Vacuolar Na+/H+ NHX-Type Antiporters Are Required for Cellular K+ Homeostasis, Microtubule Organization and Directional Root Growth

Na+/H+ antiporters (NHXs) are integral membrane transporters that catalyze the electroneutral exchange of K+ or Na+ for H+ and are implicated in cell expansion, development, pH and ion homeostasis and salt tolerance. Arabidopsis contains four vacuolar NHX isoforms (NHX1–NHX4), but only the functional roles for NHX1 and NHX2 have been assessed thus far. Colocalization studies indicated that NHX3 and NHX4 colocalize to the tonoplast. To investigate the role of all vacuolar NHX isoforms, a quadruple knockout nhx1nhx2nhx3nhx4, lacking all vacuolar NHXs, was generated. Seedlings of nhx1nhx2nhx3nhx4 displayed significantly reduced growth, with markedly shorter hypocotyls. Under high K+, but not Na+, pronounced root skewing occurred in nhx1nhx2nhx3nhx4, suggesting that the organization of the cytoskeleton might be perturbed. Whole mount immunolabeling of cortical microtubules indicated that high K+ caused significant microtubule reorganization in nhx1nhx2nhx3nhx4 root cells of the elongation zone. Using microtubule stabilizing (Taxol) and destabilizing (propyzamide) drugs, we found that the effect of K+ on nhx1nhx2nhx3nhx4 root growth was antagonistic to that of Taxol, whereas elevated K+ exacerbated the endogenous effect of propyzamide on root skewing. Collectively, our results suggest that altered K+ homeostasis leads to an increase in the dynamics of cortical microtubule reorganization in nhx1nhx2nhx3nhx4 root epidermal cells of the elongation zone.

homeostasis in microtubule organization. Here, we show that vacuolar NHX1 to NHX4 reside at the tonoplast. Using the quadruple knockout nhx1nhx2nhx3nhx4, lacking all four vacuolar NHXs, we show that vacuolar NHXs impart essential functions in K + homeostasis that affect directional root growth and cytoskeletal dynamics. The quadruple knockout exhibits more pronounced phenotypes compared to the nhx1nhx2 double knockout, including increased sensitivity to K + , suggesting that NHX3 and NHX4 have additional roles in cell expansion and K + homeostasis.

Results and Discussion
We sought to assess the contribution of all vacuolar NHXs to overall plant growth and development by generating a quadruple knockout lacking all vacuolar NHX antiporters, NHX1, NHX2, NHX3 and NHX4. Using available T-DNA insertion lines ( Figure A1), we generated the quadruple knockout nhx1nhx2nhx3nhx4 through successive crosses of single, double and triple knockouts, as described in the Methods. The level of expression of all four vacuolar NHX transcripts in the nhx1nhx2nhx3nhx4 knockout indicated that no full-length transcripts exist ( Figure A1).

NHX3 and NHX4 Are Colocalized at the Tonoplast
To determine the subcellular localization of NHX3 and NHX4, we generated constitutively expressed translational fusion proteins of both NHX3 and NHX4 with CFP and RFP, respectively. Transient co-expression of NHX3-CFP and NHX4-RFP in Arabidopsis cotyledons displayed overlapping expression at the periphery of the cell, as well as at transvacuolar strands ( Figure 1A,C). Because NHX1 and NHX2 were shown previously to localize at the tonoplast [3], we transiently expressed NHX4-CFP in plants stably expressing NHX2-YFP and observed an overlapping expression pattern between NHX2 and NHX4 that was characteristic of a vacuolar signal, including trans-vacuolar strands and vacuolar bulbs (arrowheads in Figure 1D-F). We then expressed NHX3-RFP and NHX4-CFP in aquaporin γ-TIP [14] and the R-SNARE VAMP711 [15] expressing lines, which were previously shown to colocalize with NHX1 [3]. In cotyledon epidermal cells, an overlapping expression between NHX3-RFP and γ-TIP-GFP, including in vacuolar bulbs that are characteristic of γ-TIP, was seen ( Figure 1G-I). Overlapping expression was also seen with NHX4-CFP and VAMP711-RFP ( Figure 1J-L). Taken together, these results indicated that both NHX3 and NHX4 localize to the tonoplast, as suggested by other heterologous expression studies [5,7] and sequence similarity to NHX1 and NHX2 [1].

Growth Phenotypes of nhx1nhx2nhx3nhx4 and Responses to Na + , K +
We used the etiolation response of dark-grown seedlings to examine the contribution of NHX1, NHX2, NHX3 and NHX4 to cell expansion and seedling growth. Previously, we reported that etiolated nhx1nhx2 double knockout seedlings had profound hypocotyl and root elongation phenotypes [3]. Etiolated nhx1nhx2nhx3nhx4 seedlings displayed significantly reduced growth of both roots and hypocotyls ( Figure 2). In control media, hypocotyl lengths of nhx1nhx2nhx3nhx4 were less than 50% the length of WT after 11 days of etiolated growth ( Figure 2M). The hypocotyl elongation response of nhx1nhx2nhx3nhx4 was highly responsive to both added Na + and K + . For example, hypocotyls of nhx1nhx2nhx3nhx4 grown in 30 mM NaCl were 31% longer than those of seedlings grown on control media (1 mM NaCl, 1 mM KCl), but hypocotyls of nhx1nhx2nhx3nhx4 seedlings grown in 30 mM KCl were 49% shorter ( Figure 2D,H,K,L) than those grown in control media. By contrast, little change in hypocotyl length was observed in the WT seedlings grown on either 30 mM NaCl or KCl ( Figure 2B,F,J,L). Root growth was also markedly different between nhx1nhx2nhx3nhx4 and WT vertically-grown seedlings, but only in response to either K + or Na + . Similar to the K + and Na + -dependent growth response of hypocotyls, nhx1nhx2nhx3nhx4 roots were also significantly shorter (26%) under K + and longer (13%) under Na + as compared to nhx1nhx2nhx3nhx4 grown in control media. In addition, the nhx1nhx2nhx3nhx4 seedlings began to display prominent left-handed skewing (as viewed through the agar, following the convention assigned by Rutherford and Masson [16]) after four days of growth on media supplemented with 30 mM K + ( Figure A2). Seedlings grown for another seven days on the same media displayed complete curling of roots ( Figure 2H). By contrast, WT root growth remained unchanged under 30 mM K + even after 11 days.  Etiolation response of the knockout nhx1nhx2nhx3nhx4. Etiolated wild-type (A,C,E) and nhx1nhx2nhx3nhx4 (B,D,F) seedlings grown for 11 days in the dark on either 30 mM K + (C,D) or 30 mM Na + media (E,F). No significant difference in hypocotyl or root growth was noted in the wild-type; however, in nhx1nhx2nhx3nhx4, root and hypocotyl growth was significantly inhibited in seedlings grown in 30 mM K + and increased in those grown in 30 mM Na + , as compared to control media. Quantification of hypocotyl (G) and root (H) growth normalized relative to growth on control media (1 mM Na + and 1 mM K + ). nhx1nhx2nhx3nhx4 displayed left-handed root skewing when grown on media containing 30 mM K + (D), but not 1 mM K + (B). Values represent the mean ± SE (n = 29). * p ≤ 0.05, *** p ≤ 0.001; p-values reflect significant differences between control and either K + or Na + treatments within each genotype. The yellow bar is 0.5 cm.
Previously, we reported that etiolated nhx1nhx2 plants also displayed reduced hypocotyl growth, that was 70% the length of WT hypocotyls on media containing 1 mM K + and 50% of WT on 30 mM K + [3], while in etiolated nhx1nhx2nhx3nhx4 seedlings, the average hypocotyl length was 40% and 20% of WT on 1 and 30 mM K + , respectively. A comparison of root growth responses between nhx1nhx2nhx3nhx4 and nhx1nhx2 to high K + [3] indicates that root curling is also more pronounced and root growth more sensitive to high K + , similar to that observed in hypocotyls. The dramatic differences in root and hypocotyl elongation between nhx1nhx2 and nhx1nhx2nhx3nhx4 under the higher K + suggest that NHX3 and NHX4 have additional roles in K + homeostasis and cell expansion.
To determine if the root curling observed in nhx1nhx2nhx3nhx4 roots was specific to K + , we compared the root growth of nhx1nhx2nhx3nhx4 on media supplemented with KCl, KNO3, NaCl or mannitol. Mannitol was included in order to determine whether the root curling response of nhx1nhx2nhx3nhx4 under high K + was due to an osmotic effect; thus, 60 mM mannitol (60 mOsm, equivalent to the osmolarity of the added salt) was used. We found that nhx1nhx2nhx3nhx4 root growth on mannitol was not different from the control ( Figure A3J,B), and therefore not likely due to an osmotic effect. As shown in Figure A3, both 30 mM KCl and 30 mM KNO3 supplemented media resulted in similar left-handed root skewing and curling of the root tip that was not seen when nhx1nhx2nhx3nhx4 was grown on mannitol, suggesting that the response of nhx1nhx2nhx3nhx4 was specific to K + . Collectively, experiments comparing nhx1nhx2nhx3nhx4 seedling growth (i.e., Figures 2 and A3) indicate that the knockouts' root curling response was caused by aberrant K + homeostasis.
Previous work has provided a strong link between ion stress and directional root growth. For example, sos1(nhx7) null mutants displayed a loss of normal gravitropic root growth when grown vertically on media supplemented with high (50-150 mM) Na + , with young roots exhibiting skewing and eventual upward directed growth [17]. Moreover, roots of both WT and sos1(nhx7) seedlings grown on high Na + showed delayed gravitropic responses, thereby suggesting that aberrant intracellular ion homeostasis can specifically affect directional root growth. The sos1 skewing phenotype is remarkably similar to nhx1nhx2nhx3nhx4 grown on K + supplemented media, and although effects of high cytosolic Na + have been suggested, the effects of K + on root skewing remain unclear. Furthermore, sos1(nhx7) mutants were more sensitive to microtubule reorganization, with whole mount immunostaining experiments in elongating root epidermal cells demonstrating that sos1 (nhx7) in the presence of 50 mM Na + had helically-orientated microtubules, while WT plants maintained the normal transverse array [12]. Although microtubule remodeling has been shown to be induced by high external Na + , reports of K + -induced microtubule reorganization are sparse [12,13,18]. In KCl (350 mM)-treated Zea mays, a reorganization of the cortical microtubule array from the transverse to oblique orientation in root cells of the maturation zone occurred within 3 min after application. By contrast, treatment with sorbitol did not result in the reorganization of microtubules [19]. These results suggested that cortical microtubule organization in roots was affected by high concentrations of Na + and K + . Given the phenotypes of nhx1nhx2nhx3nhx4, the knockout represents a useful tool to study the role of K + in cytoskeletal organization. The positive effect of Na + and the negative effect on K + on nhx1nhx2nhx3nhx4 growth is well supported by similar results obtained in nhx1nhx2 in which Na + was proposed to substitute for the lack of K + accumulation in vacuoles, providing the osmoticum needed to generate the turgor for cell expansion [3,10]. The sensitivity of both nhx1nhx2 and nhx1nhx2nhx3nhx4 to added K + , as well as the aberrantly high cytosolic K + concentrations of nhx1nhx2, together with previously reported effects of Na + on cytoskeleton disassembly and reorganization, suggests that aberrant K + homeostasis in the nhx1nhx2nhx3nhx4 might affect the organization of the cytoskeleton.

nhx1nhx2nhx3nhx4 Lacks Accumulation of K + in the Vacuole
Previously, we quantified the concentration of K + in vacuoles of the knockout nhx1nhx2 using the K + binding dye, PBFI [3]. Using the same approach, we quantified the vacuolar [K + ] of nhx1nhx2nhx3nhx4 in root tip cells of the elongation zone, where K + -induced root curling was observed ( Figure 2D). In nhx1nhx2nhx3nhx4, vacuolar K + was ~19 mM (±7 mM) and significantly lower than WT, which was 77 mM (±13 mM). These values are similar to those obtained in nhx1nhx2, suggesting that the lack of NHX3 and NHX4 in addition to NHX1 and NHX2 did not significantly alter root tip cell vacuolar K + . In the same nhx1nhx2 knockout, the cytosolic K + was also reported to be aberrantly high compared to WT [9]. Therefore, given that the electrochemical potential of K + favors uptake into the cell [1,3], it is possible that in nhx1nhx2nhx3nhx4, the lack of K + accumulation into the vacuole might lead to an aberrantly high K + accumulation in the cytosol.

Cells of the Root Transition Zone Lack Expansion in nhx1nhx2nhx3nhx4
To investigate the significance of NHX1 to NHX4-mediated Na + and K + transport on cell expansion and root skewing, we stained roots of nhx1nhx2nhx3nhx4 seedlings with propidium iodide and visualized cell size. Compared with WT, cells in the nhx1nhx2nhx3nhx4 root transition zone appeared to be much shorter when grown on 30 mM KCl-supplemented media ( Figure 3E,F). Cell size was similar between WT and nhx1nhx2nhx3nhx4 in the apical meristem, but epidermal cells in the transition zone were markedly smaller in nhx1nhx2nhx3nhx4. Whereas in the WT, a gradual increase in cell size from the meristem to the elongation zone ( Figure 3E, arrows) was apparent, the length of epidermal cells in the transition zone of nhx1nhx2nhx3nhx4 was aberrantly small, and as a consequence, the transition zone appeared "compressed" (Figure 3F, white arrows). This phenotype was not evident in nhx1nhx2nhx3nhx4 grown on 30 mM NaCl-supplemented media, where no difference in cell size or shape was observed between WT and nhx1nhx2nhx3nhx4 in any root zones ( Figure 3C,D).

K + Dependent Root Skewing Depends on the Dynamic Organization of the Cytoskeleton
In order to investigate whether the root growth response of nhx1nhx2nhx3nhx4 to elevated K + concentrations was associated with dynamic changes in the organization of the cytoskeleton, we tested the effect of microtubule perturbing drugs on root morphology and directional root growth in etiolated nhx1nhx2nhx3nhx4 seedlings as above. Propyzamide is a microtubule destabilizing drug that is known to induce right-handed root skewing (when viewed through the agar) without inhibiting overall growth and root elongation [12,20,21]. When applied at 3 μM, anisotropic root growth of wild-type Arabidopsis is moderately impaired without causing swelling of epidermal roots cells [22]. We measured root growth angles for both WT and nhx1nhx2nhx3nhx4 vertically-grown seedlings on minimal media (containing 1 mM KCl) and an identical media supplemented with an additional 30 mM KCl. As shown in Figure 4, both the WT and nhx1nhx2nhx3nhx4 displayed the characteristic right-handed skewing ( Figure 4E,F) when grown in the presence of 3 µM propyzamide, but only nhx1nhx2nhx3nhx4 displayed left-handed skewing in response to 30 mM K + ( Figure 4D). When both elevated K + and 3 µM propyzamide were included in the media, roots from WT seedlings continued to skew towards the right, unlike the roots from nhx1nhx2nhx3nhx4, which displayed less pronounced skewing ( Figure 4G,H).  Wang and colleagues demonstrated that WT and sos1(nhx7) seedlings exhibited left-handed root skewing when grown vertically on NaCl-enriched MS media, similar to the root skewing of nhx1nhx2nhx3nhx4 in the presence of high KCl. Their results indicated that 50 mM Na + was sufficient to initiate root skewing that eventually resulting in strong root curling and the formation of root "loops". Interestingly, Na + -dependent root skewing was exacerbated in sos1(nhx7) mutants, which displayed more prominent root skewing than the WT at lower Na + treatments, suggesting that cytosolic ion accumulation played an important role in directional root growth. NaCl-induced salt stress resulted in seedling death, which occurred more frequently in sos1(nhx7) than WT seedlings, but was partially prevented by propyzamide treatment, suggesting that microtubule reorganization was required for salt tolerance [18].
We also tested the effect of Taxol, a microtubule stabilizing drug, on K + -induced root skewing in nhx1nhx2nhx3nhx4 plants. A characteristic effect of Taxol on roots is a right-handed skewing response that is dose dependent [23]. We observed that 0.5 µM Taxol was sufficient to induce right-handed root skewing in the WT ( Figure 5E), but the same concentration did not significantly change the root growth direction of nhx1nhx2nhx3nhx4 plants ( Figure 5F). It is unlikely that the lack of nhx1nhx2nhx3nhx4 root skewing observed in the 0.5 µM Taxol treatment was caused by a lack of overall root growth, because nhx1nhx2nhx3nhx4 roots did not vary significantly in length between 0.5 µM or 1.0 µM Taxol treatments. Indeed, the addition of 30 mM KCl was sufficient to induce left-handed skewing in nhx1nhx2nhx3nhx4, despite the presence of 0.5 µM Taxol ( Figure 5J). In contrast, the added KCl had no effect on WT root growth, with seedlings continuing to skew to the right ( Figure 5I). We observed that a threshold of 1.0 µM Taxol was sufficient to overcome the K + -induced left-handed skewing in nhx1nhx2nhx3nhx4 plants and to cause right-handed skewing, as shown ( Figure 5L). Previously, it was shown that 0.5 μM Taxol counters the effect of Na + -induced root skewing, which is dose dependent [18]; a result that is strikingly similar to what we observed in Taxol-and K + -treated nhx1nhx2nhx3nhx4 ( Figure 5). To better understand the possible effect of high K + on cytoskeletal organization, we used a whole mount immunostaining approach to visualize the organization of cortical microtubules in cells of the root elongation zone, where pronounced skewing and curling occurred.
As depicted in Figure 6, no differences were observed between the microtubule organization of roots from nhx1nhx2nhx3nhx4 and the WT when grown on media containing 1 mM K + (Figure 6A,B). By contrast, nhx1nhx2nhx3nhx4 seedlings grown on media containing 30 mM K + displayed a helical orientation of the cortical microtubule array in elongating nhx1nhx2nhx3nhx4 root cells ( Figure 6D) that was observed only when roots exhibited left-handed skewing. Both WT and nhx1nhx2nhx3nhx4 displayed the characteristic Taxol-induced "bundling" of microtubules [24,25] with a highly parallel microtubule organization ( Figure 6E,F). However, when seedlings were grown on media with both 0.5 µM Taxol and 30 mM KCl, the nhx1nhx2nhx3nhx4 microtubules displayed a similar spiraling pattern that was seen when grown on 30 mM KCl plates ( Figure 6H,D). Higher K + had no observable effect on microtubule organization in the WT, even in the presence of Taxol ( Figure 6C,G). Figure 6. Potassium induces helical reorganization of cortical microtubules of nhx1nhx2nhx3nhx4 epidermal cells in the root elongation zone. Immunolabeled microtubules in WT (top) and nhx1nhx2nhx3nhx4 (bottom). Seedlings were germinated on minimal media containing 1 mM K + for three days, then transferred to plates with two concentrations of Taxol and grown for an additional five days before fixation. Normal microtubule orientation is observed in WT (A) and nhx1nhx2nhx3nhx4 (B) when grown on media containing 1 mM K + . WT seedlings grown on 30 mM K + displayed normally-orientated microtubules (C), whereas 30 mM K + induced a helical orientation of the cortical microtubules in nhx1nhx2nhx3nhx4 (D). In the presence of 1 mM K + and 0.5 µM Taxol, both WT (E) and nhx1nhx2nhx3nhx4 (F) display characteristic bundling of cortical microtubules. Whereas 30 mM K + had no effect on microtubule orientation in WT grown with 0.5 µM Taxol (G), nhx1nhx2nhx3nhx4 seedlings grown with 30 mM K + and 0.5 µM Taxol (H) exhibited a helical microtubule organization similar to that seen when grown with just 30 mM K + . Arrows point to helically-oriented microtubules. The bar is 10 μM. The upper left insets are enlarged 50% to highlight microtubule orientation.
In particular, cortical microtubules depolymerize and reorganize in a Ca 2+ -dependent manner under salt stress; preventing this response with microtubule stabilizing drugs, such as Taxol, results in seedling death [12,13], suggesting that cytoskeletal reorganization of root cells is important during salt stress. In addition, mutations of the Na + /H + antiporter SOS1 and associated SOS2 kinase were shown to suppress the cortical microtubule disruptions and subsequent helical growth associated with the Arabidopsis spiral1 mutants. This study also demonstrated that even in the absence of salt stress, loss of function SOS mutants resulted in altered cortical microtubule arrays when seedlings were treated with the same dose of propyzamide as used in the present study [12].  [26]. All single knockouts were backcrossed twice prior to crosses for generating higher order knockouts. Subsequent crosses generated double knockouts and two triple knockouts, which were then crossed to generate a tetra nhx knockout. RT-PCR confirmed no DNA amplification with allele-specific primers from leaf-derived nhx1nhx2nhx3nhx4 cDNA. Positions of T-DNA insertion sites and RT-PCR results are shown in Figure A1.

Plant Materials and Growth Conditions
Arabidopsis thaliana (ecotype Columbia [Col-0]) and nhx1nhx2nhx3nhx4 seedlings were grown at 22 °C under diurnal light conditions (8 h light and 16 h dark), unless otherwise specified. Seeds were germinated on vertically-positioned plates containing a modified Murashige and Skoog media [27] (0.5% sucrose, 1 mM Na + , 1 mM K + , 0.8% phytagel, pH 5.7) and supplemented as specified for each experiment. For experiments testing stress responses, this media was considered as a control and was modified by the addition of 30 mM NaCl or KCl or 60 mM mannitol. For drug treatment assays, media was supplemented with 3 µM propyzamide (Sigma cat.#45645) or either 0.5 µM or 1 µM Taxol (Sigma cat.#T7191).

Transient Transformation of Arabidopsis Seedlings
Transient transformation of Arabidopsis was adopted from Li, et al. [31] using 4-5-day-old seedlings grown in a sterile 24-well plate. Seedlings were co-cultivated with Agrobacterium tumefasciens (strain GV3101) for 18-24 h before observation. Co-transformation using two constructs ( Figure 1I,L) was accomplished by co-cultivating seedlings with two bacterial cultures, each diluted to one half of the final optical density used when transforming a single construct (OD 0.5).

Hypocotyl and Root Length Measurements
Seedlings grown on agar plates were scanned on a flatbed scanner as color TIFF files with 300-dpi resolution. Root and hypocotyl lengths were quantified using the ImageJ [32] plugin SmartRoot [33]. The angles of root growth were measured with the same software suite using gridlines etched on the rear of the petri dishes as horizontal and vertical reference points (see Figure A4).

Quantification of Vacuolar K +
Vacuolar K + was quantified using the K + -specific dye PBFI-AM, as described previously [3].

Whole Mount Immunostaining of Arabidopsis Seedlings
Whole mount immunostaining of root cortical microtubules was modified from the protocol of Hauser et al. [34]. Dark-grown seedlings grown for five days after germination were fixed on agar plates in a solution of 4% p-formaldehyde, 5% DMSO, 0.025% glutaraldehyde in 1× microtubule stabilizing buffer (MTSB; 50 mM PIPES, 5 mM EGTA, 5 mM MGSO4, pH 7) vacuum infiltrated at room temperature for 1 h. Seedlings were then washed three times in wash buffer (MTSB with 5% DMSO) three times for ten minutes each, and the solution was exchanged for a cell wall digestion buffer (MTSB with 1% cellulase, 0.1% pectolyase) and incubated at room temperature for 30 min. Following digestion, three washes were performed with MTSB for five minutes each, and the seedlings were incubated for ten minutes in a solution of MTSB with 1 mg/mL NaBH4 for ten minutes. After three more washes with MTSB for five minutes, each seedling was incubated for ten minutes in a solution of MTSB containing 10 mM glycine and then washed once more with MTSB three times for a total of 30 min. Severed roots were transferred to microscope slides (Thermo-Scientific Superfrost plus, cat.#4951PLUS-001) and incubated in a membrane permeabilization solution (MTSB with 1% Triton X-100) at room temperature for 15 min and then in methanol at −20 °C for 20 min.

Confocal Microscopy
Fluorescence microscopy was performed using a confocal laser scanning microscope (Zeiss LSM 710) equipped with a 40× water immersion objective. The excitation wavelength was 488 nm, with emission wavelengths of 500-535 nm for GFP or FITC, 525-615 nm for YFP/Venus, 450-500 nm for CFP and 565-605 nm for RFP. For imaging of GFP/RFP and CFP/YFP double reporter lines, sequential scanning on multiple channels was used to avoid crosstalk between fluorescence emission spectra.

Conclusions
We previously reported that NHX1 and NHX2 colocalize to the vacuole and are essential for vacuolar K + accumulation and the regulation of pH [3]. Here, we demonstrate that both NHX3 and NHX4 are also localized to the tonoplast and that vacuolar cation/H + exchange by all four vacuolar NHXs are essential for normal growth and development. Compared to both WT and the nhx1nhx2 double knockout, the quadruple knockout nhx1nhx2nhx3nhx4 has significantly reduced vacuolar K + and growth, as well as high sensitivity to added K + . nhx1nhx2nhx3nhx4 also exhibited pronounced root skewing and curling responses not observed under equimolar Na + that were correlated with an increase in cytoskeletal dynamics and a reduction in the size of the root cells of the transition zone.   F). Root growth of nhx1nhx2nhx3nhx4 does not differ between treatments, but hypocotyls are shorter on 30 mM K + (D) than 1 mM K + , 1 mM Na + (B) or 30 mM Na + (F). Growth of both roots and hypocotyls in WT is unaffected across treatments. The yellow bar is 1 cm. Figure A3. nhx1nhx2nhx3nhx4 root tip curling is K + specific. Six-days-old WT (A,C,E,G) and nhx1nhx2nhx3nhx4 (B,D,F,H). Both WT and nhx1nhx2nhx3nhx4 exhibit straight root growth on media containing 1 mM KCl (A,B). When grown on media supplemented with either 30 mM KCl (C,D) or 30 mM KNO3 (E,F), WT roots grow normally, whereas roots of nhx1nhx2nhx3nhx4 exhibit a left-handed curling. Neither 30 mM NaCl (G,H) nor 60 mM mannitol (I,J) induced root curling, although the addition of NaCl produced longer root growth in nhx1nhx2nhx3nhx4 (H) compared with other treatments. The yellow bar is 1 cm. Figure A4. Root skewing measurement methodology. The etched grid on the back of square petri plates served as a reference point for assigning the vertical and horizontal vectors. Growth vectors were measured by drawing a line along the general axis of root growth. For curling roots, the growth vector was determined before the point at which the root begins to curl.