Down-Regulation of Plasma Intrinsic Protein1 Aquaporin in Poplar Trees Is Detrimental to Recovery from Embolism    

Abstract

During their lifecycles, trees encounter multiple events of water stress that often result in embolism formation and temporal decreases in xylem transport capacity. The restoration of xylem transport capacity requires changes in cell metabolic activity and gene expression. Specifically, in poplar (Populus spp.), the formation of xylem embolisms leads to a clear up-regulation of plasma membrane protein1 (PIP1) aquaporin genes. To determine their role in poplar response to water stress, transgenic Populus tremula × Populus alba plants characterized by the strong down-regulation of multiple isoforms belonging to the PIP1 subfamily were used. Transgenic lines showed that they are more vulnerable to embolism, with 50% percent loss of conductance occurring 0.3 MPa earlier than in wild-type plants, and that they also have a reduced capacity to restore xylem conductance during recovery. Transgenic plants also show symptoms of a reduced capacity to control percent loss of conductance through stomatal conductance in response to drought, because they have a much narrower vulnerability safety margin. Finally, a delay in stomatal conductance recovery during the period of stress relief was observed. The presented results suggest that PIP1 genes are involved in the maintenance of xylem transport system capacity, in the promotion of recovery from stress, and in contribution to a plant’s control of stomatal conductance under water stress.

Long-distance water transport in vascular plants occurs in a conduit network of nonliving cells connecting roots to leaves (Sperry, 2003). Often under drought conditions, the water column within the lumen of xylem vessels or tracheids can be subjected to tensions that result in cavitation and the subsequent formation of embolisms (Holbrook and Zwieniecki, 2008). This hydraulic failure within the xylem network can cause tissue damage, loss of plant productivity, and ultimately, plant death (Tyree and Sperry, 1989; Sperry et al., 1998; Zwieniecki and Holbrook, 2009). Plants have evolved several strategies to prevent and/or mitigate the effects of hydraulic failure caused by embolism and restore xylem transport capacity after embolism occurs (Stiller and Sperry, 2002; Nardini et al., 2011; Secchi and Zwieniecki, 2012). These strategies include passive, often long-term responses, like the growth of new vessels/tracheids or dieback followed by the growth of new shoots (shrubs), or active, often fast responses that result in the restoration of hydraulic conductivity by (1) creating positive pressure through root or stem pressure in the complete transport system (xylem level; Cochard et al., 1994; Ewers et al., 1997; Yang et al., 2012) or (2) enabling positive pressures in specific, embolized conduits, despite negative pressure in the surrounding xylem (conduit level; Salleo et al., 2004; Nardini et al., 2011; Brodersen and McElrone, 2013).

Although embolism formation is a purely physical process related to the degree of tension in the water column and a wood’s physicochemical properties (Brennen, 1995; Tyree and Zimmermann, 2002), embolism removal requires that empty vessels fill with water against existing energy gradients as the bulk of water in the xylem remains under tension caused by transpiration. Thus, recovery from embolism cannot happen spontaneously and necessitates some physiological activities that promote water flow into embolized vessels (Holbrook and Zwieniecki, 1999; Thomas Tyree et al., 1999; Salleo et al., 2004; Zwieniecki and Holbrook, 2009; Secchi et al., 2011). Visual evidence from cryo-scanning electron microscopy studies, magnetic resonance imaging observations, and computed tomography scans showed that water (xylem sap) can return to empty vessels, suggesting that plants do have the ability to restore functionality in the xylem (Holbrook et al., 2001; Clearwater and Goldstein, 2005; Scheenen et al., 2007). Brodersen et al. (2010) showed that water droplets preferentially form on the vessel walls adjacent to parenchyma cells and that these droplets grow until the lumen completely refills. In addition, scientific support for the existence of embolism/refilling cycles in intact stems of Acer rubrum are provided using magnetic resonance imaging (Zwieniecki et al., 2013). Droplet formation on the walls of empty vessels that are in contact with parenchyma cells support predictions that these living cells supply both water and energy to drive the restoration of xylem hydraulic function.

Processes related to water transport across the cellular membrane involve plasma intrinsic protein (PIP; aquaporins) moderators, and thus, the role of PIPs must be considered when contemplating how plants recover from embolism formation. Plant aquaporins show a great diversity and are classified into five major homologous groups that reflect specific subcellular localizations (Prado and Maurel, 2013). Among different aquaporin gene families (26-like intrinsic proteins, tonoplast intrinsic proteins, X unrecognized intrinsic proteins, small basic intrinsic proteins, and PIPs; Danielson and Johanson, 2008), the PIPs represent the largest number of members and can be further divided into two subfamilies, PIP1 and PIP2. There is a large body of evidence that aquaporins from the PIP2 subfamily contribute to water transport. The generation of data has been multidisciplinary and involved the use of chemical blockers, the down-regulation and up-regulation of genes in plants, and the expression of these proteins in oocytes (Hukin et al., 2002; Postaire et al., 2010; Shatil-Cohen et al., 2011). Expression levels of several PIP and TIP members change after the dynamic of increasing water stress and recovery in many woody plants, including walnut (Juglans regia), poplar (Populus trichocarpa.), and grapevine Vitis vinifera; (Sakr et al., 2003; Secchi et al., 2011; Perrone et al., 2012a, 2012b; Laur and Hacke, 2013; Pou et al., 2013). Furthermore, an increase in the expression of PIP2.1 and PIP2.2 genes was observed in vessel-associated parenchyma cells in walnuts at the same time that recovery from embolism was taking place (Sakr et al., 2003). The role of genes from the PIP1 subfamily in tree responses to water stress is less well-understood. PIP1s were shown to have little to no water channel activity when expressed in oocytes on their own. However, coexpression of PIP1.1 proteins with an isoform from the PIP2 subfamily led to higher membrane permeability than that observed with the expression of a single PIP2 protein (Fetter et al., 2004; Secchi and Zwieniecki, 2010). With respect to their role in mediating water stress, it was shown that the expression level of several PIP1 genes in poplar changed significantly during the onset of stress, during recovery, during the formation of embolisms after water stress, and under no stress conditions but with induced embolism, whereas the expression of PIP2 genes remained mostly unresponsive (Secchi and Zwieniecki, 2010; Secchi et al., 2011; Secchi and Zwieniecki, 2011).

Despite significant effort invested in elucidating the contribution of aquaporins to the regulation of xylem hydraulic capacity throughout the progression of drought and recovery from water stress, evidence of their active role in vivo is only partially confirmed. Genetic approaches provide a reliable and effective strategy for determining the physiological function of aquaporin genes in plant water relations. However, most studies thus far have been conducted on herbaceous plants (Kaldenhoff et al., 1998; Postaire et al., 2010). For example, Arabidopsis (Arabidopsis thaliana) plants expressing PIP antisense genes exhibit an impaired ability to recover from water stress (Martre et al., 2002), and knockout mutants exhibit reduced leaf hydraulic conductivity (Da Ines et al., 2010). The Nicotiana tabacum aquaporin1 (NtAQP1) down-regulated tobacco plants show reduced root hydraulic conductivity and lower water stress resistance (Siefritz et al., 2002). RNA technology, although not often used for woody plants, has been adapted for grapevine (Perrone et al., 2012a, 2012b) and Eucalyptus spp. trees (Tsuchihira et al., 2010); in both cases, analysis focused on overexpressing specific isoforms of aquaporin genes. The PIP2;4 root-specific aquaporin enhanced water transport in transformed Vitis spp. plants under well-watered conditions but not under water stress (Perrone et al., 2012a, 2012b), whereas Eucalyptus spp. hybrid clones overexpressing two Raphanus sativus genes (RsPIP1;1 and RsPIP2;1) did not display any increase in drought tolerance (Tsuchihira et al., 2010). To date, no research on the recovery from embolism formation in woody plants with impaired aquaporin expression has been conducted.

In this study, we used poplar transgenic plants characterized by a strong down-regulation of PIP1 genes to test the role of this aquaporin subfamily in the plant response to water stress and subsequent recovery from stress. Although transformed poplars did not show morphologically different phenotypes compared with wild-type plants, they were found to be more sensitive to imposed water stress, resulting in increased vulnerability to embolism formation and the loss of stomatal conductance. We also noted a reduced capacity of transformed plants to restore xylem water transport.

RESULTS

Physiological Changes in Response to Water Stress

Populus tremula × Populus alba-transformed trees were previously generated using a reverse genetic approach (RNA interference) aimed at suppressing more than one gene belonging to the poplar PIP1 subfamily (Secchi and Zwieniecki, 2013). Silencing the entire subfamily was preferred to silencing particular isoforms to avoid the potential for compensation of expression within that same gene group. To estimate levels of PIP1 subfamily down-regulation in the stems of five selected transgenic lines, reverse transcription PCR analyses were performed. The expression of PIP1 genes was strongly reduced in all lines examined compared with the wild type, decreasing by 91–94% (Fig. 1; ANOVA, P < 0.001). Because there were no significant differences in PIP1 expression level among lines (Fig. 1), we pooled all lines into a single transformed group to increase the sample size in subsequent physiological experiments. After being pooled, analysis yielded a 93% reduction in expression for PIP1 subfamily gene compared with wild type (Fig. 2; Student’s t test, P < 0.001). Additionally, PIP1.1 and PIP1.3 genes (the two highest expressed genes in stems from the PIP1 subfamily [Secchi et al., 2009] and the most responsive genes to drought and embolism formation [Secchi and Zwieniecki, 2010]) showed reduced expression levels in the combined transgenic stems, confirming the successful down-regulation of multiple isoforms belonging to the same subfamily. The expression of the other genes belonging to the PIP1 subfamily was also monitored. All PIP1 genes were strongly down-regulated and resulted to be significantly different from their expression in stems of wild-type plants (Fig. 2; Supplemental Fig. S1A). To test the possible compensatory response of the PIP2 gene subfamily members in response to PIP1 down-regulation, the transcript levels of seven of eight genes belonging to the PIP2 subfamily were measured (PIP2.8 was analyzed, but its expression was not detected in the stem tissue). In general, gene expressions were not significantly different from wild-type plants, suggesting a lack of PIP2 compensatory response (Fig. 2; Supplemental Fig. S1B).

Both wild-type and PIP1 down-regulated poplars showed similar levels of native embolism in well-watered plants, averaging around 30.7% and 37.3%, respectively (Fig. 3A). An increase in water stress resulted in additional losses of stem hydraulic conductivity plateauing above 80%, below xylem pressures (P  _x) of −2.3 MPa in both groups. The relationship between the loss of hydraulic conductivity and xylem pressure was fitted using four-parameter logistic curves (dose–response curves; “Materials and Methods”). The 50% loss of effective stem conductivity (EC50; not including native embolism) described by the EC50  _{percent loss of conductance (PLC)} parameter of the curve (half-maximal effective concentration—in our case, effective xylem pressure) occurred at −1.76 MPa (se = 0.0642, t = 27.3493, P < 0.0001) for wild-type plants and −1.43 MPa (se = 0.0816, t = 17.54, P < 0.0001) for transgenic plants. The EC50  _PLC parameter was significantly different between the two groups (Z test; t = 3.120, degrees of freedom [df] = 47, P < 0.0025; Paternoster et al., 1998), indicating that the stems of transgenic lines were more vulnerable to embolism than the stems of the wild type. For the purpose of clarity, we will later refer to plants as moderately stressed when xylem pressure is above EC50  _PLC (i.e. PLC < 50% but below xylem pressure of well-watered plants) and severely stressed if xylem pressure is below EC50  _PLC (i.e. PLC > 50%).

Stomatal conductance (g  _s) was similar for nonstressed (well-watered) wild-type and PIP1 down-regulated plants (∼600 mmol m⁻² s⁻¹; Fig. 3B). A decrease in g  _s was observed with an increase in water stress, and both groups showed full stomatal closure at xylem pressures below −2.0 MPa. Changes in stomatal conductance in response to xylem pressure were fitted with the dose–response curve. An effective drop in g  _s by 50% from its observed maximum in well-watered plants (EC50  _{g  s}) was at −1.102 MPa for wild-type and −1.316 MPa for transgenic lines. There was a statistical difference between EC50  _{g  s} parameters of the two groups (Z test; t = 2.066, df = 43, P < 0.025). Importantly, wild-type plants closed their stomata by 50% at ∼0.6 MPa before EC50  _PLC, providing a relatively wide PLC safety margin, whereas transgenic plants closed stomata at only ∼0.1 MPa ahead of EC50  _PLC, providing a very narrow safety margin. However, both groups completely closed their stomata before the maximum loss of stem conductance occurred.

The total osmotic potential (estimated from combined concentration of sugars and ions) of sap collected from functional vessels was mostly accounted for by the presence of ions (Fig. 4). Ion impact on osmotic potential of xylem sap was 10 times higher than sugar content in xylem sap collected from well-watered and moderately stressed plants and 4 times higher for severely stressed plants (Supplemental Fig. S2). Osmotic potential under no stress conditions was not different between wild-type and transgenic lines. Osmotic potential increased with the increase of water stress in both groups, reaching 0.125 and 0.089 MPa, respectively, for wild-type and transgenic trees. There was no significant difference in sap osmotic potential in moderately stressed plants and only a small increase in osmoticum content in wild-type plant over transgenic plant under severe stress (Fig. 4).

Physiological Changes on Recovery from Induced Water Stress

Moderately and severely stressed plants were rewatered to their field capacity and allowed 1.5 h of recovery time (Fig. 5). Rewatering moderately stressed plants resulted in a fast increase of P  _x in both wild-type and transgenic lines. P  _x returned to the values of nonstressed control plants within the allotted time (Fig. 5). This relief of P  _x was correlated with the restoration of xylem transport capacity for wild-type plants that showed almost full embolism recovery (95.45% initial embolism incidence). However, only partial PLC recovery (43%) was observed in PIP1 down-regulated lines, despite recovery of P  _x. The recovery from embolism was significantly different between wild-type and transgenic plants (Student’s t test; t = 2.1150, df = 17, P < 0.05; Supplemental Table S1). Recovery from severe stress was not different between wild-type and transgenic trees. Both groups showed a significant recovery of xylem pressure, and both groups showed only a partial drop in the level of PLC, which was especially low in plants recovering from a drop in P  _x below −2.0 MPa (Supplemental Table S1).

Direct observation of plant recovery from stress using time-lapse imaging showed that turgor recovery in leaves could be characterized by two phases: (1) a slow phase, which lasted for 37 to 40 min, characterized by a slow, steady decrease in the angle between the petiole and the stem; and (2) a fast phase, which lasted more than 40 min, characterized by a fast change in the angle that resulted in total recovery of initial leaf position (Fig. 6). The rate of recovery in the slow phase was similar in both groups of plants. The rate of recovery during the fast phase was only similar in plants recovering from low stress levels above EC50  _PLC (P  _x was less negative than EC50  _PLC). However, recovery from stress around EC50  _PLC or lower (P  _x was equal to or more negative then EC50  _PLC) was significantly slower in transgenic plants, resulting in a delay of several minutes in restoration of turgor and prestress leaf positions (Fig. 6).

The recovery of stomatal conductance did not follow the patterns of fast recovery observed in P  _x (minutes to a couple of hours; Fig. 7), leaf turgor (Fig. 6), and PLC (Fig. 5). Full g  _s recovery did not occur until after 4 d of full irrigation. The general pattern of g  _s recovery was also different in plants recovering from moderate stress and severe stress between wild-type and transgenic plants. Moderate stress only forced full stomata closure in wild-type plants, whereas transgenic stomata remained partially open (Fig. 7 compared with Fig. 3B). Wild-type moderately stressed plants showed signs of some g  _s recovery immediately after rewatering (first day; Fig. 7, B and C). The recovery was not observed in transgenic plants, despite their tendency to have higher initial (under stress) g  _s. The recovery of g  _s continued in wild-type plants during the second day, reaching the g  _s of control, nonstressed plants during the third day. Full recovery occurred 4 d after the return of irrigation. Transgenic plants did not show significant signs of g  _s recovery in the second day, but recovery was similar to wild-type plants in the third and fourth days. Severe stress (below EC50  _PLC) forced full stomatal closure in both wild-type and transgenic plants. Recovery from severe water stress did not start during the first day for either plants group but later, showed a similar pattern to that observed in recovery from moderate stress. Again, the start of the partial recovery of g  _s in transgenic lines was delayed 1 d compared with wild type (Fig. 7, D and E). Differences of g  _s values between mornings and afternoons were related to variation in greenhouse temperature (Fig. 7A).

The osmotic potentials (sugars and ions) of sap collected from functional vessels during recovery from stress are accounted for by ion concentration, with little contribution from sugar content (Supplemental Fig. S3). Wild-type plants recovering from moderate and severe stress showed osmotic potential values similar to those values measured in well-watered plants (Fig. 8, gray bars), with the exception of higher osmotic values found under severe stress. Liquid collected from moderately and severely stressed transgenic plants had lower osmoticum concentrations than control plants (Fig. 8, white bars), suggesting that these plants may be impaired in their use of ions in their response to stress.

DISCUSSION

The significant role that PIP1 aquaporins play in stem response to both presence of water stress (Secchi and Zwieniecki, 2010) and artificially induced embolism in the xylem (Secchi and Zwieniecki, 2011) has been previously suggested. However, these studies provided association based only on changes in PIP gene expression in response to particular treatments. The presented approach of comparative response analysis between wild-type and transgenic plants, with down-regulated expression of multiple isoforms of the PIP1 subfamily, directly points at the role that PIP1 genes play in two basic physiological functions: stem hydraulics and stomatal conductance during the onset of water stress. We also show that the dynamics of recovery from water stress are significantly affected by lower levels of PIP1 expression.

Stems of transgenic poplars showed a substantial reduction in PIP1 gene expression in all selected lines (Fig. 1). The pooled transformed group did not show a significant compensatory regulation effect of PIP2 expression in response to PIP1 down-regulation, whereas different isoforms belonging to the PIP1 subfamily were strongly down-regulated (Fig. 2). Despite the substantial suppression of gene expression, down-regulating the PIP1 transcript did not affect the basic physiological functions of nonstressed trees. Both wild-type and transgenic plants had similar stomatal conductance and stem hydraulic properties, including similar levels of native xylem embolism (Fig. 3A). This lack of phenotypic or functional impairment because of the down-regulation of PIP1 genes coincides with a previous report showing that both plant groups did not significantly differ in physiological functions related to photosynthesis (Secchi and Zwieniecki, 2013). This finding is also in agreement with other studies; data reported for transgenic banana (Musa spp.) plants (constitutively overexpressing a PIP1;2 gene) showed that they were phenotypically and physiologically indistinguishable from the untransformed lines under normal growth conditions (Sreedharan et al., 2013). A similar behavior was also reported by Siefritz et al., (2002), evidencing that, despite a strong reduction in NtAQP1 expression and changes in root hydraulic conductivity, tobacco plants grown under optimal conditions in the greenhouse did not show morphological changes.

The physiological similarity of wild-type and transgenic plants did not persist under water stress conditions. Transgenic lines were more susceptible to xylem embolism, displaying a 50% loss of PLC at less negative xylem pressure (−1.4 MPa versus −1.7 MPa for transgenic and wild-type plants, respectively), indicating that the presence of PIP1 genes might be beneficial to plant cavitation resistance (Fig. 3A). This increased vulnerability to embolism in transgenic plants was associated with their reduced capacity to control stomatal conductance during stress development, suggesting a different content of abscisic acid (ABA) in the tissues and consequently, a delay of stomata closure in response to drought stress. In transgenic plants, 50% of stomatal shutdown occurred at −1.3 MPa, and in wild-type plants, 50% of stomatal shutdown occurred at −1.1 MPa, with approximately 40% of maximum g  _s at EC50  _PLC for transgenic plants and less than 10% of maximum g  _s at EC50  _PLC for wild-type plants (Fig. 3B). Because stomatal closure is one of the mechanisms that plants can adopt to limit water loss and control stress level, the behavior assumed by transgenic plants suggests that they were less likely to control transpiration rates to protect xylem from embolism formation. In other words, the down-regulation of PIP1 gene expression resulted in a significant reduction of the xylem vulnerability safety margin (Sperry and Ikeda, 1997; Choat et al., 2012; Johnson et al., 2012).

Because embolism formation is mostly a physical process, it might be hard to imagine how PIP1 genes are able to influence a plant’s vulnerability curve without also affecting morphology or xylem anatomy. However, previous studies have shown a positive correlation between PIP1 expression and stress conditions as well as the up-regulation of PIP1 expression in response to embolism formation, even in the absence of water stress (Secchi and Zwieniecki, 2010). Thus, combined with the results presented here, we can propose that the processes of embolism formation and refilling are not separated in time but happen simultaneously with respective rates that are functionally linked to plant stress level. Embolism formation rates are expected to increase with increases in stress, whereas embolism refilling rates are expected to decrease with stress. To rephrase, the current level of PLC at any given moment is a result of embolisms formed minus refilling. Because the down-regulation of PIP1 genes negatively affects the refilling rate, the result of the competition between the two processes shifts transgenic plants to an apparently higher vulnerability to embolism formation, whereas it is, in fact, a result of reduced capacity to refill. Thus, PIP1s do not directly influence cavitation thresholds but do reduce refilling rates.

Rewatering moderately stressed plants resulted in the fast recovery of xylem pressure in both groups of plants and was followed by full or partial recovery of xylem hydraulic conductivity. Full recovery of hydraulic capacity to the initial PLC was observed in wild-type plants recovering from moderate stress, whereas the xylem PLC recovery of transgenic plants was significantly impaired. PLC recovery was in the range of ∼44% within 1.5 h, indicating that only wild-type plants were capable of dealing with moderate embolism incidence over short temporal scales, despite active transpiration and the presence of tension (Fig. 5). This impaired recovery of PLC observed in transgenic plants might be partially related to the generally lower xylem sap osmotic potential of transgenic plants (Figs. 4 and 8). Although the concentration of sap collected from functional vessels is very low and cannot be used to explain the driving force required for recovery (Secchi and Zwieniecki, 2012), it can reflect the ability of plant xylem parenchyma cells to move solutes to vessels, thus suggesting that wild-type plants were more capable of loading vessels with solutes than transgenic plants, especially under increasing stress conditions. The observed delay in recovery of stem hydraulic parameters (several hours) in transgenic poplar was relatively small compared with reports of the slow recovery observed in the transgenic Arabidopsis plants with down-regulated PIP genes. A delay of several days was observed for the recovery of hydraulic conductance and transpiration rates of transgenic plants returning from stress compared with wild-type plants (Martre et al., 2002).

The dynamics of recovery from water stress in terms of P  _x and PLC did not coincide with the recovery of stomatal conductance. Although wild-type plants recover significantly faster during the first 2 d after rewatering than transgenic plants, a full recovery did not occur until 4 d of full irrigation (Fig. 7). This pattern of PLC recovery but delayed g  _s recovery from drought is a common phenomenon and has been observed in Eucalyptus pauciflora; stem hydraulic capacity was restored within 6 h, but stomatal conductance had not fully recovered even 10 d after a return to favorable water status (Martorell et al., 2013). Similar results were reported for grapevines; xylem embolism in petioles, roots, and shoots recovered during the 24 h after rehydration, whereas stomatal conductance required an additional 48 h (Lovisolo et al., 2008). Evidence of a delay in g  _s recovery from drought suggests that the regulation of stomatal conductance depends on factors beyond the supply of water through the xylem and stem water pressure, possibly the functionality/integrity of the photosynthetic system and/or ABA physiology (Lovisolo et al., 2008; Brodribb and McAdam, 2013). Thus, the delayed g  _s recovery in transgenic plants might imply that the regulation of aquaporin expression may involve an ABA-dependent signaling pathways (Wan et al., 2004).

The function of PIP1 genes in promoting tolerance to water stress has been studied through reverse genetic approaches that overexpress or underexpress aquaporin genes (Martre et al., 2002; Cui et al., 2008; Kaldenhoff et al., 2008; Zhang et al., 2008; Postaire et al., 2010; Sreedharan et al., 2013). Conclusions have been made with antisense tobacco plants; under well-watered conditions, the NtAQP1 aquaporin did not seem to be important for water uptake or management. However, in a water-limited environment, NtAQP1 antisense plants were not able to maintain turgor and seemed to be less drought tolerant compared with wild-type tobacco plants (Siefritz et al., 2002). When subjected to drought, tobacco plants with overexpressed a Brassica napus gene were more tolerant to water stress, whereas the antisense lines with reduced mRNA levels of a B. napus gene showed reduced water uptake and were more sensitive to water deficiency (Yu et al., 2005). Results presented here suggest that the major effect of the PIP1 gene subfamily on the stress physiology of woody plants is directly linked to plant management of apparent vulnerability to embolism (Secchi and Zwieniecki, 2010; Secchi and Zwieniecki, 2012). Specifically, we propose here that the observed increased vulnerability to embolism formation and the significant delay in recovery of hydraulic capacity in transgenic plants indicates that a loss of PIP1 gene expression reduces the rate of refilling and effectively shifts the balance between the rates of embolism formation and refilling, such that embolism became a dominate process at lower tensions. Apparent increased vulnerability without obvious phenotypic differences between wild-type and transgenic plants underlines the role of physiology in the maintenance of xylem hydraulic capacity and suggests that continuous competition between the processes of embolism formation and removal might be mediated by PIP1 activity (Zwieniecki and Holbrook, 2009; Zwieniecki et al., 2013).

MATERIALS AND METHODS

Plant Materials and Experimental Design

Wild-type and transgenic hybrid white poplars (Populus tremula × Populus alba; Institut National de la Recherche Agronomique France clone 717-1B4) were used for the study. Down-regulated PIP1 transgenic plants were previously generated and described in the work by Secchi and Zwieniecki (2013). Poplars were grown in a greenhouse with the following ambient conditions: temperature maintained in the range of 25–32°C and natural daylight supplemented with light from metal halogen lamps to maintain a minimum of 500–600 µmol photons m⁻² s⁻¹ during a 12-h-light/12-h-dark cycle. Plants were approximately 1.5 m tall at the onset of the experiments.

In total, 124 poplars were used (56 wild-type and 64 transgenic plants) in four different experiments.

(1) Poplar response to increasing water stress; 53 poplars (25 wild-type and 28 transgenic plants) were used in this study, and of these poplars, 12 plants were kept as controls. They were watered every day to field capacity (well-watered plants). The remaining 41 plants were gradually subjected to drought by stopping irrigation. Plants were used to construct PLC curve and relate stomatal conductance to xylem pressure. Duration of drought treatment depended on the levels of desired water stress, and it was between 1 and 5 d. Physiological measurements (PLC, xylem pressure, and g  _s) were performed between 9 am and 12 pm.

(2) Dynamics of plants recovery from embolism; 35 plants (14 wild type and 21 transgenic) were water stressed and then rewatered in the morning (∼9 am). Plants were allowed 1.5 h of recovery time followed by measurements of PLC and xylem pressure.

(3) Dynamics of stomatal conductance recovery; 20 stressed plants (11 wild type and 9 transgenic) were rewatered in the morning (∼9 am) to field capacity, and the dynamic of stomatal conductance recovery was monitored. Stomatal measurements were started just before the rewatering (∼9 am) and continued until 3 pm during 4 consecutive d. Plants were irrigated several times during a day.

(4) Dynamics of plants recovery from wilting (movies); 16 plants (8 wild type and 8 transgenic) were used. One wild-type plant and one transgenic plant were grown in the same 5.7 × 8.3-cm pot. Plants were allowed to wilt in the pots over the period of several days. When the desired wilting point was achieved, plants were rewatered, and subsequent recovery was observed. The temporal dynamics of leaf movements during increasing water stress and recovery were recorded using time-lapse video. Pictures were taken during water stress development every 5 min and every 30 s during the recovery period. Analysis of leaf motion was performed every 1 h during stress development and every 6 min during the recovery period.

Expression of Aquaporin Genes in Transgenic Poplars

Total RNA was isolated from wild-type and transgenic stems according to the protocol of Chang et al. (1993). First-strand complementary DNA was synthesized from total RNA treated with DNase I (Fermentas) using oligo(dT)_12–18 as primers (Fermentas) and SuperScript II Reverse Transcriptase (Invitrogen). The sequences of primers used for real-time PCR analysis are listed in Supplemental Table S2. Primers were tested on complementary DNA of hybrid poplar through PCR with RED Taq DNA Polymerase (Sigma) according to the manufacturer’s instructions. The transcript abundance of each gene was quantified with SYBR Green JumpStart Taq Ready Mix (Sigma) on an Eco Real-Time PCR System (Illumina). Thermocycler conditions for all real-time analyses were 95°C for 5 min followed by 40 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30s.

Data were analyzed using Eco software (Illumina), and the expression values were normalized to the geometric mean of two housekeeping genes (ubiquitin and actin). These genes were found to have, for the same poplar species, the highest amplification efficiency and most stable expression across different tissues (Carraro et al., 2012). Real-time PCR was carried out using three biological replicates per transformed line. Two technical replicates were performed for each of the three biological replicates.

Measurements of Xylem Pressure and Stem Hydraulic Conductivity

Stem water pressure was measured on nontranspiring leaves. Leaves were covered with aluminum foil and placed in a humidified plastic bag for 15 min before excision. After excision, leaves were allowed to equilibrate for an additional 20 min before the xylem pressure was measured using a Scholander-type pressure chamber (Soil Moisture Equipment Corp.).

Stem hydraulic conductivity was measured using a standard approach that was previously described (Secchi and Zwieniecki, 2010). Briefly, a 1.5-m-long shoot was cut under water. Within a few minutes, this initial cut was followed with cutting a set of three stem segments. Segments were excised under water approximately 20–30 cm from the initial cut (distance longer than 2 times the length of vessels in studied poplar). Each segment wasapproximately 4-cm long. The initial hydraulic conductance (k  _i) of each stem segment was measured gravimetrically by determining the flow rate of filtered 10 mm KCl solution. A water source was located on a balance (Sartorius ± 0.1 mg) and connected to the stem by a plastic tube. During measurements, stems were submerged in a water bath with a water level approximately 10 cm below the level of water on the balance. After a steady flow rate was reached (within a few minutes), the tube connecting the stem to the balance was closed, and a bypass tube was used to push water across the segment under approximately 0.2 MPa pressure for approximately 20 s to remove embolism. The chosen segment length used for determining PLC was short enough to have the majority of vessels open in poplar stems (vessel length is usually approximately 5 cm), thus making removal of embolism very easy and complete within a few seconds. Stem conductance was then remeasured to find maximum conductance (k  _max). The PLC was calculated as PLC = 100 × (k  _max − k  _i)/k  _max. The same procedure was used in experiments 1 and 2.

Measurements of Stomatal Conductance

Stomatal conductance was measured using an SC-1 Leaf Porometer (Decagon Devices) on fully expanded leaves. g  _s Values on control and water-stressed plants (wild type and transgenic) were measured between 9 am and 12 am with increasing water stress and from 9 am to 3 pm during recovery from stress (experiments 1 and 3). Along the process, several leaves were collected for estimation of stem water pressure.

Curve Fitting

Relationships between PLC and stomatal conductance (g  _s) in response to stem water pressure were fitted with a four-parameter logistic curve (dose response), where PLC = initial_PLC + (maximum_PLC − initial_PLC)/(1 + (P  _x/EC50  _PLC)^SlopeEC50PLC) and g  _s = minimum_{g  s} + (intial_{g  s} − minimum_gs)/(1 + (P  _x/EC50  _{g  s})^{SlopeEC50  g  s}). This function was preferred over other sigmoidal shapes, because it allows for treating xylem pressure as a treatment (dose) and allows for the fit of initial values to true preexisting conditions. EC50  _{PLC(  g  s)} is the parameter describing a 50% change in the curve between the initial value of PLC or g  _S and the corresponding final value at very low xylem pressures (PLC − maximum and g  _s − minimum). Slope_{PLC(  g  s)} describes the rate of change in PLC or g  _s at the inclination point of the curve. To compare wild-type and transgenic plants, PLC, and g  _s response to xylem water pressure, we compared EC50  _{PLC(  g  s)} parameters of fitted curves using the corrected statistical Z test for the equality of regression coefficients (Paternoster et al., 1998).

Carbohydrate and Ion Contents in Xylem Sap of Functional Vessels

The xylem sap of functional vessels was collected from the same plants that were used to determine PLC and xylem pressure (experiments 1 and 2) using the procedure previously described (Secchi and Zwieniecki, 2012). Briefly, leaves were removed, and the stem was attached through a plastic tube to a syringe needle. The needle was threaded through a rubber cork to a vacuum chamber, with the needle tip placed in the 1.5-mL plastic tube. After the generation of a vacuum, short pieces of stem were consecutively cut from the top, allowing liquid from open vessels to be sucked out of the stem and collected in the tube. Collected liquid was then analyzed for sugar and ion concentrations following the procedures described in detail by Secchi and Zwieniecki (2012). Briefly, carbohydrate content was quantified using the colorimetric anthrone-sulfuric acid assay (Leyva et al., 2008); 50 µL xylem sap was added to 150 µL fresh anthrone reagent, and samples were mixed, kept 10 min at 4°C, and then incubated 20 min at 100°C. After heating, they were cooled for 20 min at room temperature, and the absorbance at 620 nm was read with a microplate multiscan reader (Multiscan Thermo Scientific). Total carbohydrate content was calculated as milligrams per milliliter Glc, and from the deduced molal concentration of each xylem sap solution, the relative osmotic potential was calculated based on the law for perfect gases: Π = miRT, where m is molality of the solution (moles of solutes per 1,000 g H₂0), i is a constant that accounts for ionization of the solute (for Glc, i = 1), R is the gas constant (0.00831 L MPa mol⁻¹ K⁻¹), and T is the temperature (293.16 K).

Ion concentration was measured as electrical conductivity using a 5-µL capillary fitted with gold electrodes at the both ends and connected to a multimeter (True RMS digital multimeter 289; Fluka Europe). Liquid samples were sucked into the capillary using a pipettor. A series of potassium chloride solutions with different concentrations was used to establish a calibration curve. Electrical conductivity of xylem sap was translated to the equivalent concentration of potassium ions.

Movies on Drought

Digital images taken from movies (experiment 4) were analyzed using ImageJ software (http://rsbweb.nih.gov/ij/). For each plant, at least two leaves were selected, and the angle between stem axes and arm, described as a line linking leaf blade base and petiole base, was measured on a series of consecutive pictures during the increasing of water stress and recovery from stress. Dynamics of the stress development required measurements at 1-h intervals, whereas recovery from stress was much faster and required measurements at 6-min intervals.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Relative gene expression in the stems of each transgenic line analyzed.

• Supplemental Figure S2. A, Sugar and (B) ion osmotic potentials collected from the xylem sap of plants (light green, transgenic; light red, wild type) subjected to increased water stress (P  _x).

• Supplemental Figure S3. A, Sugar and (B) ion osmotic potential collected from xylem sap of well-watered, stressed, and recovered plants.

• Supplemental Table S1. Mean values (± sd) of xylem pressure (P  _x) and percent of PLC recovered for moderately and severely stressed wild-type and transgenic plants.

• Supplemental Table S2. Sequences of primers used for quantitative real-time PCR.

ACKNOWLEDGMENTS

We thank Benjamin Taylor, Ramona Hihn, and Anna Saffray for help during the experimental work and Jessie Godfrey for editorial help with the manuscript.

Glossary

 
• ABA

  abscisic acid

 
• EC50

  50% loss of effective stem conductivity

 
• PIP

  plasma intrinsic protein

 
• PLC

  percent loss of conductance

LITERATURE CITED

Author notes