PIN phosphorylation is sufficient to mediate PIN polarity and direct auxin transport

PIN1 Phosphorylation at Ser337/Thr340 Is Required for Auxin-Related Development.

A putative phosphorylation site of PIN1 had been isolated by mass spectrometry at Ser337 and/or Thr340 in the central hydrophilic loop of the PIN1 coding sequence (22). These Ser and Thr of the phosphorylation site are conserved in all plasma membrane-localized PIN proteins in Arabidopsis thaliana (Fig. 1A) and other species (Fig. S1) when compared to the endoplasmic reticulum-localized subfamily of PIN proteins (24). To test the involvement of the phosphorylation site in the polar PIN targeting in vivo, site-directed mutagenesis of the conserved residues within the PIN1 sequence was carried out (Fig. 1A). Ser and Thr were both converted into Ala, which is a nonphosphorylatable residue, and to Asp, which mimics phosphorylation. The PIN1 genes, fused to the green fluorescent protein (GFP) (PIN1::PIN1-GFP) and the hemagglutinin (HA) (PIN2::PIN1-HA), were mutagenized and the constructs were transformed into the wild type and the pin1 and pin2 mutants, respectively. PIN1-GFP(Ala) and PIN1-GFP(Asp) partially rescued a shoot phenotype of the pin1 mutant, but the rescued lines displayed various developmental defects in adult plants. The independent transgenic lines PIN1-GFP(Ala) (5/7) as well as PIN1-GFP(Asp) (7/9) caused defective phyllotaxis and floral morphology, discernible by fused flowers with two pistils, outgrowth of two siliques from the same position and nondeveloped pistils (Fig. 1B). The same range of the phenotypes (Fig. 1B) with comparable frequencies [PIN1-GFP(Ala) (6/8) and PIN1-GFP(Asp) (6/7)] was observed when PIN1-GFP(Ala) and PIN1-GFP(Asp) were transformed into the wild type, demonstrating the dominant effect of the mutations. In contrast, whereas PIN1-GFP(Asp) rescued to a large extent the shoot development, it completely failed to complement embryonic and cotyledon phenotypes of pin1; the transformants still showed a defective embryo development, reflected by monocots, tricots, or fused cotyledons at the seedling stage (Fig. 1C). To investigate why the PIN1-GFP(Asp) could not rescue the embryonic phenotype, we examined the subcellular localization of PIN1-GFP(Asp) during embryogenesis. Consistently, defects in the polar PIN1 localization were observed in triangular-stage embryos (n = 62), predominantly manifested by the loss of the basal polar localization facing the root pole in the inner embryo cells and its replacement by an apolar or preferentially apical localization (Fig. 1D). These results reveal that the PIN1 phosphorylation at the site involving Ser337/Thr340 is important for the PIN1 function, particularly during flower and embryo development. Moreover, the flower defects in the wild-type plants highlight the dominant effect of these mutations.

Phosphomimic PIN1-HA(Asp) Can Functionally Replace PIN2.

We tested the effects of PIN1-HA(Ala) and PIN1-HA(Asp) in the PIN2 expression domain (PIN2::PIN1-HA constructs) in the pin2 mutant background. PIN2 is a key regulator of root gravitropism; it localizes to the apical side of epidermis cells and, after gravistimulation, it mediates the auxin flow along the lower side of the root from the tip toward the elongation zone where the auxin response is activated and inhibits elongation (25–27). The expression of PIN1-HA in pin2 mutants was not sufficient to restore either the gravitropic response or the asymmetric auxin accumulation as monitored by the GFP-fused artificial auxin-responsive reporter DR5rev::GFP (9)(Fig. 2 A and B). PIN1-HA(Ala) slightly enhanced the agravitropic phenotype of the pin2 mutant (n = 45) and did not rescue the activation of the DR5-visualized auxin response at the lower side of roots (n = 25) (Fig. 2 A and B). In contrast, PIN1-HA(Asp) rescued to a large extent the agravitropic phenotype of the pin2 mutant (n = 39) and also restored the DR5-visualized auxin response along the lower side of the root (n = 30) during gravitropic responses (Fig. 2 A and B). These results reveal that mutation of the phosphorylation site in PIN1 is sufficient to acquire a PIN2-like function in root epidermal cells and to efficiently mediate the auxin flow for the root gravitropic response.

PIN Phosphorylation at Ser337/Thr340 Mediates PIN Polarity and Auxin Distribution.

All observations in the PIN1::PIN1-GFP and PIN2::PIN1-HA transgenic lines suggest that the Ala and Asp mutation-induced phenotypes might result from the modulated polar targeting of mutated PIN1 variants. Therefore, we investigated the subcellular polar targeting of PIN1-HA(Ala) and PIN1-HA(Asp) proteins in epidermal cells. Ectopically expressed PIN1-HA in the epidermis localizes predominantly to the basal side of cells, whereas the endogenous PIN2 localizes to the apical side (9)(Fig. 2C). PIN1-HA(Ala) localized at the basal side of epidermal cells, but, importantly, PIN1-HA(Asp) predominantly to the opposite apical side, similar to the endogenous PIN2 (Fig. 2C). Accordingly, only PIN1-HA(Asp) was able to functionally replace PIN2 in the pin2 mutant.

Similarly to PIN1-HA, the mutations in Ser337/Thr340 influenced also the PIN1-GFP targeting. In roots of the wild type, PIN1 localizes to the basal side of stele cells (28)(Fig. 3A). In comparison with PIN1-GFP, the polar localization of PIN1-GFP(Asp) was dramatically disturbed (n = 83 roots). Generally, the basal polarity of PIN1 was much less pronounced and resulted in a largely nonpolar (68%) or preferentially apical distribution (7%) (Fig. 3A). In contrast, PIN1-GFP(Ala) had a largely strict basal localization (n = 56; Fig. 3A). Importantly, these polarity changes correlated well with the incidence of the auxin response in the root tip as visualized indirectly by the auxin response reporter DR5::GUS. PIN1-GFP(Ala) lines revealed an increased DR5 activity in the root tip, contrasting with the dramatically decreased DR5 signal in the roots of PIN1-GFP(Asp) (Fig. 3B). The changes in DR5 activity were confirmed by direct auxin measurements in the 1-mm seedling root tips. Consistently with the DR5 observations, PIN1-GFP(Ala) and PIN1-GFP(Asp) showed a higher and lower auxin content than that of PIN1-GFP, respectively (Fig. 3C). These results demonstrate that preferentially the basally localized PIN1(Ala) promotes, whereas more apically localized PIN1(Asp) diminishes the auxin accumulation at the root tips. These quantitative changes in DR5 activity and auxin content show that the PIN phosphorylation-mediated PIN polarity changes have a direct impact on the auxin distribution.

PIN1 Phosphorylation at Ser337/Thr340 Is Related to the PID Action in Polar Targeting.

The observed changes of PIN1 polarity, auxin distribution, and resulting phenotypes revealed that phosphorylation of the Ser337/Thr340 site is sufficient to redirect the PIN polarity as well as directionality of the auxin flow, similarly as shown for manipulations of the PID kinase activity (18, 23). Therefore, we addressed the question of whether this particular phosphorylation site is related to the PID kinase action in the PIN polarity determination. pid gain of function (35S::PID) leads to a basal-to-apical PIN polarity shift, which causes auxin depletion from the root meristem, ultimately leading to its collapse (23). Several independent transgenic lines of PIN1-HA(Ala) and PIN1-HA(Asp) showed the consistent changes in the root collapse frequency. We observed that PIN1-HA(Ala) significantly delayed the 35S::PID-mediated root collapse (n = 96) when compared to that of the wild-type PIN1-HA (n = 85) (Fig. 4A). In contrast, the collapse of the root meristem in PIN1-HA(Asp) (n = 97) and PIN1-HA seedlings did not differ significantly (Fig. 4A), which might be due to the already near maximum effect of 35S::PID on PIN apicalization and root collapse.

Next, we tested the pid loss-of-function line. We introduced PIN1-GFP(Ala) and PIN1-GFP(Asp) transgenes into the pid^+/− wag1 wag2 allele. WAG1 and WAG2 are the closest homologs of PID in the family of the AGC serine/threonine protein kinases of Arabidopsis and the triple mutant pid wag1 wag2 displays a completely disrupted cotyledon formation (29, 30). We analyzed the F₂-segregating progeny and quantified the seedlings with defective cotyledon patterning. The frequency of the seedlings that failed to make any cotyledon was increased by PIN1-GFP(Ala) and reduced by PIN1-GFP(Asp) transgenes. We observed 16% (n = 816) and 4% (n = 454) aberrant seedlings without cotyledons in PIN1-GFP(Ala) and PIN1-GFP(Asp), respectively, when compared to 8% (n = 788) in the F₂ progeny of control crosses with the wild-type PIN1-GFP (Fig. 4B). Furthermore, the pid wag1 wag2 mutations did not affect the PIN1-HA(Ala) and PIN1-HA(Asp) localizations at the basal and apical sides of the epidermal cells, respectively (Fig. 4C). Altogether, these data show that PIN1(Ala) have antagonistic and PIN1(Asp) mutations agonistic effects with the PID action, consistently with the PID regulating the PIN polarity targeting this and/or other phosphorylation sites. Importantly, the apical localization of PIN1(Asp) does not require the function of PID and related kinases, suggesting that other potential phosphorylation sites targeted by PID or other kinases cannot override the effect of the phosphomimic mutation at the Ser337/Thr340 site. These results reveal that the Ser337/Thr340 phosphorylation site is involved in the similar mechanism of PIN polarity regulation as PID action; nonetheless these results cannot discriminate whether Ser337/Thr340 is a target of PID.

PID Might Not Directly Target Ser337/Thr340.

Next, we tested whether the Ser337/Thr340 site of the PIN proteins can be directly phosphorylated by PID. The HIS-tagged wild-type hydrophilic loop of PIN1 (HIS-PIN1HL, amino acids 288–452; Fig. S2 A and B), GST-tagged PID (GST-PID), and a 30-amino-acid stretch encompassing the Ser337/Thr340 site (GST-PIN1P-site, amino acids 321–350; Fig. S2 A and B) as well as its Ala-substituted version were heterologously expressed in Escherichia coli and coincubated in an in vitro phosphorylation reaction. After electrophoretic separation of the proteins, PID autophosphorylation was detected as well as phosphorylation of HIS-PIN1HL and the standard Ser/Thr kinase substrate myelin basic protein (MBP) (Fig. 4D and Fig. S3). In contrast, we failed to observe any PID-dependent phosphorylation of both the wild-type and mutated versions of the GST-PIN1P-site (Fig. 4D and Fig. S3). These results show that the PID kinase phosphorylates the large hydrophilic loop of PIN1 in vitro, but is probably not able to phosphorylate the Ser337/Thr340 site in the same assay.

Taken together, the results obtained in in vivo genetic and cell biological studies and in vitro phosphorlyation assays are consistent with a scenario in which the Ser337/Thr340 phosphorylation site could be a target for the other potential kinases that act together with PID in regulating the PIN phosphorylation and the apical/basal PIN polar targeting.

Materials.

The following mutants, transgenic plants, and constructs have been described previously: pin1 (pin1-1) (38), pin2 (eir1-1) (25), pid^+/− wag1 wag2 (kind gift of Remko Offringa, Leiden University, The Netherlands), PIN1::PIN1-GFP (39), PIN2::PIN1-HA (9), DR5rev::GFP (40), DR5::GUS (41), 35S::PID (19), pHIS-PIN1HL (amino acids 288–452) (42), and pGST-PID (43). Myelin basic protein (MBP) was purchased from Sigma.

The constructs pPIN1::PIN1-GFP(Ala), pPIN1::PIN1-GFP(Asp), pPIN2::PIN1-HA(Ala), and pPIN2::PIN1-HA(Asp) were generated from the original pPIN1::PIN1-GFP (39) and pPIN2::PIN1:HA (9) by converting the two residues Ser337(TCG) and Thr340(ACT) on the PIN1 fragment to Ala(GCG) and Ala(GCT) or Asp(GAC) and Asp(GAT), respectively, through the PCR-based site-directed mutagenesis kit (Stratagene Quickchange XL). The mutated strands were synthesized with the following primer combinations: 5′- CCCAGGGATGTTTGCGCCCAACGCTGGCGGTGGTGGAG-3′ and 5′-CTCCACCACCGCCAGCGTTGGGCGCAAACATCCCTGGG-3′ for the Ala mutation and 5′-CCCAGGGATGTTTGACCCCAACGATGGCGGTGGTGGAGGC-3′ and 5′-GCCTCCACCACCGCCATCGTTGGGGTCAAACATCCCTGGG-3′ for the Asp mutation. The resulting constructs were transformed into the wild-type and the pin1, pin2, and 35S::PID lines. DR5rev::GFP, DR5::GUS markers, and pid^+/− wag1 wag2 mutations were introduced into the phosphorylation transgenic mutant lines by crosses. For the in vitro phosphorylation assay, the sequences of wild-type PIN1P-site and mutant PIN1P-site(Ala) encompassing 30 amino acids surrounding the Ser337 were generated by PCR amplification with the primer pairs containing attB recombination sites (underlined): 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGCGGTGGCGGTGGAGGAGCGCAT-3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTGTCATCCTTTCGCCGCCGCGCCTCCACC-3′ from the pPIN1::PIN1-GFP and pPIN1::PIN1-GFP(Ala) constructs, respectively.

E. coli BL21 Star (DE3) cells were used for heterologous expression of wild-type HIS-PIN1HL and Rosetta cells for the other recombinant proteins.

Growth Conditions and Phenotypic Analyses.

A. thaliana (L.) Heynh. plants were grown on soil or half-strength Murashige and Skoog medium with sucrose as described (28) under a 16-h light/8-h dark cycle at 21°C or 18°C. All phenotypic comparisons were done at least in duplicate, with three to nine independent lines of a minimum of 25 seedlings each time. For evaluation of root gravitropism in 6-day-old seedlings, we used the quantification tool Image J (rsbweb.nih.gov/ij/download.html) and calculated the vertical growth index (VGI = vertical length [VL]/root length [RL]) (44). For the timing of the 35S::PID-mediated collapse of the primary root meristem, we assayed 5-day-old and 7-day-old seedlings with a Leica binocular. The root collapse becomes visually apparent a few days after germination and can be recognized by the thin and pointed appearance of the root tip (19). Results are presented as means with standard error bars and all claimed comparisons are based on statistical evaluations with t test or Fisher exact test.

In Situ Expression and Localization Analyses.

Histochemical stainings for GUS activity and whole-mount immunolocalizations in 5-day-old seedlings were carried out as described (39, 45, 46). Each experiment was done on three to nine independent lines with three repetitions (homozygous lines, n > 15; segregating lines, n > 300 individuals). Antibodies were diluted as follows: 1:1,000 for rabbit anti-PIN1 (42); 1:500 for mouse anti-HA (Babco); 1:500 for mouse anti-GFP (Roche); and 1:600 for CY3- (Sigma) and Alex- (Invitrogen) conjugated anti-rabbit and anti-mouse secondary antibodies, respectively. For the HA-tagged PIN proteins, colocalization of anti-PIN1 and anti-HA were always carried out to confirm the identity of the proteins. For in vivo GFP inspection, 6-day-old seedlings were mounted in 5% glycerol without fixation for live-cell imaging. Regarding the auxin translocation assay, confocal imaging was done in 6-day-old seedlings, immediately after 3 h of a 90° gravistimulation in the dark. All of the fluorescence signals were evaluated on a Zeiss LSM 510 or Leica TCS SP2 confocal laser scanning microscope. Always the same microscope settings were used for each independent experiment, and pixel intensities were taken into account when comparing the images between control and mutants. Images were finally processed with Adobe Photoshop 7.0 and Adobe Illustrator 8.0.

Auxin Measurements.

One-millimeter-long root tips were transferred immediately after harvest to tubes containing 1 mL of methanol kept at 0°C. To each tube, ²H₅-IAA was added at the concentration of 10 fmol per root tip. The samples were put into a freezer (−20°C) and centrifuged at 36,000 × g after 24 h. Supernatants were transferred into glass tubes, evaporated to dryness, and methylated with ethereal diazomethanol. Further processing by immunopurification was performed as described (47) and final analysis was done with a UHPLC coupled to a Waters Xevo TQ MS detector.

In Vitro Phosphorylation Assays.

Approximately 3 μg of purified protein expressed in E. coli (PID and substrates) was added to the kinase reaction mix, containing 1× kinase buffer [25 mM Tris-HCl (pH 7.5), 1 mM DTT, 5 mM MgCl₂, 2 mM CaCl₂] and 1× ATP solution (100 μM MgCl₂, 100 μM ATP-Na₂, 2 μCi ³²P-γ-ATP). Reactions were incubated at 30°C for 45 min and stopped by addition of 5 μl of 5× protein loading buffer [310 mM Tris-HCl (pH 6.8), 10% SDS, 50% glycerol, 750 mM β-mercaptoethanol, 0.125% bromophenol blue] and boiling for 5 min. Reactions were subsequently separated over 12% acrylamide gels, which were washed three times for 30 min with kinase gel wash buffer (5% trichoroacetic acid, 1% Na₂H₂P₂O₇), coomassie stained, destained, and dried. Autoradiography was done for 48 h with Hyperfilm MP films and a Hypercassette (GE Healthcare Life Sciences).