Oscillations in extracellular pH and reactive oxygenspecies modulate tip growth of Arabidopsis root hairsG. B. Monshausen*†‡, T. N. Bibikova†, M. A. Messerli§, C. Shi†, and S. Gilroy*†¶

*Department of Botany, University of Wisconsin, Madison, WI 53706; †Biology Department, Pennsylvania State University, 208 Mueller Laboratory,University Park, PA 16802; §BioCurrents Research Center, Marine Biological Laboratory, Woods Hole, MA 02543; and ‡Botanical Institute 1,University of Karlsruhe, 76128 Karlsruhe, Germany

Edited by Enrico Coen, John Innes Centre, Norwich, United Kingdom, and approved October 29, 2007 (received for review September 10, 2007)

Root hairs show highly localized cell expansion focused to theirgrowing tips. This growth pattern is accomplished through restric-tion of secretion to the elongating apex and modulation of cell wallproperties, with the wall just behind the tip becoming rigidified toresist the lateral expansive forces of turgor. In this report we showthat root hairs exhibit oscillating growth that is associated withoscillating increases in extracellular pH and reactive oxygen species(ROS), which lag growth by �7 s. Consistent with a role for thesechanges in growth control, artificially increasing extracellular pHarrested root hair elongation, whereas decreasing pH elicitedbursting at the tip. Similarly, application of exogenous ROS ar-rested elongation, whereas scavenging of ROS led to root hairbursting. Roots hairs of the root hair-defective rhd2-1 mutant,which lack a functional version of the NADPH oxidase ATRBOH C,burst at the transition to tip growth. This phenotype could berescued by elevating the pH of the growth medium to >6.0. Suchrescued root hairs showed reduced cytoplasmic ROS levels and alack of the oscillatory production of ROS at the tip. However, theyexhibited apparently normal tip growth, including generation ofthe tip-focused Ca2� gradient thought to drive apical growth,indicating that ATRBOH C is not absolutely required to sustain tipgrowth. These observations indicate that root hair elongation iscoupled to spatially distinct regulation of extracellular pH and ROSproduction that likely affect wall properties associated with thepolarized expansion of the cell.

NADPH oxidase � AtRBOH C � rhd2 � proton flux

Apically growing cells, such as root hairs, provide an impor-tant model with which to study the dynamic regulation of

growth. Root hair elongation is driven by turgor pressure andmaintained by the activity of the exocytotic machinery thatdelivers new material to the expanding point of the root hair tip(1, 2). A tube-like growth habit is then maintained by thesubapical wall resisting the expansive forces of turgor.

In the apical cytoplasm, it is well established that Ca2�, thecytoskeleton, monomeric G proteins, and a host of other cyto-plasmic factors play an important role in regulating the activityof the secretory apparatus that sustains growth (3–6). However,much less is known about how the spatial patterning in wallstructure contributes to restricting growth to the tip (e.g., ref. 7).Reactive oxygen species (ROS) produced by the plasma mem-brane NADPH oxidase ATRBOH C are thought to be indis-pensable for tip growth because they activate Ca2�-permeablechannels required to generate the tip-focused Ca2� gradient thatdrives apical growth (8). However, such extracellular ROSproduction may also directly affect cell wall structure (e.g., refs.9–11).

Changes in cell wall pH have also been reported to affect bothdiffuse (12) and localized cell expansion (13). Indeed, there is alarge, tip-localized oscillatory influx of H� accompanying thegrowth of lily pollen tubes (14), whereas small, steady, inward-or outward-directed H� f luxes have been reported aroundgrowing root hairs (15, 16). Despite such evidence that H� f luxes

are associated with tip growth, their biological function remainsto be defined.

We report here that, during root hair growth, oscillatoryincreases in extracellular pH and ROS work together to likelymodulate wall properties to support tip growth. Analysis of roothair growth of the rhd2-1 mutant indicates that the NADPHoxidase AtRBOH C is required for localized, oscillatory ROSproduction. However, the lesion in tip growth in this mutant canbe compensated for by increased extracellular pH, indicatingthat extracellular pH and ROS may be acting in a coordinatedand complementary manner to restrict expansion of the growinghair.

ResultsRoot Hair Growth Is Associated with Oscillating H� Fluxes. Wedeveloped an approach to visualize pH changes along the surfaceof growing root hairs based on supplementing the growthmedium with the pH-sensitive fluorescent dye fluorescein (con-jugated to a 10-kDa dextran). Confocal ratio imaging was thenused to generate two-dimensional maps of pH around the roothairs [Fig. 1, supporting information (SI) Materials and Methods,and SI Movie 1]. This approach revealed that root hair tip growthwas tightly associated with oscillatory changes in surface pH(Fig. 1 and SI Fig. 8). Although the amplitude of the extracellularpH oscillations varied considerably between different root hairs(0.1–1.2 pH units), their frequency was relatively constant at 2–4oscillations per minute (Fig. 1 and SI Fig. 8). The pH oscillationswere restricted to the apical 5–10 �m of the root hair tip, withthe largest changes focused around the extreme apex of thegrowing hair (Fig. 1). No changes in extracellular pH weredetected along the root hair f lanks (Fig. 1) or base or aroundnongrowing root hairs (SI Fig. 9A). Simultaneous measurementof cytosolic pH using plants expressing a pH sensitive variant ofGFP (17) revealed that a transient cytosolic acidification ac-companied each extracellular alkalinization (Fig. 1, SI Fig. 8, andSI Movie 1), suggesting oscillatory H� f luxes across the plasmamembrane. Measurement of these f luxes by using a self-referencing H�-selective microelectrode showed peak influxdensities of up to 90 pmol cm�2 s�1 (SI Fig. 9B).

High-resolution growth analysis (14) indicated that, under ourgrowth conditions, Arabidopsis root hairs consistently grew in anoscillating manner, with an average period of 18.2 � 0.6 s (n �

Author contributions: G.B.M. and T.N.B. contributed equally to this work; G.B.M., T.N.B.,and S.G. designed research; G.B.M. and T.N.B. performed research; G.B.M. and T.N.B.contributed new reagents/analytic tools; G.B.M., T.N.B., M.A.M., C.S., and S.G. analyzeddata; and G.B.M., T.N.B., M.A.M., and S.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

See Commentary on page 20649.

¶To whom correspondence should be addressed at: Department of Botany, University ofWisconsin, Birge Hall, 430 Lincoln Drive, Madison, WI 53706. E-mail: sgilroy@wisc.edu.

This article contains supporting information online at www.pnas.org/cgi/content/full/0708586104/DC1.

© 2007 by The National Academy of Sciences of the USA

20996–21001 � PNAS � December 26, 2007 � vol. 104 � no. 52 www.pnas.org�cgi�doi�10.1073�pnas.0708586104

Dow

nloa

ded

by g

uest

on

June

13,

202

0

10). Each large peak in growth was associated with a strongsurface alkalinization localized to the growing apex; smallergrowth peaks were coupled to less-pronounced pH increases(Fig. 2A). Cross-correlation analysis revealed that the oscillatorypH increases most likely lag growth oscillations by 7.0 � 0.3 s(Fig. 2B).

To investigate a potential functional role of these extracellularpH changes, we manipulated the extracellular pH by bufferingthe medium. When the medium was alkalinized to pH 8 andabove, root hair growth was rapidly inhibited, whereas decreas-ing extracellular pH to 4.5 or below caused all growing root hairsto burst at their tips within 10 min (Table 1). Nongrowing roothairs did not burst under these same low pH conditions.

Oscillating Changes in Extracellular ROS Production Are Associatedwith Root Hair Apical Growth. Oxidative cross-linking and severingof wall polymers have been proposed to regulate plant growth bymodulating the yield threshold of the cell wall (e.g., refs. 9–11).We therefore asked whether the polarized growth of root hairsis associated with localized changes in the oxidative environmentof the apoplast. To visualize extracellular ROS with high spatialand temporal resolution, we developed an imaging approachbased on OxyBURST green H2HFF, a nonfluorescent reagentthat becomes fluorescent upon oxidation. OxyBURST was ap-plied as a conjugate with BSA, making the ROS sensor cellimpermeable. Importantly, the intensity of OxyBURST fluores-cence does not vary with environmental factors such as pH orionic strength of the medium (ref. 18 and data not shown)

In growing root hairs, there was a clear gradient in Oxy-BURST wall labeling with weaker fluorescence at the apex andmuch stronger fluorescence in subapical regions (Fig. 3A). Incontrast, nongrowing root hairs showed strong homogenous

fluorescence in the cell wall along the entire root hair length(Fig. 3B). The intense labeling of the apoplast by OxyBURSTwas not due to nonspecific adsorption of the OxyBURST-BSAconjugate because preoxidized OxyBURST-BSA was not accu-mulated in the wall (SI Fig. 10A). Increasing the imagingcontrast revealed that tip growth was invariably accompanied byperiodic bursts of increased OxyBURST fluorescence that oc-curred at a frequency of 2–4 peaks min�1 (Fig. 3C, SI Fig. 11, andSI Movie 2). The strongest f luctuations were detected along thesubapical f lanks of the root hair (SI Movie 2). The sensitivity ofroot hair growth to laser irradiation coupled to the need toperform OxyBURST imaging in liquid media imposed limita-tions on the resolution of our fluorescence and growth mea-surements (see SI Materials and Methods for detailed explana-tion). However, we were able to perform cross-correlation ofOxyBURST fluorescence with growth, although with poorerresolution than for the pH measurements. Such analysis of thegrowth/ROS phase relationship indicated that the peaks ofOxyBURST fluorescence most likely lag peaks in growth rate by8.0 � 1.4 s or precede the minima in growth rates by 7.5 � 1.5 s(n � 6; Fig. 3D). Both timings are consistent with a model inwhich ROS production acts to help restrict cell expansion aftera peak of elongation. Consistent with OxyBURST reportingROS production to the apoplast of the root hair, the formationof these patterns of fluorescence was abolished by preincubatingthe root with the general ROS scavenger ascorbic acid (SIFig. 10B).

OxyBURST is documented to react with only certain forms ofROS, showing great sensitivity for superoxide and being oxi-dized by H2O2 only in the presence of peroxidases (19). There-fore, to determine which endogenous ROS species was respon-sible for OxyBURST fluorescence around the root hair, wepreincubated the root with the superoxide scavenger MCLA(methoxy cypridina luciferin analog; 20), which completelyinhibited the development of OxyBURST fluorescence (Fig.

Table 1. Effect of manipulation of extracellular pH and ROSconcentration on Arabidopsis root hair growth

Treatment Growing, % Stopping, % Bursting, %

pH 4.5 (228) 0 0 100pH 8 (54) 4 96 01 �M H2O2 (24) 0 100 0100 �M ascorbate (104) 16 0 8410 �M MCLA (100) 2 0 9825 �M DPI (40) 0 17 83

To determine the effect of extracellular pH and ROS on tip growth, roothairs were selected that exhibited stable growth rates of 1–2 �m min�1 inunbuffered minimal medium. After monitoring root hair growth for �10 min,treatment was started by adding compounds/buffers (10 mM DMGA for pH4.5, 10 mM Hepes for pH 8) listed to the minimal growth medium. Numbers ofroot hairs are in parentheses.

Fig. 1. Two-dimensional map of surface pH around a growing Arabidopsis root hair. Bright-field (leftmost) and fluorescence images of a root hair expressingcytosolic pH-sensitive GFP and immersed in fluorescein dextran, allowing measurement of cytosolic and surface pH, respectively. Numbers represent time ofobservation in seconds. The regions of interest (ROI) outlined in the bright-field image mark the areas selected for quantitative analysis of fluorescence intensityshown in SI Fig. 8. V, vacuole. Representative of n � 20 root hairs. (Scale bar, 10 �m.)

Fig. 2. Root hair surface pH oscillations lag growth oscillations. (A) Simul-taneous measurement of pH-dependent fluorescence and tip growth showedthat surface pH at the root hair tip and growth rate oscillate with the sameperiodicity but out of phase. Fluorescence intensity (F.I.) was measured in anROI close to the apex of the growing root hair (see Fig. 1). Note that peaks inalkalinization follow peaks in growth. Data are representative of n � 10 roothairs. (B) Cross-correlation analysis showed that pH oscillations lag growthoscillations by 7.0 � 0.3 s (means � SD of n � 10 root hairs).

Monshausen et al. PNAS � December 26, 2007 � vol. 104 � no. 52 � 20997

PLA

NT

BIO

LOG

YSE

ECO

MM

ENTA

RY

Dow

nloa

ded

by g

uest

on

June

13,

202

0

3E), confirming that the OxyBURST signal most likely de-pended on root hair production of superoxide.

To investigate a potential functional role of extracellular ROS,we elevated the oxidative activity in the cell wall by applying 1�M H2O2, which caused immediate cessation of root hair apicalgrowth (Table 1). Cytoplasmic streaming continued in theseH2O2-treated hairs, which also continued to label with the vitalstain fluorescein diacetate (data not shown), indicating that theROS-induced cessation of growth was unlikely to be due to anonspecific cytotoxic action of the treatment. Similarly, appli-cation of hydroxyl radicals using the Fenton reaction (10 �MH2O2, 10 �M ascorbic acid, 10 �M CuCl2) led to growth arrest(SI Fig. 12). Conversely, scavenging endogenously producedROS by adding 100 �M ascorbic acid or 10 �M less-membrane-permeant MCLA to the growth medium induced rapid (within5 min) bursting of growing, but never of nongrowing root hairapices (Table 1). The DMSO solvent control for MCLA indi-

cated no detectable effect on growth or development of Oxy-BURST fluorescence (n � 7, data not shown). These results areconsistent with the hypothesis that increasing extracellular ROSconcentration leads to cell wall rigidification, whereas a decreasein ROS concentration enhances cell wall extensibility.

The NADPH Oxidase ATRBOH C Is Important for Maintaining Root HairCell Wall Integrity. ATRBOH C, a superoxide-producing NADPHoxidase expressed in the root epidermis and root hairs ofArabidopsis, has been shown to play an important role in roothair growth (8). The lack of a functional ATRBOH C in theroot hair-defective Arabidopsis mutant rhd2-1 does not impederoot hair initiation but was reported to cause root hair devel-opment to arrest before the onset of tip growth (8, 21). On closerinspection, we observed that, in this mutant, initiating root hairsruptured at the tip upon switching to apical growth (SI Fig. 13,SI Movie 3, and ref. 22); consistent with this observation,treatment with 25 �M DPI (diphenyleneiodonium), a broad-range inhibitor of flavoproteins (such as NADPH oxidases; 23)also induced bursting of most tip growing root hairs (Table 1).These findings support the idea that decreased extracellularoxidative activity in the mutant weakens the cell wall and rendersit less resistant to turgor pressure.

The data presented above indicate that elevated apoplastic pHand ROS both contribute to stabilize the root hair cell wall at thegrowing apex. To test the hypothesis that these elements act inconcert to support root hair elongation, we investigated theeffect of increased extracellular pH on root hair development inrhd2-1. When grown on the surface of 1/4-strength Murashigeand Skoog medium buffered to pH 5, WT roots produced anabundance of long, regularly shaped root hairs (Fig. 4A). Underidentical conditions, the root hairs of rhd2-1 ruptured just aftertheir initiation (Fig. 4 C and G). This growth pattern changeddramatically, however, when the pH of the medium was in-creased to �6. At this less acidic pH, rhd2-1 root hairs developedapparently normally, showing no obvious disruption of the tipgrowth process and differing neither in morphology, density, norlength from the root hairs of WT plants grown under identicalconditions (Fig. 4 B and D–F and Table 2). The same pH-dependent rescue of the root hair-defective phenotype wasobserved in the Salk�071801 T-DNA insertional mutant ofAtrbohC (Table 2). The ROS produced by rhd2-1 have beenproposed to regulate the apical Ca2� gradient thought to drivetip growth (8). However, the Ca2� gradient observed in WT roothairs was also seen in rhd2-1 pH rescued tip growing root hairs(Fig. 5).

The observation that experimentally alkalinizing the growthmedium helped rescue the burst root hair phenotype of rhd2-1led us to investigate whether the endogenous tip growth-associated pH regulatory system was altered in the rhd2-1background. Although rescued rhd2-1 root hairs showed fluc-tuating changes in surface pH, the degree of alkalinization was

Fig. 3. Extracellular ROS around growing Arabidopsis root hair. (A and B)Distribution of OxyBURST fluorescence along the cell wall of a growing (n �30, A) and nongrowing (n � 20, B) root hair. The image was acquired shortlyafter addition of OxyBURST. (C) Time course of growth and extracellular ROSproduction of a growing root hair. Note that extracellular ROS and growthrate oscillate with the same periodicity but out of phase. Fluorescence inten-sity (F.I.) was measured in an ROI close to the apex of the root hair (see A). Dataare representative of n � 6 root hairs. (D) Cross-correlation analysis showedthat ROS oscillations lag behind growth oscillations by 8.0 � 1.4 s or precedethe minima in growth by 7.5 � 1.5 s (means � SE of n � 6 root hairs). (E)Pretreatment with MCLA prevents oxidation of OxyBURST around a nongrow-ing root hair. The root was incubated in 20 �M MCLA for 45 s before additionof OxyBURST; n � 8. Fluorescence images in A, B, and E were collected andprocessed by using the same microscope settings. (Scale bars, 10 �m.)

Fig. 4. pH-dependent rescue of rhd2-1 root hair phenotype. (A–D) Seeds of WT Arabidopsis (A and B) and rhd2-1 (C and D) were germinated for 4 d in continuouslight on agar plates containing 1/4-strength Murashige and Skoog (M&S) basal salts, 1% (wt/vol) sucrose and 5 mM Mes, titrated to pH 5 (A and C) or pH 6 (Band D). Roots were imaged by using a Leica Wild M420 stereo microscope. Representative of n � 15 WT and rhd2-1 roots, respectively. (E and F) WT and rescuedrhd2-1 root hair at higher magnification. (Scale bar in A, 1 mm for A–D; scale bar in E, 10 �m for E and F. (G) Higher magnification of the bursting phenotypeseen in rhd2-1 grown in 1/4-strength M&S medium, pH 5. * indicates burst root hairs. (Scale bar, 100 �m.)

20998 � www.pnas.org�cgi�doi�10.1073�pnas.0708586104 Monshausen et al.

Dow

nloa

ded

by g

uest

on

June

13,

202

0

higher than in WT, regularly reaching pH values �7 (Fig. 6). Theduration of these alkalinizations was also significantly increased(Fig. 6 and SI Fig. 14).

ATRBOH C Is Required for Oscillatory Changes in ROS ProductionDuring Root Hair Tip Growth. It has been reported (8) that initiatingroot hair bulges of rhd2-1 show reduced accumulation of cyto-plasmic ROS compared with WT. To verify that pH-rescued roothairs of rhd2-1 are also deficient in the production of ROS, wefirst investigated the intracellular redox status using the ROSsensor dichlorofluorecin (DCF). ROS-dependent fluorescenceof DCF was significantly greater in WT root hairs than in roothairs of rhd2-1 (Fig. 7A), whereas no such difference in fluo-rescence intensity was observed in control experiments with thestructurally similar, but ROS-independent, f luorescent dye flu-orescein diacetate (SI Fig. 15). No tip-focused accumulation ofROS was detectable in either WT or pH-rescued rhd2-1 tipgrowing root hairs (SI Fig. 16).

We then visualized extracellular ROS at the surface of rescuedrhd2-1 root hairs using OxyBURST-BSA. Similar to WT, thesegrowing rhd2-1 root hairs showed strong OxyBURST fluores-cence in the lateral walls with weak fluorescence at the apex (Fig.7B). However, in none of the monitored root hairs was growthever accompanied by detectable oscillatory changes in ROS-dependent fluorescence (Fig. 7C), indicating that the oscillatingproduction of ROS seen in WT root hairs depends on thepresence of a functional ATRBOH C.

DiscussionExtracellular pH as a Regulator of Expansion at the Growing Root HairApex. In this study, we explored the role of ROS and pH inmaintaining the highly defined growth pattern associated withroot hair elongation. Cell wall pH is thought to play an importantrole in regulating wall yielding and, therefore, the rate at whicha cell can expand (24). We found that tip growth of an Arabi-dopsis root hair is associated with cyclic changes in H� influxaround the extreme apex of the cell (Fig. 1 and SI Figs. 8 and 9)

that correlated with the periodicity of growth (Fig. 2). Similargrowth oscillations have been documented for the more rapidlytip-growing pollen tubes that are also characterized by oscillatingchanges in H� f luxes (14, 25). In pollen tubes, periods of intenseand slower growth are preceded by periods of low and high H�

influx density, respectively (14). Cross-correlation analysis ofgrowth and surface pH demonstrated a similar relationship inArabidopsis root hairs (Fig. 2), indicating that growth acceleratesafter apoplastic acidification and slows upon alkalinization. Thisis consistent with our observation that suddenly imposing astable alkaline environment of pH 8 on root hairs arrested tipgrowth, whereas acidifying the medium to pH 4.5 resulted inbursting, i.e., uncontrolled expansion, of all growing root hairs(Table 1).

Rapid changes in apoplastic pH are likely to modulate theactivity of a number of regulatory elements such as cell wallproteins, e.g., the expansins (26) and pectin methylesterases(27–30). In addition, plasma membrane proteins such as iontransporters may also be affected, either directly, such as the

Fig. 6. Surface pH dynamics at apex of growing Arabidopsis root hairsmeasured with a H�-selective microelectrode. Surface pH of WT (A) andrescued rhd2-1 (B) root hair. The gray arrows indicate the pH measured in themedium at 30-�m distance from the root hair apices. Representative of n � 20WT and rhd2-1 root hairs, respectively.

Fig. 7. Intra- and extracellular ROS of pH rescued rhd2-1 root hair. (A)Intracellular redox status of WT and rhd2-1 root hairs monitored with DCF.Higher DCF fluorescence intensity indicates higher cytoplasmic oxidative ac-tivity. Values are means � SE of n � 16 (WT) and 23 (rhd2-1) root hairs. Meansare significantly different (P � 0.001, Student t test). Fluorescence intensitywas measured in a region of interest (ROI) in the apical cytoplasm (see Fig. 1).(B) ROS-dependent OxyBURST fluorescence along a pH-rescued growingrhd2-1 root hair. The image was acquired by using the same settings as for Fig.3 A, B, and E. (Scale bar, 10 �m.) (C) Time course of extracellular ROS produc-tion of WT and rhd2-1 root hairs. OxyBURST fluorescence intensity (F.I.) wasmeasured in an ROI behind the extreme apex. Data are representative of n �14 experiments for WT and rhd2-1, respectively.

Table 2. pH-dependent rescue of root hair growth in Arabidopsisknockout mutants of Atrboh C

Genotype, treatment*

Root hair density(root hairs per mm

of root length)Root hair

length, �m

WT, pH5 25 � 7 (11)† 492 � 67 (15)WT, pH6 25 � 5 (12) 343 � 97 (24)rhd2-1, pH5 0.6 � 0.6 (16) narhd2-1, pH6 21 � 4 (20) 357 � 71 (25)Salk�071801, pH5 0.2 � 0.2 (9) naSalk�071801, pH6 25 � 8 (8) 334 � 64 (42)

na, not applicable.*Seedlings were grown on 1% (wt/vol) agar plates containing 1/4-strengthMurashige and Skoog basal salts, 1% (wt/vol) sucrose, buffered to indicatedpH with 5 mM MES/NaOH.

†Mean values � SD (n).

Fig. 5. Tip-focused Ca2� gradients measured in growing cameleon YC3.6-expressing Arabidopsis root hairs. (A) WT. (B) ‘‘pH-rescued’’ rhd2-1. CytosolicCa2� levels have been pseudocolor-coded according to the scale at the right.(Scale bar, 10 �m.)

Monshausen et al. PNAS � December 26, 2007 � vol. 104 � no. 52 � 20999

PLA

NT

BIO

LOG

YSE

ECO

MM

ENTA

RY

Dow

nloa

ded

by g

uest

on

June

13,

202

0

pH-sensitive potassium channels (31, 32) or indirectly by alter-ations in the steepness of the proton gradient across the mem-brane. The concomitant large swings in cytoplasmic pH (Fig. 1,SI Fig. 8) may also affect a myriad of cytosolic protein activities,providing a global integrator of cellular growth-related activities.

Extracellular ROS and Tip Growth. Apoplastic ROS have recentlyemerged as an important element in the regulation of plant cellexpansion (9, 10, 33–36). Although nitro blue tetrazolium (NBT)has been used to monitor extracellular ROS, in our hands, thisreagent has proven to be cytotoxic (SI Fig. 17 and SI Movie 4).We therefore developed an assay using OxyBURST greenH2HFF-BSA, a nonfluorescent reagent that becomes highlyfluorescent upon oxidation (Figs. 3 and 7 and SI Fig. 12) andwhich proved to be nontoxic to roots and root hair growth. Ourmeasurements revealed a characteristic pattern of fluorescencein the cell wall of the growing root hair with high signal intensityalong root hair f lanks and less fluorescence at the apex (Fig. 3).This observation indicates that lower levels of extracellular ROSare associated with the rapidly expanding region of the root hairtip, whereas higher ROS concentrations are present along theroot hair f lanks where wall yielding is reduced (2). Interestingly,ROS levels also strongly increased at the extreme apex when aroot hair ceased to grow (Fig. 3), and low concentrations ofexogenous ROS (1 �M H2O2 or 10 �M H2O2 driving hydroxylradical production from the Fenton reaction) immediately in-hibited root hair expansion (Table 1, SI Fig. 12). Conversely,ROS scavengers such as ascorbic acid and MCLA caused roothair bursting (Table 1). All of these findings are consistent withthe idea that growing cells generate extracellular ROS to pro-mote the rigidification of the cell wall (33, 37) and that thelocalization and intensity of ROS production is tightly regulatedto control where such rigidification will occur.

ROS have been proposed to modulate root hair tip growth byactivating a Ca2� channel required to generate the tip-focusedCa2�-gradient, with the production of these regulatory cytosolicROS mediated by the NADPH oxidase ATRBOH C (8). How-ever, NADPH oxidases generate extracellular ROS by transfer-ring electrons from cytosolic NADPH across the plasma mem-brane to reduce molecular oxygen (38). Thus, NADPH oxidasesare likely to also directly affect wall ROS status. Root hairs ofArabidopsis rhd2-1 mutants, which lack a functional ATRBOH C(8), showed normal root hair initiation but then burst (SI Fig. 13,SI Movie 3, and ref. 22), indicating that, in addition to a possiblerole for this enzyme in modulating cytosolic ROS and subse-quent signaling (8), it may also be involved in wall rigidificationduring tip growth. Significantly, this phenotype was rescued byincreasing extracellular pH. Thus, when grown at pH �6, rhd2-1root hairs developed normally with regard to morphology,length, and density (Fig. 4, Table 2). In addition, although theserescued root hairs showed reduced cytosolic ROS levels, theyformed a tip-focused Ca2� gradient (Fig. 5). This observationindicates that ATRBOH C is not absolutely essential for roothair development or gating of the Ca2� channels needed togenerate the tip focused gradient and that some of its function(possibly a role in initial wall rigidification) can be compensatedby more-alkaline pH. However, technical limitations on theanalysis of wall composition at the subcellular level have so farprevented us from directly monitoring this likely alteration in thestructure of the wall at the root hair initiation site in rhd2-1plants.

Although indistinguishable in appearance from WT, rescuedrhd2-1 root hairs differed in their pattern of extracellular ROS.Wall ROS were present in both WT and rhd2-1 root hairs (Figs.3 and 7) and presumably reflected the activity of multiple sourcesfor extracellular ROS. These likely include the range of super-oxide-producing NADPH oxidases (39), although wall-boundperoxidases (40) and apoplastic amine oxidases (41) may also

contribute. However, the cyclic ROS production seen at the tipof WT root hairs was not detected in the rhd2-1 mutant (Figs. 3and 7). This observation indicates that ATRBOH C contributesto the oscillatory generation of ROS associated with root hair tipgrowth.

ATRBOH C activity also affects the intracellular redox stateof the root hair (Fig. 7), likely through consumption of cytosolicNADPH (42) and/or generation of apoplastic ROS that thendiffuse across the plasma membrane into the cytoplasm. Thus, inaddition to growth control through modulating wall properties,ROS generated at the apex and well behind the growing root hairtip (Fig. 3A) are likely involved in many other physiologicalresponses known to be modulated by ROS (43). Interestingly, theloss of ATRBOH C-dependent ROS production in rescuedrhd2-1 root hairs was accompanied by altered pH oscillationswith very strong alkalinizations that lasted significantly longerthan in WT (Fig. 6), suggesting that feedback between these twosystems is required to support normal tip growth.

Although ATRBOH C appears to be a key element inmodulating growth-related ROS production, the molecularmechanisms responsible for the highly dynamic H� f luxes wehave described remain to be identified. It seems likely that pHchanges involve the oscillatory activation/deactivation of H�

transporters such as H�-ATPases and cation- (H�) or anion-(OH�) permeable channels. Identifying these H� transportersand the signaling elements involved in the concerted regulationof H� f luxes and ATRBOH C activity is the challenge for futureresearch.

Materials and MethodsPlant Material. Seeds of Arabidopsis thaliana (ecotype Columbia) were surfacesterilized and germinated aseptically on Murashige and Skoog medium(Sigma) supplemented with 1% (wt/vol) sucrose and 1% (wt/vol) agar at 21°Cunder continuous light conditions. Four- to 5-day-old seedlings were chosenfor experiments. Seeds of rhd2-1 were kindly provided by J. Lynch (Pennsyl-vania State University, State College, PA) and plants expressing YC3.6 by JeffHarper (University of Nevada, Reno, NV). Seeds of Salk�071801 were obtainedfrom the European Arabidopsis Stock Center.

Imaging of Intra- and Extracellular pH. Arabidopsis seedlings expressing thepH-sensitive GFP variant H148D were transferred to purpose-built cuvettes(44) and covered with 2% (wt/vol) agarose (type VII; Sigma) in minimalmedium (0.1 mM KCl, 0.1 mM CaCl2, 1 mM NaCl, 1% (wt/vol) sucrose, �pH 6).After 3 h, the agarose was carefully removed from near the root tip, theexposed surface covered with minimal medium, and the root allowed to growout of the cut surface for an additional 3 h. To measure extracellular pHaround the exposed root hairs, the medium was supplemented with 30 �gml�1 of the pH-sensitive fluorescent dye fluorescein, conjugated to 10-kDadextran (Sigma), and ratio imaged (18) with the Zeiss LSM 510 laser scanningconfocal microscope (Carl Zeiss Inc.) using a �40 water-immersion, 1.2 numer-ical aperture, C-Apochromat objective. Excitation was alternated between458 and 488 nm (switching between wavelengths after each scanned line at��1 s), emission collected by using a 488-nm dichroic mirror and 505-nmlong-pass filter. Bright-field images were acquired simultaneously by usingthe transmission detector of the confocal microscope. For time-lapse analysis,images were collected every 3 s, with each individual image scan lasting 2.2 s.For extracellular pH, calibration was performed in 100 mM Mes buffer at pH5.0, 5.25, 5.5, 6.0, and 6.5 with identical imaging parameters. Endpoint cali-bration of cytosolic pH was performed with 100 mM NH4Cl, pH 9.3, and 100mM KHCO3, pH 6.2. The pH was calculated by using the formula [H�]cyt � Kd

(Rmax � R)/(R � Rmin), with a Kd of GFP H148D of �10�7.8M (45). Data wereanalyzed by using the Zeiss LSM software.

Measurement of Root Hair Growth. Bright-field images were collected every 2 ssimultaneously with fluorescence images using single wavelength (488 nm)excitation. High-resolution growth measurements were made by using thecomputer vision tracking software as described (14).

Measurement of Intra- and Extracellular Oxidative Activity. Arabidopsis seed-lings were mounted as described for extracellular pH measurements (seeabove), incubated for 20 min with 20 �M 2,7-dichlorofluorescin diacetate

21000 � www.pnas.org�cgi�doi�10.1073�pnas.0708586104 Monshausen et al.

Dow

nloa

ded

by g

uest

on

June

13,

202

0

(HDCF-DA; Sigma, 20 mM stock in DMSO), and imaged for no more than 5 min.Root hairs were imaged with a Zeiss LSM 510 confocal microscope by using the�40 lens described above. The fluorescent dye was excited with the 488-nmline of the argon laser, and emission was collected by using a 488 dichroicmirror and 505-nm long-pass filter.

Extracellular release of ROS was determined by using the fluorogenicreagent OxyBURST green H2HFF (dihydro-2,4,5,6,7,7-hexafluorofluores-cein)-BSA (Invitrogen). Roots were incubated with 100 �g ml�1 OxyBURST, andfluorescence was recorded with the Zeiss LSM 510 microscope with the sameimaging parameters as described above for DCF.

Neither HDCF nor OxyBURST had any detectable effect on normal root hairgrowth or development (data not shown).

Imaging of Cytosolic Ca2� Levels. Arabidopsis seedlings expressing the FRET-based Ca2� sensor cameleon YC3.6 (46) were mounted as described above andratio imaged with the Zeiss LSM 510 microscope using the �40 objectivedescribed above. The sensor was excited with the 458-nm line of the argonlaser. The CFP (473–505 nm) and FRET-dependent (536–546 nm) emission werecollected by using a 458-nm primary dichroic mirror and the Meta detector ofthe microscope.

Electrophysiology. Extracellular pH and H� fluxes at the surface of Arabidopsisroot hairs were measured with H�-selective microelectrodes. The microelec-trode and reference electrode were fabricated and calibrated as described(47), see also SI Materials and Methods for detailed protocols.

To prepare plants for measurement, seedlings were transferred onto hold-ers coated with 2% (wt/vol) agar in minimal medium and fixed in place bycovering the roots with 1% (wt/vol) agarose. The holders were then placedinto cuvettes containing minimal medium. After several hours, the growingroot tip had emerged from the agar, and the developing root hairs andepidermal cells were accessible to H�-selective microelectrodes.

To measure extracellular pH, the microelectrode was moved to within 1 �mof the cell surface and the voltage recorded in 1-s intervals. To measure H�

fluxes, the microelectrode was vibrated close (1–2 �m) to the surface of theroot hair apex. One measurement cycle consisted of waiting for 1.5 s, mea-suring for 0.3 s, and then stepping the electrode 10 �m away from the previousposition (perpendicular to the surface), waiting for 1.5 s, and again measuringfor 0.3 s, followed by the next cycle. Each pair of measurements was used tocalculate the H� concentration gradient and the corresponding H� flux ac-cording to Fick’s first law of diffusion.

ACKNOWLEDGMENTS. We thank Prof. M. H. Weisenseel for his generoussupport; Dr. Adam Bertl (Technische Universitat, Darmstadt, Germany), Dr.Andreas Meyer (University of Heidelberg, Heidelberg), and Phillip Day (Penn-sylvania State University) for many helpful discussions; and Dr. Sarah Swansonfor critical reading of the manuscript. This research was supported by NationalAeronautics and Space Agency Grant NAG-1594 (to S.G.), by National ScienceFoundation Grants MCB 02-12099, MCB 0641288, IBN 03-36738, and DBI03-01460 (to S.G.) and MCB 0641288 (to G.B.M.), and by National Institutes ofHealth Grant NCRR P41 RR001395 (to M.A.M.).

1. Hepler PK, Vidali L, Cheung AY (2001) Annu Rev Cell Dev Biol 17:159–187.2. Shaw SL, Dumais J, Long SR (2000) Plant Physiol 124:959–970.3. Hussey PJ, Ketelaar T, Deeks MJ (2006) Annu Rev Plant Biol 57:109–125.4. Sieberer BJ, Ketelaar T, Esseling JJ, Emons AMC (2005) New Phytol 167:711–719.5. Brownlee C, Goddard H, Hetherington AM, Peake LA (1999) J Exp Bot 50:1001–1011.6. Fu Y, Yang Z (2001) Trends Plants Sci 6:545–547.7. Emons AM, Mulder BM (2000) Trends Plants Sci 5:35–40.8. Foreman J, Demidchik V, Bothwell JHF, Mylona P, Miedema H, Torres MA, Linstead P,

Costa S, Brownlee C, Jones JDG, et al. (2003) Nature 422:442–446.9. Schopfer P (1996) Planta 199:43–49.

10. Pedreira J, Sanz N, Pena MJ, Sanchez M, Queijeiro E, Revilla G, Zarra I (2004) Plant CellPhysiol 45:530–534.

11. Schopfer P (2001) Plant J 28:679–688.12. Cosgrove DJ (1999) Annu Rev Plant Physiol Plant Mol Biol 50:391–417.13. Bibikova TN, Jacob T, Dahse I, Gilroy S (1998) Development (Cambridge, UK) 125:2925–

2934.14. Messerli MA, Danuser G, Robinson KP (1999) J Cell Sci 112:1497–1509.15. Herrmann A, Felle HH (1995) New Phytol 129:523–533.16. Jones DL, Shaff JE, Kochian LV (1995) Planta 197:672–680.17. Fasano JM, Swanson SJ, Blancaflor EB, Dowd PE, Kao TH, Gilroy S (2001) Plant Cell

13:907–921.18. Chen CS (2002) BMC Cell Biol 3:21.19. Haugland RP (2005) Handbook of Fluorescent Probes and Research Chemicals (Mo-

lecular Probes, Eugene, OR).20. Nakano M (1998) Cell Mol Neurobiol 18:565–579.21. Schiefelbein JW, Somerville C (1990) Plant Cell 2:235–243.22. Jones MA, Raymond MJ, Yang Z, Smirnoff N (2007) J Exp Bot 58:1261–1270.23. Mori IC, Pinontoan R, Kawano T, Muto S (2001) Plant Cell Physiol 42:1383–1388.24. Cosgrove DJ (2000) Nature 407:321–326.

25. Feijo JA, Sainhas J, Hackett GR, Kunkel JG, Hepler PK (1999) J Cell Biol 144:483–496.26. Cosgrove DJ, Li LC, Cho HT, Hoffmann-Benning S, Moore RC, Blecker D (2002) Plant Cell

Physiol 43:1436–1444.27. Sherrier DJ, Vandenbosch KA (1994) Plant J 5:185–195.28. Li YQ, Chen F, Linskens HF, Cresti M (1994) Sex Plant Reprod 7:145–152.29. Bosch M, Hepler PK (2005) Plant Cell 17:3219–3226.30. Micheli F (2001) Trends Plants Sci 6:414–419.31. Ilan N, Schwartz A, Moran N (1996) J Membr Biol 154:169–181.32. Hartje S, Zimmermann S, Klonus D, Mueller-Roeber B (2000) Planta 210:723–731.33. Hohl M, Greiner H, Schopfer P (1995) Physiol Plant 94:491–498.34. Bradley DJ, Kjellbom P, Lamb CJ (1992) Cell 70:21–30.35. Wojtaszek P (1997) Biochem J 322:681–692.36. Brisson LF, Tenhaken R, Lamb C (1994) Plant Cell 6:1703–1712.37. Held MA, Tan L, Kamyab A, Hare M, Shpak E, Kieliszewski MJ (2004) J Biol Chem

279:55474–55482.38. Sumimoto H, Miyano K, Takeya R (2005) Biochem Biophys Res Commun 338:677–686.39. Torres MA, Dangl JL (2005) Curr Opin Plant Biol 8:397–403.40. Passardi F, Penel C, Dunand C (2004) Trends Plants Sci 9:534–540.41. Cona A, Rea G, Angelini R, Federico R, Tavladoraki P (2006) Trends Plants Sci 11:80–88.42. Cardenas L, McKenna ST, Kunkel JG, Hepler PK (2006) Plant Physiol 142:1460–1468.43. Mori IC, Schroeder JI (2004) Plant Physiol 135:702–708.44. Wymer CL, Bibikova TN, Gilroy S (1997) Plant J 12:427–439.45. Elsliger MA, Wachter RM, Hanson GT, Kallio K, Remington SJ (1999) Biochemistry

38:5296–5301.46. Nagai T, Yamada S, Tominaga T, Ichikawa M, Miyawaki A (2004) Proc Natl Acad Sci USA

101:10554–10559.47. Arend M, Monshausen G, Wind C, Weisenseel MH, Fromm J (2004) Plant Cell Environ

27:1288–1296.

Monshausen et al. PNAS � December 26, 2007 � vol. 104 � no. 52 � 21001

PLA

NT

BIO

LOG

YSE

ECO

MM

ENTA

RY

Dow

nloa

ded

by g

uest

on

June

13,

202

0