Salt stress leads to massive accumulation of toxic levels of Na+ and Cl− ions in plants. By using the recombinant fluorescent probe CLOMELEON, we demonstrate passive anion flux under salt stress. Chloride influx is restricted in the presence of divalent cations like Mg2+ and Ca2+, and completely blocked by La3+. The amount but not the rate of the reported chloride uptake is independent from the kind of corresponding permeable cation (K+ versus Na+), external pH and magnitude of osmotic stress. Cl− efflux however seems to involve stretch-activated transport. From the influence of Ca2+ on reported changes of cytosolic anion concentrations, we speculate that transport mechanisms of Cl− and Na+ might be thermodynamically coupled under saline conditions.
Plants have evolved mechanisms to take up, compartmentalise and/or export NaCl and thereby developed salt tolerance or resistance. The mechanisms by which Na+ ions enter the cells have been studied intensively. Also Na+ extrusion mechanisms, which aid salt tolerance, are known actually in molecular detail (reviews, e.g. Tester and Davenport, 2003; Xiong and Zhu, 2002). Such knowledge is helpful for the development of salt-tolerant crop varieties (Apse and Blumwald, 2002). Another pre-requisite for this is to also understand Cl− uptake and accumulation under saline conditions. Many anion transporters have been characterised under low-salt conditions, and much information has been obtained by using the radioisotope 36Cl− as tracer (reviews, e.g. Barbier-Brygoo et al., 2000; White and Broadley, 2001). The mechanisms of Cl− import, re-location and/or extrusion under salt stress in whole intact plants, however, are largely obscure. There are no Cl− transport mechanisms characterised yet, which can directly be linked to salt tolerance mechanisms as is the case for Na+ transport. Under ‘normal’ conditions (i.e. [Cl−]ext < [Cl−]cyt), chloride is actively taken up by a ΔpH-driven Cl−/nH+ symport (Barbier-Brygoo et al., 2000; Felle, 1994; Sanders and Hansen, 1981). In particular, Felle (1994) showed that sudden but moderate increases in [Cl−]ext lead to a transient pHcyt drop in Sinapis alba root cells. This suggests the Cl−/nH+ symport at least participating in the Cl− influx, but there is no direct evidence that this is also a route for Cl− import with salt-stress conditions. Alternatively, for high-salt conditions, an outward rectifying anion channel permeable to NO3– and Cl− has been electrophysiologically characterised in isolated wheat protoplasts (Skerrett and Tyerman, 1994). Taken together, there are still many open questions: e.g., What is the main doorway for Cl− influx under salt stress? To what extend is Cl− compartmentalised and transported over longer distances? Which route does Cl− take when salt stress is relieved? How tightly is Cl− transport coupled to the transport of its counter cation under saline conditions?
To tackle at least a few questions, we started to employ the recombinant anion indicator CLOMELEON (Kuner and Augustine, 2000) in order to directly visualise [Cl−]cyt by fluorescence resonance energy transfer (FRET). This indicator consists of a cyano-blue fluorescent protein (CFP) linked to a yellow-green fluorescent protein (YFP) and is similar to CAMELEON (Allen et al., 1999), however, without Ca2+-binding domain. As YFP is anion sensitive and CFP is not, FRET declines with increasing anion concentration, i.e. yellow fluorescence is decreasing for the benefit of cyan fluorescence (spectra given as Supplementary Material; Figure S1). We expressed the ratioable FRET-based probe in Arabidopsis thaliana (Columbia (Col-0)) and show how Cl− is loaded via the cytoplasmic pathway into intact whole plants when put under salt stress.
Results and discussion
When roots are put under salt-stress conditions, the cytoplasmic chloride concentration [Cl−]cyt dramatically increases with time (Figure 1a). This increase is dependent on the presence of divalent cations (Figure 1b,c). To block a further [Cl−]cyt increase with saline conditions, 10 mm Ca2+ is sufficient (Figure 1c).
There is no chloride transport yet characterised, which is directly inhibited by Ca2+, but it is well known that Ca2+ reduces Na+ influx with saline conditions (Essah et al., 2003; Kinraide, 1999; Tyerman et al., 1997). Also, the anion channel characterised by Skerrett and Tyerman (1994) is relatively insensitive to extracellular Ca2+. Therefore, we assume an indirect effect of Ca2+ on Cl− transport via a direct Ca2+-mediated block of cation channels. The latter would prevent the depolarization necessary for anion influx. However, for validation of such mechanisms, thorough electrophysiological experiments and estimation of driving forces of the involved ions are needed.
Low concentrations of the ion channel blocker La3+ have the most drastic effect on salt-induced [Cl−]cyt increase (Figure 1d). This is in line with findings by Lewis and Spalding (1998) who showed that La3+ is able to non-selectively block many ion channels, in particular, anion channels. It suggests that Cl− transport reported here is mainly driven through channels.
We also investigated Cl− efflux mechanisms and therefore performed experiments with repetitive salt treatments (Figure 2). When NaCl is washed away from the outer medium, Cl− is removed from the cytoplasm. However, we never observed a complete clearance of Cl− back to the initial [Cl−]cyt level within the time-frame of our experiments. This is not surprising and can be explained by salt load into the vacuoles. When vacuoles are loaded, salt and thus Cl− will probably be re-located to the cytoplasm when cleared from there by washing the tissue.
As mentioned above, the Ca2+ block of Cl− influx (Figure 1c) made us speculate that Cl− might be carried alongside with its corresponding cation through unspecific leaks produced by the strong concomitant osmotic stress. Consequently, we supplemented the washing solutions with mannitol to reach equal osmolarities and to minimise mechanical stress and primary damage (Figure 3). Uptake kinetics were fairly similar to what has been obtained before (Figure 2). Hence, the ‘leak hypothesis’ seems less likely. The efflux during the first wash period (t = 1.0–1.5 h; Figure 3a), however, seems to be inhibited. This is in line with findings by Teodoro et al. (1998) who showed that hyper-osmotic stress (i.e. 200 mm mannitol) does reduce Cl− efflux. They concluded the participation of stretch-induced Cl− efflux channels. So, it could be assumed that for effective Cl− efflux hypo-osmotic conditions are required. Subsequent washing periods suggest that other efflux mechanisms seem to come into play when higher [Cl−]cyt are reached ( t = 2, 3, and 4 h; Figure 3a).
When 10 mm Ca2+ is switched into the perifusion stream during repetitive salt treatments (Figure 3b), Cl− influx again is found inhibited and efflux is slowed down. If the experiment in Figure 3(a) is repeated with Ca2+ as the only counter cation, no Cl− at all is able to enter the cytoplasm (Figure 3c). This supports the notion of a coupling between Cl− transport and its accompanying permeable cation during high-salt treatment.
Another working hypothesis was that Cl−/nH+ carriers (Felle, 1994) do at least participate when cells are flooded with Cl−. We expected a reduced Cl− uptake with reduced ΔpH at the plasmalemma (i.e. elevated pHext in the outer medium). However, there was no important difference in the [Cl−]cyt responses when experiments at pHext 6.0 (Figure 2a) were repeated with pHext 8.3 (data given as Supplementary Material; Figure S2b). Influx kinetics appeared slightly slower at high pHext, but in view of the SDs, this seems not being a significant difference. This is in line with the long-standing observation that Cl− influx is not greatly sensitive to pH (Skerrett and Tyerman, 1994; White and Broadley, 2001; and references cited therein).
We further replaced Na+ by K+ and repeated repetitive salt treatments with KCl at pHext 6.0 and 8.3 (data given as Supplementary Material; Figure S2c,d), with only little differences to what has been shown in Figure 2. Thus, bulk Cl− influx does not depend much on the kind of the corresponding permeable cation. The main difference is that the increase of [Cl−]cyt is faster in 100 mm KCl (Figure S2c) than in 100 mm NaCl (Figure S2a). This is no surprise because it is very likely that the conductivity of the membrane potential generating channels is higher in KCl than in NaCl where the dominance of Na+ conductivity is probably caused by closure of K+ channels.
CLOMELEON is also sensitive to other physiological relevant anions like NO3– (Kuner and Augustine, 2000). When the experiment shown in Figure 2 is repeated with NO3– instead of Cl−, the indicator reports NO3– influx (Figure 4). Although the data seem similar to what has been seen with Cl− (Figure 4; a versus c) there are differences in the kinetics. This suggests differences in transport mechanisms between Cl− and NO3– and/or nitrate consumption.
The NO3– sensitivity of the probe means that absolute numbers given for [Cl−] in the figures have to be taken with caution. They are derived by using the in vitro chloride calibration curve (Supplementary Material; Figure S1). The plants have been grown on NO3– containing MS medium. Hence, the in vivo fluorescence emission ratio of the indicator is always adulterated by unknown cytoplasmic [NO3–] (e.g. Figure 4), and all [Cl−] values given thus indicate just rough estimates rather than precise figures. The basal cytosolic chloride concentration ([Cl−]cyt) has been reported with other techniques to be in the range from 2 to 20 mm (Barbier-Brygoo et al., 2000; Felle, 1994). This is also found here by using the F480/F530 ratio of CLOMELEON.
Conclusions and prospects
The results confirm that Cl− influx under saline conditions is passive through channels. The Ca2+-mediated inhibition of Cl− influx suggests that, for salt-stress conditions, anion import is coupled with the transport of corresponding cations. Most previous studies, however, investigated and regarded anion transport independent from cation transport and vice versa.
Of course, the experiments shown here produce shock responses. It can be argued that such procedure has explanatory power (Tester and Davenport, 2003). However, many previous studies demonstrated that this is a first step to explain at least some important details. The use of whole, intact, self-reporting plants is a step towards physiologically more realistic and/or biologically relevant experimentation.
Standard PCR and cloning techniques (Ausubel et al., 1999; Sambrook et al., 1989) were employed to engineer plasmid constructs. CLOMELEON was expressed in ecotype Col-0 of A. thaliana under the control of the CaMV 35S promoter using the pART7/pART27 cloning/expression system (Gleave, 1992). For Agrobacterium-mediated transformation, the floral dip method (Clough and Bent, 1998) was applied.
For bacterial expression, CLOMELEON was cloned into the bacterial expression vector pQE30 (Qiagen, Hilden, Germany). 6xHis-tagged fluorescent protein from bacteria was purified with Ni2+/NTA-column (Qiagen) and NAP-25 column (Pharmacia Biotech, Freiburg, Germany). In vitro spectra (Supplementary Material; Figure S1) of the indicator were taken with a fluorescence spectrometer (Hitachi F-2500; Kendro Laboratory Products GmbH, Langenselbold, Germany) in KCl buffers containing 50 mm HEPES–KOH adjusted to pH 7.4 and the indicated Cl− concentration. The K+ concentration was kept constant at 1 m for all Cl− concentrations by using gluconate.
Transgenic Arabidopsis were grown in 9-cm Petri dishes on vertical agar as described by Plieth and Trewavas (2002). Whole seedlings of 6–12 day old were used for experiments. [Cl−]cyt was measured in the hairy root segments near the hypocotyl. Experimental conditions, perifusion technique and fixation of plant material were as previously described by Plieth et al. (1998). Roots were placed in a volume of 1.6 ml and perifusion flow was adjusted to 2.4 ml min−1. Composition of perifused buffer solutions are given in the figure legends. Fluorescence images at emission wavelength of 480 and 530 nm were taken every 12 sec with a ratio imaging system from TILL-Photonics (TILL-Photonics, Gröfelfing, Germany; http://www.TILL-photonics.de) extended with a Ludl filter wheel (Ludl Electronic Products, Hawthorne, NY, USA; http://www.ludl.com) in the emission path and custom-tailored control software. TILL software (tillvision 3.3) was used to process raw data.
We are grateful to Valentin Stein (University of California, San Francisco, USA) for the generous gift of CLOMELEON cDNA. We thank Steffi Schnell for technical assistance, Ulf-Peter Hansen (Kiel University, Germany) for helpful advice, and anonymous referees for valuable suggestions.
(a) In vitro emissionspectra at different chloride concentrations (excitation at 434 nm). Curves were normalized by F(534ex; 513em) (b) corresponding calibration curve for fluorescence emission ratio R(434ex; 480em/530em).
Figure S2. Influx of chloride into the cytoplasm of Arabidopsis roots during repetitive salt treatments at different pH (a and c versus b and d) and with different accompanying cations (Na+ in a and b versus K+ in c and d). All curves are average of n = 6 from three independent experiments. SD bars are given at t = 0.25, 0.75, 1.25, 1.75, 2.25, 2.75, 3.25, 3.75, and 4.25 h. Experiments were carried out in pH 6.0 with 2 mm MES/KOH or in pH 8.3 with 2 mm Taps/KOH.