The Medicago truncatula HKT family: Ion transport properties and regulation of expression upon abiotic stresses and symbiosis

Soil salinity is one of the most important abiotic stresses affecting plant growth. In legumes, symbiotic nitrogen fixation in nodules is affected by salt stress, and salinity tolerance is variable among species. Genes from the High affinity K+ Transporter (HKT) family are known to play crucial roles in salt stress tolerance in different plant species. In legumes these transporters are still very poorly characterized.. Here we study the HKT transporter family from the model legume Medicago trunacatula, which is moderately tolerant to salinity. The genome of this species comprises five HKT genes, hereafter named MtHKT1;1 to MtHKT1;5. Phylogenetic analysis indicated that the MtHKT polypeptides belong to HKT subfamily 1. Three members (MtHKT1;2, MtHKT1;4 and MtHKT1;5) of the Medicago truncatula family were cloned and expressed in Xenopus oocytes. Their electrophysiological properties revealed a permeability 10 times higher for Na+ than for K+ and varying rectification properties. Expression analyses of the three MtHKT genes under different biotic and abiotic conditions suggested that MtHKT1;5 is the main transporter from this family in the root, the three genes sharing a decrease of expression in drought and salt stress conditions in non inoculated plants as well as plants inoculated with rhizobia. In the shoot, the three MtHKT would be present at similar levels independently on the applied stresses. Based on biomass and ion content analysis, the nodule appeared as the most sensitive organ to the applied salt and drought stresses. The level of expression of the three MtHKT genes was strongly decreased by both stresses in the nodule.

oocytes. Qualitative differences clearly appear between the 3 MtHKT curves. In the case of MtHKT1;5, the I-V curve is quasi linear in the whole range of recorded currents, inward or outward, as reported for the Arabidopsis AtHKT1 (Xue et al., 2011;Ali et al., 2016) and most other HKT subfamily 1 transporters characterized from dicots (Fairbairn et al., 2000;Almeida et al., 2014a) as well as from monocots (Suzuki et al., 2016). In contrast, in the case of MtHKT1;4, the shape of the I-V curve clearly shows that the ability of this transporter to mediate inward currents is stronger than its ability to mediate outward currents. In other words, this transporter behaves as an inwardly rectifying conductance. The shape of the MtHKT1;2 I-V curve indicates that this transporter too is endowed with inward rectification, but to a lesser extent than MtHKT1;4. Rectification has been rarely reported in HKT transporters so far (Jabnoune et al., 2009;Ali et al., 2016).
The magnitude of the currents mediated by the three MtHKT cannot be compared directly because the amounts of proteins expressed and functionally targeted to the cell membrane by the oocyte machinery could be different between the three transporters. displayed by Figures 3A, 3B and 3C, respectively. These figures allow to compare, for each MtHKT transporter, the relative positions of the I-V curves obtained in the different ionic conditions and the potentials at which these curves cross the x axis, i.e. the zero current potential, also named reversal potential (Erev). For each MtHKT transporter, Erev was very sensitive to the changes in Na + concentration, being strongly shifted towards more positive values when this concentration was increased. Such sensitivity to the concentration of a given ion demonstrates that the transporter is significantly permeable to this ion. Thus, the 3 MtHKT transporters are permeable to Na + . In contrast, the strong increase in K + external concentration, from 0.3 to 10 mM, poorly affected Erev in the 3 MtHKT transporters. The conclusion is thus that these transporters were not significantly permeable to K + in these ionic conditions, where both Na + and K + were simultaneously present in the external solution.
MtHKT permeability was then investigated using solutions containing a single monovalent cation, either K + , Na + or Li + , at a concentration of 10 mM. The corresponding I-V curves are displayed in Figures 4A, 4B and 4C for MtHKT1;2, MtHKT1;4 and MtHKT1;5, respectively. For each of these 3 transporters, Erev and the current magnitude were very similar in presence of K + and Li + . The current magnitude was larger and Erev was more positive in presence of Na + . These results indicated that K + and Li + were much less permeant than Na + in these ionic conditions. The shift in Erev when K + was replaced by Na + in the external solution was close to +59 mV, +51 mV and +58 mV for MtHKT1;2, MtHKT1;4 and MtHKT1;5, respectively (Figure 4). The "Goldman" equation allows to determine the ratio of the permeability of a transport system to a given ion X, PX, to its permeability to another ion Y, PY, from the magnitude of the shift in Erev when X is replaced by Y in the external solution (Dascal et al., 1984). The shifts in Erev determined from Figure 4 when K + is replaced by Na + led to the following values of the PNa to Pk permeability ratio: PNa/PK = 10.3 for MtHKT1;2, 7.5 for MtHKT1;4 and 9.9 for MtHKT1;5. Thus, the 3 MtHKT transporters were much more permeable to Na + than to K + these ionic conditions too. Figure 4B suggests that MtHKT1;4 is endowed with a distinctive functional property, not displayed by MtHKT1;2 and MtHKT1;5 in the same experimental conditions (Figures 4A and 4C). Indeed, the former Figure shows that, at membrane potential more positive than Erev, in other words when an outward current was observed, the magnitude of the current was more important in presence of 10 mM Na + (i.e., "against" 10 mM Na + ) than in presence of 10 mM K + or Li + . Based on the above results, it can be assumed that this current was mostly carried by Na + , the most permeant ion in these experimental conditions. Thus, the presence of a high concentration of Na + in the external solution would not reduce the efflux of this ion through MtHKT1;4 but instead would increase this efflux. The simplest hypothesis is that MtHKT1;4 displays a regulatory site present at its extracellular face and with which external Na + ions can interact and stimulate the transporter activity, i.e., its conductance. Within the framework of this hypothesis, the inward conductance of MtHKT1;4 too would display a similar positive regulation by external Na + .

Expression of MtHKT genes under different biotic and abiotic conditions
Expression of the 3 MtHKT genes in plants inoculated or not inoculated with rhizobia and grown in standard conditions or submitted to drought or salt stress was investigated by Q-RT-PCR ( Figure 5). When performed, inoculation was achieved by watering the plants with a solution containing rhizobia (S. Rhizobium) 7 days after germination, and the nitrogen source (urea) was then withdrawn from the standard (Fahräeus) nutrient solution. All the plants were 6-weekold when harvested. Plants submitted to drought stress were not watered during the last week of growth, a treatment that was observed to result in wilting symptoms at the end of this last growth week. Plants submitted to salt stress were watered with standard Fahräeus solution supplemented with 100 mM NaCl and thereafter 200 mM NaCl during the last growth week.
The wilting symptoms observed at the end of the drought treatment (not shown) and the large increase in plant Na + contents did not significantly affect the final biomass of the plant shoots, roots and nodules ( Figure 5A and 5B).
In shoots of non-inoculated plants, the 3 MtHKT genes were found to have similar levels of expression (transcript accumulation) and their expression was not significantly sensitive to the drought stress and salt stress ( Figure 5C, 5D and 5E). Inoculation tended to increase the level of expression of the three genes in shoots in control conditions but not upon stresses and did not modify significantly the plant sensitivity to the drought and salt stresses ( Figure 5A, 5B).
In the roots, the 3 genes differed in their expression levels: in absence of drought and salt stress, the MtHKT1;5 transcript appeared as the most abundant and the MtHKT1;4 transcript as the less abundant, whether the roots were inoculated or not. The drought stress in the case of MtHKT1;2, and both the drought and salt stresses in the case of MtHKT1;4, significantly reduced the level of expression of these genes in roots of non inoculated plants.
In inoculated plants, root expression of the 3 genes was reduced by each of the two stresses ( Figure 5C, 5D and 5E).
In nodules, the expression levels of the 3 MtHKT genes and their variations in response to the drought and salt stresses were similar to those observed in inoculated roots, i.e., in the absence as well as in the presence of drought and salt stress, MtHKT1;5 displayed the highest levels of transcripts, and MtHKT1;2 the lowest levels, and the levels of transcripts of each of the 3 genes were significantly reduced by each of the two stresses, drought or salinity ( Figure   5C, 5D and 5E).

Discussion
Phylogenetic analyses have identified 2 subfamilies within plant HKT transporters (Platten et al., 2006), and functional analyses have provided evidence that these 2 subfamilies differ in their K + and/or Na + transport ability. The distinctive property of subfamily 1 members is to be selectively permeable to Na + , when compared to K + , while subfamily 2 members are significantly permeable to both Na + and K + and have the ability to behave as Na + -K + symporters at least when heterologously expressed in animal cells (Rubio et al., 1995;Uozumi et al., 2000;Oomen et al., 2012;Sassi et al., 2012). It should however be noted that some HKT members from subfamily 1 could display a physiologically significant permeability to K + when heterologously expressed in yeast (Ali et al., 2012). Subfamily 1 is present in both dicots and monocots, while members of subfamily 2 have been identified only in monocots so far. Another difference with respect to the HKT family between dicots and monocots can be the family size, which appears to be smaller in dicots. Indeed, when the first genome sequence became available, it appeared that this family comprised a single member in Arabidopsis or poplar, or two members in tomato (Asins et al., 2013), while 8 or 9 HKT genes, depending on the cultivar, could be identified in rice, subfamilies 1 and 2 having a similar number of members, 4 or 5.
Also, 10 or so HKT genes are present in each wheat genome. It should however be noted that recent genome sequences do not support such a dichotomic view since 5 HKT genes are present in the dicot Vitis vinifera (grape vine).
The number of HKT genes seem to be rather variable in legumes. As shown in the present report, M. truncatula possesses 5 HKT (including one pseudo-gene), which are present within a cluster in chromosome 6. In silico analysis (not shown) indicates that the Lotus japonicus genome harbors 5 HKT genes, forming 2 clusters, the first one located in chromosome 5 and comprising 2 genes, (Lj5g3v0196960 and Lj5g3v0196990), and the second one located in chromosome 2 and comprising 3 genes (Lj2g3v0914290, Lj2g3v0914260 and Lj2g3v0914190) (Sato et al., 2008). In soybean (Glycine max), the HKT family displays 4 members, at four different locations in the genome (Glyma.01g002300, Glyma.06g271600, Glyma.07g130100 and Glyma.12g133400) (Schmutz et al., 2010). In contrast, common bean (Phaseolus vulgaris) has only two HKT genes, (Phvul.004G177200, Phvul.004G177300), which form a cluster (Schmutz et al., 2014). The Chickpea (Cicer arietinum) possesses one or two HKT genes depending of the sequence database (Varshney et al., 2012;Varshney et al., 2013). Altogether, these data indicate that the size of this transporter family in dicots is not as reduced as initially thought and can be rather variable even within a same family.
Based on phylogenetic criteria, the 4 HKT transporters identified in M. truncatula belong to HKT subfamily 1. Three from them, whose transcripts could be easily detectable in shoots and roots, have been functionally characterized by electrophysiological analyses in Xenopus oocytes. They all appear as typical members from subfamily 1, being much more permeable to Na + than to K + , which is consistent with the hypothesis that only HKT subfamily 1 is present in dicots and that the members of this subfamily display strong selectivity for Na + against K + .
In HKT subfamily 1, a serine residue located in the pore loop of the first MPM domain has been MtHKT1;4 appears to be endowed with another original property, besides rectification.
This transporter seems to be activated by external Na + , as suggested by Figure 4B. Such activation, together with the inward rectification, could render this transporter especially dedicated to Na + uptake from high Na + concentrations (e.g., in Na + retrieval from the xylem sap?). Conversely, the absence of strong rectification in MtHKT1;2 could allow this transporter to be more specifically involved in Na + secretion (e.g., in Na + release from the phloem vasculature into the root stele and thereby in Na + recirculation from shoots to roots?).
The roles of HKT transporters is probably more documented in monocots than in dicots due to the strong interest that these transporters have for the cereal breeders interested in crop adaptation to soil salinity. Indeed, evidence has been clearly provided that members from HKT subfamily 1 play major roles in the control of Na + shoot contents, and thereby in salt tolerance, in agronomically important cereals such as rice and wheat. For instance, a major salt tolerance QTL in rice, named SKC1 (shoot K + content) which contains the subfamily 1 transporter OsHKT1;5 (also named OsHKT8) (Lin et al., 2004;Ren et al., 2005), is presently used by plant breeders for increasing salt tolerance in rice cultivars (Ashraf et al., 2012). Major QTL of salt tolerance have been shown to correspond to HKT genes in wheat and maize too (Horie et al., 2009;Zhang et al., 2017). In dicots, the most detailed information is available for the Arabidopsis AtHKT1, which has been shown to be expressed in plant vasculature and to play a major role in salt tolerance by controlling Na + translocation from roots to shoots via the xylem sap and/or Na + recirculation from shoots to roots via the phloem sap (Berthomieu et al., 2003;Horie et al., 2009). Contribution to control of Na + long distance transport in the plant vasculature has also been reported for SlHKT1;2 in tomato (Almeida et al., 2014b), whose gene is located, together with SlHKT1;1, in the major tomato QTL involved in Na + /K + homeostasis (Asins et al., 2013). In legumes, information on the role of an HKT transporter has been obtained for GmHKT1 from soybean: overexpression of its encoding gene enhances    Oocytes were injected with 50 nl of either water (control oocytes; left panels) or MtHKT cRNA solution (right panels). Currents (Means ± SD) were recorded in solutions differing in their Na + and K + concentrations, so that the concentration of K + was constant (0.3 mM) and that of Na + varied (0.3, 3 or 10 mM), or the concentration of Na + was constant (0.3 mM) and that of   Plants were grown for 6 weeks in total in vermiculite/sand soil. They were inoculated or not inoculated with rhizobia at the end of the first week, and submitted or not submitted (control plants) to either salinity or drought stress during the last growth week. For salinity stress, they were watered with 100 mM NaCl the first day of the 6 th week, and with 200 mM NaCl two days later. For drought stress, they were not watered during the last 7 days. Roots, shoots and nodules (in inoculated plants) were then harvested for biomass measurements, Na + content assays and Q-RT-PCR analysis of MtHKT gene expression. Left panels: non inoculated plants. Right panels: inoculated plants. Light grey, white and dark grey bars: control, salinity and drought treatments. Means ± SD of at least three biological replicates. Statistical comparison was performed within each kind of organs, roots, shoots or nodules, using Turkey's test. Different letters indicate statistically significant differences between the results from the control, salinity and drought treatments at the level of 5% (Tukey's test). Normal, bold and italic letters: comparison of the response in shoot, root and nodules, respectively. A : Analysis of dry weight (mg) of non inoculated and inoculated plants.