Full Text

Turn on search term navigation

ARTICLE

Received 22 Sep 2016 | Accepted 16 Mar 2017 | Published 15 May 2017

DOI: 10.1038/ncomms15300 OPEN

Low phosphate activates STOP1-ALMT1 to rapidly inhibit root cell elongation

Coline Balzergue1,2,3,*,w, Thibault Dartevelle1,2,3,*, Christian Godon1,2,3,*, Edith Laugier1,2,3,*,Claudia Meisrimler1,2,3,*,w, Jean-Marie Teulon4,5,6, Audrey Creff1,2,3,w, Marie Bissler1,2,3, Corinne Brouchoud1,2,3, Agns Hagge7,w, Jens Mller8, Serge Chiarenza1,2,3, Hlne Javot1,2,3, Nolle Becuwe-Linka1,2,3,Pascale David1,2,3, Benjamin Pret1,2,3,w, Etienne Delannoy1,2,3,w, Marie-Christine Thibaud1,2,3, Jean Armengaud9, Steffen Abel8, Jean-Luc Pellequer4,5,6, Laurent Nussaume1,2,3 & Thierry Desnos1,2,3

Environmental cues profoundly modulate cell proliferation and cell elongation to inform and direct plant growth and development. External phosphate (Pi) limitation inhibits primary root growth in many plant species. However, the underlying Pi sensory mechanisms are unknown. Here we genetically uncouple two Pi sensing pathways in the root apex of Arabidopsis thaliana. First, the rapid inhibition of cell elongation in the transition zone is controlled by transcription factor STOP1, by its direct target, ALMT1, encoding a malate channel, and by ferroxidase LPR1, which together mediate Fe and peroxidase-dependent cell wall stiffening. Second, during the subsequent slow inhibition of cell proliferation in the apical meristem, which is mediated by LPR1-dependent, but largely STOP1ALMT1-independent, Fe and callose accumulate in the stem cell niche, leading to meristem reduction. Our work uncovers STOP1 and ALMT1 as a signalling pathway of low Pi availability and exuded malate as an unexpected apoplastic inhibitor of root cell wall expansion.

1 Laboratoire de Biologie du Dveloppement des Plantes, Institut de Biosciences et Biotechnology Aix-Marseille, Commissariat lEnergie Atomique et aux nergies alternatives, Saint-Paul-Lez-Durance 13108, France. 2 Centre National de la Recherche Scientique, UMR 7265 Biol. Vgt. & Microbiol. Environ., Saint-Paul-Lez-Durance, France. 3 Aix-Marseille Universit, UMR 7265, Marseille, France. 4 CNRS, IBS, Grenoble F-38044, France. 5 CEA, IBS, Grenoble F-38044, France. 6 Universit Grenoble Alpes, IBS, Grenoble F-38044, France. 7 Commissariat lEnergie Atomique et aux nergies alternatives, Service de Biologie et de Toxicologie Nuclaire, Laboratoire dEtude des Protines Cibles, 30200 Bagnols sur Cze, France. 8 Department of Molecular Signal Processing, Leibniz Institute of Plant Biochemistry, Halle (Saale) 06120, Germany. 9 CEA, DRF, JOLIOT/DMTS/SPI/Li2D, Laboratory Innovative Technologies for Detection and Diagnostics, Bagnols-sur-Cze F-30200, France. * These authors contributed equally to this work. w Present addresses: Laboratoire de

Recherche en Sciences Vgtales, Universit de Toulouse, CNRS, UPS, 24 chemin de Borde Rouge, Auzeville, BP42617, 31326 Castanet Tolosan, France (C.B.); Plant-Microbe Interactions, Department of Biology, Faculty of Science, Utrecht University, Padualaan 8, Hugo R. Kruytgebouw, 3584 CH Utrecht, The Netherlands (C.M.); Laboratoire Reproduction et Dveloppement des Plantes, ENS de Lyon, UCB Lyon 1, CNRS, INRA, F-69342 Lyon, France (A.C.); CNRS, Universit Claude Bernard Lyon 1, Ens de Lyon, Institut des Sciences Analytiques, UMR 5280, 5 rue de la Doua, F-69100 Villeurbanne, France (A.H.); Centre National de la Recherche Scientique, Equipe Dveloppement et Plasticit du Systme Racinaire, UMR Biochimie et Physiologie Molculaire des Plantes, Institut de Biologie Intgrative des Plantes, Campus INRA/Montpellier SupAgro, 2 Place Pierre Viala, 34060 Montpellier, France (B.P.); Institute of Plant Sciences Paris Saclay IPS2, CNRS, INRA, Universit Paris-Sud, Universit Evry, Universit Paris-Saclay, Btiment 630, 91405 Orsay, France et Institute of Plant Sciences Paris-Saclay IPS2, Paris Diderot, Sorbonne Paris-Cit, Btiment 630, 91405 Orsay, France (E.D.). Correspondence and requests for materials should be addressed to T.D. (mailto:[email protected]

Web End [email protected]).

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

Phosphate limitation ( Pi), prevailing in around 70% of global cultivated land, is a common abiotic stress that profoundly alters root growth with critical consequences for

crop yields1,2. However, the underlying mechanisms of root growth inhibition by low Pi availability remain largely unknown.

To address this problem at the cellular and molecular level, we chose the primary root of Arabidopsis thaliana seedlings as the experimental system because it displays a transparent and simple anatomical structure. Roots longitudinally and indeterminately grow at their apex. Decades of laboratory work have shown that cells proliferate in the root apical meristem (RAM) in concentric layers (epidermis, cortex and endodermis surrounding the stele). These cell types originate in the stem cell niche (SCN), enclosing the quiescent centre (QC), divide as transit-amplifying cells before rapid and extensive cell expansion ensues in the elongation zone (EZ)3,4 (Supplementary Fig. 1a). A complex set of regulatory networks and developmental genes coordinate cell proliferation and cell elongation required for root extension4.

Previous studies with Arabidopsis seedlings showed that acute Pi stress is locally sensed at the primary root apex in

a Fe-dependent manner and inhibits root growth by affecting both cell elongation and cell proliferation512. The local response is independent of internal root Pi status13 and is therefore not caused by a metabolic deciency syndrome.

To date, several mutants specically altered in the local root response have been isolated5,6,1418, but only a few corresponding mutated genes have been identied. Much has been learned from LOW PHOSPHATE ROOT 1 (LPR1) and PHOSPHATE DEFICIENCY RESPONSE 2 (PDR2), which genetically interact and are expressed in cell type-specic but overlapping domains of the root apex, comprising the RAM and early EZ8,9,12. While LPR1 encodes a cell wall (CW) targeted multicopper oxidase with ferroxidase activity, PDR2 codes for AtP5A, the single P5-type ATPase in Arabidopsis, which is thought to repress LPR1 biogenesis or activity8,9. Under Pi, the genetic PDR2LPR1

module mediates accumulation of Fe in the SCN and EZ within 20 h, which triggers generation of reactive oxygen species (ROS) as well as deposition of callose and pectin in CWs of the RAM and EZ12,19. Callose deposition in the SCN obstructs cell-to-cell communication as demonstrated by impaired movement of SHORT-ROOT, a key transcription factor of root patterning and cell fate specication4, which is followed by reduced stem cell maintenance and RAM activity12.

Here we identify and characterize two key players of a new signalling pathway, which is activated under Pi to rapidly

inhibit root growth by targeting cell expansion in the EZ: the transcription factor STOP1 (SENSITIVE TO PROTON TOXI-CITY1) and the malate efux channel ALUMINUM-ACTIVATED MALATE TRANSPORTER1 encoded by ALMT1, a direct STOP1 target gene. Surprisingly, mutations in either gene spatially and temporally uncouple the rapid inhibition of cell elongation in the EZ (o2 h) from the slower inhibition of cell proliferation in the SCN and RAM (o4 days), which are both dependent on LPR1 ferroxidase activity. Based on our data, we propose that malate efux by internal cell les of the root tip facilitates Fe chelation in the apoplast. LPR1-dependent Fe redox cycling stimulates peroxidase-catalysed CW stiffening followed by inhibition of root cell elongation.

Results

STOP1 and ALMT1 repress primary root growth under Pi.

To discover new genes important for primary root growth inhibition under Pi, we screened B4 105 ethylmethane

sulfonate-mutagenized WT seedlings (see Methods). We isolated 85 independent mutants with a signicantly longer primary root

when grown under Pi. Genetic mapping and molecular

analysis identied 26 lpr1 alleles, suggesting that our screen approached saturation for isolating complete or partial loss-of-function mutations conferring a lpr1-like root phenotype. We also identied 24 mutant alleles of the transcription factor STOP1 and 14 alleles of the malate efux channel ALMT1 (see Methods, Fig. 1ac and Supplementary Fig. 1b), we describe here a representative subset of both complementation groups. All almt1 and stop1 mutations are recessive, except for Stop125 and Stop148, which are dominant negative mutations displaying the same phenotype as seedlings homozygous for stop1 null alleles (Supplementary Fig. 1c,d). The presumably truncated mutant STOP1 proteins dominantly inhibit ALMT1 expression, possibly by titrating a limiting factor required for ALMT1 activation.

ALMT1 promotes primary root growth arrest under Pi.

Previous studies showed that STOP1 promotes root exudation of malate and citrate by stimulating expression of ALMT1 and MATE efux transporters, respectively2024. Root exudation of organic acids such as malate and citrate is thought to improve plant growth by solubilizing Pi anions complexed with metal cations like iron (Fe) and aluminium (Al) in soil25,26. We were therefore intrigued by the observation that stop1 and almt1 mutants develop a longer primary root than the WT in Pi

conditions. We thus analysed in more detail the role of STOP1 and ALMT1 in Pi-dependent root growth inhibition.

First, chemical complementation by exogenous malate application restored in a dose-dependent manner primary root growth inhibition on Pi medium in almt1 mutants (Fig. 1d); whereas

no reduced root growth was observed in Pi conditions

(Supplementary Fig. 2a). This result is consistent with the hypothesis that the ALMT1-dependent malate exudation inhibits WT root growth under Pi.

We then tested whether expression of ALMT1 and STOP1 is regulated by Pi availability. One hour after transfer from the Pi

to Pi medium the expression level of ALMT1 increases and

reaches a plateau at 24 h (Fig. 1e). However, on both media ALMT1 expression is abrogated in strong stop1 mutants (Supplementary Fig. 2b). MATE is also upregulated on Pi

medium (Fig. 1e) in a STOP1-dependent manner (Supplementary Fig. 2b). However, under Pi the mateKO mutant does not grow

a long primary root, and the main root of the almt161;mateKO double mutant is as long as the root of the almt161 single mutant (Supplementary Fig. 2c). The pattern of MATE expression in the primary root is different from that of ALMT1, in particular it is not expressed in the root tip27, and this may explain why the mateKO single mutant is not altered in the root growth under

Pi. We thus tested whether expressing MATE under the control of the ALMT1 promoter could restore root growth inhibition by Pi of the almt1KO;mateKO double mutant. First,

we conrmed that in this line27 the pALMT1::MATE construct is functional and complements the Al3 sensitivity of the almt1KO;mateKO double mutant (Supplementary Fig. 2d).

However, expression of pALMT1::MATE in almt1KO;mateKO plants did not restore primary root growth inhibition on Pi

medium (Supplementary Fig. 2d). Together, these results indicate that citrate exudation by MATE cannot replace malate exudation by ALMT1 to inhibit primary root growth under Pi.

A prominent role of ALMT1 as an effector downstream of STOP1 for root growth inhibition under Pi is supported

by the following observations: (i) the root length of stop1 and almt1 mutants are similar (Fig. 1c), and (ii) the constitutive expression of ALMT1 fully rescues both almt132 (Supplementary Fig. 2e) and the dominant negative Stop148 mutant phenotypes (Supplementary Fig. 2f).

2 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

We quantitatively showed in yeast one-hybrid assays that STOP1 recognizes a small region of the ALMT1 promoter including the STOP1-binding site found previously by others28 (Supplementary Fig. 2g). Interestingly, the low Pi-stimulated expression of ALMT1 is independent of PHR1 and PHL1 (Supplementary Fig. 3a), two master transcription factors controlling many Pi-starvation responses29.

Pi enhances expression of ALMT1 in the root apex. To determine the spatial expression pattern of ALMT1 and STOP1, we established several independent transgenic lines carrying the GUS reporter gene placed under the control of their respective promoters (pALMT1::GUS and pSTOP1::GUS). The pALMT1::-GUS lines (WT background) showed ALMT1 expression in the whole root (Supplementary Fig. 3b). This histological marker

ALMT1 STOP1

+Pi Pi

Black = missence mutations Blue = splicing mutations Red = nonsense mutations

almt169 (M1I)

almt145 (G71E)

almt161

almt1140 (C176Y)

almt1154

almt151

almt11

almt132

stop160 , stop167

stop1127

stop1115

stop1136 (H262Y)

stop1148 (G265R)

stop1139 (E353K)

stop1141

Stop125

Stop148

ATG TAA

stop190

ATG TAA

b c

almt1 alleles stop1 alleles

almt1

amt1

Stop1

stop1

Primary root length (mm)

****

Col

+Pi

Coler105

almt1KO

almt11

almt132

almt145

almt151

almt161

almt169

almt1140

almt1154

stop1KO

Stop125

Stop148

stop160

stop167

stop190

stop1115

stop1136

stop1139

stop1141

stop1148

almt1 alleles stop1 alleles

****

** NS

***

**** ****

**** **** ****

Relative expression

er105

Primary root length (mm)

+Pi

STOP1 ALMT1

0 50 100 150 200

Malate (M)

1 6 24 48

MATE

SPX1

Time after transfer (h)

+Pi

pSTOP1::GFP-STOP1

+Pi +Pi

****

pALMT1::GUS #2

Fluorescence intensity (a.u.)

+ Pi

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

Root growth arrest under Pi is rapid. To characterize with

higher temporal resolution, the early phase of primary root growth inhibition under Pi, we performed a time course

experiment. WT seedlings grown under Pi were transferred to

either Pi (control) or Pi agar before primary root length was

measured at 20 min intervals for 7 h (Fig. 2a). When compared to the control, we detected root growth inhibition 2 h after transfer to Pi medium, a fast morphological response (Fig. 2b). In

contrast, the main root of stop1, almt1 and lpr1;lpr2 mutants did not respond during the entire 7.5 h time course (Fig. 2c). The short primary root length of WT under Pi correlated with

a strong reduction of root epidermal cell length (Fig. 2d). In the mutants, epidermal cell length was not or only slightly reduced under Pi (Fig. 2d). Our data show that STOP1, ALMT1 and the

LPRs mediate a rapid inhibition of root cell elongation that contributes to root growth inhibition on Pi medium.

stop1 and almt1 uncouple the Pi response of the RAM and EZ.

Previous work showed that the root growth arrest under Pi is

a Fe-dependent response7,8,12. Inductively coupled plasma mass spectrometry analysis did not show signicant differences in total Fe accumulation in roots of mutants compared to WT (Supplementary Fig. 7a). However, using histological Perls/DAB (3,30-diaminobenzidine) staining33, it was reported that labile (non-haem) Fe accumulates in Pi-deprived WT but not lpr1;lpr2 root tips12,19. We thus tested whether the stop1 and almt1 mutations alter the pattern of Fe accumulation. By using Perls/DAB staining, we conrmed that WT but not lpr1;lpr2 root tips accumulated Fe in the EZ and SCN on Pi medium

(Fig. 3a and Supplementary Fig. 7b). Interestingly, stop1 and almt1 mutants did not show Fe accumulation in the EZ (Fig. 3a and Supplementary Fig. 7b), but, to our surprise, both mutant lines still accumulated Fe in the SCN similar to the WT (Fig. 3a and Supplementary Fig. 7b).

Callose accumulation in the EZ and SCN is another root response of the WT, but not of the lpr1;lpr2 double mutant, when grown under Pi (ref. 12; Fig. 3b). Mirroring the

Fe accumulation (Fig. 3a), the stop1 and almt1 mutants still deposited callose in the SCN but not in the EZ (Fig. 3b). Note that exogenous malate application complemented iron and callose deposition in the EZ of the almt1 root tip (Supplementary Fig. 8).These unexpected observations indicate that loss of STOP1 or ALMT1 uncouples the two major regions of Fe and callose accumulations in the root apex.

Because inhibition of cell elongation under Pi is a rapid

process (Fig. 2b) and occurs well before the onset of root meristem size reduction12, we hypothesized that primary root

Figure 1 | Phenotype of almt1 and stop1 mutants and expression of ALMT1. (a) Position and nature of the almt1 and stop1 mutations. (b) Appearance of representative WT and mutant seedlings after germination and growth for 6 days on Pi or Pi medium. (c) Primary root length of WT (Coler105) and

mutant seedlings grown for 6 days in Pi or Pi conditions. Means.d., n 914 seedlings per line and condition; unpaired two-tailed t-test;

****Po0.0001). (d) Malate complementation. WT and almt1 seedlings grown under Pi for 2 days were transferred for 6 days to Pi medium with or

without the indicated concentrations of malate (means.d., n 714 seedlings per condition; two-way ANOVA, ****Po0.0001; ***Po0.001; **Po0.01;

*Po0.05; NS, not signicant). (e) Gene expression analysis. WT seedlings grown under Pi for 5 days were transferred to Pi or Pi medium for

148 h, total root RNA was extracted and expression of the indicated genes analysed by qRTPCR. SPX1 was used as a positive marker for the Pi stress.

Relative expression levels are normalized to that of the Pi at 1 h (means.e.m., n 6 independent experiments; unpaired two-tailed t-test; **Po0.01;

*Po0.05; NS, not signicant). (f) Pattern of ALMT1 expression in primary root tip. Transverse (top) and longitudinal (bottom) sections of the root tip of pALMT1::GUS (WT). Seedlings were grown 5 days under Pi and transferred to Pi or Pi medium for 48 h prior to GUS staining. (g) GFP-uorescence

of GFP-STOP1 in root tips. Four-day-old stop1KO;pSTOP1::GFP-STOP1 (# B10A) seedlings grown under Pi were transferred to Pi or Pi plates for 24 h,

and GFP-uorescence was visualized by confocal microscopy. Left panel: the two roots were mounted side by side on the same microscope slide( 10 objective). Middle panel: magnication of the root tips ( 20 objective). See also Supplementary Fig. 6c for additional lines. Right panel: intensity of

the GFP-uorescence (a.u., arbitrary units) in nuclei of the root tip (means.d., n 54 and 43 nuclei in Pi and Pi, respectively; unpaired two-tailed

t-test; ****Po0.0001). Two (c,d,f) or three (g) independent experiments were performed with consistent results and one representative experiment is shown. Scale bars, 1 cm (b), 100 mm (f,g).

likely reports the genuine expression pattern of endogenous ALMT1, conrming our quantitative PCR with reverse transcription (qRT PCR) analysis (Fig. 1e). As shown in Fig. 1f

(and Supplementary Fig. 3b), pALMT1::GUS was more highly expressed in Pi than in Pi conditions, but was not detectable

in the Stop148 mutant background (Supplementary Fig. 3c).

We previously showed that the root growth arrest does not occur under Pi at pH 46.3, or when iron is omitted in the

medium8. We therefore assessed whether this observation is due to the lack of ALMT1 expression. Our results revealed that ALMT1 expression was not detectable in roots on Pi medium

at pH 7.1 (Supplementary Fig. 4a), whereas Fe omission did not impair stimulation of ALMT1 expression by Pi (Supple

mentary Fig. 4a). On the other hand, WRKY46, a transcription factor whose expression is enhanced in roots by Fe deciency30 and which represses ALMT1 expression31, does not seem to modulate the root growth inhibition under Pi because the

wrky46KO mutant behaved like the WT control (Supplementary Fig. 5). These results support the view that Fe does not play a crucial role in the induction of ALMT1 by Pi.

As shown in longitudinal and transverse sections of the root meristem (Fig. 1f), ALMT1 is expressed during Pi limitation in

the SCN, including the QC, and in all cell les, although at a lower level in the external root cap layer and columella cells. We therefore demonstrated here that, in addition to other cues32,

Pi stimulates ALMT1 expression in the root tip.

The pSTOP1::GUS marker revealed that STOP1 is expressed in the entire root tip (Supplementary Fig. 6a); however, in contrast to ALMT1, STOP1 expression is not enhanced by Pi exposure

(Fig. 1e and Supplementary Figs 2b, 3a and 6a). This suggests that STOP1 is post-transcriptionally regulated to stimulate ALMT1 expression under Pi. We thus tested whether nuclear

localization of STOP1 is regulated. We fused the Green Fluorescent Protein (GFP) to the STOP1 N terminus and the GFP-STOP1 coding region was placed under control of the pSTOP1 or pUBQ10 promoter. The translational fusions are functional because they complement the root growth response of the stop1KO mutant (Supplementary Fig. 6b). Confocal microscopy of root tips revealed nuclear localization of GFP-STOP1 under Pi (Fig. 1g and Supplementary Fig. 6c), which

interestingly intensied on seedling transfer to Pi medium

(Fig. 1g and Supplementary Fig. 6c). These results indicate that STOP1 is post-translationally regulated and suggest that the stimulation of ALMT1 expression under Pi is, at least partly,

due to increased accumulation of STOP1 in the nucleus. Overall, our results show that the STOP1ALMT1 module is a component of a new, PHR1-independent regulatory pathway activated by

Pi stress.

4 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

a b

Transfer

Coler105

2,500

600

2,000

Root length (m)

1,500

****

Root length

1,000

500

110

130

150

170

190

210

230

250

270

290

310

330

350

370

390

410

430

450

470

**** **** ********

****

Time after transfer (min)

700

Epidermal cell length (m)

600

500

400

300

200

100

er105

2,500

2,000

****

**** ****

Root length (m)

1,500

1,000

500

0 + + + + + +

n= 137

n= 423

n= 92

n= 135

n= 100

n= 133

n= 121

n= 144

n= 96

n= 187

n= 258

n= 199

110

130

150

170

190

210

230

250

270

290

310

330

350

370

390

410

430

450

470

Time after transfer (min)

CoIer105

Ipr1,Ipr2

Stop148

stop167

aImt132

aImt1KO

Figure 2 | Rapid primary root growth arrest under Pi. (a) Scheme depicting the D primary root length measured in seedlings transferred from a Pi plate to a Pi or Pi medium. (b) Time course of the D primary root length of WT seedlings transferred to Pi or Pi medium. Inset shows

magnication of the rst time points. Three independent experiments were performed with consistent results and one representative experiment is shown (means.e.m., n 10 seedlings per condition; two-way ANOVA, multiple comparisons; ****Po0.0001). (c) Time course of the D primary root length of

WT and mutant seedlings transferred to Pi. Two independent experiments were performed with consistent results and one representative experiment is

shown (means.e.m., n 810 seedlings per line and condition; two-way ANOVA, multiple comparisons; ****Po0.0001). (d) Seven-day-old seedlings of

the indicated genotype were transferred to Pi or Pi medium for 24 h before measuring the nal length of root epidermal cells (medianinterquartile;

Tukey whiskers; MannWhitney test: ****Po0.0001; n, number of cells).

growth inhibition of stop1 and almt1 mutants will be detected only after long-term exposure to Pi. Indeed, when grown for

more than 4 days on Pi medium, the root length of stop1 and

almt1 mutants was shorter when compared to lpr1 and lpr1;lpr2 roots (Fig. 3c), which correlated with a shorter RAM, a proxy of cell proliferation activity (Fig. 3d). Together with root length and growth kinetics analysis, our results show that the absence of rapid inhibition of root growth of stop1 and almt1 under Pi

correlates with reduced accumulation of Fe and callose in the EZ, whereas the late root growth inhibition correlates with accumulation of Fe and callose in the SCN and reduced RAM size. Thus, inactivation of the STOP1ALMT1 module uncouples the effect of Pi on root cell elongation and proliferation,

whereas loss of LPR function impacts both processes (Fig. 3e).

Pi stress rapidly stiffens CW in the transition zone. We next asked how Pi inhibits cell elongation. In plants, one mechan

ism that could alter cell expansion is the modication of the mechanical properties of their surrounding CW34. We reasoned that CW stiffness increases early after the onset of Pi stress to

prevent expansion of the encased root cell symplast. To test this hypothesis, we used atomic force microscopy (AFM) with a microindentation probe, which allows measuring stiffness of small plant cell surfaces in living organs35,36. We measured

the stiffness of the root surface of the transition zone (located between the RAM and EZ, Supplementary Fig. 1a), where cells are predestined to rapidly elongate when conditions are permissive.

AFM measurements were performed on WT seedlings on transfer from Pi to Pi or Pi for up to 2 h (see Methods).

Interestingly, CW stiffness increased as early as 30 min after transfer to Pi and continued to increase at later time points

(Fig. 4a); by contrast, we observed a reduced stiffness under

Pi (Fig. 4a). A higher CW stiffness (see also WT in Fig. 4b) was not detectable for the stop1, almt1 and lpr1;lpr2 lines (Fig. 4b) in which root cell expansion is not restrained under

Pi (Fig. 2d), nor in WT grown in Pi-Fe condition during

which root growth is not inhibited8,12 (Fig. 4c). Thus, our results reveal a negative reciprocal relation between CW stiffness in the root transition zone and the nal epidermal cell length. This suggests a simple causal scenario where Pi triggers

CW stiffening of pre-elongated cells to restrict their elongation.

Peroxidase activity increases CW stiffening under Pi. Class

III peroxidases are thought to inhibit cell expansion by catalysing crosslinks between some polysaccharides or proteins, thereby tightening the CW34,37. The Arabidopsis genome contains 73 genes of class III peroxidases38; many of which are expressed

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

100

lpr1; lpr2

Primary root length (mm)

lpr154

almt161 stop1127

Coler105

Stop148

stop167

almt1KO

almt151

lpr1;lpr2

lpr1130

stop167

almt151

Stop148

+Pi

0 0 2 4 6 8

Time after transfer (d)

d Coler105

Stop148

almt1KO

1,000

lpr1;lpr2

Root meristem length (m)

800

****

600

****

400

200

Coler105

Stop148

stop167

almt1KO

lpr1;lpr2

+Pi Pi +Pi Pi

4 10 Days after sowing

+Pi

WT stop1

& almt1 lpr1;lpr2

SCN

Growthinhibitionin Pi: Rapid Slow

Figure 3 | The stop1 and almt1 mutations largely uncouple the effects of Pi on the EZ and RAM. (a) Fe accumulation and distribution in primary root

tips. Three-day-old seedlings of the indicated genotypes were transferred to Pi or Pi plates for 48 h prior to Perls/DAB staining. Note the accumulation

of Fe in the EZ of the WT (white arrowhead) and in the SCN (black arrowhead). Bottom panel: magnication of the root tip of seedlings under Pi. Images

are representative of three independent experiments. Scale bar, 100 mm. (b) Four-day-old seedlings grown under Pi were transferred 20 h in Pi or Pi

before callose staining. The white arrowhead points to callose accumulation in the EZ; the red arrowheads point to callose accumulation in the SCN. Two independent experiments were performed with consistent results and one representative experiment is shown. Scale bar, 100 mm. (c) Time course of the

D primary root length of WT and mutant seedlings transferred to Pi. Two independent experiments were performed with consistent results and one

representative experiment is shown (means.d., n 814 seedlings per line). (d) Primary root meristem length of WTand mutant seedlings grown for 4 or

10 days in Pi or Pi medium (means.d., n 1425 meristems per condition; two-way ANOVA, multiple comparisons; ****Po0.0001). (e) Schematic

diagram of the consequences of stop1, almt1 and lpr1;lpr2 mutations on accumulation of Fe and callose, and growth of the primary root under Pi. The WT

root (left) accumulates Fe and callose in the EZ and SCN (red strips and dots, respectively) and its growth is rapidly inhibited. In the lpr1;lpr2 mutant (right), there is no accumulation of Fe and callose and no inhibition of root growth. In the stop1 and almt1 mutants (middle), Fe and callose accumulates only in the SCN and the root growth inhibition is slow.

in root tissues39,40. To test our hypothesis that the Pi cue leads

to CW stiffening in root tips, we used a pharmacological approach to inhibit peroxidase activity and to overcome a presumed functional redundancy among these enzymes. Salicylhydroxamic acid (SHAM) and methimazole are two peroxidase inhibitors41,42 and DAB is a classical peroxidase substrate43, which we used as a competitive inhibitor. These three

chemicals restored the WT root growth under Pi in a dose-

dependent manner (Fig. 5ad). Consistent with our hypothesis, SHAM strongly decreased CW stiffness in Pi-challenged WT

(Fig. 5e), and SHAM as well as DAB treatment signicantly increased root epidermal cell length (Fig. 5f).

These results prompted us to histologically visualize peroxidase activity in the root apex of the WT and mutant lines using

6 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

****

600

****

400

Stiffness (kPa)

200

0 +Pi 30 min 1 h 2 h

0 30 min 1 h 2 h

+Pi

Coler105 lpr1,lpr2

300

****

300 ****

Stiffness (kPa)

200

100

Coler105

150

****

0 +Pi Pi

Stop148 almt132

100

Stiffness (kPa)

300

Stiffness (kPa)

200

0 Pi Pi-Fe

100

0 +Pi Pi

Figure 4 | Pi induces cell wall stiffening in the root transition zone. (a) WT (Coler105) seedlings were transferred to Pi or Pi medium for the

indicated time prior to measuring the stiffness of the cell surface by AFM in the transition zone of the primary root (see Methods). (b) Seedlings of the indicated genotypes were transferred to Pi or Pi medium for 30 min prior to measuring the stiffness of the cell surface by AFM in the transition zone of

the primary root. (c) WT (Coler105) seedlings were transferred to Pi or Pi-Fe medium for 30 min prior to measuring the stiffness of the cell surface by

AFM in the transition zone of the primary root (medianinterquartile, MannWhitney test. ****Po0.0001; **Po0.01; *Po0.05; NS, not signicant (P40.05)). The experiment was performed twice with consistent results and one representative experiment is shown.

4-chloro-1-naphtol staining. We detected peroxidase activity in the WT root apex, from the root cap to the early EZ (Fig. 5g). Supporting our hypothesis, peroxidase staining was more intense in Pi-deprived than Pi-replete WT root tips (Fig. 5g). SHAM-treated seedlings displayed reduced staining, thus conrming SHAM-dependent inhibition of peroxidase activity in planta (Fig. 5g and Supplementary Fig. 9a). Remarkably, in the lpr1;lpr2 double mutant, peroxidase activity was strongly reduced in both Pi and Pi

conditions, whereas the stop1 and almt1 mutants displayed intermediate activities (Fig. 5h). Note, when compared to the WT in both Pi and Pi conditions, the stained zone of the stop1 root

tips was shorter and particularly weaker in the EZ (Fig. 5h).

To further challenge the idea that high peroxidase activity correlates with reduced root growth under Pi, we performed

peroxidase staining on WT seedlings grown at pH 7.1 or in a Fe-decient medium, which are two permissive conditions for root growth under Pi (ref. 8). As shown (Supplementary

Fig. 9b,c) in either condition, peroxidase activity of WT root tips was highly reduced and indistinguishable to that of the mutants. Note that at pH 7.1 no ALMT1 expression was detected (Supplementary Fig. 4a). These results convincingly show that under Pi, LPR, and to lesser extent STOP1 and

ALMT1, stimulate peroxidase activity in the root apex, which in turn inhibits root growth. Combined with our AFM and

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

pharmacological results, these observations support the view that peroxidase activity catalyses crosslinks in some CW components, thereby reducing cell expansion under Pi.

It is worth noting that the CW stiffness of SHAM-treated WT (Fig. 5e), as well as of WT grown under Pi-Fe (Fig. 4c) and in

lpr1;lpr2 double mutant grown under Pi (Fig. 4b), is lower than

in the corresponding controls grown under Pi. This softening

is correlated with reduced peroxidase activity observed in these roots (Fig. 5g, Supplementary Fig. 9b and Fig. 5h). These observations may indicate that in conditions permissive for root

growth, the LPR ferroxidases and some peroxidases participate to maintain a basal strength of the CW.

Under Pi, peroxidase activity also represses cell proliferation

and accelerates cell differentiation in the RAM (Fig. 5d, pictures of the root tips). This is reminiscent to the role of UPB1, which regulates the transition from cell proliferation to differentiation in the root tip via the expression of some class III peroxidases44. However, primary root growth of the upb1 mutant and an UPB1-overexpressor line is not altered in response to Pi exposure (Supplementary

Fig. 9d), indicating that UPB1 is not essential in this process.

a b c d

OH H N

H N

N SH

SHAM (15 M)

DAB (100 M)

Methi. (500 M)

+Pi

NH CH

Methimazole (M)

+Pi Pi

Primary root length (mm)

1 10 100

0 1

100

10 1,000

SHAM (M)

DAB (M)

****

400 ****

****

400 ****

Stiffness (kPa)

Epidermal cell length (m)

300

SHAM

300

200

100

+SHAM

n= 196

n= 262

n= 187

n= 266

+Pi

Pi + SHAM

+Pi

-Pi +SHAM

Pi +DAB

Coler105

Stop148

stop167

stop1127

stop1KO

almt132

almt1KO

lpr1;lpr2

+Pi

8 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

necessary, for STOP1-dependent root growth inhibition by Pi.

This indicates that ALMT1-exported malate locally diffuses in the apoplast of root tip cells.

DiscussionIdentication of the cellular and molecular mechanisms by which environmental cues affect organ growth is a current challenge for plant biology. Pi starvation is one of the most common limiting factors for plant growth and crop productivity. Previous studies identied Arabidopsis LPRs and PDR2 as part of a pathway controlling primary root growth in response to Pi availability surrounding the root tip, the so-called local response10,11.

Our work uncovers a novel pathway in local Pi sensing. We identied in a forward genetic screen two functionally interacting components with roles in primary root growth arrest under Pi:

the transcription factor STOP1 and its direct target gene ALMT1.

Absence of a reciprocal regulation of pALMT1 and pLPR1 promoter activity (Supplementary Fig. 12a,b) suggests that STOP1ALMT1 and LPR1 function in distinct pathways.However, the similar unrestricted root cell elongation in stop1, almt1 and lpr1 mutants (Fig. 2c,d) indicates that the two pathways converge on a common inhibition mechanism. Therefore, the PDR2LPR1 and STOP1ALMT1 modules dene the rst components of two separate pathways targeting the root transition zone to inhibit cell elongation under Pi (Fig. 7).

To our knowledge, STOP1 is the rst plant transcription factor showing a regulation reminiscent of yeast PHO4 (ref. 49), which accumulates more intensely in the nucleus upon Pi starvation.Thus, STOP1 and ALMT1 dene a new Pi regulatory pathway,

which is supported by PHR1/PHL1-independent activation of ALMT1 (Supplementary Fig. 3a). In addition to low pH and Al3 (refs 21,22,24,32), low Pi is thus a new environmental stimulus activating the STOP1ALMT1 pathway in Arabidopsis and most probably in crops50. Indeed, the primary root of the Brassica napus, a crop plant known to exude malate under Pi (ref. 51),

is also inhibited under Pi (Supplementary Fig. 13). This

pathway is thus likely of general relevance.Our work further uncovers that root growth inhibition under

Pi results from rapid inhibition (o2 h) of cell elongation and subsequent inhibition (o4 days) of cell proliferation, previously referred to as meristem exhaustion5,9,12,52,53. Interestingly, our data demonstrate that both processes can be genetically uncoupled in stop1 and almt1 mutants. Both inhibitory processes correlate with a higher Fe accumulation, callose deposition, and peroxidase activity in the EZ and SCN, as well as with CW stiffening at least in the EZ. Recently, a comparative

Figure 5 | Cell wall stiffening and inhibition of root cell elongation under Pi are peroxidase-dependent. (ac) Doseresponse relationship of the

D primary root length plotted as a function of concentration of peroxidase inhibitors (means.d.). For each data point, 3-day-old WT (Coler105) seedlings were transferred for 7 days to Pi plates containing different concentrations of inhibitor. The structure of the inhibitors is shown above the corresponding

curves. (a) Salicylhydroxamic acid (SHAM). Blue data points: Pi; yellow data point, Pi without inhibitor (n 1214 seedlings per condition).

(b) 3,30-diaminobenzidine (DAB) (n 812 seedlings per condition). (c) Methimazole (n 1113 seedlings per condition). (d) Picture of the effect of

peroxidases inhibitors on WT seedlings. Three-day-old WT seedlings were transferred for 7 days to Pi or Pi plates containing the indicated inhibitors.

Top: picture of one representative seedling for each treatment (Methi., methimazole). Arrowheads indicate the primary root tips. Scale bar, 1 cm. Bottom: pictures of the root tips corresponding to the seedlings shown above. Note that inhibitors of peroxidases suppress the Pi-induced early differentiation of

root hairs. (e) Three-day-old WT seedlings were transferred for 30 min to Pi or Pi medium with or without 15 mM SHAM prior to measuring the

stiffness of the cell surface by AFM in the transition zone of the primary root (medianinterquartile, Tukey whiskers, MannWhitney test: ****Po0.0001). (f) Effect of SHAM and DAB on root epidermal cell length. Three-day-old WTseedlings were transferred for 7 days to Pi or Pi medium with or without

15 mM SHAM or 100 mM DAB prior to measuring the root epidermal cell length (medianinterquartile; Tukey whiskers, MannWhitney test: ****Po0.0001, n number of cells). (g) Peroxidase activity revealed by 4-chloro-1-naphtol staining of the WT primary root tip (the colourless dye turns

black with peroxidases activity). Four-day-old WT seedlings were transferred for 48 h to Pi or in Pi plates with or without 15 mM SHAM (see also

Supplementary Fig. 9a). (h) Peroxidase activity revealed by 4-chloro-1-naphtol staining of the primary root tip of the indicated genotypes. Four-day-old seedlings were transferred for 48 h to Pi or in Pi plates prior to staining. Single and double-headed arrows indicate the position of the elongation zone.

The experiments were performed three times with consistent results and one representative experiment is shown (ac,g,h). Scale bars, 100 mm (g,h).

Pi promotes ROS accumulation in the EZ. Correlated with the deposition of Fe and callose, the generation of ROS in CWs of the RAM and EZ was observed in WT, but not in the lpr1;lpr2 double mutant, when transferred to Pi medium

for 1 day12,19. We also observed accumulation of ROS (H2O2 and peroxidases-mediated oxidation of CM-H2DCFDA) in the transition and differentiation zones of WT seedlings transferred to Pi for 24 h (Supplementary Fig. 10). Under Pi,

accumulation of ROS is reduced when iron is omitted or when SHAM is added to the growth medium (Supplementary Fig. 10). This result conrms that Pi induces accumulation of ROS

in the root tip and indicates that peroxidase activity could play a role.

Malate efuxed into the RAM inhibits root growth under Pi.

Specic cell types of the RAM control root growth in Arabidopsis (see for example ref. 45). To test whether STOP1 activity in specic cell les inhibits root growth under Pi, we used cell

type-specic promoters to conne and restore STOP1 expression (and thus ALMT1 expression) in a stop1KO mutant (Fig. 6a). As shown in Fig. 6b, we fully rescued root growth inhibition by transformation with a pSCR::STOP1 construct, which drives STOP1 expression specically in the QC and endodermal cell lineage46. The pCO2 promoter46 restored STOP1 expression essentially in young cortex cells of the RAM and complemented stop1KO almost fully. The deposition of Fe in the EZ was also complemented in the pSCR::STOP1 and pSCR::Co2 lines (Fig. 6c).

In contrast, we observed only partial complementation using the pPEP promoter, which is active in cortex cells of the elongation and maturation zones47 and slightly in the endodermis47, and no complementation with the pCASP1 promoter, which is active in mature endodermal cells48 (Fig. 6b). Using a pALMT1::GUS reporter, the expected pattern where STOP1 is functional in the root and root tip of all these lines was conrmed (Supplementary Fig. 11a,b). Taken together, our results show that STOP1 expression conned to the endodermis (and QC) or to the early cortex is sufcient for rescuing the stop1KO root phenotype.

To test whether these two cell types are necessary for STOP1-dependent growth arrest, we generated WT lines expressing the dominant negative Stop148 gene under the pSCR or pCo2 promoter. Our results show that the two constructs, expressed either alone (Supplementary Fig. 11c) or in combination (Fig. 6d), do not release root growth inhibition under Pi.

Taken together, these experiments show that expression of STOP1 in endodermis or cortex of root tips is sufcient, but not

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

pSCR b

pCASP1

pCo2

pPEP

****

Cortex

Endodermis

****

Primary root length (mm)

Quiescent centre

Coler105

****

stop1KO

pSCR::STOP1

pCo2::STOP1

pPEP::STOP1 pCASP1::STOP1

****

Primary root length (mm)

pSCR::STOP1 pCo2::STOP1

Col-0 stop1KO lpr1;lpr2

#2A #10A #6A #17A

Coler105

pSCR::Stop148 #2C

pCo2::Stop148 #1G

pSCR::Stop148 #2C

Stop148

pCo2::Stop148 #1G

Figure 6 | Expressing STOP1 in the endodermis or early cortex of the stop1 mutant is sufcient to restore root growth arrest under Pi. (a) Schematic

diagram showing the position of cortex, endodermis and quiescent centre, and the tissue-specic expression (blue) conferred by the indicated promoters. Only one longitudinal half of the root is shown. (b) Primary root growth of stop1KO mutant lines with tissue-specic expression of STOP1. Three-day-old seedlings were transferred for 4 days to Pi and the D primary root length measured in 56 independent transgenic lines par construct (means.d., n 17

seedlings per line; box and Tukey whiskers; statistics by unpaired two-tailed t-test: ****Po0.0001; NS, not signicant). The experiment was performed

twice with consistent results; one experiment is shown. (c) Fe accumulation and distribution in primary root tips of pSCR::STOP1 and pCo2::STOP1 lines.

Three-day-old seedlings of the indicated genotypes (two independent lines for each construct) were transferred from Pi to Pi or Pi plates for 48 h

prior to Perls/DAB staining; one representative seedling is shown. Note the accumulation of Fe in the EZ of the complemented lines (white arrowhead) compared to the parental stop1KO line. Scale bar, 100 mm. (d) The dominant negative Stop148 was expressed in the endodermis (pSCR::Stop148), in the cortex (pCo2::Stop148) or in both cell types (F1 progeny from the cross pSCR::Stop148 X pCo2::Stop148) in Coler105. Seedlings were grown for 5 days under Pi and

the primary root length measured (means.d., n 714 seedlings per line; one-way ANOVA; ****Po0.0001; NS, not signicant). Note that the primary

root of transgenic lines behaves like in WT control.

transcriptomics/proteomics study revealed that Pi deprivation profoundly alters expression of CW-related genes (o20 h)

and leads to accumulation of callose and non-esteried pectins in WT root tips at sites of Fe deposition (EZ, SCN), which are not detected in lpr1;lpr2 roots12,19. Crosslinking of non-esteried pectins by Ca2 reduces CW extensibility and regulates cell expansion54.

It is quite unexpected that malate efux inhibits elongation of root cells under Pi, because exudation of small organic acids is

thought to promote plant growth by solubilizing scarcely available Pi and Fe nutrients, or by forming non-toxic complexes

with Al3 cations25,26,55,56. However, malate and Fe form chelates participating in plant Fe homeostasis57. Interestingly, when complexed with malate, Fe3 reduction is faster when compared to Fe3 citrate or Fe3 EDTA complexes58.

Therefore, a low Pi/Fe ratio, as under Pi, may allow

formation of Fe3 malate complexes in CWs of the RAM and EZ, promoting Fe accumulation, LPR1 ferroxidase-dependent Fe redox cycling, ROS generation and callose accumulation (Fig. 7).

The role of Fe might not be limited to ROS production; one might speculate that Fe3 also activates ALMT1 to control malate efux, as Al3 regulates ALMT1 channels from different

10 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

STOP1

ALMT1

Malate

Malate::Fe

Elongation zone Meristem

PDR2

LPR1

Stem cells

Figure 7 | Working model depicting the effect of Pi on cell proliferation and elongation in the primary root tip. A low Pi/Fe ratio, that is, Pi stress12

(yellow star), post-transcriptionally activates STOP1 which induces ALMT1 expression followed by malate exudation into the apoplast. The interaction of malate-altered iron chelation and LPR1 ferroxidase activity inhibits cell elongation by peroxidase-dependent cell wall stiffening. The P5-type ATPase PDR2 restricts LPR1 function12. The LPR1PDR2 module interacts with an unknown pathway parallel to STOP1ALMT1 (question mark) to inhibit cell division in the SCN by a similar mechanism involving accumulation of Fe and callose, peroxidase-activity, and cell wall thickening. Mutations in STOP1 and ALMT1 restore cell elongation under Pi but cell divisions are still inhibited.

species20,5963. However, the latter mechanism is still debated for AtALMT1 (ref. 64). Gamma-aminobutyric acid, pH and external anions also regulates the activity of wheat ALMT1 (ref. 62) and we do not exclude the possibility that small molecules also stimulate AtALMT1 activity under Pi.

As shown at pH 7.1 (Supplementary Fig. 4a), ALMT1 is not expressed in Pi and Pi conditions (maybe because the

STOP1 protein is not activated at this pH). This could partly explain the lack of root growth inhibition under Pi at this pH

(ref. 8). However, even when ALMT1 is constitutively expressed, there is no growth inhibition (Supplementary Fig. 4b); one explanation might be that at higher pH Fe precipitates in a form that cannot inhibit root growth.

Peroxidases produce or consume ROS37. Restoring root growth under Pi with the competitive articial substrate

DAB (Fig. 5b,d) supports the view that peroxidases inhibit growth via their peroxidative activity that consumes H2O2. This

activity of class III peroxidases is known to catalyse intermolecular covalent bonds between CW components like monolignols, feruloylated pectins or structural proteins such as hydroxyproline-rich glycoprotein and extensins34,37. These crosslinks are thought to impede CW expansion and therefore prevent cell enlargement. The increased (early) CW stiffness was conrmed by nanoindentation experiments (AFM), which fully support a causal link between peroxidase activity and the arrest of cell elongation. Likewise, it was recently reported that Pi deprivation deregulates the expression of about 30 class III peroxidases at the mRNA or protein level19 and causes increased lignication in WT but not in lpr1;lpr2 root tips65.

Interestingly, peroxidase-mediated strengthening of extracellular brillar proteins (for example, collagen) by crosslinking also operates in animals, pointing to an evolutionary conserved mechanism66,67.

In conclusion, we propose a new pathway and mechanism for local Pi sensing by root tips. Pi limitations enhance the nuclear abundance of STOP1 to transcriptionally activate in a PHR1-independent manner ALMT1 expression, which promotes malate efux by internal cell layers of the root apex. Malate-facilitated Fe chelation and LPR1-dependent Fe redox cycling stimulate both the peroxidase-dependent CW stiffening, which rapidly inhibits root cell elongation, and the accumulation of callose in the EZ (Fig. 7). With its dual role in promoting or attenuating

root growth, exuded malate may be considered a genetically controlled intercellular diffusible factor that adjusts root cell elongation to the local chemical environment of the rhizosphere.

Methods

Plant growth. Unless specied, plants and seedlings were grown as previously described68. Briey, surface-sterilized seeds were sown on a horizontal line in plates that were vertically disposed in a growing chamber (16 h photoperiod; 24.5 C/ 21 C day/night, respectively). The growth medium contained 0.15 mM MgSO4,

2.1 mM NH4NO3, 1.9 mM KNO3, 0.34 mM CaCl2, 0.5 mM KI, 10 mM FeCl2, 10 mM H3BO3, 10 mM MnSO4, 3 mM ZnSO4, 0.1 mM CuSO4, 0.1 mM CoCl2, 0.1 mM Na2MoO4, 0.5 g l 1 sucrose. The agar (0.8 g l 1) for plates was from Sigma-

Aldrich (A1296 #BCBL6182V). The Pi and Pi agar medium contained 10 and

500 mM Pi, respectively; the media were buffered at pH 5.65.8 with 3.4 mM 2-(N-morpholino)ethane sulfonic acid.

Growth media at pH 7.1 were made using 2 mM 1,4-piperazinediethanesulfonic acid pH 7.1. Growth media at pH 4.4 were prepared by mixing equal amountof autoclaved 2 liquid medium (composition as above, without Pi and Fe,

buffered at pH 4.4) with 2 agar (dissolved in H2O); Pi and Fe were subsequently

added according to media. For the pH experiment (Supplementary Fig. 4a), Fe was at 15 mM.

Brassica napus seeds (cv Bohme) were sterilized and seedlings grown on the same media, as for Arabidopsis, except that the growth media contained 100 mM

FeCl2. Briey, seeds were sown on a Pi plate, left 24 h at 4 C and let germinate

for 12 h. Germinating seeds (with the primary root just protruding) were then transferred in Pi or Pi plates for 3 days before measuring the primary root

length.

For testing resistance to Al3, seeds were sown in 2 ml liquid medium (NH4NO3 4.1 mM, KNO3 3.8 mM, CaCl2 200 mM, MgSO4, 0.3 mM, H3BO3 20 mM, MnSO4 20 mM, KI 1 mM, ZnSO4 6 mM, Na2 MoO4 0.2 mM, CoCl2 0.2 mM, CuSO40.2 mM, sucrose 0.5 g l 1, Homopiperazine-1,4-bis(2-ethanesulfonic acid) 2 mM pH 4) and grown for 4 days before measuring the primary root length.

Mutagenesis and mutants screening. For the mutagenesis, we used Coler105,a Columbia background (Col-0) with the null allele erecta-105 (ref. 69) that confers a stunted habitus facilitating handling of adult plants in the growth chambers.

About 8 104 Coler105 seeds were mutagenized with 0,2% of ethylmethane

sulfonate (EMS, Sigma-Aldrich) for 16 h, under gentle agitation. The EMS was inactivated with 100 mM sodium thiosulfate for 10 min followed by washing the seeds with tap water (10 times). Seeds were sown on soil and the M2 generation was collected in 156 pools of B400500 M1 plants each. From each pool,about 2,500 seeds were sown on two plates (12 12 cm) containing Pi medium

(each plate contained two horizontal lines of about 625 seeds each). After 12 h at 4 C plates were placed on a 30 vertical angle in a growth chamber. After 7 days, plates were screened for seedlings with a long primary root (in comparison to neighbouring roots), and putative mutants were transplanted into soil. The mutant phenotype was checked in the M3 generation and backcrossed to the WT Coler105 line. When several putative mutants appeared in a pool, only one was kept for further study. From a total of about 390,000 M2 seeds sown on Pi medium,

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

we isolated 85 viable putative mutants with a long primary root. The mutant phenotype was conrmed in the M3 generation.

Mapping almt132 and Stop148. Two non-allelic mutants (ems32 and ems48), that do not carry mutations LPR1 (see below), were chosen to identify the corresponding mutant genes by map-based cloning approaches.

The almt132 mutant was crossed with the Sha accession, which has a short root when grown under Pi (ref. 6). In the F2 population grown under Pi, we

selected seedlings with a mutant phenotype. DNA of individual lines was extracted, and a combination of high-resolution melting (HRM) and microsatellites molecular markers were tested for genetic linkage analysis. The mutation was mapped between molecular markers MSAT12.62 and PMSha-2.78. For the ne mapping, we developed new microsatellite and HRM markers (Supplementary Table 1), allowing localization of almt132 between MSAT12.62 and HRM2745575 markers, a 120 kb interval.

The same procedure was applied to Stop148 except that we selected F2 seedlings with a WT phenotype (because Stop148 is dominant). The locus was mapped between MSAT112.22 and CWI1 in a 6 Mb interval. Among B1,440 chromosomes, 7 and 107 were recombinant for MSAT112.22 and for CWI1, respectively.

Identication of the almt132 and Stop148 mutations. The recessive ems32 and the dominant ems48 mutations were mapped on the top arm of chromosome I in regions of 120 kb and B2 Mb, containing 31 and B6700 genes, respectively. To identify the corresponding mutant genes, we followed a double gene candidate approach reasoning that the ems32 and ems48 mutations could be located in functionally related genes. By looking at the predicted genes in the 120 kb of chromosome interval containing ems32 we found ALMT1 (At1g08430) and in the2 Mb region containing ems48 we found STOP1 (At1g34370), encoding a putative transcription factor regulating the expression of ALMT1 (ref. 22). Sequencing ALMT1 in the ems32 mutant and STOP1 in the ems48 mutant conrmed our conjecture. Homozygous T-DNA and Ds insertion mutant alleles of these genes conferred the same long root mutant phenotype as the EMS alleles (Fig. 1c) thus demonstrating that almt1 and stop1 mutants have a long root under Pi.

Allele screening. To identify the potential lpr1 alleles in our collection of EMS-mutagenized mutants, we used a molecular test based on the restriction enzyme ENDO1 that cuts PCR products containing a DNA mismatch.

The ENDO1 restriction enzyme was used according to instruction manual (SERIAL GENETICS). Briey, the LPR1 gene was amplied with primers 23010cDNAF plus 23010R2 and with 23010F4 and RT10R (Supplementary Table 1) in the mutants and in the WT. A 1/1 ratio of mutant and WT amplicons were mixed in a test tube. After denaturation at 95 C, DNA was allowed to renature by cooling (1 C min 1) to 5 C. 0.4 unit of ENDO1 reaction mix containing 2.5 mg of renatured DNA was added and the tubes transferred at37 C for 20 min. DNA was electrophoretically resolved on a 2% agarose gel and stained with ethidium bromide to detect amplicons partially cut by ENDO1. To identify the mutation, we sequenced the amplicon obtained in an independent PCR reaction with DNA of mutant lines.

To nd potential alleles of almt1 and stop1 in our collection of EMS-mutagenized mutants, we rst tested all the mutants to identify those with a primary root growth hypersensitive to Al3 (ref. 22). Seedlings of these mutantswere grown 24 h in liquid medium containing 50 mM Al3 and the expression of ALMT1, MATE and STOP1 was measured by qRTPCR (primers in

Supplementary Table 1). We then sequenced the STOP1 gene of mutants in which the expression of both MATE and ALMT1 was reduced (compared to WT) and we sequenced the ALMT1 gene of mutants in which the expression of both STOP1 and MATE was not reduced.

Other Arabidopsis lines. The stop1KO (SALK_114108, NASC reference N666684), almt1KO (SALK_009629, NASC reference N680999). lpr1;lpr2 (ref. 8), phr1;phl1 (ref. 29), wrky46KO (SALK_134310C, NASC reference N667392), pLPR1::GUS line12, almt1KO;mateKO carrying the pALMT1::MATE construct27, and upb1-1 and 35S::YFP-UP1 line44 are in the Col-0 background. The mateKO (SK3573, NASC reference N1000719) is in Col-4 background.

Malate complementation. A volume of 200 ml per well of MS/10 medium8 without FeCl2, KH2PO4 and agar were distributed in a 12-well plate and used to germinate seeds for 48 h (20 seeds per well and 2 wells per line). The plate was placed in the same conditions (light and hygrometry) as agar plates (see above). After 48 h, when the primary root and cotyledons emerge from the seed coat, the medium was replaced by 2 ml of a fresh MS/10 medium containing 10 mM FeCl2 and 10 mM ( Pi) or 500 mM ( Pi) of KH2PO4 and different concentrations

of L-malate (0200 mM) and adjusted at pH 5.5 with KOH. After 60 h, the medium was replaced by fresh medium of identical composition, and seedlings grown for additional 60 h. At the end of this time each seedling was placed on an agar plate to measure the primary root length.

Chemical stock solutions. SHAM (Sigma-Aldrich, France) was dissolved at10 mM in absolute ethanol, DAB (Sigma-Aldrich, France) was dissolved at 100 mM in dimethyl sulfoxide (Sigma-Aldrich, France) and methimazole (Sigma-Aldrich, France) was dissolved at 100 mM in H2O.

Quantitative RTPCR. Total RNA was extracted from whole roots using the RNeasy Plant Mini Kit (Qiagen, France) and treated with the RNase-free DNase Set (Qiagen, France) according to the manufacturers instructions. Reverse transcription was performed on 400 ng total RNA using the SuperScript VILO DNA Synthesis Kit (Invitrogen). Quantitative PCR (qRTPCR) was performed on a 480 LightCycler thermocycler (Roche) using the manufacturers instructions with Light cycler 480sybr green I master (Roche) and with primers listedin Supplementary Table 1. We used ROC3 (AT2G16600) as a reference genefor normalization.

Transgene constructs. To study promoter activity, promoter::GUS lines were constructed for pALMT1 and pSTOP1. The 1.7 and 2.1 kb region upstream of the start codon of the ALMT1 and STOP1 gene were amplied respectively from WT genomic DNA. The PCR product was then cloned into the pDONR P4-P1R by a BP clonase reaction (LifeTechnologies). GUS was cloned into a pENTR vector and nal constructs were obtained by multisite gateway reaction with pB7m24GW.3 destination vector as described elsewhere70.

GFP-reporter lines using promoter UBQ (UBIQUITIN10) were constructed for ALMT1 and STOP1. For localization of STOP1 additionally GFP-reporter lines with the native promoter were constructed. pEN-L1-F-L2 and pEN R2-GFP-L3 were used for GFP-fusion constructs by multisite gateway cloning. The 640 bp upstream region of the start codon of UBQ10 and 2.1 kb upstream region of the STOP1 gene was amplied from genomic DNA and cloned into the pDONR P4-P1R. The full-length coding region of ALMT1 and STOP1 were ampliedby PCR and inserted into the pDONR P2R-P3. The nal gateway reactionwas accomplished with pB7m34GW as destination vector, resulting in pUBQ10::GFP-STOP1, pUBQ10::ALMT1-GFP and pSTOP1::GFP-STOP1. The resulting vectors were transformed into Arabidopsis by oral dip method71. Primers used for vector constructions are listed in Supplementary Table 1.

Several independent lines segregating in 3:1 (in the T2 generation) for the resistance to the antibiotic carried by transgene were selected for all our constructs.

To introgress pALMT1::GUS#2 in the Stop148 background, pollen of Stop148 was used to cross the pALMT1::GUS#2 WT line and the cross was checked in F1 grown under Pi (long primary root). F2 seedlings with a long primary root under Pi were selected and transferred in phosphinothricin (PPT) plates. Several PPT-resistant plants were transferred in soil for selecting the double homozygous (pALMT1::GUS#2/pALMT1::GUS#2; Stop148/Stop148) by checking the root phenotype under Pi and the resistance to PPT.

The pLPR1::GUS line12 has been crossed with Stop148 and stop1127 mutants and lines homozygous for both the transgene and the stop1 mutation selected in the progenies.

Cell type-specic complementation. For cell type-specic complementation, we used pCO2, pPEP, pCASP1 and pSCR promoters as described4648,72. Constructs were generated by multisite gateway reaction with pB7m24GW.3 as destination vector for STOP1WT and Stop148. Vectors were then used to transform the stop1KO or the Coler105 (containing pALMT1::GUS#2) background. F2 seedlings with PPT or kanamycin resistance were selected and transferred to soil.

Measurement of the primary root length. Pictures of the plates were taken with a camera and root lengths measured using the plugin NeuronJ73 for the ImageJ software74. For the early kinetic analysis, seedlings were grown 7 days under

Pi and then transferred onto a Pi or a Pi medium and pictures of the root

tip were taken every 20 min for 7.5 h with an Axiozoom-V16 macoscope (Zeiss). D root lengths were measured with Image-J software74 and a special M macro developed on purpose (Jrme Mutterer, IBMP, Strasbourg).

Measurement of root epidermal cell and RAM lengths. To stain the CWs of epidermal and cortex cells, seedlings were dipped, respectively, 3 min in 15 mM propidium iodide (PI) or 5 min in 30 mM PI, at room temperature, and rinsed twice with H2O.

To measure RAM length, based on the les of cortex cells (from the QC to the rst elongated cell), images were collected on a Zeiss LSM780 confocal microscope (Carl Zeiss, France) using a 10 (no immersion) or 20 (water immersion)

objective. For PI visualization, the DPSS laser (561 nm) was used. Emitted light was collected from 600 to 675 nm for PI, using the MBS 561 lter.

Images of epidermal cells, in root zones where elongation has ended, were collected with an Axiozoom-V16 macroscope imaging system (mRFP lter, excitation at 572 nm, emission at 629 nm) using the Apotome functionality and processing with Zen blue (Zeiss). Cell lengths were measured with the ImageJ software (Rasband74). For each growth condition, cells were measured in three independent zones, along each of the ve roots that were stained.

12 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

Inductively coupled plasma mass spectrometry analysis. Seeds were sown in line on a sterile nylon mesh (1.5 12 cm, 100 mm). They were grown for 5 days

under Pi before transfer for 24 h under Pi or Pi thanks to nylon mesh.

60 root tips per genotype (E5 mm length) were harvested with Teon razor blades and tweezers into cryovial tubes.

Samples were mineralized at 80 C for 3 h using concentrated HNO3/H2O2 75/25 (v/v). Subsequently, they were diluted 10 times by ultrapure water and injected via a peristaltic pump at 400 ml min 1 ow rate in the 7,700 inductively

coupled plasma mass spectrometry analysis (ICP/MS) (Agilent). ICP conditions were the following: nebulization gas ow rate 1 l min 1, plasma gas ow rate 15 l min 1, auxiliary gas ow rate 1 l min 1. Power plasma was set to 1,550 W.

The collision cell was activated to decrease polyatomic ion interferences using He at 4 ml min 1 ow rate. Other conditions were adjusted to both maximize analyte signals and minimize oxide and doubly charged ion formation. Quantication of Fe was performed at m/z 54 and 56 using external standards. Between analyses, the

system was rinsed for 3 min by HNO3 5% a blank was injected to control the absence of any memory effect. Acquisition was performed for 3 s in triplicate.

Yeast one-hybrid assay. The Saccharomyces cerevisiae strains used were YPH501 (diploid, ura352, lys2801, ade2101, trp1-D 63, his3-D 200, leu2-D 1) and

YPH499 (MATa, ura352, lys2801, ade2101, trp1-D 63, his3-D 200, leu2-D 1).

For yeast transformation, cells were grown in the YPD (yeast extractpeptone dextrose) medium, and for the yeast one-hybrid assay, cells were grown in rafnose synthetic complete medium minus uracil and tryptophan, supplemented with1% of galactose synthetic complete medium minus uracile and trytophan at 30 C.

DNA constructions of bait::reporter fusions were made as follow: the ALMT1 promoter (1.7 kb and sub-fragments) was PCR-amplied (DNA primers in Supplementary Table 1) and the amplicons inserted in front of the lacZ reporter gene by homologous recombination within the XmaI-linearized pSH1834 vector (Invitrogen), using the yeast strain YPH501. To make the effector construct, theE. coli B42 activation domain was fused by homologous recombination to the PCR-amplied (Supplementary Table 1) AtSTOP1 C-terminal coding sequence into the pB42AD (Clontech) vector; expression of this AtSTOP1-B42 fusionwas driven under the pGAL inducible promoter. All constructs were checked by DNA sequencing. The primers used for plasmid construction are shown in Supplementary Table 1.

Yeast strain YPH499 transformed with both the bait::reporter and effector constructs were used to quantify DNA-binding activity of STOP1; b-Galactosidase activity assays were performed as described75. For all pALMT1 fragments tested, there was two strains (each with two independent clones): a test strain containing the pALMT1 vector and the STOP1::B42 construct and a control strain with the same pALMT1 vector and the B42 construct. The b-galactosidase relative activity was calculated as follow: the mean uorescence of 4-methylumbelliferone (360 nm) per OD600nm measured from the two independent clones (containing the pALMT1 fragment STOP1::B42) was divided by the mean uorescence of the

corresponding control strain (containing the pALMT1 fragment B42).

GFP imaging. To stain root CWs, seedlings were dipped 5 min in 30 mM PI solution at room temperature and rinsed twice with H2O. Images were collected on a Zeiss LSM780 confocal microscope (Carl Zeiss, France) using a Plan Apochromat

10/0.45 objective and a Plan-Apochromat 20/0.8 objective. GFP and PI were

excited sequentially with the blue argon ion laser (488 nm) and the DPSS laser (561 nm), respectively. Emitted light was collected from 491 to 535 nm for GFP, and from 600 to 675 nm for PI, using the MBS 488/561 lter. Images were processed using Zen black software (Zeiss).

Quantication of GFP uorescence. To quantify GFP uorescence in nuclei, images were obtained with a LSM780 confocal microscope (Carl Zeiss, France) using a 20 dry objective (Plan Apo NA 0.80). GFP was excited with the Argon

laser (488 nm), and emitted light was collected between 493 and 538 nm with a pinhole setting of 1.51 a.u.

All nuclei were imaged using the same conditions of gain, offset and resolution (with zoom set to 3), at a constant distance (350 mm) from the root tip. A Z-stack of 15 images (separated by 1 mm distance) imaged representative nuclei on the surface of the root. Images were acquired in 16 bits using Zen black software (Zen black 2012 SP2 Version 11.0), then converted to a maximal projection image. Average nuclei uorescence intensities were quantied using the Zen blue software(Zen 2 blue edition, version 2.0.0.0), by drawing identical ROI (region of interest) over 310 visible (in focus) nuclei for each maximal projection image.

PERLs-DAB staining. PERLs-DAB staining was performed as described12 with minor modications. DAB intensication was performed for 5 min with a nal concentration of 0.00625% DAB. Clearing solution was modied to a higher concentration (8:2:1 hydro-chlorate:water:glycerol (w/v)). Roots were observed and documented with the Axiozoom-V16 macroscope (Zeiss) using the 2.3 objective.

GUS histochemical staining. For Arabidopsis seedlings, the GUS staining was conducted as previously described76. Briey, the samples were submersed in a GUS

staining solution (50 mM sodium phosphate buffer pH 7.0, 0.01% Triton X-100, 1 mM K3Fe(CN)6, 1 mM K4Fe(CN)6, 1 mg.ml 1 5-bromo-4-chloro-3-indolyl b-D-glucuronide) and incubated 1 h at 37 C. Plants were observed and documented with a DM6000B upright microscope (Leica; 10 objective) after

clearing with chloral hydrate solution as described12.

For histological sections samples were stained for 2025 min at 37 C. Samples were xed (1% glutaraldehyde, 4% paraformaldehyde in 50 mM phosphate buffer, pH 7.0) for 24 h, followed by stepwise dehydration in ethanol. Embedding was accomplished with Technovit 7100 (Kulzer, Germany). Sections (10 mm) were prepared and stained with ruthenium red. Sections were observed and documented with the axio Zoom.V16 (Zeiss) using the 2.3 objective.

Visualization of peroxidase activity. 4-Chloro-1-naphtol staining was used to visualize peroxidase activity in roots. A volume of 1 ml of a 3% 4-chloro-1-naphthol stock solution (30 mg ml 1 in absolute ethanol) and 1 ml of a 3% H2O2 stock solution were added to 10 ml of 250 mM Tris buffer pH 7.4 resulting in the nal staining solution. Plants were stained for 60 s. Staining solution was exchanged with Pi or Pi buffer, depending on the sample, and washed three times.

Staining was observed and documented with the Leica 6000D ( 10 objective) for

all samples in the same conditions.

ROS and callose staining. ROS and callose staining were accomplished as described12. ROS were stained using CM-H2DCFDA (6-chloromethyl-2,7-dichlorodihydrouorescein diacetate; this compound is oxidized by peroxidases in presence of H2O2) with minor modications (for example, the phosphate-buffer was replaced by 20 mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES) buffer, pH 6.0). Images were collected on Zeiss LSM780 confocal microscope (Carl Zeiss, France).

AFM analysis. Mounting root samples: AFM data were acquired on the primary root from whole seedlings mounted under the microscope. Seedlings were grown 6 days under Pi prior to transfer under Pi or Pi during 0.52 h. Seedlings

were then glued on a microscope slide (one seedling per slide) as follow: a layer of silicone (MED11,356, NuSil Technology LLC, USA), with the thickness ofa coverslip slide (130170 mm), was spread on the slide; silicone was let to dry out for 1 min prior to laid down the seedling on it. To securely anchor the root, a silicone band covering o100 mm of the root tip was stretched using a syringe needle. Then, a droplet of liquid medium68 with 10 mM ( Pi) or 500 mM KH2PO4

( Pi), corresponding to the medium (minus agar) in which the seedlings were

transferred before AFM analysis, was deposited to cover entirely the seedling to prevent drying. The mounted seedling was then positioned under the AFM instrument. Three seedlings were used per line and per condition.

Data acquisition and apparent Youngs modulus cartography: Forcedistance curves were obtained with an AFM Dimension 3100 (Bruker, Santa Barbara, USA) with a nanoscope ve controller running the Nanoscope 7.3 software. Triangular pyrex nitride cantilever with pyramidal tips of nominal radius of 15 nm, a half-opening angle of 35, and a nominal spring constant k 0.06 N m 1 were used

(PNP-TR, NanoWorld AG, Neuchatel, CH). Calibration of photodiode sensitivity was performed using approach-retract curves in liquid on a glass substrate; an average value of 55 nm V 1 was commonly obtained. Cantilever spring constant was determined using the thermal tune module of nanoscope and an average value of 0.04 N m 1 was commonly found. Positioning of AFM cantilever on plant roots was performed using the Dimension 3100 optical camera so that the targeted zone was at B500 mm from the end root, as measured with the crosshair (that is, after the optically dense zone corresponding to the meristem). Indentation of plant root surface was performed with approach-retract curves using the picoforce module of nanoscope with a ramp size of 34 mm, a scan rate of 1 Hz, and 2,000 points per curve. Ramp size was adjusted using a relative set-point threshold so that an indentation of about 0.5 mm was achieved. Each measurement includes an area of 10 40 mm2 using nine zones (5 mm in the direction of the width and 10 mm in the

direction of the root axis) into which about 10 force curves were recorded. Thus, a total of about 90 curves were collected per area. Force curves were translated in ASCII text les using YieldFinder77 to t the le format of software FORCE v1.1.0 (ref. 78). Indentation curves were tted according to the authors instructionsby a model of Sneddon79. The two adjustable parameters, contact point and indentation depth, were manually adjusted to optimize the tting curve. The cell stiffness as dened by the cell elastic modulus (Ecell) was obtained according to the formula:

2 tan a

Ecell p 1 m

Dz2

where F is the applied force in nN, Dz is the indentation in nm, a is the half-opening angle of the tip and m is the Poissons coefcient for typical cell, which is taken as 0.5. Thus, Ecell is obtained using the following rearrangement:

Ecell

p 1 m2

2 tana

where a FDz (obtained from the tting curves). In this form, a has a unit of

106 kPa.

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 13

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

Statistics and reproducibility of experiments. Statistical analyses were performed using Prism 6.0 (GraphPad). Sample size was not predetermined using statistical methods, but took into account the variability of the traits measured, assessed by the s.d.; no samples were excluded from the analyses. Data sets with normal distributions were analysed using one-way or two-way ANOVA, or using two-sided, unpaired Students t-tests and presented as the meanss.d. except for Figs 1e and 2b,c and Supplementary Fig. 3a where they are presented as the meanss.e.m. Data sets with non-normal distribution (for example, AFM measurements) were analysed with MannWhitney tests and presented as the medianinterquartile. Data sets with dichotomous distributions (Supplementary Fig. 1d) were analysed with a binomial test. Tukey whiskers extend to the extreme values included in the interval calculated as 1.5 IQR (interquartile range), where the IQR is calculated as the third quartile minus the rst quartile. The results for statistical signicance tests are included in the legend of each gure; n values represent the number of independent experiments performed or the number of seedlings, cells or meristems per condition. The samples were not randomized and the investigators were not blinded to allocation during experiments and outcome assessment. All micrographs are representative of at least two independent experiments.

Data availability. All data supporting the ndings of this study are available from the corresponding author on request.

References

1. Lynch, J. P. & Brown, K. M. Topsoil foragingan architectural adaptation of plants to low phosphorus availability. Plant Soil 237, 225237 (2001).

2. Lopez-Arredondo, D. L., Leyva-Gonzalez, M. A., Gonzalez-Morales, S. I., Lopez-Bucio, J. & Herrera-Estrella, L. Phosphate nutrition: improving low-phosphate tolerance in crops. Annu. Rev. Plant Biol. 65, 95123

2014:

3. Heidstra, R. & Sabatini, S. Plant and animal stem cells: similar yet different. Nat. Rev. Mol. Cell. Biol. 15, 301312 (2014).

4. Petricka, J. J., Winter, C. M. & Benfey, P. N. Control of Arabidopsis root development. Annu. Rev. Plant Biol. 63, 563590 (2012).

5. Sanchez-Calderon, L. et al. Characterization of low phosphorus insensitive mutants reveals a crosstalk between low phosphorus-induced determinate root development and the activation of genes involved in the adaptation of Arabidopsis to phosphorus deciency. Plant Physiol. 140, 879889

2006:

6. Reymond, M., Svistoonoff, S., Loudet, O., Nussaume, L. & Desnos, T. Identication of QTL controlling root growth response to phosphate starvation in Arabidopsis thaliana. Plant Cell Environ. 29, 115125 (2006).

7. Ward, J. T., Lahner, B., Yakubova, E., Salt, D. E. & Raghothama, K. G. The effect of iron on the primary root elongation of Arabidopsis during phosphate deciency. Plant Physiol. 147, 11811191 (2008).

8. Svistoonoff, S. et al. Root tip contact with low-phosphate media reprograms plant root architecture. Nat. Genet. 39, 792796 (2007).

9. Ticconi, C. A. et al. ER-resident proteins PDR2 and LPR1 mediate the developmental response of root meristems to phosphate availability. Proc. Natl Acad. Sci. USA 106, 1417414179 (2009).

10. Abel, S. Phosphate sensing in root development. Curr. Opin. Plant Biol. 14, 303309 (2011).

11. Peret, B., Clement, M., Nussaume, L. & Desnos, T. Root developmental adaptation to phosphate starvation: better safe than sorry. Trends Plant Sci. 16, 442450 (2011).

12. Muller, J. et al. Iron-dependent callose deposition adjusts root meristem maintenance to phosphate availability. Dev. Cell 33, 216230 (2015).

13. Thibaud, M. C. et al. Dissection of local and systemic transcriptional responses to phosphate starvation in Arabidopsis. Plant J. 64, 775789

2010:

14. Chen, D. L., Delatorre, C. A., Bakker, A. & Abel, S. Conditional identication of phosphate-starvation-response mutants in Arabidopsis thaliana. Planta 211, 1322 (2000).

15. Ticconi, C. A., Delatorre, C. A., Lahner, B., Salt, D. E. & Abel, S. Arabidopsis pdr2 reveals a phosphate-sensitive checkpoint in root development. Plant J. 37, 801814 (2004).

16. Gonzalez-Mendoza, V. et al. APSR1, a novel gene required for meristem maintenance, is negatively regulated by low phosphate availability. Plant Sci. 205, 12 (2013).

17. Karthikeyan, A. S. et al. Arabidopsis thaliana mutant lpsi reveals impairment in the root responses to local phosphate availability. Plant Physiol. Biochem. 77, 6072 (2014).

18. Dong, J. et al. An Arabidopsis ABC Transporter Mediates Phosphate Deciency-Induced Remodeling of Root Architecture by Modulating Iron Homeostasis in Roots. Mol. Plant 10, 244259 (2017).

19. Hoehenwarter, W. et al. Comparative expression proling reveals a role of the root apoplast in local phosphate response. BMC Plant Biol. 16, 106 (2016).

20. Hoekenga, O. A. et al. AtALMT1, which encodes a malate transporter, is identied as one of several genes critical for aluminum tolerance in Arabidopsis. Proc. Natl Acad. Sci. USA 103, 97389743 (2006).

21. Kobayashi, Y. et al. Characterization of AtALMT1 expression in aluminum-inducible malate release and its role for rhizotoxic stress tolerance in Arabidopsis. Plant Physiol. 145, 843852 (2007).

22. Iuchi, S. et al. Zinc nger protein STOP1 is critical for proton tolerance in Arabidopsis and coregulates a key gene in aluminum tolerance. Proc. Natl Acad. Sci. USA 104, 99009905 (2007).

23. Liu, J., Magalhaes, J. V., Shaff, J. & Kochian, L. V. Aluminum-activated citrate and malate transporters from the MATE and ALMT families function independently to confer Arabidopsis aluminum tolerance. Plant J. 57, 389399 (2009).

24. Sawaki, Y. et al. STOP1 regulates multiple genes that protect arabidopsis from proton and aluminum toxicities. Plant Physiol. 150, 281294

2009:

25. Kochian, L. V., Hoekenga, O. A. & Pineros, M. A. How do crop plants tolerate acid soils? Mechanisms of aluminum tolerance and phosphorous efciency. Annu. Rev. Plant Biol. 55, 459493 (2004).

26. Ryan, P., Delhaize, E. & Jones, D. Function and Mechanism of Organic Anion Exudation from Plant Roots. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52, 527560 (2001).

27. Liu, J. et al. A promoter-swap strategy between the AtALMT and AtMATE genes increased Arabidopsis aluminum resistance and improved carbon-use efciency for aluminum resistance. Plant J. 71, 327337

2012:

28. Tokizawa, M. et al. SENSITIVE TO PROTON RHIZOTOXICITY1, CALMODULIN BINDING TRANSCRIPTION ACTIVATOR2, and other transcription factors are involved in ALUMINUM-ACTIVATED MALATE TRANSPORTER1 expression. Plant Physiol. 167, 9911003 (2015).

29. Bustos, R. et al. A central regulatory system largely controls transcriptional activation and repression responses to phosphate starvation in Arabidopsis. PLoS Genet. 6, e1001102 (2010).

30. Yan, J. Y. et al. A WRKY Transcription Factor Regulates Fe Translocation under Fe Deciency. Plant Physiol. 171, 20172027 (2016).

31. Ding, Z. J., Yan, J. Y., Xu, X. Y., Li, G. X. & Zheng, S. J. WRKY46 functions as a transcriptional repressor of ALMT1, regulating aluminum-induced malate secretion in Arabidopsis. Plant J. 76, 825835 (2013).

32. Kobayashi, Y. et al. Characterization of the Complex Regulation of AtALMT1 Expression in Response to Phytohormones and Other Inducers. Plant Physiol. 162, 732740 (2013).

33. Roschzttardtz, H., Conjro, G., Curie, C. & Mari, S. Identication of the endodermal vacuole as the iron storage compartment in the Arabidopsis embryo. Plant Physiol. 151, 13291338 (2009).

34. Wolf, S., Hematy, K. & Hofte, H. Growth control and cell wall signaling in plants. Annu. Rev. Plant Biol. 63, 381407 (2012).

35. Milani, P. et al. In vivo analysis of local wall stiffness at the shoot apical meristem in Arabidopsis using atomic force microscopy. Plant J. 67, 11161123 (2011).

36. Peaucelle, A. et al. Pectin-induced changes in cell wall mechanics underlie organ initiation in Arabidopsis. Curr. Biol. 21, 17201726 (2011).37. Passardi, F., Penel, C. & Dunand, C. Performing the paradoxical: how plant peroxidases modify the cell wall. Trends Plant Sci. 9, 534540

2004:

38. Welinder, K. G. et al. Structural diversity and transcription of class III peroxidases from Arabidopsis thaliana. Eur. J. Biochem. 269, 60636081

2002:

39. Tognolli, M., Penel, C., Greppin, H. & Simon, P. Analysis and expressionof the class III peroxidase large gene family in Arabidopsis thaliana. Gene 288, 129138 (2002).

40. Valerio, L., De Meyer, M., Penel, C. & Dunand, C. Expression analysis of the Arabidopsis peroxidase multigenic family. Phytochemistry 65, 13311342 (2004).

41. Rich, P. R., Wiegand, N. K., Blum, H., Moore, A. L. & Bonner, Jr. W. D. Studies on the mechanism of inhibition of redox enzymes by substituted hydroxamic acids. Biochim. Biophys. Acta 525, 325337 (1978).

42. Balazs, C., Kiss, E., Levey, A. & Farid, N. R. The immunosuppressive effect of methimazole on cell-mediated immunity is mediated by its capacity to inhibit peroxidase and to scavenge free oxygen radicals. Clin. Endocrinol. 25, 716 (1986).

43. Herzog, V. & Fahimi, H. D. A new sensitive colorimetric assay for peroxidase using 3,3-diaminobenzidine as hydrogen donor. Anal. Biochem. 55, 554562 (1973).

44. Tsukagoshi, H., Busch, W. & Benfey, P. N. Transcriptional regulation of ROS controls transition from proliferation to differentiation in the root. Cell 143, 606616 (2010).

45. Ubeda-Tomas, S. et al. Root growth in Arabidopsis requires gibberellin/DELLA signalling in the endodermis. Nat. Cell Biol. 10, 625628 (2008).

14 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300 ARTICLE

46. Heidstra, R., Welch, D. & Scheres, B. Mosaic analyses using marked activation and deletion clones dissect Arabidopsis SCARECROW action in asymmetric cell division. Genes Dev. 18, 19641969 (2004).

47. Vermeer, J. E. et al. A spatial accommodation by neighboring cells is required for organ initiation in Arabidopsis. Science 343, 178183 (2014).

48. Pster, A. et al. A receptor-like kinase mutant with absent endodermal diffusion barrier displays selective nutrient homeostasis defects. Elife 3, e03115 (2014).

49. ONeill, E. M., Kaffman, A., Jolly, E. R. & OShea, E. K. Regulation of PHO4 nuclear localization by the PHO80-PHO85 cyclin-CDK complex. Science 271, 209212 (1996).

50. Liang, C. et al. Low pH, aluminum, and phosphorus coordinately regulate malate exudation through GmALMT1 to improve soybean adaptation to acid soils. Plant Physiol. 161, 13471361 (2013).

51. Hofand, E., Findenegg, G. R. & Nelemans, J. A. Solubilization of rock phosphate by rape I. Evaluation of the role of the nutrient uptake pattern. Plant Soil 113, 155160 (1989).

52. Ticconi, C. A. & Abel, S. Short on phosphate: plant surveillance and countermeasures. Trends Plant Sci. 9, 548555 (2004).

53. Sanchez-Calderon, L. et al. Phosphate starvation induces a determinate developmental program in the roots of Arabidopsis thaliana. Plant Cell Physiol. 46, 174184 (2005).

54. Wolf, S. & Greiner, S. Growth control by cell wall pectins. Protoplasma 249(Suppl 2): S169S175 (2012).

55. Raghothama, K. G. Phosphate Acquisition. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 665693 (1999).

56. Hinsinger, P. Bioavailability of soil inorganic P in the rhizosphere as affected by root-induced chemical changes: a review. Plant Soil 237, 173195 (2001).

57. Grillet, L. et al. Ascorbate efux as a new strategy for iron reduction and transport in plants. J. Biol. Chem. 289, 25152525 (2014).

58. Grillet, L. Spciation, transport et localisation subcellulaire du fer chez Pisum sativum et Arabidopsis thaliana. Thse de Doctorat, SupAgro Montpellier (2012).

59. Sasaki, T. et al. A wheat gene encoding an aluminum-activated malate transporter. Plant J. 37, 645653 (2004).

60. Pineros, M. A., Cancado, G. M. & Kochian, L. V. Novel properties of the wheat aluminum tolerance organic acid transporter (TaALMT1) revealed by electrophysiological characterization in Xenopus Oocytes: functional and structural implications. Plant Physiol. 147, 21312146 (2008).

61. Furuichi, T. et al. An extracellular hydrophilic carboxy-terminal domain regulates the activity of TaALMT1, the aluminum-activated malate transport protein of wheat. Plant J. 64, 4755 (2010).

62. Ramesh, S. A. et al. GABA signalling modulates plant growth by directly regulating the activity of plant-specic anion transporters. Nat. Commun. 29, 7879 (2015).

63. Sasaki, T., Tsuchiya, Y., Ariyoshi, M., Ryan, P. R. & Yamamoto, Y. A chimeric protein of aluminum-activated malate transporter generated from wheat and Arabidopsis shows enhanced response to trivalent cations. Biochim. Biophys. Acta 1858, 14271435 (2016).

64. Sasaki, T. et al. A domain-based approach for analyzing the function of aluminum-activated malate transporters from wheat (Triticum aestivum) and Arabidopsis thaliana in Xenopus oocytes. Plant Cell Physiol. 55, 21262138 (2014).

65. Ziegler, J. et al. Non-targeted proling of semi-polar metabolites in Arabidopsis root exudates uncovers a role for coumarin secretion and lignication during the local response to phosphate limitation. J. Exp. Bot. 67, 14211432 (2016).

66. Hall, H. G. Hardening of the sea urchin fertilization envelope by peroxidasecatalyzed phenolic coupling of tyrosines. Cell 15, 343355 (1978).

67. Thein, M. C. et al. Combined extracellular matrix cross-linking activityof the peroxidase MLT-7 and the dual oxidase BLI-3 is critical for post-embryonic viability in Caenorhabditis elegans. J. Biol. Chem. 284, 1754917563 (2009).

68. Bonnot, C. et al. A chemical genetic strategy identify the PHOSTIN, a synthetic molecule that triggers phosphate starvation responses in Arabidopsis thaliana. New Phytol. 209, 161176 (2016).

69. Torii, K. U. et al. The Arabidopsis ERECTA gene encodes a putative receptor protein kinase with extracellular leucine-rich repeats. Plant Cell 8, 735746 (1996).

70. Karimi, M., Depicker, A. & Hilson, P. Recombinational cloning with plant gateway vectors. Plant Physiol. 145, 11441154 (2007).

71. Clough, S. J. & Bent, A. F. Floral dip: a simplied method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, 735743 (1998).

72. Mustroph, A. et al. Proling translatomes of discrete cell populations resolves altered cellular priorities during hypoxia in Arabidopsis. Proc. Natl Acad. Sci. USA 106, 1884318848 (2009).

73. Meijering, E. et al. Design and validation of a tool for neurite tracing and analysis in uorescence microscopy images. Cytometry A 58, 167176 (2004).

74. Rasband, W. S. ImageJ U. S. National Institutes of Health, Bethesda, Maryland, USA https://imagej.nih.gov/ij/ (19972016).

75. Bahassou-Benamri, R. et al. Subcellular localization and interaction network of the mRNA decay activator Pat1 upon UV stress. Yeast 30, 353363 (2013).

76. Misson, J., Thibaud, M. C., Bechtold, N., Raghothama, K. & Nussaume, L. Transcriptional regulation and functional properties of Arabidopsis Pht1;4, a high afnity transporter contributing greatly to phosphate uptake in phosphate deprived plants. Plant Mol. Biol. 55, 727741 (2004).

77. Odorico, M. et al. An integrated methodology for data processing in dynamic force spectroscopy of ligand-receptor binding. Ultramicroscopy 107, 887894 (2007).

78. Lekka, M. et al. Elasticity of normal and cancerous human bladder cells studied by scanning force microscopy. Eur. Biophys. J. 28, 312316 (1999).

79. Sneddon, I. N. The relation between load and penetration in the axisymmetric boussinesq problem for a punch of arbitrary prole. Int. J. Eng. Sci. 3, 4757 (1965).

Acknowledgements

We thank B. Alonso (CEA, cadarache) for preparing samples in AFM experiments andJ. Mutterer (IBMP, Strasbourg) for Image-J M macro design; P. Milani (ENS, Lyon) andA. Peaucelle (INRA, Versailles) for advices in mounting samples in AFM experiments;J.-L. Montillet and D. Rumeau (CEA, cadarache), S. Mari and C. Curie (INRA, Montpellier) and C. Dunand (Univ. Paul Sabatier, Castanet-Tolosan) for fruitful discussions; J. Vermeer (IPB, Zrich) for providing the clones containing the promoters for pSCR, pCo2, pPET and pCASP1 constructs; J. Paz-Ares (CSIC, Madrid) for the phr1;phl1 double mutant; J. Liu and L. Kochian (Cornell Univ., USA) for the almt1KO;mateKO double mutant carrying the pALMT1::MATE construct; P. Benfey (Duke Univ., USA) for the upb1-1 & 35S::YFP-UP1 lines; G. Desnos for rapeseed seeds; the Nottingham Arabidopsis Stock Centre for providing Arabidopsis KO mutants; the Groupe de Recherches Appliques en Phytotechnologie (GRAP, cadarache) for plant care. Support for the microscopy equipments was provided by the Rgion Provence Alpes Cte dAzur, the Conseil Gnral des Bouches du Rhne, the French Ministry of Research, the European Union (European Regional Development Fund), the HlioBiotec platform, the CEA and the CNRS; the qRT-PCR machine was funded by Hliobiotech. This work was funded by CEA (APTTOX021401, APTTOX021403), Agence Nationale de la Recherche (ANR-09-BLAN-0118; ANR-12-ADAP-0019) and Investissements davenir (DEMETERRES). C.Ba. was supported by Agence Nationale de la Recherche (ANR-12-ADAP-0019); T.Da. was supported by Aix-Marseille-Universit, CEA (APTTOX021401) and Investissements davenir (DEMETERRES); C.M. was supported by Investissements davenir (DEMETERRES); E.L. and E.D. were supported by Agence Nationale de la Recherche (ANR-09-BLAN-0118); M.B. was supported by CEA (APTTOX021401); B.P. was supported by Agence Nationale de la Recherche(ANR Retour Post-doc EmPhos PDOC00301), EMBO Long-Term Fellowship and European Reintegration Grant under the 7th Framework Program of the European Commission (ERG-2010-276662).

Author contributions

All experimenters contributed to the design of individual experiments and analysis of results, and participated to parts of manuscript preparation. C.Ba. designed and carried out short-term kinetic experiments, contributed to callose and iron-staining experiments, cell length measurements, ICP-MS and AFM experiments. T.Da. prepared plasmids and transgenic plants, performed qRTPCR and contributed to confocal microscopy. C.G. designed and performed AFM experiments, designed and performed the malate complementation experiment, designed and supervised yeast experiments and carried out the pH experiment with GUS staining. E.L. characterized the EMS mutants and identied most almt1 and stop1 mutations, prepared plasmids and performed the yeast experiments. C.M. designed and performed some GUS staining, the ROS and peroxidase histology and meristem length experiments, participated to confocal microscopy. J.-M.T. designed and performed AFM experiments. A.C. performed mapping of almt132 and screened EMS mutants for lpr1 alleles. M.B. made plasmids and transgenic plants, and assisted the yeast experiments and contributed to ICP-MS experiment. C.Br. carried out iron staining experiments, contributed to measurements of roots and cell lengths and genotyping, participated to confocal microscopy. A.H. performed the ICP-MS analysis. J.M. contributed to iron and callose staining experiments and to malate complementation experiments. S.C. assisted with GUS histology and qRT-PCR. H.J. assisted and supervised confocal microscopy imaging and quantication. N.L.-B. contributed to genotyping. P.D. contributed to genotyping. B.P. provided materials and tips related to this study. E.D. assisted with qRT-PCR and the design of HRM markers. M.-C.T. assisted with confocal microscopy. J.A. contributed for funding. S.A. helped with discussions about hypotheses, manuscript writing and provided protocols. J.-L.P. supervised, designed and analysed AFM experiments. L.N. contributed for funding, lab administration and to Supplementary Fig. 1a. T.De. gained funding, conceived and supervised the entire project, participated to the design and analysis of all experiments, performed the mutagenesis, mutants screening, genetics analysis and mapping, identied almt132 and Stop148, performed experiments on drugs, rapeseed, root length and GUS staining, generated all the gures and tables; wrote the manuscript with feedback from S.A., C.G., H.J., C.M., C.Ba., T.Da., J.-L.P., B.P., L.N., M.-C.T. and E.D.

NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 15

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15300

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Balzergue, C. et al. Low phosphate activates STOP1-ALMT1 to rapidly inhibit root cell elongation. Nat. Commun. 8, 15300 doi: 10.1038/ncomms15300 (2017).

Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2017

16 NATURE COMMUNICATIONS | 8:15300 | DOI: 10.1038/ncomms15300 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

Word count: 14762

Show less

Abstract

Translate

Environmental cues profoundly modulate cell proliferation and cell elongation to inform and direct plant growth and development. External phosphate (Pi) limitation inhibits primary root growth in many plant species. However, the underlying Pi sensory mechanisms are unknown. Here we genetically uncouple two Pi sensing pathways in the root apex of Arabidopsis thaliana. First, the rapid inhibition of cell elongation in the transition zone is controlled by transcription factor STOP1, by its direct target, ALMT1, encoding a malate channel, and by ferroxidase LPR1, which together mediate Fe and peroxidase-dependent cell wall stiffening. Second, during the subsequent slow inhibition of cell proliferation in the apical meristem, which is mediated by LPR1-dependent, but largely STOP1-ALMT1-independent, Fe and callose accumulate in the stem cell niche, leading to meristem reduction. Our work uncovers STOP1 and ALMT1 as a signalling pathway of low Pi availability and exuded malate as an unexpected apoplastic inhibitor of root cell wall expansion.

Details

Title

Low phosphate activates STOP1-ALMT1 to rapidly inhibit root cell elongation

Author

Balzergue, Coline; Dartevelle, Thibault; Godon, Christian; Laugier, Edith; Meisrimler, Claudia; Teulon, Jean-marie; Creff, Audrey; Bissler, Marie; Brouchoud, Corinne; Hagège, Agnès; Müller, Jens; Chiarenza, Serge; Javot, Hélène; Becuwe-linka, Noëlle; David, Pascale; Péret, Benjamin; Delannoy, Etienne; Thibaud, Marie-christine; Armengaud, Jean; Abel, Steffen; Pellequer, Jean-luc; Nussaume, Laurent; Desnos, Thierry

Pages

15300

Publication year

2017

Publication date

May 2017

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms15300

ProQuest document ID

1898652626

Low phosphate activates STOP1-ALMT1 to rapidly inhibit root cell elongation

Jump to:

Full Text

Abstract

Details

Suggested sources