S. portulacastrum (Aizoaceae) is a dicotyledonous perennial halophyte thriving in sandy soils. The plants used in this study were obtained by multiplication of selected plants by cuttings. The mother plants were cultivated at our experimental station, close to the sea shore (35 km north-east of Tunis; 10°10′E, 36°48′N; temperature and mean annual rainfall 19.4 °C and 456 mm, respectively), in outdoor containers filled with a sandy soil and organic matter mixture, and irrigated with tap water. Three-centimetre-long stems having one node and two opposite leaves were cut from mother plants, disinfected for 5 min in a saturated calcium hypochlorite solution, thoroughly rinsed with distilled water, and rooted in a Hewitt nutrient solution [23]. Individual rooted cuttings were then put in 2-L plastic pots filled with silty-sandy soil containing 0.30 mmol Na⁺ and 0.95 mmol of K⁺ per 100 g dry soil. The pH and the electrical conductivity of the aqueous extract (1/10) were 7 and 0.1 mS cm⁻¹, respectively. The field capacity (FC) of soil, measured according to Bouyoucos [24], was 11.25%. Plants were grown in a greenhouse at 30 ± 5 °C/16 ± 2°C (day/night) temperature, 60 ± 5%/90 ± 5% (day/night) relative humidity, and 14 h light/10 h dark regime.

Individuals presenting homogeneous development and size were selected for this study. An initial harvest was achieved after a pre-treatment period with optimal watering (100% FC). Plants were then divided in two lots: the first one was irrigated with nutrient solution lacking salt, while the second lot was irrigated with the same nutrient solution supplemented with 100 mM NaCl. One week later, plants of each treatment were divided into two further batches subjected or not for three months to water-deficit stress (25% and 100% FC). After 45 days of treatment, one lot of dehydrated plants was rewatered at 100% FC. Thus, there were finally six different treatments: (i) no water-deficit with or without salt; (ii) water-deficit, with or without salt; (iii) plants previously submitted to water-deficit, then rewatered with or without salt. Eighteen pots were weighed each two days and rewatered to 100% FC or to 25% FC by replacing the corresponding amount of transpired water. All the plants received the same quantity of nutrients, independently of the level of watering. For each treatment, six pots without plants were used to monitor the evaporative water loss from the soil surface throughout each watering regime.

At the final harvest, plants (six per treatment) were divided into leaves, stems and roots, and their respective fresh weight (FW) was determined. Leaves were counted and their surface area was measured using a portable area meter (LI-3000A). Dry weight (DW) was obtained after oven drying at 60 °C until a constant weight was reached. The tissue water content (TWC) was determined as TWC (ml g⁻¹ DW) = (FW − DW)/DW.

Leaf succulence index was calculated as LS (mg cm⁻²) = leaf FW/leaf surface area. Leaf water potential was measured immediately after the plant sampling (six replicates per treatment) using the pressure chamber method [25]. Water use efficiency (WUE) was calculated for each plant as the ratio of biomass production to transpired water during the experiment. While calculating the amount of water transpired during the experiment, evaporative loss from the pots was taken into account by subtracting the average amount of water loss from the control pots.

Free proline was spectrophotometrically quantified according to Bates et al. [26]. Na⁺ and K⁺ were assayed by flame emission spectrophotometry after extraction in 0.5% HNO₃ of the finely grounded dry matter. For each plant, potassium use efficiency (KUE) was calculated as the ratio of the biomass produced to the amount of K⁺ used during the experiment. Potassium absorption efficiency (KAE) was the ratio of the amount of K⁺ used during the experiment to the mean root biomass.

Analysis of variance (ANOVA) using AV1W MSUSTAT program with orthogonal contrasts and mean comparison procedures was performed to detect differences between treatments. Mean separation procedures were carried out using the multiple range tests with Fisher's least significant difference (LSD) (P<0.05).

Both soil drying and salinity significantly reduced plant growth, with a more pronounced effect for drought in our conditions (Fig. 1). The effects of the two factors were not additive. Indeed, the same whole plant biomass production was obtained under drought, with or without salt (ca. 40% of control). Rewatering plants after 45 days partially restored plant growth, reaching approximately 75% of that of the control plants. Furthermore, there was no significant difference between the plants previously submitted to water-deficit stress alone, or to the combined effects of water-deficit stress and salinity.

Changes in shoot DW were similar to those observed in the whole plant (Fig. 2A). Both salinity and soil drying significantly affected root DW (Fig. 2B). Yet, unlike shoots, root growth was more sensitive to salt than to soil drying. Moreover, drought significantly exacerbated the adverse effect of salinity on root growth. The effects of water-deficit stress and salinity were additive on these organs, so that the root/shoot ratio (Fig. 2B) increased by 28% in plants submitted to water-deficit versus only 6% in plants simultaneously exposed to drought and salt stresses. There was no significant difference between rewatered plants previously submitted to water-deficit stress alone or to the combined effects of water-deficit stress and salinity.

Both salinity and drought reduced leaf number (Fig. 3A) and total leaf surface area (Fig. 3B), their effects being partially additive only on the latter parameter. The individual leaf surface area was systematically higher in plants subjected to a single stress, or to the combined effects of the two constraints (Table 1). Thus, the reduction of the total leaf area of these plants was more due to a reduction in the leaf number.

Plant rehydration after a long-term water-deficit stress significantly improved both of the leaf number and the individual leaf surface area. In plants previously submitted to both stresses, rehydration only increased leaf area, without any effect on leaf number (Fig. 3B). Thus, the plant ability to recover its leaf development was better preserved after water-deficit stress than after the exposure to the combination of both drought and salinity.

Both drought and salinity significantly reduced the leaf water content, with a much more marked impact for the first constraint (Fig. 4A). Yet, the effects of the two constraints when combined were not additive. On the contrary, salt presence in the culture medium of plants subjected to drought was partially shown to prevent leaf tissue dehydration. Transferring the plants previously subjected to water-deficit (with or without salt) improved tissue hydration, this trend being much less pronounced following the exposure to water-deficit alone than when both stresses were combined.

In all treatments, leaf dehydration caused by salinity, water-deficit, and their interaction was accompanied by a significant decrease in leaf water potential (Fig. 4B). However, the more negative values of the water potential were not associated with larger dehydration, but were rather due to the presence of salt, suggesting that salt accumulation contributed more efficiently than passive dehydration in lowering the leaf's water potential.

Salinity and water-deficit when separately applied led to similar increases in WUE and showed additive effects when being combined (Fig. 5A). Rewatering plants after drought or drought combined with salinity did not significantly modify WUE. Drought and salinity had opposite and additive effects on leaf succulence index (Fig. 5B). In plants simultaneously submitted to both constraints, the increase of WUE due to salinity plants largely balanced the decrease due to water-deficit. Rewatering plants after drought partially restored succulence to the value observed in controlled plants, but had no effect on plants previously submitted to combined drought and salinity.

3.4.1 Potassium

Both salinity and water-deficit restricted K⁺ uptake, as estimated by the whole plant K⁺ amount (Fig. 6B). Salinity had a significantly more drastic effect on K⁺ content than water-deficit stress, and there was no significant interaction between the two constraints. Potassium concentration in leaf tissues (mmol g⁻¹ DW) was poorly affected by water-deficit stress, but was largely restricted when salt was present in the medium (Fig. 6A). Potassium status, i.e. whole plant content and leaf K⁺ concentration, was partially restored in plants rewatered after a water-deficit stress alone, but not after combined drought and salinity treatments.

3.4.2 Sodium

A preferential accumulation of Na⁺ in shoots rather than in roots was observed regardless of the treatments applied (not shown). High sodium concentration was observed in the leaves of salt-treated plants, and was significantly increased by the simultaneous exposure to drought and salinity of this parameter (Fig. 7). Rewatering the plants in the presence of salt slightly reduced Na⁺ concentration in the leaves.

3.4.3 Proline accumulation

In both leaves and roots, drought and salinity applied separately increased proline concentration, and their combination resulted in higher proline accumulation (Fig. 8). Proline was more abundant in leaves than in roots, especially in plants cultivated in the presence of NaCl, and the stress-induced changes presented larger amplitude in the former organs. The high capacity to accumulate proline in leaves of salt-treated plants was maintained in the rewatered plants after a combined stress treatment.

Consistently with our findings, it has been reported that water-deficit stress was more inhibiting for shoot than for root biomass, resulting in an increase of the root/shoot DW ratio. This behaviour is considered a criterion of adaptation to drought [22,29–31], although correlation between root development and water extraction ability is not always clear [32,33]. In our study, the preferential biomass allocation to roots of plants subjected to water-deficit was not maintained when plants were simultaneously submitted to salinity and water shortage. The higher inhibition of root growth by salinity, whether alone or combined with drought, led to a decreased root/shoot ratio. Thus, our results support the assumption [34] that enhancing root growth cannot confer resistance to water-deficit under salinity conditions. Indeed, high root/shoot ratio may not be a favourable trait under salt stress, since it might enhance the accumulation of toxic ions into the shoot and, consequently, might anticipate the oncoming of the salt tolerance threshold [35]. Consistently with this statement, the decreased root/shoot ratio of salt-treated S. portulacastrum, which contrasted with the increased ratio that occurred in water-deficit conditions, would enable the plant to control salt accumulation. One may hypothesize that the large reduction in root growth observed in plants subjected to salt alone or in combination with water-deficit limited their capacity to explore soil and to absorb water. However, we showed that under these conditions, leaf water content was significantly enhanced by comparison with the plants subjected to water-deficit stress alone. This suggests that the plants may have relied on other mechanisms to ensure an appropriate water supply, particularly by lowering its leaf water potential.

Salinity and water-deficit stress separately applied or in combination caused significant reduction of the leaf area, and a remarkable decrease of the plant dry matter accumulation. Together with the stomatal closure, the leaf area reduction under drought and saline stresses can be considered an avoidance mechanism, which minimizes water losses [16,36].

When subjected to water-deficit stress, plants reduce their transpiration as a consequence of both stomatal control and decrease in total leaf area, which enhances water use efficiency (WUE) for dry matter production [37]. WUE, traditionally defined as the ratio of the dry matter accumulation to the water consumption over a season, significantly increases under water-deficit stress, and is regarded as an important adaptive trait [38–40]. In our study, despite WUE was calculated over a relatively short period, it may match the same concept as in the traditional definition. WUE significantly increased in plants subjected to water-deficit stress or to salinity, the highest value being observed in plants subjected to the combined effects of both salinity and drought. Hence, these plants were as productive as the plants submitted to water-deficit stress alone, but with a lesser water consumption. Tissue water content represents a negligible fraction of water consumption. Therefore, water loss by transpiration was definitely lower in plants subjected to the combined effects than in those subjected to water-deficit alone. This result can be explained by a high retention of water in leaves, likely resulting form their high Na⁺ and organic solute content. Other mechanisms may have contributed to the improvement of WUE in plants subjected to combined salinity and drought. For instance, Na⁺ seems to play important role in photosynthesis [41,42].

Drought and salinity, either alone or combined, similarly reduced the total leaf surface area to ca. 40% of that observed in control plants. Plants submitted to combined constraints showed the largest reduction in leaf number, but this trend was compensated by salt-induced increase in the mean surface area of individual leaves. In addition, salt stressed plants were much more sensitive in terms of leaf area development, as compared to the aboveground dry matter accumulation. At the end of the experiment, a 40% reduction of the total leaf surface area in plants submitted to water-deficit coincided with a 60% decline in the aboveground dry matter. In contrast, in salinized plants, a 50% leaf area reduction corresponded only to a 20% reduction of the aboveground dry matter accumulation. Thus, salinity may have promoted the reallocation of photosynthates in stems and leaves, which did not occur in drought-stressed plants. One of the consequences of this reallocation was the increase in leaf thickness (Fig. 5B).

Despite water-deficit substantially decreased the leaf water content, this effect was mitigated when salt was present in the irrigation solution of water-stressed plants. This trend was associated with a decrease of the leaf water potential and a high accumulation of Na⁺ and proline in leaf tissues (Fig. 9), suggesting the involvement of these solutes in osmotic adjustment [43]. S. portulacastrum may behave as a typical Na⁺ includer, which compartmentalizes sodium within the leaf cell vacuoles where it could be used as an osmoticum to lower the osmotic potential necessary for the maintenance of the plant hydric status. The absence of leaf toxicity symptom, in spite of the high tissue Na⁺ concentration, supports the hypothesis that Na⁺ was excluded from the cytoplasm. Finally, Na⁺ sequestration in the vacuole where it could replace potassium might explain the increase of K⁺ use efficiency (KUE) that was observed in plants grown in the presence of salt.

When combined, drought and salinity significantly increased proline concentration in both leaves and roots. The high negative correlation between proline and leaf water potential suggests a potent role of proline in the osmotic adjustment. Although the role of proline in osmotic adjustment and drought tolerance is still debated [31,44,45], we have shown in a previous work [46,47] that in S. portulacastrum subjected to a short-term water-deficit, the osmotic balance between vacuole and cytoplasm may be primarily achieved through accumulation of proline.