, and EF- cDNA probes.

Vacuolar Invertase Protein and Ivr2 mRNA Show the Same in Situ Localization as Induced Vacuolar Invertase Activity and Glc Content

In situ localization of Glc and invertase activity was carried out on hand-sliced fresh material by using a Glc oxidase/peroxidase-diaminobenzidine (DAB) staining method. In well-watered mature leaf only traces of Glc staining were detectable (data not shown), whereas under stress a brown DAB precipitate indicating the presence of Glc was seen in bundle sheath cells and vascular cells (Fig. 8A). Invertase activity staining was also localized within vascular and perivascular cells of water-stressed leaves (Fig. 8C), which contrasted with a cell wall localization in the various tissues of well-watered leaf (Fig. 8B). Such cell wall staining is consistent with measured activity from whole leaf extracts (Fig. 1) indicating that cell wall-bound invertase activity, although at a low level, is the only Suc cleaving enzyme in control mature leaf. The colocalization of vacuolar acid invertase activity and Glc strongly suggest that accumulated Glc is the product of the invertase activity. The non-substrate histochemical controls indicate that DAB-labeling of xylem elements in mature stressed leaf is likely to be artifactual (Fig. 8D).

View larger version (133K): [in this window] [in a new window]   Figure 8.   In situ histochemical staining for Glc and invertase activity in hand-sliced fresh mature leaf (A-D) and primary root (E-K) sampled on d 7 from well-watered and water-stressed plants. A, Stressed mature leaf; Glc localized within bundle sheath and vascular bundle cells (arrows). B, Watered mature leaf; acid invertase activity only detected in cell walls. C, Stressed mature leaf; acid invertase activity localized within bundle sheath and vascular bundle cells (arrows). D, Stressed mature leaf; histochemical control (tissue flushed, no sugar addition). E, Stressed root; strong signal for Glc in cytosol of cortex and outer central cylinder. F, Watered root; weak signal for Glc in both cell walls and cytosol of cortex and central cylinder. G, Stressed root; negative histochemical control showing artifactual xylem staining. H and I, Stressed root; strong intracellular acid invertase activity in epidermis and central cylinder (arrow), weaker signal in cortex (arrow). J and K, Watered root; moderate cell wall signal and weak intracellular signal for acid invertase activity in epidermis, cortex and central cylinder (arrow). bs, Bundle sheath; C, cortex; cc, central cylinder. Bars in A through D, 50 µm; E through G and I and J, 100 µm; H and K, 25 µm.

In primary roots marked differences in tissue staining for Glc and invertase activity were visible between stressed and watered conditions. In water-stressed root the cortex and outer cylinder cells were heavily Glc-stained in the intracellular compartment (Fig. 8E) compared with a weak apoplastic and cytoplasmic staining in well-watered root (Fig. 8F). The staining pattern of acid invertase activity was again very similar to that of Glc; in stressed primary roots a strong intracellular staining in cortex and central cylinder cells was detected (Fig. 8, H and I) corresponding to the observed induction of vacuolar invertase activity (Fig. 1). In well-watered primary roots a moderate cell wall-staining and a weaker intracellular signal for invertase activity was detected in the central cylinder and in epidermal cells (Fig. 8, J and K), which is again consistent with enzyme assays showing a higher cell wall invertase activity than vacuolar activity (Fig. 1).

In situ immunolocalization of the vacuolar invertase protein showed an intracellular tissue specific localization, namely in bundle sheath and vascular cells of mature leaf (Fig. 9A). In contrast, immunolocalization of cell wall invertases was confined to cell walls in all leaf tissues (epidermal, mesophyll, and vascular; Fig. 9B). The nonimmune serum-treated leaf was immunonegative (Fig. 9C). The cellular distribution of vacuolar invertase protein was similar to that of invertase activity and accumulated Glc in stressed leaf (compare with Fig. 8, A and C). To determine if the location of Ivr2 gene transcripts was the same as that of vacuolar invertase protein and activity, in situ hybridization using antisense (Fig. 9D) and sense (Fig. 9E) Ivr2 probes was carried out. Indeed strong signals indicated that under water stress Ivr2 transcripts were detected in bundle sheath cells and to a lesser extent in vascular cells (Fig. 9D).

View larger version (96K): [in this window] [in a new window]   Figure 9.   In situ immunolocalization of invertase protein (A-C and F-J) and in situ hybridization of Ivr2 mRNA transcripts (D-E and K-L) in mature leaves (A-E) and roots (F-L) sampled on d 7 from water-stressed plants. Epipolarization optics of water-stressed mature leaf section exposed to A, Vacuolar invertase antibodies showing intracellular labeling in the vascular bundle; B, cell wall invertase antibodies showing cell wall labeling (arrow); and C, nonimmune serum. D, leaf section of water-stressed mature leaf hybridized to Ivr2 probe, in antisense orientation; E, leaf hybridized to Ivr2 probe in sense orientation. F and I, Root section of water-stressed plants incubated with vacuolar invertase antibodies showing intracellular localization in cortex and central cylinder (arrows); G and J, water-stressed roots exposed to anti-cell wall invertase serum showing immunopositive cell walls in all root tissue (arrows in J); H, root section incubated with nonimmune serum yielded no labeling; K, root section of water-stressed plants hybridized to Ivr2 antisense probe; L, root hybridized to sense probe. c, Cortex cells; cc, central cylinder; ep, epidermis; bs, bundle sheath cells; vb, vascular bundle cells. Bars in A through C, I, and J, 25 µm; D through H, K, and L, 100 µm.

Vacuolar invertase proteins were also immunodetected in the epidermis, central cylinder, and cortex cells of primary roots. The signal was intracellular in cortex cells, supporting a vacuolar location (Fig. 9, F and I, arrows). Using anti-cell wall invertase serum, immunopositive cell walls were seen throughout the central cylinder and cortex cells (Fig. 9, G and J). Stressed root sections were hybridized in situ; the antisense probed roots showed a strong staining for Ivr2 transcripts in epidermal, cortex, and central cylinder cells (Fig. 9K) in comparison with sense-probed roots, which remained unstained (Fig. 9L). Taken together these results suggest that the induction of vacuolar invertase is due to the cell-specific expression of the Ivr2 gene.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Vacuolar Invertase Expression under Water Stress

The most remarkable response to water stress in maize organs is the induction of vacuolar invertase activity. This invertase activity is mainly induced in mature leaf, leaf sheath, and primary root and is accompanied by an increase in vacuolar invertase protein and Ivr2 transcripts. In situ localization data support the view that Ivr2 gene expression is linked to the change in vacuolar invertase proteins under water stress. Two genes, Ivr1 and Ivr2, encoding vacuolar invertases show different organ/tissue-specific and developmentally specific expression patterns (Xu et al., 1996). Ivr2 is expressed in various vegetative and reproductive organs and its expression is related to an increase in soluble sugars, whereas Ivr1 shows a narrower expression pattern, being essentially limited to roots and reproductive organs, and inhibited by high sugar level (Xu et al., 1996). The accumulation of soluble sugars by water stress could trigger the induction of Ivr2 gene expression. However, this is unlikely since the induction of Ivr2 expression is not well correlated with soluble sugar content, with the exception of young leaf. Moreover, the expression of other sugar regulated genes (Ivr1, Sus1, and Sh1) encoding Suc cleavage enzymes did not change under the same condition (Kim, 1998). Furthermore, Ivr2 expression showed diurnal variations, whereas the sugar-induced gene regulation tends to override light, tissue-dependent, or developmental regulations (Sheen, 1990). Alternatively the Ivr2 gene can be induced via a water stress-specific pathway such as ABA-related signal transduction, since exogenous ABA treatment can induce an accumulation of Ivr2 transcripts (K.E. Koch, personal communication; J. Trouverie and J.L. Prioul, unpublished data). This effect is consistent with the fact that the expression of Rab17, an ABA responsive gene, was also induced under water stress, with a diurnal and temporal pattern similar to that of Ivr2 (Figs. 6 and 7). However, it remains to be elucidated whether or not diurnal fluctuation of Ivr2 and Rab17 transcripts in shoots is regulated by ABA fluctuation.

To our knowledge, no reference about immunodetection of invertase on two-dimensional IEF-SDS gel is available to date. The increase in size of immunospots a and b (Fig. 5) under water stress would suggest that at least one of these groups corresponds to the IVR2 protein. The root specific spot c (Fig. 5) may correspond to the IVR1 protein since its pI fits with the theoretical pI of 5.0 for prepropeptide deduced from the Ivr1 sequence and since both spots and transcripts are absent in mature leaf, but present in roots. However, we cannot exclude the possibility of unknown gene(s) encoding vacuolar invertase(s) considering Ivr1 transcript levels in roots were constant. In fact we have detected two invertase groups in mature leaf and four groups in primary roots. RFLP map analysis using Ivr2 gene probe showed two to four gene copies of which two were localized on chromosome 4L and 5S, respectively (Xu et al., 1996). There are also two copies for Ivr1 on chromosome 2S and 10L. These copies may code for proteins with different molecular masses and pI properties. Post-translational glycosylation could also result in forms with altered M_r and possibly pI. Indeed maize vacuolar invertases are N-glycosylated proteins. The maize IVR1 and IVR2 invertase protein sequences, deduced from their cDNAs, show at least five potential glycosylation sites. The d spots could be derived from the glycosylation of the b spots, since they present the same pI pattern with a shift toward higher M_r, possibly due to multiple glycosylation sites.

Ivr2 expression was induced in water-stressed adventitious roots, but the vacuolar invertase activity remained unchanged, implying that there would be some post-transcriptional/translational regulation for vacuolar activity in this organ. Indeed the existence of proteinaceous invertase inhibitor has been reported for cell wall-bound invertase and more recently for the vacuolar form (Krausgrill et al., 1996).

Invertase as a Putative Factor of Hexose- Signaling System

Under water stress, hexose accumulation was associated with an increase in vacuolar invertase activity in mature leaf and primary roots. Glc accumulation in both organs colocalized in situ with vacuolar invertase activity, invertase protein, and Ivr2 transcripts. It is interesting that our results showed that RbcS expression, a bundle sheath cell specific-photosynthetic gene, was significantly diminished by water stress when leaf hexose content was higher than the threshold level (6 mg g¹ fresh weight) reported by Herbers et al. (1996; Fig. 7). This observation would support the hypothesis of a feedback inhibition of RbcS expression by excess hexoses. In a maize mesophyll protoplast system hexose treatment resulted in the repression of photosynthetic gene expression via hexokinase-associated signal transduction (Jang et al., 1997).

However, the sugar response is difficult to separate from the ABA one. Pelleschi (1997) reported that an accumulation of Glc and Fru in maize mature leaf is associated to and slightly precedes xylem ABA concentration in water-stressed maize. We also detected that the induction of Ivr2 transcript and the accumulation of hexose preceded Rab17 transcript induction in mature leaf under water stress. In addition since exogenous ABA and sugar treatments could stimulate Ivr2 expression, we suggest that ABA- and sugar-signaling pathways could crosstalk under water stress. This hypothesis is supported by Smeekens (1999), who observed that sugar repression of Cab gene expression is inhibited in Abi1 and Abi2 ABA-insensitive mutants in Arabidopsis.

Vacuolar Invertase Activity Plays a Role in Establishing Sink Organs and in Suc Mobilization under Water Stress

In maize sink organs the association of a high hexose/Suc levels with a high invertase activity supports the role of Suc-cleaving enzyme in Suc partitioning. During the day Suc is mainly synthesized in leaf mesophyll cells from current photosynthates, whereas at night Suc mainly originates from starch hydrolysis in the bundle sheath cells. In well-watered leaf with low vacuolar invertase activity Suc is exported or stored principally in the vacuole of photosynthetic cells (Huber, 1989; Heineke et al., 1994). This vacuolar Suc is remobilized for export in the early hours of the dark period, before starch remobilization. Under water stress the physiological significance of IVR2 diurnal variations may be explained by the following model (Fig. 10). In mature leaf at the beginning of the dark period under water stress, Suc stored in mesophyll vacuoles would be remobilized as in watered plants. However, Suc stored in bundle sheath cells would be hydrolyzed by stress-induced vacuolar invertase activity during the first hours of the dark period. Resulting hexoses could not be exported until resynthesis of Suc via hexokinase and Suc-P synthase, thus forming a futile cycle. These events would result in altered starch remobilization as in transgenic plants expressing yeast-derived vacuolar invertase (Heineke et al., 1994; Büssis et al., 1997). Accordingly, Pelleschi (1997) observed larger starch granules in bundle sheath chloroplasts of water-stressed maize than in watered plants at the beginning of the light period.

View larger version (40K): [in this window] [in a new window]   Figure 10.   Flowchart of Suc and hexose mobilization among different organs, cells, and compartments in maize. Suc is synthesized in leaf mesophyll cells during the daytime from current photosynthates and in bundle sheath cells at night from starch hydrolysis products. Suc is loaded into phloem or temporally stored in the vacuole and remobilized. Induction of invertase activity in bundle sheath cells during the night in water-stressed maize slows down the Suc export and produces a Suc degradation/synthesis futile cycle. In young leaves Suc is imported without hydrolysis, whereas it is imported with or without hydrolysis in roots. Black squares represent plasmodesmata and black circles represent membrane translocation. Hex, hexoses (Glc or Fru); Hex-P, hexose phosphates (Glc- or Fru-Phosphate); TP, triose-P; INVv, INVc, and INVcw: vacuolar, cytosolic, and cell wall-bound invertases respectively; SPS, Suc-P synthase.

Suc from the phloem could be unloaded into sink organs via an apoplastic or symplastic pathway. In young sink leaves with minimal wall invertase activity Suc is unloaded symplastically or apoplastically without cleavage. Compared with young leaves, roots not only contain Susy, vacuolar, and cytosolic invertase, but also a high level of cell wall invertase activity, which could itself explain the higher hexose to Suc ratio. However, in some sink organs such as maize primary roots and carrot tap roots, vacuolar invertase and/or cytosolic Susy are considered to be more important than cell wall invertase in maintaining Suc gradient (Duke et al., 1991; Strum et al., 1995). Our in situ data showed the induction of Ivr2 expression in root tips under water stress suggesting that it may play a stimulating role in carbohydrate partitioning in active sink tissues. Activation of Suc hydrolysis via the induced vacuolar invertase activity under water stress would stimulate Suc unloading from phloem, sequestration into vacuoles, and consequently contribute to an osmotic adjustment by vacuolar hexose accumulation. The increase in root to shoot dry weight ratio under water stress (Pelleschi et al., 1997; Kim, 1998) is consistent with an enhanced unloading due to a stronger induction of vacuolar acid invertase activity in primary roots than in young leaves.

In conclusion, under water stress Ivr2 gene expression is likely to coordinately regulate changes in vacuolar invertase proteins and activity in spatial and temporal specific manners. The induced vacuolar invertase results in hexose accumulation that in turn may be involved in a signal transduction pathway or in an increase of osmotic pressure leading to water stress resistance. The modulations of vacuolar invertase activity between source and sink and between sink organs, concurrent with the diurnal regulation of IVR2 activity in bundle sheath cells, would alter the competition for assimilates and thus modulate source-sink strength.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material and Sampling

Maize (Zea mays; F₂) plants were grown in a green house with a 16-h light period supplemented with fluorescent lamps (400 µmol quanta m² s¹, Philips Sun-T Agro, Paris) as described by Pelleschi et al. (1997). Plants were watered with a nutrient solution (Hydrocani C2 + Hydroplus Fer H23, HURELARC) eight times a day by an automatic system. Nutrition solution was interrupted in the water-stressed group when the ligule of the fourth leaf was visible, about 15 d after germination. In the water-stressed and control groups, five vegetative organs (fourth mature leaf, fourth leaf sheath, young leaf, primary, and adventitious roots) of four plants were sampled once or twice a day (3 h after the beginning of the photoperiod and/or 1 h before the dark period) until the 11th d after water interruption. At this time leaves were severely wilted. For young leaf, the basal yellow part of the youngest visible leaf was sampled.

Biochemical Measurements

Susy and invertase activities, and Suc and hexose contents were measured as described by Pelleschi et al. (1997) except for the following modifications. HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid]-NaOH buffer was replaced by 50 mM sodium phosphate buffer (pH 7) for neutral/alkaline invertase assays because the HEPES buffer inhibits this activity. Insoluble cell wall-bound invertase activity was assayed from well-mixed cell debris suspension. Protein concentrations were determined using a Bio-Rad (Hercules, CA) protein assay kit with bovine serum albumin (BSA; Sigma, St. Louis) as standard.

Northern Blots

Samples were harvested, quick frozen in liquid nitrogen, and stored at 80°C. Total RNA was isolated by the phenol-chloroform method as described by Mahé et al. (1992). Total RNA samples (20 µg/lane) were separated in denaturing formaldehyde 1.2% (w/v) agarose gel (Sambrook et al., 1989) and transferred to a nylon membrane in 2× SSC as a transfer buffer. Membranes were hybridized to ³²P-labeled DNA probes generated using an oligo-labeling kit (Pharmacia Biotech, Piscataway, NJ). Probes, encoding two vacuolar invertases (Ivr1, accession no. U16123: 1-kb NcoI/EcoRI cDNA fragment; and Ivr2, accession no. U31451: 0.8-kb EcoRI/PstI 3'-cDNA fragment, including a 0.45-kb 3'-coding and a 0.34-kb 3'-untranslated region), four cell wall invertases (Incw1, accession no. U17695: 0.55-kb EcoRI/SalI cDNA fragment; Incw2, accession no. AF050631: 0.8-kb EcoRI/PstI 3'-cDNA fragment; Incw3, accession no. AF043346: 0.75-kb EcoRI/HindIII fragment consisting of 3' end of exon 3 to intron 3, exon 4, and to the 5' part of exon 5; and Incw4, accession no. AF043347: 0.9-kb EcoRI/HindIII fragment consisting of 0.1-kb intron 2 and 0.8-kb exon 3), two Susy (Sh1, Pww110/1 from Werr et al. [1985]: 1.2-kb PvuII:PvuII fragment; and Sus1, PshD13 from Gupta et al. [1988]: 1.4-kb Eco:PstI fragment), a small subunit of Rubisco (RbcS, pC1 from Broglie et al. [1984]: PstI/PstI 0.5-kb cDNA fragment), an ABA-responsive gene (Rab17, from Vilardell et al. [1990]: .8-kb-EcoRI/XhoI cDNA fragment), and a translation elongation factor (EF-, 0.8-kb EcoRI/BamHI cDNA fragment) were used in succession. The loading differences were corrected with respect to the average between the EF- probe signal and 18S rRNA ethidium bromide staining intensity. These probes were specific under the hybridization conditions used: i.e. 50% (v/v) formamide, 42°C with either 2× SSC for Ivr1, Incw2, and Incw3; 3× SSC for Ivr2; 4× SSC for Sh1, Sus1, Rab17, RbcS, and EF-; or 5× SSC for Incw1.

In Situ Glc and Invertase Activity Staining

Tissue slices (200 µL) were hand-cut from fresh roots and leaves previously embedded in 4% (w/v) agarose for 5 to 10 min. Slices were washed extensively to remove soluble sugars for in situ detection of acid invertase activity or left unwashed for in situ Glc detection. Activity was visualized by an incubation in the reaction medium consisting of 200 mM citrate/tris [tris(hydroxymethyl)aminomethane], pH 6.0, 100 mM Suc, 20 µg mL¹ Glc oxidase (Sigma), 250 µg mL¹ horseradish peroxidase (Sigma), and 300 µg mL¹ DAB (Sigma). Suc was omitted for Glc detection. The controls for activity and for Glc detection were performed with flushed and denatured (3 min at 65°C) preparations, respectively, using a reaction medium without Suc. Under our conditions only acid invertase staining was detected because 200 mM Tris totally inhibits neutral invertase activity (Kim, 1998).

Immunoblots

A rabbit antiserum raised against a synthetic oligopeptide (KRVVGSAAVPVLEDEA) of IVR2 invertase deduced from Ivr2 gene (GenBank accession no. U31451) was used the for vacuolar invertase detection. The oligopeptide is vacuolar invertase-specific (IVR1 or IVR2), but has no homology with corresponding parts of cell wall invertase. Proteins were separated on a 10% (w/v) SDS-polyacrylamide gel. For two-dimensional IEF-SDS polyacrylamide gel proteins were separated on IEF gel using pH 3.0- to 10-range ampholites (Pharmacia Biotech). Proteins on polyacrylamide gel were transferred to nitrocellulose membranes by using an electrotransfer system (Bio-Rad). The resulting membranes were blocked in 5% (w/v) non-fat dry milk in Tris-buffered saline (TBS)-T (100 mM Tris-HCl, pH 7.5, 300 mM NaCl, and 1% [w/v] Tween 20), incubated with the primary antibody and horseradish peroxidase-conjugated secondary antibody, and detected with enhanced chemiluminescence reagents as recommended by the manufacturer (Amersham, Buckinghamshire, UK).

Immunolocalization

Immunolocalization of vacuolar and cell wall-bound invertases was carried out either on tissue slices of fresh material prepared as described above and/or on sections of chemically fixed and resin-embedded material. Leaf and root tissues were harvested on the morning of the 7th d after water interruption and were fixed by immersion in a solution of 0.1 M phosphate buffer (pH 7.2), 4% (w/v) paraformaldehyde, and 0.5% (w/v) glutaradehyde. Each sample was cut into small pieces with a razor blade and fixed overnight at 4°C. The samples were washed four times for 20 min with 50 mM phosphate buffer, pH 7.2, dehydrated in a graded ethanol series (30%, 50%, 70%, 80%, 90%, 95%, and 100%), embedded in LR white resin (Electron Microscopy Science, Fort Washington, PA), and polymerized at 55°C for 48 h. Semi-thin sections (5 µm) were cut with a rotary microtome and affixed to Superfrost plus glass slides. Antiserums for carrot vacuolar and Arabidopsis wall-bound invertases (gifts from Dr. A. Sturm and Dr. A. Kingston-Smith) were used for immunodetection as follows: pre-incubation was carried out for 2 h with TBS-T containing 0.1% (w/v) gelatin, 2% (w/v) BSA, and 5% (w/v) non-fat powdered milk. Incubation with each primary antiserum (cell wall-bound or vacuolar invertase) 1/1,000 and 1/500 diluted in TBS-T containing 0.1% (w/v) BSA and 0.1% (w/v) gelatin was carried out overnight at 4°C; sections were then washed in TBS-T and incubated with 1 nm of gold-conjugated antirabbit immunoglobulin G diluted 1/100 for 2 h at room temperature. Control preparations were treated in the same manner, except that rabbit nonimmune immunoglobulin G was used as the primary serum. The sections were washed in TBS-T, in distilled water and amplified for 20 min using a silver enhancement kit according to the manufacturer's (British Biocell International, Cardiff, UK) instructions.

In Situ Hybridization

Root and leaf tissues were cut into small pieces and immediately placed in a solution of formaldehyde-acetic acid (50% [v/v] ethanol, 5% [v/v] acetic acid, and 10% [v/v] formaldehyde). The fixative solution was changed twice and samples were stored at 4°C for several days. Samples were dehydrated in a series of increasing concentration of ethanol-tert-butanol washes and embedded in Paraplast Plus (Monoject Scientific, Fort Washington, PA) according to the manufacturer's instructions. Samples were cut in 8-µm sections and affixed to Superfrost plus glass slides. Paraffin was removed by soaking in Histoclear (National Diagnostics, Albania, GA) for 10 min followed by 5-min washings in Histoclear:ethanol (1:1) and 100% (v/v) ethanol. For in situ hybridization sections were prehybridized for 2 h at 42°C with 4× SSC, 50% (w/v) formamide, 5× Denhart, and 5% (w/v) dextran sulfate containing 1 mg/mL salmon sperm DNA. Hybridization with single-stranded digoxygenated antisense or sense DNA probes (produced by using single primer-specific PCR amplification with digoxigenin-UTP) was carried out overnight at 42°C in the prehybridization buffer. Hybridized RNA was revealed indirectly by incubation with an antidigoxigenin-alkaline phosphatase conjugate (Boehringer Mannheim, Basel) and nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate color substrates as described in the manufacturer's instructions.