Potassium (K⁺) gradients serve as a mobile energy source in plant vascular tissues

To obtain a better understanding of AKT2 function, we tested the effect of loss of AKT2 function in plants [akt2-1 knockout plants (18)] (Fig. S2) grown at different day lengths. Upon cultivation in normal soil in a 16-h day/8-h night cycle (in the greenhouse or in growth chambers), no phenotypic differences were detectable between the akt2-1 mutant and the wild type at any developmental stage (Fig. 1A). However, when the day length was reduced to 12 or 8 h, the plants displayed pronounced phenotypic differences. Compared with the wild type, the akt2-1 line developed fewer leaves and showed a delay in bolting (Fig. 1 B and C). The wild-type phenotype was restored when the akt2-1 line was complemented with the wild-type AKT2 allele, indicating that the phenotypic effects are indeed due to a disruption of the AKT2 locus (Fig. S2; Fig. S3, A–C).

The akt2-1 line was also rescued when AKT2 was expressed under the control of the phloem companion cell-specific AtSUC2 promoter (19) (Fig. S3, D and E). Thus, the observed phenotype of the akt2-1 plants can be attributed to a loss of AKT2 function in phloem tissues rather than, for example, in guard cells.

We next wanted to test whether the unique phosphorylation-dependent characteristics of AKT2 are important for its proposed physiological role. To do this, we complemented the akt2-1 knockout plant with mutant versions of the AKT2 channel protein that had been previously identified as being affected in posttranslational modifications (3, 4) (Fig. S4) (20): Whereas the mutant AKT2-S210N-S329N can be more easily converted into a nonrectifying channel than the wild type, the mutants AKT2-S210A-S329A and AKT2-K197S cannot be converted into a nonrectifying channel, so they function as inward-rectifying channels only.

The different mutant alleles of the channel had distinct effects on the physiology of the plants. Plants expressing inward-rectifying mutant AKT2 channels (AKT2-S210A-S329A or AKT2-K197S) behaved very similarly to the akt2-1 knockout. They produced fewer leaves and showed a delayed development under short-day conditions. In contrast, plants expressing the mutant AKT2-S210N-S329N were, like wild-type plants, not negatively affected by the applied stress (Fig. 2).

Our wet-lab experimental data indicate that the importance of AKT2 in phloem tissues is tightly connected to its ability to be converted into a nonrectifying channel. To investigate further the role of this posttranslational modification in phloem transport, we carried out dry-lab experiments. Previous studies have proposed a very detailed model for a sieve element/companion cell (SE/CC) complex that expresses a transporter network of H⁺-ATPase, sucrose/H⁺ carriers, and AKT2-like K⁺ channels (5, 21). We took this model as the basis and also included the leakage of sucrose from the phloem (e.g., by pH-independent facilitators) (22), as well as the sequestration of K⁺ from the apoplast (Fig. 3A) to better reflect the in vivo situation. The features of all transporters were described mathematically (Fig. S5) (21, 23–25) and computational simulations were carried out using Virtual Cell software.

First inspection of the thermodynamical flexibility of the ohmic network revealed that it can act in three different states with respect to values of the membrane voltage V_m. In only one of them (E_K < V_m < E_H/Suc) did regulation of AKT2 have a clear physiological impact (SI Text). To achieve such an intermediate membrane voltage, the activity of the H⁺-ATPase needs to be reduced compared with that of the other transporters. We therefore started our detailed analyses with an intermediate H⁺-ATPase activity and allowed the system to equilibrate when all AKT2 channels were inward-rectifying (Fig. 3 B–F, i). Under this condition, the apoplastic sucrose concentration remained stable (Fig. 3B, i), and there was no net transport of sucrose across the membrane (Fig. 3C, i). This indicates that all sucrose molecules leaking from the phloem were retaken up by the sucrose/H⁺ carrier, fueled by the H⁺-ATPase (Fig. 3E, i). Next, AKT2 was switched from an inward-rectifying into a nonrectifying channel (Fig. 3 B–F, arrows). This change resulted in a rapid short-term net uptake of sucrose (Fig. 3C, ii), causing a decrease of the apoplastic sucrose concentration (Fig. 3B, ii). The energy driving this additional sucrose uptake was not provided by the H⁺-ATPase; the switch of AKT2 caused a small hyperpolarization (Fig. 3D, ii) that, in turn, actually reduced the H⁺-ATPase activity (Fig. 3E, ii). Instead, the energy for the additional sucrose uptake was provided by the electrochemical K⁺ gradient (Fig. 3F, ii). The newly opened K⁺ channels allowed a rapid flux of K⁺ along the outwardly directed electrochemical gradient. This charge transport was largely compensated by a rapidly increased H⁺/sucrose influx (Fig. 3C, ii). The fraction that is not compensated affects the membrane voltage, causing the slight hyperpolarization (Fig. 3D, ii). Thus, our mathematical simulations reveal that K⁺ serves as an energy carrier for phloem (re)loading processes and that posttranslational regulation of AKT2 taps this potassium battery.

Our dry-lab experimental data indicate that regulation of AKT2 has a clear physiological impact on phloem reloading when the activity of the H⁺-ATPase is reduced, especially when the local energy provided by the H⁺-ATPase is no longer sufficient for K⁺ loading. Consequently, the observed plant phenotype (akt2-1 plants are affected under short-day but not under long-day conditions) may be interpreted as resulting from the different energy status of the plants in the two tested conditions. Indeed, the light dependence of the akt2-1 phenotype was not due to altered photosynthetic performance, as shown by measurements of electron transport rates based on chlorophyll fluorescence (Fig. S6). Interestingly, phenotypic differences between the akt2-1 mutant and the wild type could also be induced when, instead of day length, light intensity was reduced. Whereas under normal light conditions (100 μE·m⁻²·s⁻¹ for 16 h), knockout and wild-type plants were indistinguishable, a reduction of light intensity to 15 μE·m⁻²·s⁻¹ (low light) was less well tolerated by the akt2-1 mutant. The knockout plants developed fewer leaves and siliques, and the bolting time was not affected (Fig. S7).

To challenge the “energy status hypothesis” independently from light conditions, we tested plant growth at different atmospheric oxygen concentrations. With decreasing oxygen availability, the respiratory energy (ATP) production declines. This has severe effects on highly metabolically active tissues such as the phloem (26). Wild-type and akt2-1 plants did not differ under normal atmospheric conditions (21% O₂, 16 h day, 100 μE·m⁻²·s⁻¹). However, when the oxygen concentration was reduced to 10%, the knockout plants developed a similar phenotype to that observed under short-day conditions: they had fewer leaves, exhibited a delayed bolting time, and produced less biomass (fresh weight and dry weight) (Fig. 4). Here, also, the importance of AKT2 could be correlated to its convertibility into a nonrectifying channel. Similar to the akt2-1 knockout plants, akt2-1 plants expressing inward-rectifying mutant AKT2 channels (AKT2-S210A-S329A or AKT2-K197S) were affected by the O₂ reduction, whereas plants expressing the mutant AKT2-S210N-S329N were not significantly affected (Fig. 4). Thus, even without knowing the underlying signaling cascade, this experiment shows that Arabidopsis reacts on low O₂ concentrations by posttranslational control of the AKT2 channel.

Encouraged by the congruence between dry-lab predictions and wet-lab observations, we simulated another scenario that allowed us to also expose long-term effects of AKT2 channel regulation on phloem loading (Fig. 5; Fig. S8). As in the first scenario, we started again with an equilibrated system with only inward-rectifying AKT2 channels (Fig. 5, i). Then the activity of the H⁺-ATPase was slightly reduced (by as little as 10%) to simulate, for example, varying ATP levels in the phloem or varying H⁺-ATPase regulation (Fig. 5, arrow 1), and the system was again allowed to relax (Fig. 5, ii). The reduction in H⁺-ATPase activity resulted in an increase of the apoplastic sucrose concentration. This indicates a net sucrose leakage from the phloem that lasted until a new equilibrium between efflux and influx at a higher apoplastic sucrose concentration had been reached. Following this, AKT2 was switched from an inward-rectifying into a nonrectifying channel (Fig. 5, arrow 2). This change resulted in a rapid short-term net uptake of sucrose (Fig. 5, iii), which, interestingly, was followed by a long-term stimulatory effect on sucrose uptake. This became more apparent when rerunning the simulation with a different sequestration rate of K⁺ from the apoplast (red and blue curves in Fig. 5). This rate had no influence on phloem loading when the AKT2 channels were inward-rectifying (Fig. 5, i and ii). However, when the channels were posttranslationally modified into nonrectifiers, the K⁺-sequestration rate had a determining influence on the apoplastic sucrose concentration. A higher K⁺-sequestration rate resulted in an increased long-term loading of sucrose into the phloem (Fig. 5, iv). Also here, switching the operation mode of AKT2 channels taps a potassium battery, thus providing a rapidly available, albeit rapidly exhausted, energy source for phloem loading. Thereafter, surrounding cells can stimulate the (re)loading of sucrose into the phloem by sequestering K⁺ from the apoplast, thus restoring a chemical K⁺ gradient between the phloem and apoplast (Fig. S8). In this way, the energy needed for the loading process is supplied in a decentralized manner by the K⁺ ions pumped from source tissues into the phloem sap and flowing with it and by the surrounding cells that invest energy (ATP) to take up K⁺ from the apoplast for their own use.

In addition to tapping a potassium battery, the simulated scenario also points toward a role for AKT2 in K⁺ supply of growing tissues (SI Text, ii). We therefore tested the dependence of akt2-1 and wild-type plants on K⁺ availability in early development. Indeed, compared with the wild type, akt2-1 mutants showed decreased root growth, lower total fresh weight, and a decreased root-to-shoot biomass ratio when sufficient K⁺ was in the medium (2,500 μM) (Fig. 6A). Surprisingly, in contrast to observations on knockout plants of inward-rectifying K⁺ channels in roots (27–30), this phenotypical difference between wild type and akt2-1 knockout disappeared when we lowered the external K⁺ concentration. Under K⁺-deficient conditions (10 μM; 100 μM), root growth, fresh weight, and root-to-shoot ratio were almost identical in mutant and wild-type plants (Fig. 6 B and C). Thus, although both wild type and akt2-1 performed better in high K⁺ than in low K⁺, the knockout plant was not able to use the additional resources as efficiently as the wild type. Controlling the potassium battery and K⁺ supply therefore represents two sides of the physiological role of AKT2.

Molecular biology methods were performed according to standard procedures. Further details are provided in the legend of Fig. S2 (31–33).

Seeds were cold-treated for 24 h at 4 °C. Plants were grown on a standardized soil substrate (type GS-90, Einheitserde). After 2 wk, seedlings were transferred to individual pots. For all experiments, plants were grown in 60% relative humidity at 21 °C during the day and 18 °C during the night. For analyses under different light regimes, plants were grown in three different diurnal cycles with day lengths of 8, 12, or 16 h. The response to different light intensities was observed at 100 μE·m⁻²·s⁻¹ and 15 μE·m⁻²·s⁻¹ (16 h day/8 h night). The illumination system was composed of Philips TLD58W840 or Philips HPI-T+400W/645 and Philips Master Son-T PIA Agro 400-W bulbs. All phenotypical analyses were carried out in the middle of the day. For experiments under different oxygen concentrations, 2-wk-old plants were transferred into plastic containers, sealed, and supplied with a gas mixture containing 21% and 10% oxygen, respectively. The experiment was performed under standard light conditions (16 h day/8 h night). For dry weight measurements, shoots were dried for 24 h at 60 °C. Root-to-shoot ratios were calculated on the basis of fresh weight measurements.

Sterile seeds were plated on Arabidopsis medium (AM) containing 50% Murashige and Skoog salts supplemented with 1% sucrose and solidified with 1% agar. Seven-day-old seedlings were transferred to a synthetic medium [2.5 mM NaNO₃, 2.5 mM Ca(NO₃)₂, 2 mM NH₄(H₂PO₄), 2 mM MgSO₄, 0.1 mM FeNaEDTA, 25 μM CaCl₂, 25 μM H₃BO₃, 2 μM ZnSO₄, 2 μM MnSO₄, 0.5 μM CuSO₄, 0.2 μM Na₂MoO₄, 0.01 μM CoCl₂, 0.5% sucrose, pH 5.7, solidified with 1% agar] with 0.01, 0.1, or 2.5 mM KCl. Plants were vertically grown under standard light conditions (16-h day/8-h night; 100–150 μE·m⁻²·s⁻¹).

The behavior of the transporter networks was mathematically simulated using Virtual Cell Modeling and Analysis Software developed by the National Resource for Cell Analysis and Modeling, University of Connecticut Health Center. Biophysical properties of the transporters were mathematically modeled to reflect their measured in vivo characteristics. Details are outlined in the legend of Fig. S5.