Chen HY , Chen YN , Wang HY , Liu ZT , Frommer WB , Ho CH .

105-2311-B-001-045 Ministry of Science and Technology, Taiwan, 106-2311-B-001-037-MY3 Ministry of Science and Technology, Taiwan, EXC-2048/1 - project ID 390686111 Deutsche Forschungsgemeinschaft, 1208 - Project-ID 267205415 Deutsche Forschungsgemeinschaft, Alexander von Humboldt Professorship Alexander von Humboldt-Stiftung

             ammonium import across plasma membrane

	Fig. 1. CIPK15 inhibits AMT1 activity in Xenopus oocytes. a, b Co-expression of CIPK15 inhibited NH4+-triggered inward currents of AMT1;1 in Xenopus oocytes. Oocytes were injected with water only, 50 ng cRNA of AMT1;1 only, 50 ng AMT1;1 + 50 ng CIPK15, or 50 ng CIPK15 only, and perfused with NH4Cl at the indicated concentrations (a) or (b) 0.2 mM for current recordings (a) and IV curve (b). Oocytes were voltage clamped at a − 120 mV or b − 40 mV and stepped in − 20-mV increments between − 20 and − 200 mV for 300 ms. b Currents (nA) were background subtracted (difference between currents at + 300 ms in the cRNA-injected AMT1;1 only/AMT1;1 + CIPK15/CIPK15 only and water-injected control of the indicated substrates). The data are the mean ± SE for three experiments. c TVEC traces of oocytes injected with water only, 5 ng cRNA of AMT1;1 only, or 5 ng AMT1;1 + 0.5 ng CIPK15, and perfused with NH4Cl at the indicated concentrations. Similar results were obtained in at least three independent experiments using different batches of oocytes
	Fig. 2. CIPK15 inhibited AmTryoshka1;3 LS-F138I activity in yeast. a Schematic representation of AmTryoshka1;3 LS-F138I [15]. b CIPK15 reduced NH4+-triggered AmTryoshka1;3 LS-F138I [15] responses in yeast. Amtryoshka1;3 LS-F138I was co-expressed with control (vector only), CIPK15, and CIPK15m (inactive mutant). Results of normalized fluorescence ratio (normalized to buffer control = 1, λexc 440 nm, ratio = FI510nm/570nm) after addition of NH4Cl as represented by box and whiskers (mean ± SE, n = 8). Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. p, significant change as shown in the figure (two-way ANOVA followed by Tukey’s post-test). PM, plasma membrane
	Fig. 3. NH4+-triggered CIPK15 mRNA accumulation. qRT-PCR analyses of AMT1;1 and CIPK15 mRNA levels in roots after over 10 h after addition of 1 mM NH4+. Levels were normalized to UBQ10 [mean ± SE for four independent experiments (each experiment n > 50, total n > 200)]. p, significant change for mRNA levels of AMT1;1 and CIPK15 at 1, 2, and 10 h compared to at 0 h (two-way ANOVA followed by Tukey’s post-test)
	Fig. 4. CIPK15 can interact with AMT1;1. Interaction growth assay (a), β-galactosidase staining (b), and filter (c) assay in yeast; and split-fluorescent protein interaction assay (d) in tobacco leaves for AMT1.1 and CIPK15 protein interactions. a Plasmids expressing AMT1;1 and CIPKs were expressed in yeast. Interaction indicated by growth on SD-Trp -Leu -His. Growth on SD-LT as control (Figure S5). Comparable results were obtained in three independent experiments. b, c Interaction of CIPKs and AMT1;1 in a split-ubiquitin system detected by X-Gal staining and filter assays using full-length AMT1;1-Cub-PLV as bait and NubG, NubI, and NubG-full-length CIPKs as prey. NubI/NubG served as positive (blue color) and negative controls, respectively. d Split-fluorescent protein interaction assay for AMT1;1 and CIPK15. YFP/chlorophyll, merged image of fluorescence and chloroplast. Reconstitution of YFP fluorescence from nYFP-AMT1;1 + CIPK15-cCFP and nYPF-AMT1;1 + cCFP (negative control). Comparable results with different combinations shown in Figure S7
	Fig. 5. AMT1;1-T460 phosphorylation is reduced in cipk15 mutant plants. Plant seedlings were germinated and grown for 7 days in half-strength MS medium with 5 mM KNO3 as the sole nitrogen source, then starved for 2 days in half-strength MS medium without nitrogen. Seedlings were treated with 1 mM NH4Cl for 1 h, membrane fractions were isolated and probed with anti-AMT1-P antibodies (a) and anti-AMT1;1 antibodies (b) [25]. Ponceau S staining served as a loading control. Quantification of phosphorylation of AMT1-P levels normalized to Ponceau S staining and relative to wild-type shown in c. Corresponding data and replications were obtained in three independent experiments. Data (c) are the mean ± SD for three experiments. p, significant change compared to wild-type as shown in figure (two-way ANOVA followed by Tukey’s post-test)
	Fig. 6. NH4+ content and transport in cipk15 mutants. Plant seedlings were germinated and grown for 7 days in half-strength MS medium with 5 mM KNO3 as the sole nitrogen source, then all seedlings were starved for 2 days in half-strength MS medium without nitrogen. For NH4+ content analyses, a seedlings were collected after being starved for 2 days (−N), or after with 1 mM NH4Cl (NH4+), and 1 mM KNO3 (NO3−) for 1 h. For 15N-labeled uptake, b seedlings were collected after being starved for 2 days (−N) (1 mM 15NH4Cl was used for 15 mins for 15N-labeling), or after treatment with 1 mM NH4Cl (NH4+), and 1 mM KNO3 (NO3−) for 1 h (1 mM 15NH4Cl was used for last 15 mins for N15-labeling for conditions of NH4+ and NO3−). Each data point represents different experiments, in which seedlings n > 15, total n > 60) in Col-0 and two cipk15 knockout mutants and presented as box and whiskers. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. p, significant change compared to before submergence (two-way ANOVA followed by Tukey’s post-test)
	Fig. 7. Hypersensitivity of cipk15 mutant plants to NH4+and MeA. Representative images (a) and quantification results of primary root length of plants grown on plates containing 20 mM NHCl4 (b) or 20 mM MeA (c). Primary root length in wild-type (Col-0), qko mutant, and cipk15 mutants on 20 mM NH4Cl (b) or on 20 mM MeA (c) are presented as box and whiskers. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots (means ± SE; n ≥ 15). p, significant change of qko mutant and cipk15 mutants compared to wild-type plants (two-way ANOVA followed by Tukey’s post-test). Scale bar: 0.1 cm
	Figure S1. Screen for CIPK effects on AMT1;1 activity in Xenopus oocytes. Different combinations of CIPKs were analyzed for effects on AMT1:1 transport activity. Shown are currents recorded in Xenopus oocytes injected with water (control), AMT1;1 alone, AMT1;1 + CIPK3/8/9/23/24, or AMT1;1 + CIPK2/10/15/20/26 (50 ng for each gene). cRNAs of AMT1;1 and CIPKs or water as control were injected into oocytes, subjected to TVEC and perfused with NH4Cl at the indicated concentrations. Oocytes were clamped at −120 mV (similar results were obtained in multiple independent experiments using different batches of oocytes). The results indicate that either one or several CIPKs in combination from the mixture of CIPK2, 10, 15, 20 and 26 may have impaired AMT1;1 activity, while CIPK3, 8, 9, 23 and 24 had no major impact on AMT1;1 activity.
	Figure S2. The activity of AMT1;1 co-expressed with CIPKs or CBL in Xenopus oocytes. Five nanograms of cRNA of AMT1;1, and different nanograms of cRNAs of CBL1, CIPK15, CIPK19-CA as indicated in the figure, or control (H2O) were injected into oocytes. AMT1;1 activity was considerably reduced when co-injected with different amounts of CIPK15 cRNA, whereas, the AMT1;1 activity was not affected when co-expressed with CBL1 and CIPK19-CA in Xenopus oocytes. Currents recorded in single oocytes injected with water, AMT1;1, AMT1;1 + CBL1, AMT1;1 + CIPK15, AMT1;1 + CBL1 + CIPK15, or AMT1;1 + CIPK19-CA. Oocytes were perfused with square pulses of 0.1 mM and 0.2 mM NH4Cl as indicated, respectively. Oocytes were clamped at −120 mV (independent data from three different oocytes were recorded from three different batches with comparable results).
	Figure S3. Protein gel blots for AMT1;1 protein level in Xenopus oocytes as control for Figure 1. Immunodetection of AMT1 protein using affinity-purified peptide antisera against a domain in the unphosphorylated cytosolic C-terminus (marked AMT1;1) [25]. 10% SDS PAGE. AMT1;1 protein levels were detected in membrane fractions of Xenopus oocytes, which were injected with water only, cRNA of AMT1;1 only (50 ng), AMT1;1 (50 ng) + CIPK15 (50 ng), and CIPK15 (50 ng) only. Ponceau S staining of filters before transfer served as the loading control. Comparable results were obtained in three independent experiments.
	Figure. S4. Alignment of the TMH XI and C terminus of five members of the AMT1 family. The C-terminal part of TMH XI is boxed in blue and the peptide sequence, specifically against the AMT1-P phospho-antiserum, is boxed in red. Phosphorylation sites of AMTs recognized by AMT1-P are boxed in green.
	Figure S5. CIPK19 has no effect on AmTryoshka1;3 LS-F138I activity in yeast. Unlike CIPK15, CIPK19 did not impair NH4+-triggered AmTryoshka1;3 LS-F138I [15] responses in yeast. Amtryoshka1;3 LS-F138I was co-expressed with control (vector only) and CIPK19. As described in Fig. 2, results of normalized fluorescence ratio (normalized to buffer control=1, λ exc 440 nm, ratio= FI510nm/570nm) after addition of NH4Cl are represented by box and whiskers (mean ± SE, n=8). Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. p, significant change as shown in the figure (Two-Way ANOVA followed by Tukey’s post-test).
	Figure S6. Mating-based split-ubiquitin assay, controls for Figure 4a. Growth assay in non-selective conditions on SD- Leu/Trp (LT) medium with identical cultures as shown in Fig. 4a and performed in parallel. Plasmids expressing AMT1;1, CIPK15, and CIPK19 were expressed in yeast. After overnight growth, colonies were serially diluted fourfold and grown for 2 days on SD-LT medium. As described in Fig. 4, full-length AMT1;1-Cub-PLV as bait and NubG, NubI, and NubG-full-length CIPK15 or CIPK19 as prey. No difference in growth of yeast was observed for yeast cells expressing AMT1;1-Cub control plasmids or NubG-CIPK15 or -CIPK19 expressing cells. Comparable results were obtained in three independent experiments.
	Figure S7. CIPK15 can interact with AMT1;1, AMT1;2, and AMT1;3 in yeast. Yeast mating-based split-ubiquitin assay for CIPK15 interaction with AMT1;1, AMT1;2, AMT1;3, or AMT1;5. Plasmids expressing AMT1;1, AMT1;2, AMT1;3, AMT1;5 with CIPK15 were expressed in yeast. Interactions were monitored using qualitative and quantitative ß-galactosidase activity assays. (a) Blue X-gal staining on plates, and (b) ß-galactosidase assay. Expression of pDL2-NubG was used as a negative control; pDL2-NubI was used as a positive control. For quantitative ß-galactosidase assays, cells were grown to an OD600 = 0.8. Activity was measured as described in the Materials and Methods section and normalized to NubI as 100% (mean ± SE). The numbers above the error bars represent the results of interaction strength after subtracting the NubG negative control (NubG + AMT1). Values from five samples were averaged for each bait-prey combination. p, significant change for ß-galactosidase activity compared to vector only (Two-Way ANOVA followed by Tukey’s post-test).
	Figure S8. Split-fluorescent protein interaction assay for AMT1;1 and CIPKs in Nicotiana benthamiana leaves. Split-fluorescent protein interaction assay as described in [52]. Different combinations of reconstitution of YFP fluorescence for AMT1;1, CIPK15, and CIPK19 (nYFP-AMT1;1 + cCFP, nYFP + AMT1;1-cCFP, nYFP-CIPK15 + cCFP, nYFP + CIPK15-cCFP, nYFP-AMT1;1 + CIPK15-cCFP, nYFP-CIPK15 + AMT1;1-cCFP, nYFP-CIPK15 + cCFP-AMT1;1, AMT1;1-nYFP + CIPK15-cCFP, nYFP-AMT1;1 + CIPK19-cCFP) are shown. Chloroplast panels show chlorophyll autofluorescence; YFP/DIC, merged image of fluorescence, chloroplast, and brightfield images.
	Figure S9. Arabidopsis T-DNA insertion mutants of cipk15-1 and cipk15-2. (a) Schematic map of the cipk15-1 and cipk15-2 mutants showing the positions of T-DNA insertions in CIPK15. F, CIPK15 forward primer; R, CIPK15 reverse primer. Gene transcribed from left to right. Solid black box: exon, white boxes 5′-UTR and 3′-UTR, respectively. cipk15-1 carries an insertion in the 5′-UTR, cipk15-2 carries an insertion in the coding region. (b) RT-PCR analysis of CIPK15 transcript levels in cipk15-1 and cipk15-2 mutant lines. H3G1 (At4g40040) was used as the loading control. None of the mutant lines showed detectable amounts of CIPK15 mRNA, indicating that both are knockout mutants. Note that CIPK15 contains only a single intron in the 5′-UTR. (c) Col-0 and cipk15 mutants showed no obvious growth differences when grown in soil or axenically in MS medium. Representative images of rosette leaves of Col-0 and cipk15 mutants. Plants were grown in soil under MS medium in a 16/8 h light/dark period at 22ºC for 17 days. Scale bar: 1cm.
	Figure S10. Protein gel blots for AMT1;1 protein and AMT-P phosphorylation levels in wild-type and cipk15 mutant plants under half-strength MS medium. Immunodetection of AMT1 protein using affinity-purified peptide antisera against a domain in the unphosphorylated cytosolic C-terminus (marked AMT1;1) and affinity-purified antisera against the phosphorylated cytosolic C-terminus (marked AMT1-P) [25]. 10% SDS PAGE. AMT1;1 protein and AMT-P levels were detected in roots of plants grown on half-strength MS medium for 7 days as described [25]. Ponceau S staining of filters before transfer served as the loading control. Comparable results were obtained in three independent experiments.
	Figure S11. Primary root length of control (wild-type) and qko mutant on half-strength MS medium containing NH4Cl, KNO3, or MeA. Seedlings were grown on half-strength MS medium containing NH4Cl, KNO3 or on MeA. Primary root length was measured in wild-type (Col-0) and qko mutant. (a and b) Primary root length 1, 2, or 3 days after transferring wild-type and qko mutant plants to 20 mM NH4Cl, MeA or KNO3. Data are mean ± SE; n ≥5. p, significant change between wild-type and qko mutant (Two-Way ANOVA followed by Tukey’s post-test).
	Figure S12. Primary root length of Col-0, qko, and cipk15 mutant plants on media containing KNO3 as sole nitrogen source. Representative images (a) and quantification results of primary root length of plants grown on plates containing 20 mM KNO3. Scale bar: 0.1 cm. Primary root length in wild-type (Col-0), qko mutant and cipk15 mutant plants on 20 mM KNO3 (n =11) presented as box and whiskers. Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. Data are mean ± SE. No significant change between wild-type, qko, and cipk15 mutants (Two-Way ANOVA followed by Tukey’s post-test).
	Figure S13. cbl4 and cipk19 mutants do not show ammonium hypersensitivity. Primary root length of control (Col-0), cbl4 (At5g24270, SALK_113101) [48] and cipk19 (At5g45810, SALK_044735) [47] knockout mutant plants on half-strength MS medium containing NH4Cl and KNO3. Seedlings were germinated and grown on half-strength MS medium containing 5 mM KNO3 for 3 days, and then, primary root length of wild-type (Col-0) and mutants were measured after transferral to 20 mM NH4Cl or KNO3 conditions for 5 days and presented as box and whiskers. Data are mean ± SE; n=14. p, significant difference between NH4Cl and KNO3 (Two-Way ANOVA followed by Tukey’s post-test). The center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by Prism software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles.
	Figure S14. CIPK23 mRNA accumulated by NH4+. qRT-PCR analyses of CIPK23 mRNA levels in roots after over 10 h after addition of 1 mM NH4+ in Col-0. Levels were normalized to UBQ10 [mean ± SE for four independent experiments (each experiment n >50, total n >200)]. p, significant change in mRNA levels of CIPK23 at 1, 2, and 10 h compared to at 0 h (Two-Way ANOVA followed by Tukey’s post-test).

Albrecht, Glutamine: a Trojan horse in ammonia neurotoxicity. 2006, Pubmed

Albrecht, Glutamine: a Trojan horse in ammonia neurotoxicity. 2006, Pubmed
Ast, Ratiometric Matryoshka biosensors from a nested cassette of green- and orange-emitting fluorescent proteins. 2017, Pubmed
Brito, Yeast filamentation signaling is connected to a specific substrate translocation mechanism of the Mep2 transceptor. 2020, Pubmed , Xenbase
Cheong, Two calcineurin B-like calcium sensors, interacting with protein kinase CIPK23, regulate leaf transpiration and root potassium uptake in Arabidopsis. 2007, Pubmed
Dennison, Functions of AKT1 and AKT2 potassium channels determined by studies of single and double mutants of Arabidopsis. 2001, Pubmed
Di Cera, A structural perspective on enzymes activated by monovalent cations. 2006, Pubmed
Duan, Root zone-specific localization of AMTs determines ammonium transport pathways and nitrogen allocation to shoots. 2018, Pubmed
Elmlinger, Glutamine synthetase in Scots pine seedlings and its control by blue light and light absorbed by phytochrome. 1992, Pubmed
Fuglsang, Arabidopsis protein kinase PKS5 inhibits the plasma membrane H+ -ATPase by preventing interaction with 14-3-3 protein. 2007, Pubmed
Fukao, A variable cluster of ethylene response factor-like genes regulates metabolic and developmental acclimation responses to submergence in rice. 2006, Pubmed
Gazzarrini, Three functional transporters for constitutive, diurnally regulated, and starvation-induced uptake of ammonium into Arabidopsis roots. 1999, Pubmed
Gojon, Nitrate Reduction in Roots and Shoots of Barley (Hordeum vulgare L.) and Corn (Zea mays L.) Seedlings: I. N Study. 1986, Pubmed
Guo, A calcium sensor and its interacting protein kinase are global regulators of abscisic acid signaling in Arabidopsis. 2002, Pubmed
Hess, Ammonium toxicity and potassium limitation in yeast. 2006, Pubmed
Ho, CHL1 functions as a nitrate sensor in plants. 2009, Pubmed
Jones, Border control--a membrane-linked interactome of Arabidopsis. 2014, Pubmed
Kim, Xanthomonas T3S Effector XopN Suppresses PAMP-Triggered Immunity and Interacts with a Tomato Atypical Receptor-Like Kinase and TFT1. 2009, Pubmed
Kolukisaoglu, Calcium sensors and their interacting protein kinases: genomics of the Arabidopsis and rice CBL-CIPK signaling networks. 2004, Pubmed
Kronzucker, Ammonium toxicity and the real cost of transport. 2001, Pubmed
Lalonde, A membrane protein/signaling protein interaction network for Arabidopsis version AMPv2. 2010, Pubmed
Lanquar, Feedback inhibition of ammonium uptake by a phospho-dependent allosteric mechanism in Arabidopsis. 2009, Pubmed
Leisgen, The roboocyte: automated electrophysiology based on Xenopus oocytes. 2007, Pubmed , Xenbase
Leleu, Unusual regulatory nitrate reductase activity in cotyledons of Brassica napus seedlings: enhancement of nitrate reductase activity by ammonium supply. 2004, Pubmed
Lemaire, Glucose and sucrose act as agonist and mannose as antagonist ligands of the G protein-coupled receptor Gpr1 in the yeast Saccharomyces cerevisiae. 2004, Pubmed
Lima, Ammonium triggers lateral root branching in Arabidopsis in an AMMONIUM TRANSPORTER1;3-dependent manner. 2010, Pubmed
Loqué, Pore mutations in ammonium transporter AMT1 with increased electrogenic ammonium transport activity. 2009, Pubmed , Xenbase
Loqué, A cytosolic trans-activation domain essential for ammonium uptake. 2007, Pubmed
Ludewig, Molecular mechanisms of ammonium transport and accumulation in plants. 2007, Pubmed
Ludewig, Homo- and hetero-oligomerization of ammonium transporter-1 NH4 uniporters. 2003, Pubmed , Xenbase
Ludewig, Uniport of NH4+ by the root hair plasma membrane ammonium transporter LeAMT1;1. 2002, Pubmed , Xenbase
Maierhofer, Site- and kinase-specific phosphorylation-mediated activation of SLAC1, a guard cell anion channel stimulated by abscisic acid. 2014, Pubmed , Xenbase
Marini, Cross-talk between ammonium transporters in yeast and interference by the soybean SAT1 protein. 2000, Pubmed
Marini, Cloning and expression of the MEP1 gene encoding an ammonium transporter in Saccharomyces cerevisiae. 1994, Pubmed
Ninnemann, Identification of a high affinity NH4+ transporter from plants. 1994, Pubmed
Pehl, Automated higher-throughput compound screening on ion channel targets based on the Xenopus laevis oocyte expression system. 2004, Pubmed , Xenbase
Qin, CALCIUM-DEPENDENT PROTEIN KINASE 32-mediated phosphorylation is essential for the ammonium transport activity of AMT1;1 in Arabidopsis roots. 2020, Pubmed , Xenbase
Qiu, Regulation of SOS1, a plasma membrane Na+/H+ exchanger in Arabidopsis thaliana, by SOS2 and SOS3. 2002, Pubmed
Ramer, Dominant genetics using a yeast genomic library under the control of a strong inducible promoter. 1992, Pubmed
Rogato, Characterization of a developmental root response caused by external ammonium supply in Lotus japonicus. 2010, Pubmed
Schiestl, High efficiency transformation of intact yeast cells using single stranded nucleic acids as a carrier. 1989, Pubmed
Slayman, Net uptake of potassium in Neurospora. Exchange for sodium and hydrogen ions. 1968, Pubmed
Song, Role of an Arabidopsis AP2/EREBP-type transcriptional repressor in abscisic acid and drought stress responses. 2005, Pubmed
Straub, The Kinase CIPK23 Inhibits Ammonium Transport in Arabidopsis thaliana. 2017, Pubmed , Xenbase
Svoboda, L-Glutamine-induced apoptosis in microglia is mediated by mitochondrial dysfunction. 2009, Pubmed
Tang, A calcium signalling network activates vacuolar K+ remobilization to enable plant adaptation to low-K environments. 2020, Pubmed
Tang, Tonoplast CBL-CIPK calcium signaling network regulates magnesium homeostasis in Arabidopsis. 2015, Pubmed
Tremblay, Of blood, brains and bacteria, the Amt/Rh transporter family: emerging role of Amt as a unique microbial sensor. 2009, Pubmed
Wang, Ammonium Uptake by Rice Roots (II. Kinetics of 13NH4+ Influx across the Plasmalemma). 1993, Pubmed
Wu, Ammonium and nitrate regulate NH4+ uptake activity of Arabidopsis ammonium transporter AtAMT1;3 via phosphorylation at multiple C-terminal sites. 2019, Pubmed
Xu, A protein kinase, interacting with two calcineurin B-like proteins, regulates K+ transporter AKT1 in Arabidopsis. 2006, Pubmed , Xenbase
Yuan, The organization of high-affinity ammonium uptake in Arabidopsis roots depends on the spatial arrangement and biochemical properties of AMT1-type transporters. 2007, Pubmed
Zhou, A calcium sensor-regulated protein kinase, CALCINEURIN B-LIKE PROTEIN-INTERACTING PROTEIN KINASE19, is required for pollen tube growth and polarity. 2015, Pubmed