Phosphorylation state of S280 and S283 influences AtPIP2;1 localisation in yeast
Mimicking changes in C-terminal phosphorylation states of AtPIP2;1 not only altered Na+ and H2O conductance in X. laevis oocytes (Figure 2 and 3), but also influenced the sub-cellular localisation of AtPIP2;1 in S.cerevisiaeaqy1/aqy2 double mutant yeast strain (Figure 5). Sub-cellular localisation tendencies of the AtPIP2;1 phospho-mutants were monitored in yeast using both N- or C-terminal GFP fusions. Fusion of AtPIP2;1 wild-type or AtPIP2;1 phospho-mutants to GFP, redirected GFP from a diffuse cytosolic pattern (Figure 5A) to a predominantly sharp ring around the cell perimeter coinciding with the plasma membrane (PM; Figure 5C-I). Weaker GFP signal associated with the tonoplast of the vacuole was also frequently observed. In addition, a proportion of cells showed internal and patchy perimeter GFP signal, matching the localisation pattern of the SEC63::RFP endoplasmic reticulum (ER) marker (Figure 5B). The detectable frequency and intensity of the ER co-localisation differed between the AtPIP2;1 wild-type and some of the AtPIP2;1 280/283 phospho-mimic mutants (Figure 5J). A difference was observed for the AtPIP2;1 S280D mutant, which had frequently occurring intense GFP signal co-localising to the ER (Figure 5D, E and J), relative to the trend observed for the AtPIP2;1 S283D and AtPIP2;1 D/D mutants, which had distinct GFP signal around the perimeter consistent with PM localisation, and less frequent or intense ER co-localisation (Figure 5F, G, H and J). The tendency of the AtPIP2;1 S280A and A/D mutants to co-localise to the ER was more likely than for wild-type (Figure 5I and J). GFP localisation patterns for the phospho-deficient AtPIP2;1 S283A and A/A mutants along with the D/A mutant were equivalent to that of wild-type AtPIP2;1 (Figure 5J). There was no discernible difference in the localisation patterning whether the GFP was fused to the N- or C- terminal (data not shown).
Comparisons between the phospho-mutants reveal co-ordinated effects of positions S280 and S283 in determining sub-cellular localisation. The more prominent PM targeting of the AtPIP2;1 D/D mutant in comparison to the A/A, D/A, or A/D mutants, indicates that mimicking of phosphorylation of both S280 and S283 was required to promote more distinctly PM localisation (Figure 5J). The distinct ER co-localisation of AtPIP2;1 S280D was not observed in either of the two other S280D phospho-mimic mutants (D/A or D/D) (Figure 5J), indicating that a serine at position 283 could be specifically required in combination with the phospho-mimic aspartic acid at position 280 to achieve the distinct ER co-localisation.

Discussion

AtPIP2;1 is able to facilitate water and monovalent cation transport activity in vivo , and this function is influenced by the phosphorylation states of CTD residues S280 and S283 (Figures 1-4). InX. laevis oocytes, S280 and S283 phosphorylation status influenced the transport function of AtPIP2;1, such that relatively greater water transport occurred when these sites mimicked an un-phosphorylated state and greater ion transport occurred when they mimicked a phosphorylated state (Figure 2 and 3). A phosphorylation-dependent inverse relationship was observed for AtPIP2;1-facilitated water transport relative to ion transport (Figure 3 and Figure S3, 4), where there was approximately ten-fold changes in both permeabilities (Figure S3). Yeast expressing the different AtPIP2;1 S280 and S283 mutation combinations accumulated different amounts of intracellular Na+ following incubation in a buffer containing 70 mM NaCl (Figure 4). S280 and S283 phosphorylation status also influences the distribution and abundance of AtPIP2;1 protein localising between the ER and PM in yeast (Figure 5).