Differential TOR activation and cell proliferation inArabidopsis root and shoot apexesXiaojuan Lia,b,1, Wenguo Caic,1, Yanlin Liua,b,1, Hui Lia, Liwen Fua, Zengyu Liuc, Lin Xud, Hongtao Liud, Tongda Xua,2,and Yan Xionga,2

aShanghai Center for Plant Stress Biology, Chinese Academy of Sciences Center for Excellence in Molecular Plant Sciences, Shanghai Institutes for BiologicalSciences, Chinese Academy of Sciences, Shanghai 201602, People’s Republic of China; bUniversity of Chinese Academy of Sciences, Shanghai 201602, People’sRepublic of China; cHorticultural Biology and Metabolomics Center, Haixia Institute of Science and Technology, Fujian Agriculture and Forestry University,Fujian Province 350002, People’s Republic of China; and dNational Key Laboratory of Plant Molecular Genetics, Chinese Academy of Sciences Center forExcellence in Molecular Plant Sciences, Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences,Shanghai 200032, People’s Republic of China

Edited by Natasha V. Raikhel, Center for Plant Cell Biology, Riverside, CA, and approved December 29, 2016 (received for review November 13, 2016)

The developmental plasticity of plants relies on the remarkable abilityof the meristems to integrate nutrient and energy availability withenvironmental signals. Meristems in root and shoot apexes share highlysimilar molecular players but are spatially separated by soil. Whetherand how these two meristematic tissues have differential activationrequirements for local nutrient, hormone, and environmental cues (e.g.,light) remain enigmatic in photosynthetic plants. Here, we report thatthe activation of root and shoot apexes relies on distinct glucose andlight signals. Glucose energy signaling is sufficient to activate target ofrapamycin (TOR) kinase in root apexes. In contrast, both the glucoseand light signals are required for TOR activation in shoot apexes.Strikingly, exogenously applied auxin is able to replace light to activateTOR in shoot apexes and promote true leaf development. A relativelylow concentration of auxin in the shoot and high concentration ofauxin in the root might be responsible for this distinctive lightrequirement in root and shoot apexes, because light is required topromote auxin biosynthesis in the shoot. Furthermore, we reveal thatthe small GTPase Rho-related protein 2 (ROP2) transduces light-auxinsignal to activate TOR by direct interaction, which, in turn, promotestranscription factors E2Fa,b for activating cell cycle genes in shootapexes. Consistently, constitutively activated ROP2 plants stimulateTOR in the shoot apex and cause true leaf development even withoutlight. Together, our findings establish a pivotal hub role of TOR signal-ing in integrating different environmental signals to regulate distinctdevelopmental transition and growth in the shoot and root.

TOR | light | glucose | auxin | ROP2

Lacking the capability to escape from hostile environments, sessileplants adapt to their growth environments with remarkable

growth plasticity to form and regenerate organs during their life cycle.Shoot apical meristem (SAM) and root apical meristem (RAM) areresponsible for the aboveground and underground organ growth,respectively. Cell proliferation in root meristem and leaf primordiumis important for the rapid growth and production of new organs, andis controlled strictly by internal developmental cues, nutrients, hor-mones, and external environmental signals (1, 2). Understanding howgrowth and organogenesis in shoots and roots are regulated andcoordinated in complex signaling networks is a central question, andone of the most challenging questions, in plant biology.In the past decades, large efforts have been made to investigate

the key regulatory components in these processes. The most well-studied growth and organogenesis regulators in plants are phyto-hormones, which have been shown to control the meristematic celldivision activity at both shoot and root apexes (3–5). For example,in the shoot apex, auxin has been reported to trigger cell pro-liferation and differentiation, which is essential for organogenesisin leaf primordia. Cytokinin is an important cell cycle inducer tomaintain the stem cell pool by activating the expression ofWUSCHEL (WUS) in SAM (6–11). In addition, gibberellins andbrassinosteroids are implicated in determining meristem size andcell division activity (3, 12, 13).

Light is considered as one of the most important environmentalfactors. In addition to its role in activating photosynthesis, light is adirect signal that can be perceived by various photoreceptors toregulate plant development, such as leaf organogenesis and floweringtransition (14–17). Under dark conditions, the plant triggers skoto-morphogenesis, where the cell proliferation activities in shoot androot apexes are dormant and the etiolated growth is mainly by cellelongation rather than cell division. In contrast, light induces pho-tomorphogenesis of plants after they grow above the ground. Boththe shoot and root meristems are activated, which leads to rapid rootgrowth and leaf formation by promoting cell division and expansion.Recent studies have suggested that light directly influences auxin andcytokinin signaling for SAM maintenance and leaf primordia acti-vation (9, 10), although the underlying mechanism is still unclear.Besides the hormone and light signaling complexes, meristems

need to integrate nutrients and energy availability to coordinate cellproliferation activity. This integration is at least partially achieved bythe protein kinase target of rapamycin (TOR). TOR has beenknown to function as a central growth regulator that modulatesnutrient status and energy signaling to promote cell proliferation andgrowth in both animals and plants (18, 19). In Arabidopsis, TORexpression can be detected in primary meristem regions (20, 21).

Significance

The modulation of growth and development is central to sessileplants. Shoot and root meristems share almost identical molecularmachineries but are responsible for different types of above-ground and underground organogenesis. How two such highlysimilar biological systems respond to different environmental cuesand regulate distinct developmental programs remains largelyunknown. We report that, in harmony with nature, the shoot, butnot the root, demands light for the activation of cell proliferationduring organogenesis. We decipher the underlying mechanismthat light is necessary for continuous auxin biosynthesis only inthe shoot, but not the root, to activate target of rapamycin (TOR)kinase signaling. We further elucidate an auxin-Rho-related pro-tein 2 (ROP2)-TOR-transcription factors E2Fa,b signaling cascadethat integrates different environmental signals to orchestrateplant growth and development.

Author contributions: L.X., H. Liu, T.X., and Y.X. designed research; X.L., W.C., Y.L., H. Li,L.F., and Z.L. performed research; X.L., W.C., Y.L., T.X., and Y.X. analyzed data; and W.C.,T.X., and Y.X. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1X.L., W.C., and Y.L. contributed equally to this work.2To whom correspondence may be addressed. Email: yanxiong@sibs.ac.cn or tdxu@sibs.ac.cn.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1618782114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1618782114 PNAS | March 7, 2017 | vol. 114 | no. 10 | 2765–2770

PLANTBIOLO

GY

Previous work illustrated that TOR transduces photosynthesis-derived glucose energy signals specifically to control the proliferationof stem/progenitor cells in the root meristem (22). TOR directlyphosphorylates and activates the transcription factor E2Fa for rootmeristem activation (22, 23). These findings suggest that TOR kinaseplays an essential role in root meristem regulation. Interestingly, avery recent study suggested that TOR is also involved in regulatingWUS expression in SAM by transducing both light and glucose sig-naling (10), but the detailed molecular mechanism needs to beinvestigated further.Notably, different parts and stages of a plant may trigger distinct

developmental programs under varying environmental conditions.One of the key developmental transitions of plant is the heterotro-phic-to-photoautotrophic conversion when plants break up the soiland form aboveground and underground organs. A vital question iswhether and how the spatially separated shoot and root respond tothe dynamic environmental changes differentially in nature. In thisstudy, we demonstrated that the activation of cell division at theaboveground shoot apex requires both glucose and light signals butthe underground root meristem requires only glucose energy sig-naling. Strikingly, light activation of cell division at the shoot apexdepends on promoting auxin biosynthesis and accumulation. Ourwork further revealed that auxin is responsible for the distinctive lightrequirement of shoot and root activation through the Rho-relatedprotein 2 (ROP2)-TOR-E2Fa,b signaling transduction pathway.

ResultsShoot and Root Have Different Demands for Light and Glucose.Postembryonic growth starts when plant seedlings penetrate theseed coat. The shoot and root are then spatially separated after theplant breaks through the soil. The shoot and root need to sensedifferent environmental signals, including light and nutrition, to de-termine their own developmental patterns. To illustrate whether theshoot and root would respond differently to light and glucose, themitotic quiescent seedlings generated by depletion of endogenousglucose (22) were preincubated in darkness for 16 h and then treatedwith glucose and light (Fig. S1). Without glucose, the whole seedlingsremained quiescent regardless of light illumination (Fig. 1A). How-ever, glucose treatments resulted in differential responses to lightin shoot and root development. In darkness, the shoot formed anetiolated hypocotyl but could not form true leaves, whereas lighttriggered the typical photomorphogenesis of seedlings with a pair oftrue leaves and short hypocotyls (Fig. 1 A and B). In contrast, glucosesupply efficiently promoted rapid primary root growth, even in theabsence of light (Fig. 1 A and C).Because cell division is required for both root growth and leaf

organogenesis, we examined mitotic activity in the root and shootapexes in seedlings with a transcriptional mitotic reporterpCYCB1;1::GUS (24) (Fig. S2A). In agreement with the responsesobserved in seedling growth, pCYCB1;1::GUS expression reveals thatroot meristem was reactivated by glucose independent of light(Fig. 1D), consistent with our previous finding using a thymidineanalog, 5-ethynyl-2′-deoxyuridine, as a marker for S-phase entry ofthe cell cycle (22) (Fig. S2B). Furthermore, pCYCB1;1::GUS ex-pression suggests that both glucose and light are essential for mitoticactivation in the shoot apex (Fig. 1 D and E and Fig. S2A). Incontrast, glucose or light alone can only slightly activate the shootapex, suggesting that glucose and light may trigger two independentsignaling pathways to activate the shoot apex cooperatively andsynergistically (Fig. 1 D and E).Interestingly, 15 min of light exposure was sufficient to activate

pCYCB1;1::GUS expression in the shoot apex (Fig. S3A). To explorewhether light signaling is required, we further examined the re-sponses of the blue-light photoreceptor mutant cry1,cry2 (25) andred-light photoreceptor phyA,phyB (26) to blue and red light, re-spectively. In cry1,cry2 and phyA,phyB mutants, true leaf develop-ment promoted by blue or red light was greatly compromisedcompared with WT (Fig. S3B). These data indicated that light

functions as a direct signal to promote true leaf development, in-dependent of its role in photosynthesis. Taken together, thesefindings suggest that the activation of the root meristem necessaryfor rapid root growth only requires glucose, whereas both light sig-naling and glucose are essential for the activation of the shoot apexfor true leaf formation.

Both Light and Glucose Are Essential to Activate TOR Kinase in theShoot Apex. To understand how plants transduce different environ-mental signals to determine the growth patterns in different zones,we investigated the molecular mechanism of shoot and root activa-tion by light or glucose signals. Our previous studies suggested thatglucose controls TOR signaling relays to activate cell proliferation inroot meristems (22). Therefore, we tested whether TOR signalingalso controls cell division in the shoot apex. To avoid embryonic le-thality of tor null mutants, we applied chemical genetics and specificinhibitors to examine the role of TOR in shoot apex activation.Rapamycin and torin2, two chemicals that specifically inhibit TORactivity (27–29), as well as estradiol-inducible tormutants, blocked thetrue leaf formation in the presence of light and glucose (Fig. 2A andFig. S4). Furthermore, pCYCB1;1::GUS expression was also com-pletely blocked by rapamycin and torin2 (Fig. 2B), suggesting thatTOR activity is necessary for mitotic activation in the shoot apex.We then examined whether glucose and light regulate TOR sig-

naling in root and shoot apexes independently or cooperatively. Weused the rapamycin-sensitive phosphorylation of T449 in the TOR-substrate protein ribosomal S6 kinase 1 (S6K1) as a conserved in-dicator of endogenous TOR kinase activity (27). Glucose is sufficientto activate TOR kinase in the root apex (Fig. 2C). Instead, neitherlight nor glucose alone can efficiently activate TOR kinase in theshoot apex. TOR kinase could be strongly activated only when bothglucose and light are present (Fig. 2D). These data reveal that plantsintegrate light and glucose signals to control TOR kinase activitytightly for shoot development.

0.1

0.0

0.2

0.3

0.4

0.5

0.6

Fir

st t

rue

leaf

len

gth

(m

m)

L - - + +Glc - + - +

**

B

10

23456789

Ro

ot

len

gth

(m

m)

L - - + +Glc - + - +

n.s.

C

D- - + +L- + - +Glc

pC

YC

B1;

1::G

US

E

L - - + +Glc

Shoot Root- + - +

- - + +- + - +

GU

S s

tain

ing

ar

ea (

×100

µm

2 )

24

100908070

110**

** **

*

AL - - + +

Glc - + - +

Fig. 1. Shoot and root respond differently to light and glucose. (A) Differentialresponses to light and glucose in shoot and root development. The mitotic qui-escent seedlings were treated with glucose (15 mM) or light for 5 d. The redarrow indicates true leaf development. (Scale bar: 1 mm.) Measurement ofthe true leaf (B) and root (C) lengths of seedlings in A (n = 30). ** P < 0.01(ANOVA test). n.s., not significant. (D) Differential requirement of light andglucose for cell proliferation activation in shoot and root apexes. (Scale bar:100 μm.) (E) Quantification of the CYCB1-GUS intensity in D (n = 20). *P < 0.05(ANOVA test); **P < 0.01. The quiescent pCYCB1,1::GUS seedlings weretreated with glucose (15 mM) or light for 6 h. Glc, glucose; L, light.

2766 | www.pnas.org/cgi/doi/10.1073/pnas.1618782114 Li et al.

Auxin Mediates Light-Dependent Activation of TOR in the Shoot. Wethen sought to understand how light can activate TOR and why lightis required for this activation in the shoot apex but not the root apex.We speculated the involvement of the phytohormone auxin, becauseauxin is required for cell division and is reported to mediate TORsignaling in plants (30). In addition, light is one of the upstreamsignals to activate auxin signaling, and light triggers pDR5::YFP (anauxin transcriptional reporter) expression in the shoot apex (9).Interestingly, pDR5::GUS expression showed that the transcrip-tional auxin response is high in the root tip but low in the shootapex (Fig. S5A). We hypothesized that auxin is at least one of thecritical players for the activation of TOR in both the shoot androot apexes, whereas light is required for auxin accumulation inthe shoot apex but not in the root tip.To test these hypotheses, we first studied whether auxin can replace

light or glucose to activate TOR, which, in turn, promotes cell pro-liferation in the shoot apex. Exogenously applied auxin indole-3-aceticacid (IAA) or 1-naphthaleneacetic acid (NAA), together with glucose,were sufficient to activate TOR kinase, as confirmed by the phos-phorylation of T449 in S6K1, and to promote pCYCB1;1::GUS ex-pression independent of light. This activation could be completelyblocked by the TOR inhibitors rapamycin and torin2 (Fig. 3 A and Band Fig. S6). These results suggest that auxin acts downstream of thelight signaling pathway for TOR activation. Application of exogenousauxin did not further enhance mitotic activity in roots (Fig. 3A),possibly due to the saturation of endogenous auxin signaling in rootapexes. However, auxin cannot substitute glucose to activate TORand pCYCB1;1::GUS expression in the shoot apex (Fig. 3 A and B),suggesting that both glucose and auxin are indispensable for TORkinase activation. Neither glucose nor auxin alone can effectively ac-tivate TOR kinase in the shoot apex.We then examined the relation between light and auxin accu-

mulation in the shoot and the root. We used the DII-VENUStransgenic line to monitor auxin levels (31). Mitotic quiescent

seedlings were pretreated in darkness for 16 h and then transferredto the light condition. The VENUS signal in leaf primordia rapidlydecreased under the light condition (Fig. 3C), indicating that lightinduces auxin accumulation in the shoot apex. In contrast, VENUSsignal in the root tip was unaffected by light treatment (Fig. 3C).Light has been reported to promote the expression of severalYUCCAs (YUC) (32), which encode flavin monooxygenases in auxinbiosynthesis (33). Quantitative RT-PCR (QRT-PCR) analysis con-firmed that light specifically up-regulated the expression of YUC2,YUC4, and YUC7 in shoot apexes (Fig. S5C), suggesting that lighttriggers auxin accumulation partially through auxin biosynthesis.Consistent with this hypothesis, the YUCCA-specific inhibitoryucasin (34), completely blocked the light activation of cell pro-liferation in shoot apexes, but this process can be rescued by applyingNAA in the growth medium (Fig. 3D). These data provided directevidence that the light signal triggers shoot activation through pro-moting auxin biosynthesis and accumulation.Interestingly, yucasin treatment also strongly reduced auxin

accumulation and pCYCB1;1::GUS expression in the root meri-stem (Fig. 3D and Fig. S5B), even in the presence of glucose.NAA treatment restored pCYCB1;1::GUS expression afteryucasin treatment (Fig. 3D), which indicated that auxin also playsa key role in regulating cell division at the root meristem. Inaddition, yucasin strongly inhibited TOR activity, which can berescued by NAA (Fig. 3E).

L - - + + + +Glc - + - + + +Inh - - - - Rap Torin2

P-T499

S6K1

Root apex

B

A

C D

Torin2RapamycinDMSO

pC

YC

B1;

1::G

US

P-T499

S6K1

L - - + + + +Glc - + - + + +Inh - - - - Rap Torin2

Shoot apex

+ ++ L+ + +Glc

7d

Inh - Rapamycin Torin2XVE-TOR RNAi

Fig. 2. Light and glucose cooperatively activate TOR in the shoot apex.(A) TOR-dependent true leaf development. The mitotic quiescent WT and XVE-TOR-RNAi seedlings were treated with light and glucose (15 mM) with orwithout rapamycin (10 μM), torin2 (25 μM), or estradiol (10 μM) for 7 d. (Scalebar: 1 mm.) (B) TOR-dependent shoot apex activation. The mitotic quiescentpCYCB1,1::GUS seedlings were treated with glucose and light with or withoutrapamycin (10 μM) or torin2 (25 μM) for 6 h. (Scale bar: 100 μm.) (C) Effect ofglucose and light on TOR activity in the root apex. (D) Effect of glucose andlight on TOR activity in the shoot apex. The mitotic quiescent 35S::S6K1-HAseedlings were subject to 30 min of light exposure and glucose treatmentswith or without the inhibitor (Inh) rapamycin (10 μM) or torin2 (25 μM).

YucasinNAA

++

+-

--

pC

YC

B1;

1::G

US

E

B

C

Yucasin ++-NAA +--

P-T499

S6K1

L --- - +-- ++Glc ++- + -+- -+

Torin2RapInh - - --NAA +++ + +-

---

---

P-T499

S6K1

DII

mD

IID

IIm

DII

Light

- +

D

Rel

ativ

e D

II/m

DIIi

nte

nsi

ty

LightShoot apex

n.s.

*

Root apex

1.01.2

0.80.60.40.2

0

1.41.6

+- +-

A-Glc ++ ++ +-- -

L - -+ -- --+ +

-NAA -- ++ ++-- +-Inh -- Torin2- Rap-- -

pC

YC

B1;

1::G

US

Fig. 3. Light-auxin activation of TOR for shoot development. (A) Auxin substitutedlight to activate pCYCB1,1::GUS expression through the TOR-dependent pathway.(Scale bar: 100 μm.) (B) Auxin substituted light to activate TOR in the shoot apex.(C) Light increased the auxin level in the shoot, but not root, apexes. The mitoticquiescent DII-VENUS and mDII-VENUS seedlings were subjected to light treatmentfor 4 h (Upper), and the relative DII/mDII-VENUS signals were quantified (Lower)(n = 10). *P < 0.05 (ANOVA test); n.s., not significant. (Scale bar: 50 μm.) (D) Yucasinsuppressed pCYCB1,1::GUS expression in both the shoot and root apexes, whichwas rescued by auxin. (Scale bar: 100 μm.) (E) Yucasin suppressed TOR activity, whichwas rescued by auxin. In A, B, D, and E, the mitotic quiescent pCYCB1,1::GUS or35S::S6K1-HA seedlingswere treatedwith light or glucosewithout orwith inhibitorsas indicated (10 μM rapamycin, 25 μM torin2, 100 nMNAA, 15mMglucose, 100 μMyucasin).

Li et al. PNAS | March 7, 2017 | vol. 114 | no. 10 | 2767

PLANTBIOLO

GY

All of the above data strongly suggest that besides glucose, auxinsignaling is another key upstream regulator for TOR kinase activa-tion. Both glucose and auxin are essential for activating TOR kinase.In the root meristem, a relatively high concentration of auxin withthe presence of glucose is enough to active TOR kinase in the dark.However, in shoot apexes, light is indispensable for continuouslypromoting auxin accumulation for TOR activation.

Rho GTPase Mediates the Auxin Activation of TOR Kinase. We nextexamined how auxin activates TOR. ROP GTPases in plants havebeen considered as central molecular switches to regulate manyplant developmental processes (35). Interestingly, in Arabidopsis,ROPs are activated either by light to regulate stomata opening orby auxin for leaf pavement cell development (36–38). Arabidopsisseedlings expressing a constitutively active (CA) form of ROP2exhibit constitutively photomorphogenesis phenotypes in the dark(39). Furthermore, both ROP2 and TOR are expressed in theshoot apex (20, 39). These results suggest that ROP2 maytransduce the light-auxin signal to activate TOR kinase.To test the functional link between ROP2 and TOR kinase, we

first interrogated whether ROP2 directly interacts with TOR.We found that ROP2 was coimmunoprecipitated with TOR fromshoot apexes of seedlings (Fig. 4A). We then tested whetherROP2 activates TOR kinase activity. In Arabidopsis leaf cells,S6K1-Flag was coexpressed with either ROP2-HA or CA-ROP2-HA, and the TOR activity was monitored based on the phos-phorylation status of T449 in S6K1. In nonstressed leaf cells,TOR kinase is already well activated and CA-ROP2-HA onlymildly increased TOR kinase activity compared with ROP2-HA,possibly due to the saturated TOR activation (Fig. 4B andFig. S7A). In contrast, when yucasin treatment inhibited TORkinase activity, NAA and CA-ROP2-HA, but not ROP2-HA,were able to reactivate TOR kinase (Fig. 4B and Fig. S7A).These data suggest that auxin activates ROP2 to stimulate TORkinase activity.We then examined whether ROP2 activation relies on light, auxin,

or glucose. GTP-bound active ROP2 in Arabidopsis can be coim-munoprecipitated by ROP-interactive CRIB motif-containing pro-tein 1 (RIC1), which only binds active ROPs (37). ROP2 activity washigher in seedlings grown under the light condition than in darkness(Fig. 4C and Fig. S7B). Notably, glucose had no effect on ROP2activation, further demonstrating that light and glucose trigger twoindependent signaling pathways but that both are required for TORkinase activation (Fig. 4C and Fig. S7B). We then examined whetherauxin mediates the light activation of ROP2. In yucasin-treatedseedlings, light lost the capability to activate ROP2. On the otherhand, ROP2 was activated in the dark-grown seedlings when treatedwith NAA (Fig. 4D and Fig. S7C). These results indicate thatROP2’s activation by light relies on auxin.The above results indicated that light-auxin signal can activate

ROP2, and active ROP2 then promotes TOR kinase activity.Therefore, we anticipated that the 35S::CA-ROP2 seedlings wouldhave constitutive mitotic activation in shoot apexes under darkness.Our results clearly showed that CA-ROP2 overexpression indeedactivates the expression of pCYCB1;1::GUS in shoot apexes in thedark (Fig. 4E). Moreover, true leaf development was observed in thedark-grown 35S::CA-ROP2 seedlings, resembling the light-grownWT seedlings, as reported previously (39) (Fig. 4F). Importantly,rapamycin and torin2 strongly inhibited the true leaf developmentinduced by CA-ROP2 (Fig. 4F), indicating that TOR kinase actsdownstream of CA-ROP2 signaling. Notably, glucose is still requiredfor TOR kinase and shoot apex activation even with CA-ROP2overexpression (Fig. 4 E and F). These data suggest that CA-ROP2can replace the light-auxin signaling, but not the glucose signaling, toactivate TOR kinase activity and cell proliferation.

TOR Kinase Targets E2Fa and E2Fb to Control Shoot Development. Tounderstand how TOR kinase regulates the activation of both shoot

and root apexes following the perception of different environ-mental signals, we investigated the specific downstream targets ofthe TOR kinase in shoots and roots. Our previous studies revealedthat TOR controls root meristem activation through directlyphosphorylating transcription factor E2Fa and promotes its activityin activation of S-phase genes. However, no obvious difference wasobserved between WT and e2fa in shoot development (Fig. 5A).There are six members in the Arabidopsis E2F gene family (40),and E2Fa is only predominantly expressed in root apexes but not inshoot apexes (22). Interestingly, similar to E2Fa, ectopic expressionof E2Fb alone was sufficient to activate S-phase–specific markergenes, and the S-phase gene activation was inhibited by rapamycintreatment (Fig. S8A). Furthermore, endogenous TOR kinase fromArabidopsis plants phosphorylated E2Fb in vitro, which wascompletely blocked by torin2 (Fig. 5B). E2Fb could also beimmunoprecipitated with TOR protein in plant leaf cells (Fig.5C). These results indicate that E2Fb is another direct TOR ki-nase substrate. However, there was no obvious difference betweenWT and the e2fb mutant in true leaf development (Fig. 5A),suggesting a functional redundancy with E2Fa for shoot apex ac-tivation. We crossed e2fa and e2fb single-null mutant but could not

Active ROP2

- - + +L- + - +Glc

Total ROP2

C

D L - + - +Glc + + + +

- - + -NAA- - - +Yucasin

Active ROP2

Total ROP2

Glc +-+ + +L -++ - -

Rap --- + -Torin2 --- - +

wild type

CA-ROP2

A B

E

F

pCYCB1;1::GUSCA-ROP2×

pCYCB1;1::GUSL - - - -

Glc - + - +

pC

YC

B1;

1::G

US

++ - -++ +sGFP + + + - --+ + ++- +S6K1-Flag + + + + +-- + -- - -ROP2 - - - + --- - +--- -CA-ROP2 - - - - +-+ - --- -Torin2 + - - - -- - - + + + +- - ---Yucasin- - - - + - -- - ---NAA

P-T449S6K1-FlagROP2-HA

ROP2

Input

CoIPTOR

ROP2TORTOR - +

Fig. 4. ROP2 mediated the light-auxin activation of TOR. (A) ROP2 directlyinteractedwith TOR kinase as tested by immunoprecipitation from shoot apexes ofseedlings andWestern blot analysis. (B) CA-ROP2 activated TOR kinase. Protoplastscoexpressing S6K1 with ROP2-HA or CA-ROP2-HA were treated with 100 nMyucasin or 20 nM NAA. TOR activation was detected using an anti–phospho-T449antibody of S6K1. (C) Light, but not glucose, activated ROP2. (D) Auxin replacedlight to activate ROP2 in the dark. The mitotic quiescent seedlings were pretreatedwith 100 μM yucasin or 100 nM NAA for 1 h before the 15-min light/glucosetreatment. (E) CA-ROP2 activated pCYCB1,1::GUS expression in the dark. (Scale bar:100 μm.) (F) CA-ROP2 promoted true leaf development in the dark. The mitoticquiescent WT and 35S::CA-ROP2 seedlings were incubated under light or darkconditions without or with rapamycin (10 μM) and torin2 (25 μM) for 5 d. The redarrow indicates true leaf development. (Scale bar: 1 mm.)

2768 | www.pnas.org/cgi/doi/10.1073/pnas.1618782114 Li et al.

obtain the homozygous e2fa,e2fb null mutant because of siliquedefects (Fig. S8B) and embryo lethality, which indicated thatE2Fa/b has essential roles in plant development. Therefore, wegenerated the E2Fb RNAi transgenic plants in a null e2fa mutantbackground. Although there was only about 50% transcript re-duction of E2Fb in the e2fa/E2Fb RNAi double mutants (Fig.S8C), the true leaf development promoted by light and glucosetreatment was significantly compromised (Fig. 5A). QRT-PCRanalysis demonstrated that e2fa/E2Fb RNAi mutants displayedspecifically diminished glucose/light sensitivity in TOR activationof S-phase genes in the shoot apex (Fig. 5D), providing geneticevidence for a key role of E2Fa and E2Fb in TOR kinase tran-scriptional networks governing shoot apex activation (Fig. S9),although other key transcription factors may also function down-stream of TOR signaling.

DiscussionAll living organisms must integrate internal and external cues toensure survival and to dictate their growth and development. Thisintegration is of particular importance for sessile plants becausethey need to deal with intensely changing environments withoutthe option of moving to conditions optimal for their growth. Plantssolve this dilemma by generating different organs over their lifecycle, such as the aerial shoot organs and underground root organs(41, 42). Shoots and roots are spatially separated by soil, whichphysically forces them to sense distinctive environmental sig-nals from aboveground or underground. How the abovegroundshoot system and underground root system are regulated andcoordinated in complex signaling networks remains a crucialquestion in plant biology. TOR kinase has been known to be aconserved central growth regulator in animals and plants. In thisstudy, we found that TOR kinase is essential for the maintenanceof mitotic activity in both the shoot and root apexes. By inte-grating light, phytohormone auxin, and glucose energy signal,TOR kinase directly phosphorylates transcription factors E2Faand E2Fb, followed by promoting the transcriptional activation

of S-phase genes for cell proliferation activation in both the shootand root apexes.The shoot and the root apexes play different roles in plant

development but share highly similar genetic programs, withalmost the same molecular players encoded by closely related,tissue-specific homologs (43). One key question is whether andhow different environment signals from aboveground andunderground can determine the differential organogenesis inshoot and root systems. Our findings suggested that develop-ment of the shoot and root has distinct requirements for glu-cose and light signals. Glucose energy signaling is essential forboth shoot and root development through activating TORkinase and its downstream targets. However, light signal is onlynecessary for the shoot, but not the root. To explain this dif-ferential requirement of light, we further discovered that auxinis able to replace light to activate TOR and regulate shootdevelopment, which suggested that auxin is another in-dispensable upstream signal of the TOR signaling pathway.Neither glucose nor auxin alone can effectively activate TOR.We proposed that there might exist a dosage requirement forauxin to activate TOR signaling in the shoot and root apexes.The low concentration of auxin in the shoot apex is not enoughfor the continuous activation of TOR; thus, light is required topromote auxin synthesis at the shoot apex. In addition, theTOR kinase in the shoot apex would not be activated until lightincreases the auxin level. However, in the root meristem, a highconcentration of auxin alone can activate TOR kinase withoutlight. During evolution, plants seem to have developed a subtlemechanism to modulate TOR kinase and meristem activitybased on the unequal local distribution of auxin and the envi-ronmental conditions. Interestingly, recent evidence revealed thatTOR might also be involved in brassinosteroid signaling, becauseTOR kinase controls the accumulation of the brassinosteroid sig-naling transcription factor BZR1 by inhibiting autophagy (44) andTOR kinase substrate S6K2 directly phosphorylates BIN2 (45),although the direct link between brassinosteroid signaling andTOR needs to be verified experimentally. It would be intriguing toinvestigate whether other hormones, such as cytokinin and bras-sinosteroid, also have such unequal local distribution in differentmeristems and mediate TOR signaling to modulate plant growthand development in response to the dynamic environment.TOR kinase has been identified as a central coordinator of nu-

trients, energy, hormones, and stress signaling networks in bothplants and animals (18, 19, 46). However, how a single protein kinasecould act as a master regulator to incorporate various upstreamsignals to modulate a myriad of cellular activities remains largelyunknown. In contrast to recent progress in discovering the down-stream effectors of plant TOR signaling [e.g., S6K1, S6K2, E2Fa,E2Fb, RPS6, BZR1, BIN2 (19, 22, 27, 44, 45, 47)], the upstreamregulators of plant TOR signaling remain poorly defined. The smallGTPases RHEB and RAG function as the key upstream regulatorsfor mTOR activation in mammalians, but no orthologs of RHEB orRAG GTPases have been identified in plants (18, 19). Instead,plants possess a unique family of Rho-like small GTPases calledROPs (39). We found that ROP2 interacts with TOR in vivo andCA-ROP2 can rescue the inhibitory effect of yucasin to promoteTOR kinase activity strongly. Previous work showed that both lightand auxin can activate ROP2 (36, 39), and our work proved thatoverexpression of CA-ROP2 activates TOR kinase and mitotic ac-tivity in shoots even without light. These results strongly suggestedthat ROP2 is a key upstream regulator of TOR kinase in Arabidopsisthat can transduce light-auxin signals to TOR kinase-dependentmeristem activation. However, glucose cannot activate ROP2 andCA-ROP2 cannot replace glucose for TOR and meristem activa-tion, indicating that the glucose signal is sensed and transduced byother upstream regulator(s) for TOR activation. In mammalian cells,AMP-activated kinase (AMPK) is another upstream regulatorfor mTOR activation. Recent work found that the plant AMPK

wild typeA

DB

C

TORanti-TOR - + - +

TORE2Fb

E2Fb

Control

E2Fb

Input

IP

TOR-- +

-++Torin2

[32P]E2FbE2Fb

e2fbe2fa e2fa/ E2Fb RNAi

wild type -light/-glucosewild type +light/+glucosee2fa/ E2Fb RNAi -light/-glucosee2fa/ E2Fb RNAi +light/+glucose

MCM3

MCM5

ORC2

ETG1

PCNA10

50

100

150

Fo

ld c

han

ges

****

**

****

Fig. 5. TOR kinase regulates shoot and root growth through E2Fa and E2Fb.(A) First pair of true leaf growth was detected in e2fa/E2Fb RNAi plants.(Scale bar: 1 mm.) The mitotic quiescent seedlings were treated with glucose(15 mM) and light for 5 d. The red arrow indicates true leaf development.(B) TOR kinase directly phosphorylated E2Fb. (C) Interaction of E2Fb withTOR was examined by immunoprecipitation (IP) and Western blot analysis.(D) S-phase gene activation by glucose and light was diminished in e2fa/E2FbRNAi shoot apexes (n = 3). **P < 0.01 (ANOVA test).

Li et al. PNAS | March 7, 2017 | vol. 114 | no. 10 | 2769

PLANTBIOLO

GY

ortholog KIN10 directly interacts with and phosphorylates Raptor,one of the main components in the TOR complex (48). It is worthtesting whether KIN10 or its homologs act as other TOR upstreamregulators to mediate glucose signaling cascades.The findings that glucose and auxin cooperate to control TOR

signaling reveals a missing link in nutrient/phytohormone regulationof shoot and root meristems. It will be of great interest to explore therole of TOR signaling in supporting meristem activities in othersink organs (e.g., tubers, flowers, fruits, seeds), which is core to ag-ricultural and bioenergy productivity. Elucidating the molecularregulatory network underlying the central functions of TOR in plantgrowth may enable a more targeted genetic modification of plantarchitecture, carbon allocation, and resistance to stress and dis-eases. Furthermore, understanding the novel molecular mechanismsof plant TOR signaling will enhance our fundamental knowledge of

this master regulator in other organisms, including humans, wheredysregulation of mTOR has been linked to various human diseases.

Materials and MethodsPlant materials and growth conditions, immunoblots, in vitro kinase as-says, coimmunoprecipitation, QRT-PCR, and primers (Table S1), are describedin SI Materials and Methods.

ACKNOWLEDGMENTS. We thank Mark D. Curtis for the estradiol-induciblebinary vector, John L. Celenza for pCYCB1;1::GUS, Ji�rí Friml for pDR5::GFPseeds, and Jen Sheen for critical reading of the manuscript. This re-search is supported by the Horticultural Biology and Metabolomics Cen-ter, Fujian Agriculture and Forestry University (W.C.), by NationalNatural Science Foundation of China (Grant 31422008 to T.X.), by theRecruitment Program of Global Experts (People’s Republic of China, T.X.and Y.X.), and funding from the Chinese Academy of Sciences (T.X.and Y.X.).

1. Petricka JJ, Winter CM, Benfey PN (2012) Control of Arabidopsis root development.Annu Rev Plant Biol 63(1):563–590.

2. Traas J, Bohn-Courseau I (2005) Cell proliferation patterns at the shoot apical meri-stem. Curr Opin Plant Biol 8(6):587–592.

3. Vanstraelen M, Benková E (2012) Hormonal interactions in the regulation of plantdevelopment. Annu Rev Cell Dev Biol 28(1):463–487.

4. Chen L, et al. (2016) YUCCA-mediated auxin biogenesis is required for cell fatetransition occurring during de novo root organogenesis in Arabidopsis. J Exp Bot67(14):4273–4284.

5. Moubayidin L, et al. (2010) The rate of cell differentiation controls the Arabidopsisroot meristem growth phase. Curr Biol 20(12):1138–1143.

6. Yanai O, et al. (2005) Arabidopsis KNOXI proteins activate cytokinin biosynthesis. CurrBiol 15(17):1566–1571.

7. Kurakawa T, et al. (2007) Direct control of shoot meristem activity by a cytokinin-activating enzyme. Nature 445(7128):652–655.

8. Gordon SP, Chickarmane VS, Ohno C, Meyerowitz EM (2009) Multiple feedback loopsthrough cytokinin signaling control stem cell number within the Arabidopsis shootmeristem. Proc Natl Acad Sci USA 106(38):16529–16534.

9. Yoshida S, Mandel T, Kuhlemeier C (2011) Stem cell activation by light guides plantorganogenesis. Genes Dev 25(13):1439–1450.

10. Pfeiffer A, et al. (2016) Integration of light and metabolic signals for stem cell acti-vation at the shoot apical meristem. eLife 5:e17023.

11. Dello Ioio R, et al. (2007) Cytokinins determine Arabidopsis root-meristem size bycontrolling cell differentiation. Curr Biol 17(8):678–682.

12. González-García MP, et al. (2011) Brassinosteroids control meristem size by pro-moting cell cycle progression in Arabidopsis roots. Development 138(5):849–859.

13. Fu X, Harberd NP (2003) Auxin promotes Arabidopsis root growth by modulatinggibberellin response. Nature 421(6924):740–743.

14. Liu H, et al. (2008) Photoexcited CRY2 interacts with CIB1 to regulate transcriptionand floral initiation in Arabidopsis. Science 322(5907):1535–1539.

15. Li J, Terzaghi W, Deng XW (2012) Genomic basis for light control of plant develop-ment. Protein Cell 3(2):106–116.

16. Jiao Y, Lau OS, Deng XW (2007) Light-regulated transcriptional networks in higherplants. Nat Rev Genet 8(3):217–230.

17. Nemhauser J, Chory J (2002) Photomorphogenesis. Arabidopsis Book 1:e0054.18. Xiong Y, Sheen J (2014) The role of target of rapamycin signaling networks in plant

growth and metabolism. Plant Physiol 164(2):499–512.19. Dobrenel T, et al. (2016) TOR signaling and nutrient sensing. Annu Rev Plant Biol

67(1):261–285.20. Menand B, et al. (2002) Expression and disruption of the Arabidopsis TOR (target of

rapamycin) gene. Proc Natl Acad Sci USA 99(9):6422–6427.21. Ren M, et al. (2012) Target of rapamycin signaling regulates metabolism, growth, and

life span in Arabidopsis. Plant Cell 24(12):4850–4874.22. Xiong Y, et al. (2013) Glucose-TOR signalling reprograms the transcriptome and ac-

tivates meristems. Nature 496(7444):181–186.23. Xiong Y, Sheen J (2013) Moving beyond translation: Glucose-TOR signaling in the

transcriptional control of cell cycle. Cell Cycle 12(13):1989–1990.24. DiDonato RJ, et al. (2004) Arabidopsis ALF4 encodes a nuclear-localized protein re-

quired for lateral root formation. Plant J 37(3):340–353.25. Mockler TC, Guo H, Yang H, Duong H, Lin C (1999) Antagonistic actions of Arabidopsis

cryptochromes and phytochrome B in the regulation of floral induction. Development126(10):2073–2082.

26. Smith H, Xu Y, Quail PH (1997) Antagonistic but complementary actions of phyto-chromes A and B allow seedling de-etiolation. Plant Physiol 114(2):637–641.

27. Xiong Y, Sheen J (2012) Rapamycin and glucose-target of rapamycin (TOR) proteinsignaling in plants. J Biol Chem 287(4):2836–2842.

28. Montané MH, Menand B (2013) ATP-competitive mTOR kinase inhibitors delay plantgrowth by triggering early differentiation of meristematic cells but no developmentalpatterning change. J Exp Bot 64(14):4361–4374.

29. Liu Q, et al. (2012) Kinome-wide selectivity profiling of ATP-competitive mammaliantarget of rapamycin (mTOR) inhibitors and characterization of their binding kinetics.J Biol Chem 287(13):9742–9752.

30. Schepetilnikov M, et al. (2013) TOR and S6K1 promote translation reinitiation ofuORF-containing mRNAs via phosphorylation of eIF3h. EMBO J 32(8):1087–1102.

31. Brunoud G, et al. (2012) A novel sensor to map auxin response and distribution athigh spatio-temporal resolution. Nature 482(7383):103–106.

32. Yokawa K, Koshiba T, Baluška F (2014) Light-dependent control of redox balance andauxin biosynthesis in plants. Plant Signal Behav 9:e29522.

33. Zhao Y, et al. (2001) A role for flavin monooxygenase-like enzymes in auxin bio-synthesis. Science 291(5502):306–309.

34. Nishimura T, et al. (2014) Yucasin is a potent inhibitor of YUCCA, a key enzyme inauxin biosynthesis. Plant J 77(3):352–366.

35. Craddock C, Lavagi I, Yang Z (2012) New insights into Rho signaling from plant ROP/Rac GTPases. Trends Cell Biol 22(9):492–501.

36. Jeon BW, et al. (2008) The Arabidopsis small G protein ROP2 is activated by light inguard cells and inhibits light-induced stomatal opening. Plant Cell 20(1):75–87.

37. Xu T, et al. (2010) Cell surface- and rho GTPase-based auxin signaling controls cellularinterdigitation in Arabidopsis. Cell 143(1):99–110.

38. Xu T, et al. (2014) Cell surface ABP1-TMK auxin-sensing complex activates ROP GTPasesignaling. Science 343(6174):1025–1028.

39. Li H, Shen JJ, Zheng ZL, Lin Y, Yang Z (2001) The Rop GTPase switch controls multipledevelopmental processes in Arabidopsis. Plant Physiol 126(2):670–684.

40. Mariconti L, et al. (2002) The E2F family of transcription factors from Arabidopsisthaliana. Novel and conserved components of the retinoblastoma/E2F pathway inplants. J Biol Chem 277(12):9911–9919.

41. Fankhauser C, Christie JM (2015) Plant phototropic growth. Curr Biol 25(9):R384–R389.42. Ge L, Chen R (2016) Negative gravitropism in plant roots. Nature Plants 2:16155.43. Sarkar AK, et al. (2007) Conserved factors regulate signalling in Arabidopsis thaliana

shoot and root stem cell organizers. Nature 446(7137):811–814.44. Zhang Z, et al. (2016) TOR signaling promotes accumulation of BZR1 to balance

growth with carbon availability in Arabidopsis. Curr Biol 26(14):1854–1860.45. Xiong F, et al. (2016) Brassinosteriod Insensitive 2 (BIN2) acts as a downstream ef-

fector of the Target of Rapamycin (TOR) signaling pathway to regulate photoauto-trophic growth in Arabidopsis. New Phytol 213(1):233–249.

46. Xiong Y, Sheen J (2015) Novel links in the plant TOR kinase signaling network. CurrOpin Plant Biol 28:83–91.

47. Dobrenel T, et al. (2016) The Arabidopsis TOR kinase specifically regulates the ex-pression of nuclear genes coding for plastidic ribosomal proteins and the phosphor-ylation of the cytosolic ribosomal protein S6. Front Plant Sci 7(1611):1611.

48. Nukarinen E, et al. (2016) Quantitative phosphoproteomics reveals the role of theAMPK plant ortholog SnRK1 as a metabolic master regulator under energy depri-vation. Sci Rep 6:31697.

49. Ulmasov T, Murfett J, Hagen G, Guilfoyle TJ (1997) Aux/IAA proteins repress expres-sion of reporter genes containing natural and highly active synthetic auxin responseelements. Plant Cell 9(11):1963–1971.

2770 | www.pnas.org/cgi/doi/10.1073/pnas.1618782114 Li et al.