Report

A PINOID-Dependent Binary Switch in Apical-Basal PIN Polar Targeting Directs Auxin Efflux

See allHide authors and affiliations

Science  29 Oct 2004:
Vol. 306, Issue 5697, pp. 862-865
DOI: 10.1126/science.1100618

Abstract

Polar transport–dependent local accumulation of auxin provides positional cues for multiple plant patterning processes. This directional auxin flow depends on the polar subcellular localization of the PIN auxin efflux regulators. Overexpression of the PINOID protein kinase induces a basal-to-apical shift in PIN localization, resulting in the loss of auxin gradients and strong defects in embryo and seedling roots. Conversely, pid loss of function induces an apical-to-basal shift in PIN1 polar targeting at the inflorescence apex, accompanied by defective organogenesis. Our results show that a PINOID-dependent binary switch controls PIN polarity and mediates changes in auxin flow to create local gradients for patterning processes.

The plant signaling molecule auxin plays a central role in a wide variety of development processes. A major determinant in auxin-mediated plant growth is the directed transport of auxin from foci of biosynthesis to sites of action. This polar auxin transport mediates vectorial gradients that underlie tropic growth responses and provide positional cues for apical-basal patterning, organogenesis, and vascular differentiation (14). The molecular characterization of the Arabidopsis thaliana pin-formed (pin1) mutant, which is defective in auxin transport and develops pin-like inflorescences, led to the identification of the PIN family of transporter-like membrane proteins. A substantial amount of data demonstrates that PIN proteins are important regulators of polar auxin transport that possibly function as auxin efflux carriers (4). PIN proteins display asymmetric subcellular localization at the plasma membrane, which determines the direction of polar auxin transport and thus establishes the local auxin gradients that influence different developmental processes. The polarity of PIN proteins can be rapidly modulated in response to external or developmental cues (1, 3, 5), a process that is enabled by continuous GNOM ARF GEF–dependent cycling of PINs between endosomes and the plasma membrane (6) (GNOM, Arabidopsis GNOM protein; ARF, ADP ribosylation factor; GEF, guanine nucleotide exchange factor).

Loss-of-function mutants of the protein serine-threonine kinase PINOID (PID) display apical organogenesis defects similar to those of the pin1 mutant (7). Constitutive overexpression of PID (35S::PID), but not of the kinase-negative MPID (35S::MPID), leads to hypocotyl and root agravitropy and to loss of the primary root meristem function (8, 9). The collapse of the root meristem in 35S::PID seedlings, which is characterized by the loss of meristem initials followed by terminal differentiation, is restricted to the primary root and is preceded by a reduction in auxin-responsive DR5::GUS expression (Fig. 1, A and B) (9). Measurements of indole-3-acetic acid (IAA) in intact root tips showed that IAA levels are significantly reduced in 35S::PID primary root tips as compared to wild-type root tips (Fig. 1E). In contrast, free IAA concentrations in lateral root tips of 8- to 11-day-old 35S::PID and wild-type seedlings do not differ significantly (Fig. 1E), and accordingly the DR5::GUS expression peak is unchanged (Fig. 1, C and D). These data confirmed that PID overexpression results in reduced auxin accumulation in the primary root tip, thereby causing a reduction in the DR5 expression peak in the root meristem and eventually the collapse of this structure. Treatment with the auxin efflux inhibitor naphthylphtalamic acid (NPA) restores the DR5 expression peak and prevents root meristem collapse (9), whereas treatment with auxin itself has no effect. These data suggest that PID is a regulator of NPA-sensitive polar auxin transport.

Fig. 1.

Ectopic PID expression results in PIN-mediated changes in auxin homeostasis in the primary root meristem. (A to D) DR5::GUS expression in wild-type [(A) and (C)] and 35S::PID [(B) and (D)] primary [(A) and (B)] and lateral [(C) and (D)] roots. GUS stainings were performed for 1 hour in (A) and (B) and for 4 hours in (C) and (D) (21). (E) Auxin measurements in 1-mm tips of the primary and lateral roots of wild-type and 35S::PID seedlings (21). Dpg, days post germination. The asterisk marks values that are significantly different from the wild type (t test, P < 0.05). (F) Collapse of 35S::PID-21 primary root meristems is significantly delayed by NPA and in pin2/eir1-1 and pin4 mutant backgrounds (21). Scale bar, 60 μM in (A) and (B), 0.3 mm in (C), and 0.2 mm in (D).

We used the timing of root collapse as an assay to address whether PID action on auxin transport occurs through PIN efflux regulators. 35S::PID plants were crossed with loss-of-function mutants of PIN genes known to mediate root development, these being PIN2, PIN3, and PIN4 (5, 10, 11). In the pin2/eir1-1 and pin4 mutant backgrounds, the 35S::PID-mediated root collapse was significantly delayed, whereas the pin3 mutation resulted in a mild delay in root collapse around 4 to 5 days after germination (Fig. 1F). Both pin2 and pin4 mutations result in increased auxin concentrations in the root: pin2 elevates auxin concentrations because of the lack of redistribution of auxin via basipetal transport from the root tip to the elongation zone (12), and pin4-elevated auxin levels result from the absence of a focused PIN4-driven auxin sink in the first columella tier (11).

These results imply that PID gain of function changes auxin concentrations in the root tip through the PIN proteins. Conceivably, PID could regulate either the expression of PIN proteins, the polarity of their subcellular localization, or their activity. Because an activity assay for PIN proteins is so far not available, we focused on testing the expression and subcellular localization of PIN proteins, by immunolocalizing various PIN proteins in primary and lateral root tips of wild-type and 35S::PID lines. This showed that the tissue-specific expression domains of the PIN proteins are unchanged in 35S::PID root tips. Although quantitative changes cannot be excluded, this suggests that regulation of PIN gene expression or protein stability is not a primary target of PID action. 35S::PID expression did, however, lead to a basal-to-apical shift in the subcellular polarity of PIN proteins. This apical shift was most apparent in cells where PINs normally show basal subcellular polarity, such as PIN2 in the cortex (Fig. 2, A and B versus C); PIN4 in the epidermis, cortex, endodermis, and vascular cells (Fig. 2, E and F versus G and H); and PIN1 in the endodermis and stele (fig. S1, A versus B). PIN localization in 35S::PID lines remains unchanged in cells, where PIN proteins are localized apically in the wild type; for example, PIN2 in epidermal cells (Fig. 2, A and B versus C).

Fig. 2.

PID overexpression induces a basal-to-apical shift in PIN polarity. (A to H) Immunolocalization of PIN2 [(A) to (D)] and PIN4 [(E) to (H)] in the primary root [(A) to (C) and (E) to (H)] and the lateral root (D) of wild-type control seedlings [(A), (B), (E), (F)] and of seedlings of the intermediate and strong PID overexpression lines 35S::PID-10 (G) and 35S::PID-21 [(C), (D), and (H)], respectively (21). (I to L) The effect of inducible overexpression of PID on the subcellular localization of PIN1 in roots of 3-day-old seedlings. (I) pINTAMGUS control seedlings after 48 hours of incubation in tamoxifen-containing medium. pINTAMPID seedlings after 12 (J), 24 (K), and 36 (L) hours of tamoxifen treatment are shown (21). Arrowheads indicate apical or basal subcellular PIN polarity [(B) to (D) and (F) to (L)]. Images are representative of at least 10 observed samples. Scale bar, 40 μM in (A), 3 μM in (B) to (D) and (F) to (L), and 15 μM in (E).

The 35S::PID-induced basal-to-apical shift requires the active PID protein kinase, because it is not observed in the 35S::MPID line (fig. S1C), and the shift is dose-dependent, because it is most pronounced in root tips of the strong overexpression line 35S::PID-21 (Fig. 2H) as compared to the intermediate line 35S::PID-10 (Fig. 2G). To exclude the possibility that the basal-to-apical PIN polarity shift is a secondary consequence of changes in patterning before meristem collapse, we examined PIN localization in NPA-rescued root tips of 3-day-old 35S::PID-21 seedlings. NPA treatment itself did not have any pronounced effect on the subcellular polarity of PIN localization (fig. S1D), and the basal-apical shift was still observed in the rescued 35S::PID root meristems (fig. S1E). Also, a clear 35S::PID-induced basal-to-apical shift of PIN2 could be observed in the cortex of lateral root meristems that, in contrast to the primary root meristem, remain functional and do not collapse (Fig. 2D). To provide additional evidence for the immediate causality between ectopic PID expression and the shift in subcellular PIN polarity, we used an estrogen(4-hydroxytamoxifen)–inducible two-component expression system to drive ectopic PID expression. Tamoxifen-dependent expression of a coregulated GUS reporter gene was detectable in roots after 12 hours of induction (13), followed by strong constitutive expression 24 hours after induction (fig. S1G). The basal-to-apical shift of PIN polarity was clearly visible in seedlings 12 hours after treatment with tamoxifen (Fig. 2J) and became more pronounced after 24 hours (Fig. 2, K and L). Neither incubation of pINTAMPID seedlings in medium alone (fig. S1H) nor treatment of seedlings of the pINTAM-GUS control line with tamoxifen (Fig. 2I) induced changes in PIN polarity.

These data demonstrate that PID overexpression leads to apical retargeting of basally localized PIN proteins, and they explain the effect of PID overexpression on root meristem activity. In the wild-type root meristem, basal localization of PIN1, PIN2, and PIN4 in subepidermal cell types directs the auxin supply to the root tip and creates the auxin maximum needed for maintenance of root meristem activity. The 35S::PID-induced basal-to-apical polarity shift of both PIN2 and PIN4 results in a unidirectional auxin flux away from the root tip. This depletes the primary root meristem of auxin (Fig. 1E) resulting in loss of the auxin maximum and root meristem collapse. The reason for the insensitivity of lateral roots to the PID-mediated shift in PIN polarity is unclear, but it could be due to increased auxin biosynthesis in lateral root meristems (14). Alternatively, the exclusive collapse of 35S::PID primary root mersitems could relate to their embryonic origin.

PIN-dependent auxin gradients are responsible for both the establishment and the maintenance of the primary root meristem during embryogenesis (1). Root pole initiation depends on an apical-to-basal auxin flow in young globular embryos, which requires basally localized PIN proteins and GNOM activity (15). As a consequence, auxin accumulates in the hypophysis, which in response is specified as the root meristem precursor (1). As in postembryonic 35S::PID roots, a basal-to-apical switch in PIN1 localization was also observed in 35S::PID embryos; however, this switch occurred only in post-globular stages (Fig. 3B) after hypophysis specification. In line with this, 35S::PID embryos showed mild defects in root meristem patterning only at maturity (fig. S2, A to C). Because these mild embryonic phenotypes may be due to low 35S promoter activity during early embryo development, we expressed PID under the control of the RPS5A promoter, which is highly active in young embryos (16, 17). Accordingly, young globular RPS5APID embryos showed severe defects in PIN polar localization (Fig. 3C). The complete basal-to-apical shift of PIN1 polarity in RPS5APID embryos is clearly different from the random PIN1 polar localization observed in gnom mutants (15). The auxin-responsive DR5rev::GFP reporter (1) showed that RPS5APID globular- and heart-stage embryos are incapable of establishing an auxin maximum in the hypophysis. In fact, green fluorescent protein (GFP) expression is excluded from the hypophysis and instead accumulates in the embryo proper (Fig. 3, D and E, and fig. S2, D and E). Altered PIN1 polarity and DR5rev::GFP misexpression in RPS5APID embryos is accompanied by misspecification of the hypophysis, as demonstrated by defective cell divisions (Fig. 3G). After germination, RPS5APID seedlings showed strong patterning defects, including the absence of a root and defective cotyledons (Fig. 3H), which are strikingly similar to gnom and quadruple pin1, pin3, pin4, pin7 loss-of-function phenotypes (1). These data imply that the PID-mediated basal-to-apical shift in PIN polarity in the embryo reverses the apical-basal auxin flow, leading to accumulation of auxin in the apical regions instead of the hypophysis. As a consequence, the apical-basal axis (root pole) and bilateral symmetry (cotyledons) of the embryo are not properly specified.

Fig. 3.

Ectopic PID expression in embryos. (A to C) PIN1 immunolocalization in triangular-stage wild-type (A) and 35S::PID (B) embryos and in a globular-stage RPS5APID (C) embryo. The absence of basally located PIN1 in cells above the hypophyseal cell group (arrowhead) is indicative of the basal-to-apical shift in PIN1 polarity. (D and E) DR5rev::GFP expression in globular wild-type (D) and RPS5APID (E) embryos (21). (F and G) Differential interference contrast images of wild-type (F) and RPS5APID (G) globular embryos (21). The plane of cell division in the hypophyseal region is indicated by an arrowhead. (H) One-week-old RPS5APID F1 seedlings showing defects in basal patterning and bilateral symmetry formation. Images are representative of at least five observed samples. Scale bar, 10 μM in (A) to (E), 5 μM in (F) and (G), and 3 mm in (H).

The data described above show that PID gain of function induces a basal-to-apical shift in PIN polarity, resulting in alterations in auxin distribution that lead to corresponding developmental defects. Next we examined the effect of PID loss of function on PIN polarity, using transgenic lines carrying a functional PIN1:GFP fusion construct (3) (fig. S3, A to E). In both pid and pin1 mutants, inflorescence development, including organ initiation at the apex, is severely disrupted. PIN1 is localized apically in epidermal cells below the very apex in these regions in the wild type (3, 18) (fig. S3, D and E). Strikingly, PIN1:GFP is localized to the basal side of epidermal cells in the same regions in different pid mutant allele backgrounds (Fig. 4, A to C). This apical-to-basal reversal of PIN1 polarity is unlikely to be an indirect consequence of the altered inflorescence structure, because PIN1:GFP does not localize basically in morphologically comparable, NPA-induced, pin-shaped inflorescences of wild-type or 35::PID plants (Fig. 4, D to G). These data indicate that pid loss of function results in basal targeting of PIN proteins, leading to failure in the establishment of local auxin accumulation required for proper apical organ formation (3). The identification of PID as a regulator of polar PIN1 localization at the shoot apex explains the similarity in development of the pid and pin1 mutants and places PID at a central position in polar auxin transport regulation.

Fig. 4.

Localization of the PIN1:GFP fusion protein in inflorescence apices in wild-type and mutant Arabidopsis. (A to G) Confocal sections of pin-formed inflorescences of the pid::En310 [(A) and (B)] and pid::En197 (C) loss-of-function alleles and of NPA-grown wild-type [(D) and (E)] and 35S::PID-21 [(F) and (G)] plants (21). Arrowheads indicate apical or basal subcellular PIN1:GFP polarity [(B), (C), (E), and (G)]. Images are representative of at least 10 observed samples. Scale bar, 20 μM in (A), (D), and (F), and 5 μM in (B), (C), (E), and (G).

The phenotypes of PID gain- and loss-of-function plants initially suggested a role for this protein kinase in the regulation of auxin signaling (8) or auxin transport (9); however, the mechanism and target(s) of PID action remained elusive. Data presented here now indicate that PID is a regulator of polar auxin transport at the level of PIN protein localization. Rapid changes in the polarity of PIN proteins in response to developmental or environmental cues have been shown to redirect auxin flow and mediate multiple developmental processes (1, 3, 5), but the mechanism underlying polarity control is largely unresolved. These rapid relocations are enabled by constitutive cycling of PIN proteins. GNOM, an endosomal regulator of vesicle budding that mediates this process, is required for the recycling but does not seem to directly determine the polarity of PIN targeting, because gnom mutants show randomized changes in PIN1 polar localization (15). In line with this more general function of GNOM in polar auxin transport, most, if not all polar auxin transport–dependent developmental processes are disturbed in gnom mutants (19). In contrast, the role of PID in controlling PIN polarity is more specific. It appears that PID acts as a binary switch, with subthreshold PID levels leading to basal PIN localization and above-threshold PID levels leading to apical PIN localization. Consistently, the defects observed in both loss- and gain-of-function of PID lines can each be ascribed to a reversion in auxin flow. The observation that auxin controls cellular PID levels (9) raises the intriguing possibility that PID is involved in a feedback mechanism by which auxin influences the polarity of its own flow. Such a regulatory network has been proposed in the classical “canalization” model to explain the self-organizing properties of vascular tissue differentiation (20) and would also account for the dynamic auxin-dependent changes in auxin redistribution underlying phyllotaxis (3, 18). Even though endogenous PID acts in a limited subset of PIN-dependent development processes, many more cells are competent to respond to PID overexpression. In Arabidopsis, PID belongs to a family of 23 protein kinases, the members of which are differentially expressed. It is therefore likely that ectopic PID expression reveals functions that are normally represented by these different PID family members. Although further experimentation is needed to determine whether PIN proteins are direct targets of PID, it is clear that PID-dependent phosphorylation is an essential intermediate step in the control of polar PIN protein targeting, and thus of directional auxin flow regulating patterning processes.

Supporting Online Material

www.sciencemag.org/cgi/content/full/306/5697/862/DC1

Materials and Methods

Figs. S1 to S3

References

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article