Phosphorylation status of C-terminal S280 and S283 sites regulate
AtPIP2;1 facilitated water and ion transport
The influence of mutating AtPIP2;1 S280 and S283 sites to phospho-mimic
and phospho-deficient versions on channel water and ion transport
function was tested in oocytes (Figures 2-3, Figure S4). Oocytes
expressing the single phospho-mimic versions of AtPIP2;1, S280D and
S283D or the phospho-mimic double mutants A/D, D/A and D/D had greater
ionic conductance in solutions containing Na+ and
K+ and increased internal Na+content compared to oocytes expressing either AtPIP2;1 WT or the single
phospho-deficient variants S280A and S283A or the double
phospho-deficient mutant A/A (Figure 2). Oocytes expressing the single
mutant S280A and double mutant A/A had greater Pos than
the other versions (Figure S4). When the phosphorylation status of both
CTD sites were controlled but only one site mimicked a phosphorylated
state the effect of the phospho-mimicked residue presided somewhat over
the effect of the phosphor-deficient site; for example, both the D/A and
A/D phospho-mutants had water and ion transport more similar to that of
the D/D mutants than the A/A mutants (Figure 2 and 3). Previously the
Rattus AQP2 was reported to be phosphorylated at two serines on its CTD
in renal epithelial cells in response to vasopressin, and the effect of
phosphorylation of one residue presided over the channel function and
trafficking (Table 1) (Lu et al., 2008). The phosphorylation of
several CTD residues in Rattus AQP2 also exhibit a hierarchy where the
phosphorylation of particular residues does not occur unless the
phosphorylation of another site has preceded it (Hoffert et al.,2008).
AtPIP2;1 facilitated the transport of K+ and the
single phospho-mimic mutants conferred greater K+associated conductance than the other versions, similar to the trend
observed for phospho-mimic versions for Na+conductance (Figure 2). AtPIP2;1 and AtPIP2;2 have been proposed as
molecular candidates for the elusive non-selective cation channel that
have been observed in planta (Byrt et al., 2017; McGaugheyet al., 2018; Munns et al., 2019; Demidchik and Tester,
2002; Essah et al., 2003; Roberts and Tester, 1997). The
observation that AtPIP2;1 can facilitate K+ transport
in vivo adds support to this hypothesis. The NSCCs observed by
Demidchik and Tester, (2002), had greater K+conductance relative to Na+ conductance (with a
selectivity ratio of 1.49:1.00), which is similar to the trend in
K+ relative to Na+ conductance for
AtPIP2;1 when expressed in X. laevis oocytes (Figure 2c). The
regulation of AtPIP2;1 ion transport by cGMP treatments (Figure 1) is
also relevant to previous NSCC observations, because exogenous
application of cGMP was previously shown to inhibit NSCCs in
planta (Essah et al., 2003; Maathuis and Sanders, 2001), and
intracellular cGMP concentrations have been reported to increase in
response to salinity and osmotic stress treatments (Donaldson et
al., 2004; Rubio et al., 2003). Interestingly, a recent review
hypothesised that Na+ influx via AtPIP2;1 may be
inhibited by cGMP under salt stress, which is an idea worthy of follow
up investigation in planta (Isayenkov and Maathuis, 2019). The
observation that AtPIP2;1 facilitates transport of the physiologically
important element K+, and the potential for AtPIP2;1
transport of other monovalent ions such as
NH4+, indicates that a potential role
for PIPs in nutrient acquisition under normal conditions is also worthy
of testing in planta .
A negative correlation between AtPIP2;1 facilitated water and ion
transport was linked to the CTD phosphorylation state (Figure 3a).
AtPIP2;1 mutants including S280D, S283D, D/A and D/D, had a greater
tendency to facilitate the transport of ions over water compared to that
of the phosphorylation deficient mutant A/A (Figure 3b, d). The variance
seen for the ionic conductance and Pos of the D/D mutant
indicates that there are likely to be other additional regulatory sites
that were not controlled for in these experiments. Further research is
needed to test how many other AtPIP2;1 regulatory sites influence water
and ion transport functions and explore whether these sites have any
sort of dependence on the status of the CTD sites.
Several General Regulatory Factors (GRFs; also known as 14-3-3 proteins)
were recently reported to interact preferentially with AtPIP2;1 when the
S280 and S283 sites were phosphorylated, and co-expression of AtPIP2;1
D/D mutant with GRFs 3,4 and 10 in oocytes increased their
Pos compared to AtPIP2;1 A/A (Prado et al.,2019). In the current study it cannot be excluded that AtPIP2;1 could
have interacted with an endogenous oocyte GRF protein, or an endogenous
aquaporin interacting ion channel. However, the trends observed for
AtPIP2;1 CTD status and associated ionic conductance do not appear
common to all aquaporins. There are commonalities for CTD
phosphorylation trends among some Arabidopsis PIPs, but not all PIPs
with these commonalities confer ionic conductance in oocytes. For
example, in Arabidopsis AtPIP2;1, AtPIP2;2, AtPIP2;3, AtPIP2;4 and
AtPIP2;7 were found to be unphosphorylated, singly phosphorylated at
S280 or diphosphorylated at S280 and S283 (Prak et al., 2008),
but AtPIP2;7 did not facilitate ion transport when expressed in oocytes
(Kourghi et al., 2017) as confirmed here also from a lack of
Na+ uptake into yeast expressing AtPIP2;7. There is
also a precedent for plant aquaporins having ion channel functions in
the absence of any potential interacting partners, and associations with
the CTD status. The soybean (Glycine max ) Gm-NOD26, produced ion
channel activity when reconstituted in lipid bilayers (Weaver et
al., 1994). The water and ion channel function of Gm-NOD26 was also
found to be regulated by the phosphorylation of a CTD residue S262
(Guenther et al., 2003; Lee et al., 1995).
The exact physiological role of dual water and ion transporting
aquaporins in plants remains unknown and may differ in different tissues
(McGaughey et al., 2018). When Arabidopsis roots were exposed to
a NaCl treatment the phosphorylation states of AtPIP2;1 S280 and S283
residues was observed to change (Prak et al., 2008).
Specifically, when plants were treated with 100 mM NaCl the abundance of
the S280/S283 disphosphorylated form decreased. Since phosphorylation of
S280 and S283 increase AtPIP2;1 ion channel function, this reduction in
S280/S283 diphosphorylated AtPIP2;1 may be a mechanism to reduce
Na+ influx (and possibly K+ efflux)
under salt stress. Salt treatment has also been reported to increase
AtPIP2;1 location-cycling (Li et al., 2011; Luu et al.,2012), and induce AtPIP2;1 internalisation from the plasma-membrane into
intra-cellular vesicles in root cells (Boursiac et al., 2005;
Prak et al., 2008; Ueda et al., 2016) where
internalisation was reported to be dependent on S283 phosphorylation
state (Prak et al., 2008). By manipulating the phosphorylation
state of the AtPIP2;1 CTD serine residues, we were also able to alter
trafficking and abundance of the AtPIP2;1 protein between the PM and ER
in yeast (Figure 5). In the yeast system we found that the phospho-mimic
S280D mutation resulted in a more consistent localisation of AtPIP2;1 to
the ER rather than trafficking to the PM. This feature specifically
required the presence of a serine residue at position 283 and could not
be replicated by mimicking a phospho-deficient state using alanine. We
also observed that consistency in PM targeting was not only dependent on
S283 phosphorylation, but also required dual phosphorylation of both
S280 and S283, with the presence of a phospho-mimic residue at position
283 potentially influencing the phosphorylation state of S280.
Interestingly the localisation of the double phospho-deficient mutant
A/A was similar to other phospho-mimic mutants, such as D/A and A/D
(Figure 5), where these other mutant versions exhibited much greater
ionic conductance than A/A when expressed in X. laevis oocytes
(Figure 2). This data, alongside the increased Pos of
A/A relative to the other mutants (Figure 3), indicates that a
mis-localisation of the A/A mutant in oocytes is not likely to be the
cause of its lower ionic conductance. Furthermore, the fact we could
make these observations in yeast using an aquaporin from the distant
taxa of plants, indicates the potential for there to have been a shared
evolutionary origin for the process of CTD phosphorylation influencing
aquaporin trafficking.
The sophisticated relationship between AtPIP2;1 phosphorylation state
and AtPIP2;1 trafficking, localisation, and water and ion transport
function may be part of a mechanism for rapidly, reversibly and
co-ordinately adjusting water and Na+ or
K+ flux into or out of the cell under salt and osmotic
stress.