High Boron-induced Ubiquitination Regulates Vacuolar Sorting of the BOR1 Borate Transporter in Arabidopsis thaliana*

Boron homeostasis is important for plants, as boron is essential but is toxic in excess. Under high boron conditions, the Arabidopsis thaliana borate transporter BOR1 is trafficked from the plasma membrane (PM) to the vacuole via the endocytic pathway for degradation to avoid excess boron transport. Here, we show that boron-induced ubiquitination is required for vacuolar sorting of BOR1. We found that a substitution of lysine 590 with alanine (K590A) in BOR1 blocked degradation. BOR1 was mono- or diubiquitinated within several minutes after applying a high concentration of boron, whereas the K590A mutant was not. The K590A mutation abolished vacuolar transport of BOR1 but did not apparently affect polar localization to the inner PM domains. Furthermore, brefeldin A and wortmannin treatment suggested that Lys-590 is required for BOR1 translocation from an early endosomal compartment to multivesicular bodies. Our results show that boron-induced ubiquitination of BOR1 is not required for endocytosis from the PM but is crucial for the sorting of internalized BOR1 to multivesicular bodies for subsequent degradation in vacuoles.

Although boron is an essential micronutrient for plants, excess boron is toxic. The homeostasis of boron is important for plants. In fact, both boron deficiency and toxicity have negative impacts on agricultural production (1,2). The only established function of boron in vascular plants is cross-linking of the pectic polysaccharide rhamnogalacturonan II. This cross-linking mediated by borate is important to maintain cell wall architecture and is necessary for normal growth and de-velopment (3,4). In contrast, the molecular mechanism of boron toxicity is still unclear, despite several efforts at applying biochemical, genetic, and molecular approaches (5)(6)(7)(8)(9).
Boron is taken up by plant roots, mainly in the form of boric acid. Arabidopsis thaliana possesses two distinct types of boron-transporting proteins, BOR1 (10) and NIP5;1 (11), for effective boron uptake into roots and transport to shoots under boron-limiting conditions (12). BOR1 is an efflux-type borate transporter with similarities to animal bicarbonate transporters (13). In roots, BOR1 is polar localized to the inner (stele-facing) plasma membrane (PM) 4 domain of various cells (14) and functions in inward transport of boron toward the stele under low boron conditions (10). NIP5;1 is a member of the MIP (major intrinsic protein) family and functions as a boric acid channel for efficient boron uptake into the root under boron-limiting conditions (11). NIP5;1 is polar localized to the outer (soil-facing) PM domain of root epidermal and root cap cells (14) and increases the permeability of the outermost PM domain to boric acid for efficient boron uptake.
These transporters have been used successfully to generate boron stress-tolerant plants. BOR1 overexpression in A. thaliana results in enhanced boron translocation from the root to shoot and improved shoot growth and seed yields under boron-limiting conditions (15). Enhancing NIP5;1 expression by activation tag improves root growth under boron-limiting conditions (16). Overexpression of BOR4, a BOR1 paralog, reduces boron concentration in both roots and shoots and confers tolerance to toxic levels of boron (17).
Accumulations of NIP5;1 and BOR1 are regulated by boron availability via different regulatory mechanisms. NIP5;1 is regulated at the level of mRNA accumulation in response to boron concentration in the media (11), whereas BOR1 accumulation is regulated through protein degradation. In the presence of high boron concentrations, BOR1 is rapidly removed from the PM and transferred to the lytic vacuole through the endocytic pathway for prompt degradation (18). The endocytic degradation system is assumed to be beneficial for plants to quickly shut down boron transport and avoid excess accumulation of boron.
Endocytic membrane trafficking is a key mechanism for controlling the amount of protein in the PM. During endocytic degradation, the target protein is internalized from the PM to the early endosomal compartments by endocytosis, sorted into late endosomes (LEs)/multivesicular bodies (MVBs), and transferred into vacuoles/lysosomes for degradation (19). Although the details of this mechanism have been studied intensively in yeast and mammals (19 -22), much less knowledge has been accumulated in plants. A recent report demonstrated that the A. thaliana brassinosteroid insensitive-1 (BRI1) receptor and BOR1 are sorted into LEs/MVBs for vacuolar degradation through the trans-Golgi/early endosome (EE) network (23).
Ubiquitination is a key signal for the homeostatic regulation of proteins. Two distinct forms of ubiquitination are known: polyubiquitination is used in the proteosomal protein degradation system, and mono-and multimonoubiquitination are required for endocytic internalization of transmembrane proteins for lysosomal/vacuolar turnover of membrane proteins in yeast and mammals (19). In addition, monoubiquitination has also been reported as a sorting signal of the membrane proteins to MVBs in mammals (24). This ubiquitination is required for binding to the endosomal protein sorting complex (ESCRT). In A. thaliana, ESCRT-related proteins CHMP1A and CHMP1B are required for MVB sorting of the three auxin carriers: PIN1, PIN2, and AUX1 (25). PIN2 turnover, which is related to root gravitropism, is regulated by endocytic membrane trafficking and proteasome activity (26). This study showed that PIN2 is modified by ubiquitin and that ubiquitination depends on proteasome activity; however, whether polyubiquitination or multimonoubiquitination is required remains unclear. In addition, no direct evidence exists that ubiquitination is involved in PIN2 degradation (26).
Here, we identified a ubiquitination site in BOR1, which is required for boron-induced vacuolar sorting of BOR1. We reported previously that BOR4, unlike BOR1, is not degraded by high boron (ϩB) supplementation (17). Thus, in this study, we searched out amino acid residues involved in boron-induced degradation of BOR1 by differences between the BOR1 and BOR4 sequences and found that lysine 590 (Lys-590) was critical for the degradation. Supplementation with high boron induced mono-or diubiquitination of GFP-tagged BOR1 (BOR1-GFP), and a Lys-590 mutation completely inhibited ubiquitination. Furthermore, the Lys-590 mutation did not inhibit the endocytosis step from the PM but prevented MVB sorting of BOR1 for degradation in the vacuole.

MATERIALS AND METHODS
Plasmid Construction and Plant Transformation-Binary plasmids for plant transformation to express GFP-fused BOR mutants under control of the BOR1 promoter were generated using Gateway technology (Invitrogen). Construction of BOR1-BOR4 chimeras and site-directed mutagenesis were conducted using the in vitro overlap extension PCR method (27). The BOR1-GFP (10) or BOR4-GFP (17) construct was used as a template, and the primers used are listed in supplemental Table S1. The PCR fragments of GFP-tagged genes were subcloned into pENTR/D-TOPO (Invitrogen) according to the manufacturer's instructions and sequenced to confirm that that no PCR errors occurred. The subcloned GFP-tagged genes were mobilized into pAT100 (14), a destination binary vector carrying the BOR1 promoter, with LR Clonase (Invitrogen) according to the manufacturer's instructions.
To generate a T-DNA insertion mutant for both BOR1 and BOR2, T2 seeds of bor1-3 (SALK_37312) and bor2-1 (SALK_56473) were obtained from the Salk Institute (30). Plant Growth Conditions-MGRL medium (31) was prepared with slight modifications according to Takano et al. (18) and used with various concentration of boric acid as the boron source. MGRL solid medium was made with 1.5% gellan gum (Wako Pure Chemical, Inc., Osaka, Japan) and 2% sucrose in a sterile square Petri dish (140 ϫ 100 mm; Eiken, Tokyo).
A. thaliana seeds were surface-sterilized by soaking in 99.5% ethanol for 2 min and allowed to air-dry for a few hours; then they were sown onto MGRL solid medium. After a 2-day cold acclimation at 4°C, the plates were placed vertically, and the plants were grown at 22°C under a 16-h light/ 8-h dark cycle.
For microsome isolation, plants were grown on vertically placed MGRL solid medium containing 30 M boron for 16 days and grown hydroponically with MGRL liquid medium containing 30 M boron for 4 days and then with the medium containing 3 M boron (ϪB) for 4 days under a 10-h light/ 14-h dark cycle followed by treatment with the medium containing 100 M boron (ϩB, an optimal condition). The method for hydroponic culture was as follows. Twenty-five plants were placed between two formed polyethylene sheets (15 ϫ 100 ϫ 4 mm), and four sets of the sheets were floated on 400 ml of MGRL liquid medium in a light-shielded plastic container. The liquid medium was aerated continuously by an air pump and was changed every 4 days.
Confocal Fluorescent Imaging-Confocal imaging was performed as described previously with slight modifications (18). The transgenic plants were grown on MGRL solid medium supplemented with 3 M boron (ϪB) for 4 days. The roots were cut and incubated with MGRL liquid medium containing boric acid, inhibitors, or fluorescent dye at room temperature, as described in the figure legends (Figs. 2, 3, and 5; supplemental Figs. S2, S3, and S4). A stock solution of cycloheximide or FM4-64 (Molecular Probes, Carlsbad, CA) was prepared with water at 25 or 10 mM respectively, and brefeldin A (BFA) (Sigma) was prepared with dimethyl sulfoxide at 50 mM.
Confocal imaging was performed using an LSM510 (Carl Zeiss, Oberkochen, Germany) or FV1000 laser scanning confocal microscope (Olympus, Tokyo). The excitation wavelength was 488 nm, and emitted fluorescence was detected with a 505-530-nm band-pass filter for GFP or a 650-nm long-pass filter for FM4-64. Signal intensity was monitored to avoid pixel saturation. At least two independent transgenic lines (T 2 or T 3 plants) from each construct were used to confirm the reproducibility of the results.
Preparation of Microsomal Fractions from Arabidopsis Tissues-Microsomal proteins were prepared as described previously with some modifications (18). bor1-3/2-1 and transgenic plants (T 3 homozygous-lines) were grown hydroponically, as described above. The roots and shoots were harvested after treatment with 100 M boron (ϩB) MGRL liquid medium. Approximately 1 g of tissues was homogenized in 5 ml of ice-cold homogenization buffer (250 mM Tris-Cl (pH 7.8), 25 mM EDTA, 290 mM sucrose, 75 mM 2-mercaptoethanol, 10 mM N-ethylmaleimide, 1 mM PMSF, and one tablet/50 ml protease inhibitor mixture (Complete, EDTA-free; Roche Applied Science)) using a Polytron PT-2100 homogenizer (Kinematica Inc., Bohemia, NY) at 26,000 rpm for 5 s. This homogenization was repeated four times. The homogenates were centrifuged at 8,000 ϫ g for 5 min at 4°C, and the supernatant was centrifuged at 100,000 ϫ g for 10 min at 4°C. The microsome pellets were resuspended in 200 l of ice-cold storage buffer (50 mM Tris-Cl (pH 8.0), 150 mM NaCl, 10 mM N-ethylmaleimide, 1 mM PMSF, and one tablet/50 ml protease inhibitor mixture (Complete, EDTA-free; Roche Applied Science)) using a pestle. The protein concentration was determined using the Bradford method (Quick Start Protein Assay Kit I; Bio-Rad). The samples were frozen in liquid nitrogen and stored at Ϫ80°C.

RESULTS
The Amino Acid Sequence Involved in Boron-induced Degradation of BOR1-Our previous studies demonstrated that BOR1 and BOR4 are distinct in terms of boron-induced endocytic degradation and polar localization (14,17). BOR1 and BOR4 localize to the inner and outer PM domains of various root cells, respectively, and BOR1 is degraded by ϩB supplementation (Fig. 2), but BOR4 is not. Therefore, we searched amino acid residues involved in boron-induced degradation by means of differences between BOR1 and BOR4 sequences.
To identify the BOR1 region responsible for boron-induced degradation, we constructed two BOR1 and BOR4 chimeras. The C-terminal region of BOR1 was replaced with the corresponding region of BOR4 from Leu-468 or Glu-641 and called chimera 1 and chimera 2, respectively (Fig. 1). BOR1, BOR4, and chimeras, C-terminally fused to GFP and placed under the control of the BOR1 promoter, were generated. To exclude the possibility that an endogenous BOR1 or BOR2 (the closest homolog of BOR1) might disturb the mutational effects of introduced BOR mutants through protein-protein interactions, the constructs were introduced into bor1-3/2-1, a double BOR1 and BOR2 mutant without BOR1 and BOR2 accumulation. The growth defect in the bor1-3/2-1 mutant under boron-limiting conditions was almost fully restored by expressing BOR1-GFP and was partially restored by BOR4-GFP, chimera 1-GFP, and chimera 2-GFP (supplemental Fig.  S1), indicating that these expressed proteins possessed borontransporting activity. As shown in Fig. 2, BOR1-GFP was localized to the PM of the root cap and epidermal cells in the root tip with inward polarity and degraded upon treatment with ϩB for 2 h. In contrast, chimera 1-GFP and BOR4-GFP did not show inward polarity and did not disappear with ϩB. Although a part of the 1-GFP chimera was observed in the cytosol or endomembrane, most likely endoplasmic reticulum, PM localization was confirmed by co-localization with a specific membrane dye, FM4-64 (33) (supplemental Fig. S4). Chimera 2-GFP showed no or weaker polarity but was degraded by ϩB supplementation. These results suggest that the BOR1 sequence from Leu-469 to Glu-641 and Phe-642 to the last amino acid residue (Asn-704) contains the motif(s) for boron-induced degradation and polar localization, respectively.
Lysine 590 Is Crucial for Boron-induced Degradation of BOR1-We focused on potential ubiquitination sites to identify an amino acid residue required for boron-induced endocytic degradation of BOR1. Ubiquitin conjugation to the lysine residue in yeast is involved in endocytosis and degradation of PM proteins such as receptors, channels, and trans-porters (22,34). Therefore, we searched for a lysine residue in the region from Leu-469 to Glu-641 of BOR1 and found only one lysine residue, Lys-590 that is not conserved in BOR4 (Fig. 1B). We introduced a mutation to change Lys-590 to A (K590A). The GFP-tagged K590A mutant (BOR1 K590A -GFP) was expressed under the control of BOR1 promoter. Several independent transgenic plants were generated, and the transgenic plant lines displayed almost full restoration of the bor1-3/2-1 growth defect under a boron-limiting condition (supplemental Fig. S1). The BOR1 K590A -GFP showed inward polarity similar to the wild-type BOR1-GFP; however, boroninduced degradation was not observed (Fig. 3A). GFP fluorescence was seen in the PM 4 days after treatment with ϩB (100 M). This result indicates that Lys-590 is involved in BOR1 degradation. For further confirmation, microsomes isolated from shoots and roots of these transgenic plants after treatment with ϩB were subjected to immunoblot analysis using an anti-GFP antibody (Fig. 3C). A single band corresponding to the expected size of full-length BOR1-GFP or BOR1 K590A -GFP (ϳ106 kDa) was detected in all root samples, indicating that the GFP fluorescence seen in Fig. 3A was derived from the full-length BOR1-GFP protein. Boron-induced degradation of BOR1-GFP was observed in the root samples, but degradation was not detected for BOR1 K590A -GFP, consistent with the results of GFP imaging. Degradation was not observed in either BOR1-GFP or BOR1 K590A -GFP in the shoot during this short period of time (30 -120 min). In the shoot, another band was observed at a higher weight than 250 kDa, which was assumed to be a trimer or tetramer.
We also tested the effects of other lysine mutations in the BOR1 C-terminal region (see Fig. 1B), including K649A, K689A, and a K649A/K689A double mutant. All three mutant proteins, BOR1 K649A -GFP, BOR1 K689A -GFP, and BOR1 K649A,K689A -GFP, fully complemented the hypersensitivity to boron deficiency of the bor1-3/2-1 mutant (supplemental Fig. S1) and exhibited normal polarity and boron-induced degradation (supplemental Fig. S2). These results further confirmed the importance of Lys-590 but not of other nearby lysine residues.
We then tested whether Lys-590 is sufficient for boroninduced degradation of BOR proteins by introducing an N590K or N590K/P591G double mutation into chimera 1-GFP and N596K or N596K/P597G into BOR4-GFP. These changes represent the introduction of BOR1-type amino acid residues at these specific positions in the BOR4-type sequences. These constructs partially rescued the hypersensitivity to boron deficiency of the bor1-3/2-1 mutant, confirming boron transport activity of the BOR-GFP derivatives (supplemental Fig. S1). These amino acid substitutions did not induce boron-dependent degradation in either the 1-GFP chimera or BOR4-GFP (supplemental Fig. S3). The chimera 1 derivatives, chimera 1 N590K -GFP and chimera 1 N590K, P591G -GFP, as well as chimera 1-GFP, were localized in the PM and partly in the endomembrane, most likely the endoplasmic reticulum. The PM localization of these proteins was further confirmed by co-localization with FM4-64 (supplemental Fig.  S4). These results indicate that Lys-590 is necessary but not sufficient for boron-induced degradation of BOR proteins.
BOR1 Is Ubiquitinated Under ϩB Conditions, and Lys-590 Is the Principal Ubiquitination Site-We assumed that BOR1 ubiquitination at Lys-590 was involved in boron-induced degradation. Therefore, we attempted to detect ubiquitin conjugated to BOR1-GFP and compared that with BOR1 K590A -GFP. BOR1-GFP, BOR1 K590A -GFP, and BOR1 K649A,K689A -GFP were immunoprecipitated using the anti-GFP antibody from root extracts of transgenic plants after a time course treatment with ϩB, and ubiquitination was detected by immunoblotting using an anti-Ub antibody (Fig. 4). Two specific ubiquitinated protein bands were observed from BOR1-GFP and BOR1 K649A,K689A -GFP within 30 min after treatment with ϩB. The lower and upper bands corresponded to the size of BOR1-GFP conjugates with one and two ubiquitin molecules, respectively. No apparently ubiquitinated protein was observed for BOR1 K590A -GFP. In addition, immunoblotting using anti-GFP detected a relatively small amount of diubiquiti-  nated BOR1-GFP compared with non-ubiquitinated protein in a long exposure image, indicating that the ubiquitinated BOR1 protein is rapidly degraded and does not accumulate in large amounts within plant cells. These results suggest that BOR1 is mono-or diubiquitinated and that Lys-590 is essential for ubiquitination under a ϩB condition. Thus, Lys-590 is likely to be the ubiquitination site.
Lysine 590 Is Essential for Vacuolar Sorting of BOR1 but Not for Endocytosis-Boron-induced turnover of BOR1 is mediated by endocytosis and subsequent degradation in the vacuole (18). To determine which step of the endocytic degradation of BOR1 requires Lys-590 ubiquitination, we compared the endocytic trafficking of BOR1-GFP and BOR K590A -GFP using a general endocytosis marker, FM4-64 (33), and the membrane traffic inhibitor BFA. BFA inhibits the transition of membrane proteins from EEs to MVBs and to the PM but allows endocytosis (35)(36)(37).
BFA treatment was performed in the presence of cycloheximide to prevent de novo protein synthesis. After 30 min of BFA treatment, both BOR1-GFP and BOR1 K590A -GFP accumulated with endocytosed FM4-64 in the BFA compartment of root epidermal cells (Fig. 5A). The same volume of dimethyl sulfoxide, the BFA stock solution solvent, was used as the control, and no effect on traffic was observed. This result indicated that Lys-590 is not involved in endocytosis of BOR1 from the PM; however, it may be involved in subsequent vacuolar sorting steps (supplemental Fig. S6).
To exclude the possibility that K590A simply enhanced protein synthesis and overaccumulated BOR1 at the PM, we monitored GFP accumulation in the root vacuoles of transgenic plants expressing BOR1 K590A -GFP and compared that with BOR1-GFP to confirm that the K590A mutation inhib-ited BOR1 transport to the lytic vacuole. In the absence of light, GFP is resistant to vacuolar proteolysis; thus, dark treatment is an effective method of visualizing vacuolar targeting of GFP-tagged proteins (14,36,38). The roots of transgenic plants expressing BOR1-GFP and BOR1 K590A -GFP were incubated in ϪB and ϩB medium in the dark for 2 h, and GFP fluorescence in the epidermal cells of the root tip was monitored (Fig. 5B). After ϪB treatment, fluorescence from both BOR1-GFP and BOR1 K590A -GFP was observed in the PM. After ϩB treatment, GFP fluorescence in the BOR1-GFPexpressing plant was observed in the vacuole as reported previously (14), whereas fluorescence in the BOR1 K590A -GFPexpressing plant was not detected in the vacuole but was retained in the PM. This observation confirmed that the K590A mutation prevents sorting of BOR1 into the vacuole under a ϩB condition.
We also investigated the effect of wortmannin (WM) on trafficking of BOR1-GFP and BOR1 K590A -GFP. WM is a specific inhibitor of phosphatidylinositol 3-kinase and induces homotypic fusions and enlargement of the prevacuolar compartment/LE/MVB in plants (39 -41). In agreement with a previous study (40), BOR1-GFP was observed in several round structures after 90 min of treatment with 16.5 M WM both in ϩB and ϪB medium (Fig. 5, C and D) A part of the GFP signal in the round structures was co-localized with the endocytic dye FM4-64, indicating that the structures were enlarged prevacuolar compartment/LE/MVBs. In contrast, BOR1 K590A -GFP was rarely observed in enlarged MVBs. This result suggested that BOR1 K590A -GFP is scarcely able to enter the MVB pathway. Taken together, these data suggest that Lys-590 is not involved in the endocytosis of BOR1 from the PM but is essential for sorting in MVBs for subsequent vacuolar degradation (supplemental Fig. S6).

DISCUSSION
We showed that BOR1 is biochemically modified with mono-or diubiquitin in response to boron availability and that a Lys-590 residue is required for both boron-induced ubiquitination and boron-induced endocytic degradation. Mutational analysis showed that a K590A mutant, BOR1 K590A -GFP, was not degraded under ϩB conditions unlike BOR1-GFP (Fig. 3), indicating that Lys-590 is required for boron-induced endocytic degradation. Our most recent work revealed the other BOR1 residues involved in endocytic degradation (14): three putative tyrosine-based sorting signals (YXXØ, where Y is tyrosine, X is any amino acid, and Ø is a bulky hydrophobic residue) in the predicted largest cytosolic loop region. The tyrosine-based sorting signal is recognized by the clathrin-adaptor protein complex for sorting transmembrane proteins into clathrin-coated vesicles and is involved in many post-Golgi trafficking steps including endocytosis and polar PM trafficking (42). Mutations in all of the corresponding tyrosines (Y373A/Y398A/Y405A) of BOR1 prevent boron-induced degradation. Additionally, the tyrosine mutations inhibit polar localization in the root tip cells. In contrast, the K590A mutation inhibits only endocytic degradation and does not affect polar localization (Fig. 3), indicating that Lys-590 is specifically required for BOR1 vacuolar trafficking. We also showed that introducing an N590K or N590K/P591G mutation into the chimera 1-GFP and N596K or N596K/P597G into BOR4-GFP did not promote boron-induced degradation in either chimera 1-GFP or BOR4-GFP (supplemental Fig. S3), indicating that Lys-590 is necessary but not sufficient for endocytic degradation. Although BOR4 lacks two tyrosine residues corresponding to the putative tyrosine-based sorting signals of BOR1, Tyr-373 and Tyr-405 (supplemental Fig. S5), chimera 1-GFP mutants contain all of the putative tyrosinebased sorting signals. This suggests the existence of extra component(s) other than the three tyrosine-based sorting signals and indicates that Lys-590 functions in BOR1 endocytic degradation.
Immunoprecipitation and immunoblotting analyses showed that BOR1 is mono-or diubiquitinated in a ϩB-dependent manner and that Lys-590 is likely a principal ubiquitin-acceptor site (Fig. 4). Our data are consistent with studies in yeast showing that several nutrient transporters are posttranslationally down-regulated by substrate-dependent mono-or multi-monoubiquitination (22,34). Diverse forms of ubiquitination are involved in the distinct membrane protein degradation pathways in yeast and mammalian systems (19); polyubiquitination mainly regulates proteasomal degradation, and mono-or multi-monoubiquitination is required for the endocytic degradation pathway. In mammals, Lys-63linked short-chain ubiquitination and di-or triubiquitination at the Lys-270 lysine residue regulate endocytosis of the aquaporin-2 water channel (43). As a PM localized protein in plants, PIN2 has been shown to be ubiquitinated (26). Another report demonstrated that endocytically internalized PIN2 is degraded mainly in the vacuole (36). However, for PIN2, signal-inducing ubiquitination, which is a direct link between ubiquitination and turnover, the form of ubiquitination, and the ubiquitination site are still unknown. Another example is that down-regulation of PIP2, a PM aquaporin, is regulated via ubiquitination catalyzed by drought stressinduced Rma1, a homolog of a RING domain containing E3 ligase in A. thaliana (44). In addition, A. thaliana ironregulated transporter 1 (IRT1), a high affinity iron transporter, is also posttranslationally regulated in response to iron availability, and two lysine residues in the intercellular loop region are necessary for iron-induced turnover (45). However, IRT1 ubiquitination has not yet been reported. Our results show the ubiquitination form, an ubiquitination site, and inducibility by substrate availability (Fig. 4), thus filling a gap in the current knowledge regarding posttranslational turnover of PM proteins in plants. Our findings suggest that plants are able to use mono-or diubiq-uitination as critical labeling to identify endocytic degradation of PM proteins.
Our previous report showed that after treatment with ϩB for 30 -60 min, BOR1-GFP is translocated into dot-like structures corresponding to MVBs (18,23). However, BOR1 K590A -GFP was not observed as dot-like signals after ϩB treatment (Fig. 3). Furthermore, although BOR1-GFP was observed in enlarged MVBs following WM treatment even under a ϪB condition, BOR1 K590A -GFP was much less frequently observed in enlarged MVBs. This observation suggests that a pool of BOR1-GFP is transferred into MVBs even under a ϪB condition and that this MVB sorting is largely dependent on Lys-590.
In yeast, many PM proteins require mono-or multi-monoubiquitination for the endocytic internalization step (22). In contrast, ubiquitination of the epidermal growth factor receptor (EGFR) in animals is not necessary for endocytic internalization from the PM but is required for lysosomal sorting (46). Similar to the EGF receptor, Lys-590 of BOR1 seems not to be involved in endocytosis from the PM but is essential for transport to the MVB pathway and then to the vacuole ( Fig. 5 and supplemental Fig. S6). This result agrees with the observation that BOR1 K590A -GFP and wild-type BOR1-GFP showed inward polarity (Figs. 2 and 3), because endocytic cycling is required to maintain polar localization of PM proteins (14,42). Substrate-induced degradation of a number of yeast nutrient transporters is regulated at the endocytic internalization step (22). It is considered an efficient way because unicellular yeast does not require endocytically generated polarity of nutrient transporters for efficient nutrient uptake. However, plants might require continuous endocytic recycling of their nutrient transporters to maintain polar localization for directional nutrient transport under nutrient-deficient conditions. One may reasonably assume that plants have independent regulatory mechanisms for endocytosis (internalization from PM) and vacuolar sorting.
In conclusion, the data presented here show that boroninduced mono-or diubiquitination of BOR1 at Lys-590 is an essential transport step to MVBs and that ubiquitination does not control the rate of endocytosis from the PM (supplemental Fig. S6). This mechanism is distinct from that of the yeast system and may be important in enabling polar localized plant nutrient transporters to achieve a balance between the polar localization and the vacuolar degradation systems. BOR1 is a good model for studying endocytic degradation of transmembrane proteins in plants because degradation and ubiquitination can be controlled by only one amino acid residue. Further studies may show how plants sense the concentration of boron and regulate ubiquitination in a boron-dependent manner.