Cell wall integrity controls root elongation via a general 1-aminocyclopropane-1-carboxylic acid-dependent, ethylene-independent pathway.

Cell expansion in plants requires cell wall biosynthesis and rearrangement. During periods of rapid elongation, such as during the growth of etiolated hypocotyls and primary root tips, cells respond dramatically to perturbation of either of these processes. There is growing evidence that this response is initiated by a cell wall integrity-sensing mechanism and dedicated signaling pathway rather than being an inevitable consequence of lost structural integrity. However, the existence of such a pathway in root tissue and its function in a broader developmental context have remained largely unknown. Here, we show that various types of cell wall stress rapidly reduce primary root elongation in Arabidopsis (Arabidopsis thaliana). This response depended on the biosynthesis of 1-aminocyclopropane-1-carboxylic acid (ACC). In agreement with the established ethylene signaling pathway in roots, auxin signaling and superoxide production are required downstream of ACC to reduce elongation. However, this cell wall stress response unexpectedly does not depend on the perception of ethylene. We show that the short-term effect of ACC on roots is partially independent of its conversion to ethylene or ethylene signaling and that this ACC-dependent pathway is also responsible for the rapid reduction of root elongation in response to pathogen-associated molecular patterns. This acute response to internal and external stress thus represents a novel, noncanonical signaling function of ACC.

Cell proliferation and cell expansion are the two aspects of growth that determine cell, tissue, and ultimately organ size in multicellular organisms. Unlike in metazoans, cell expansion in plants is an important contributor to organ size. An osmolyte-filled vacuole exerts hydraulic pressure against a mechanically strong wall. Controlled relaxation of these walls allows for cell expansion via water influx into the vacuole. Many factors, including environmental conditions and the physical stability of the wall, determine how far cell wall polymer remodeling can proceed and thus how big the cell can become (De Cnodder et al., 2005). Cell elongation in roots is negatively controlled by ethylene, which in turn requires auxin biosynthesis and transport (Rů zicka et al., 2007;Stepanova et al., 2007;Swarup et al., 2007) as well as the production of reactive oxygen species (ROS;De Cnodder et al., 2005). ROS have both a signaling function and a direct effect on cell wall elasticity; apoplastic peroxidases can cross-link Hyprich glycoproteins in the presence of ROS to stiffen walls (Passardi et al., 2004), while hydroxyl radicals can sever polysaccharide chains (Mü ller et al., 2009).
Plant tissues undergoing rapid expansion are particularly vulnerable to cell wall defects. Etiolated hypocotyls and root tips, therefore, have been used in forward genetic screens to identify genes with a role in cell wall biogenesis (Baskin et al., 1992;Hauser et al., 1995;Desnos et al., 1996). Reduced elongation and a loss of growth anisotropy in roots that manifests as radial swelling are hallmarks of cell wall-deficient mutants. Other characteristic phenotypes, particularly of cellulose-deficient mutants, are ethylene-and jasmonic acid-dependent defense responses such as ectopic lignification (Cañ o-Delgado et al., 2000;Ellis et al., 2002). These responses are phenocopied by inhibitors of cellulose biosynthesis like isoxaben and dichlobenil (2,6-dichlorobenzonitrile). It could be argued that reduced cell expansion is an inevitable consequence of the loss of cell wall integrity. However, the characteristic defense-like responses that occur in addition to growth defects suggest a more complex situation and a dedicated signaling pathway. In order to take external factors into account, the regulatory circuit that controls cell expansion must include a mechanism to feed back information about cell wall integrity into the cytoplasm. Early evidence for such a pathway came from the observation that in dark-grown Arabidopsis (Arabidopsis thaliana) seedlings, commitment to rapid elongation (up to 48 h after imbibition) but not elongation itself was sensitive to the cellulose synthase inhibitor isoxaben (Refrégier et al., 2004).
Our knowledge of this postulated signaling pathway is limited. Fungal cell walls, despite their different chemical composition, have a mechanical function very similar to that of plant cell walls. In yeast, a cell wall integrity signaling pathway has been characterized in detail (Levin, 2005) that requires a family of highly glycosylated transmembrane sensors (WSC1-WSC3, MID2, MTL1). No obvious plant ortholog of the WSC family sensor proteins can be identified because of their low sequence complexity. Several families of plant receptor-like kinases (RLKs) have been proposed to perform cell wall-sensing functions and/or mediate responses to cell wall damage (Humphrey et al., 2007;Seifert and Blaukopf, 2010;Steinwand and Kieber, 2010). Mutations in some of these genes cause reduced cell expansion up to severe dwarfism and other phenotypes typical for cell wall defects, but evidence for direct binding of the extracellular domain of RLKs to cell wall polysaccharides remains limited (Kohorn et al., 2009). The clearest genetic evidence for an active signaling mechanism that communicates cell wall status and controls growth comes from the theseus mutant. Loss of the CrRLK family member THE1 partially suppresses the short-hypocotyl phenotype of a weak cellulose-deficient mutant, procuste 1-1 (Hématy et al., 2007), without restoring its cell wall defect.
Cell wall defects typically lead to altered wall composition and transcriptional changes that resemble those triggered in defense responses (Humphrey et al., 2007). Turgor pressure and sugars are required for at least some of the transcriptional responses to inhibition of cellulose biosynthesis (Hamann et al., 2009); the ectopic lignification is additionally modulated by the NADPH oxidase, RbohD, and jasmonate. The crosswiring of apoplastic surveillance for pathogens and cell wall integrity is so strong that genetic screens for altered pathogen resistance have uncovered cell wall biosynthetic genes (Nishimura et al., 2003;Vogel et al., 2004), and vice versa, some mutants identified as cell wall deficient are more resistant to pathogens (Hernández-Blanco et al., 2007). This overlap is perhaps not surprising because for plant pathogenic fungi, oomycetes, and some bacteria, access to host resources requires breaking of the cell wall barrier. During cell expansion, however, feedback control of cell wall integrity should act rapidly and before the manifestation of large-scale structural damage.
We are interested in the acute response to perturbation of cell wall integrity. The yeast cell wall integrity pathway is triggered by inhibitors of glucan and chitin synthases (e.g. echinocandin and nikkomycin, respectively) and by compounds that bind to cell wall polysaccharides and inhibit their higher order assembly (Congo red and Calcofluor white; Levin, 2005). We have used analogous tools to trigger cell wall damage in plants, including inhibitors of cellulose biosynthesis (isoxaben, thaxtomin A, and dichlobenil) and a cellulosebinding dye (Congo red). We show that the accelerated elongation phase of root cells is rapidly inhibited by both types of cell wall-damaging agents. This inhibition is 1-aminocyclopropane-1-carboxylic acid (ACC) dependent but, strikingly, is not dependent on ethylene perception. It represents a general rapid-response pathway of root growth control that also underlies at least the initial response to microbial pathogen-associated molecular patterns (PAMPs). Auxin signaling and superoxide production are required downstream of ACC to reduce elongation, in agreement with the previously described effect of ethylene on root elongation.

RESULTS
It takes several days for newly produced root cells to traverse the cell division zone. In contrast, expansion to their mature size in the elongation zone only takes 6 to 8 h (Beemster and Baskin, 1998). This 10-fold or higher increase in length and thus cell surface area places a great demand on cell wall biosynthesis. In 4-d-old seedlings, expression of the primary wall cellulose synthase genes CesA1, CesA3, and CesA6 is strongest around the transition zone and continues into the elongation zone ( Fig. 1A; Scheible et al., 2001;Desprez et al., 2007). We have used the herbicide isoxaben to inhibit cellulose biosynthesis (Heim et al., 1990) and so phenocopy the type of cell wall damage seen in cellulose-deficient mutants. CesA3 and -6 are targets of isoxaben (Scheible et al., 2001;Desprez et al., 2002); CesA6-yellow fluorescent protein-labeled particles disappear from the cell surface within 20 min of isoxaben treatment (Paredez et al., 2006). Over 16 h, isoxaben induces root swelling and strongly reduces elongation, as evident in the "crowding" of root hairs due to shortening of trichoblasts (Fig. 1B). To analyze more acute effects of cellulose biosynthesis inhibition, we used a simple proxy for the complex spatial variation of elongation rates along the root. The length of the first epidermal cell with a visible root hair bulge (LEH; Fig. 1C; Le et al., 2001;De Cnodder et al., 2005) is a parameter that reflects rapid effects on elongation much more sensitively than macroscopic root length measurements. Isoxaben reduced the LEH to about 35% of control within 8 h; a significant effect was evident from about 1 h (Fig. 1D). For the following experiments, we chose 3-h treatments, which led to a robust response. With this time window, the differentiating trichoblast cells measured for the LEH parameter would have already committed to rapid elongation at the time of treatment (Beemster and Baskin, 1998). The fact that the LEH drops continuously over several hours shows that root cells are sensitive to relevant environmental signals throughout elongation, in contrast to cells in etiolating hypocotyls, which lose sensitivity to cell wall damage once committed to rapid elongation (Refregier et al., 2004). Although the study of root growth is complicated by the contribution of both cell division and expansion, the short-term experiments shown here are unlikely to be affected by changes in cell division.
Many plant hormones negatively affect root growth, including abscisic acid, auxin, and jasmonate, but the best-characterized effect specifically on elongation is exerted by ethylene (Benková and Hejátko, 2009). To find out if the isoxaben effect is mediated by ethylene, we applied isoxaben together with a chemical inhibitor of ethylene biosynthesis, aminoethoxyvinylglycine (AVG), or with silver ions (as silver thiosulfate) to block ethylene perception. While silver had no significant effect on cell length in control or isoxaben-treated roots, AVG fully restored elongation in the presence of isoxaben ( Fig. 2A). Often, roots treated with both isoxaben and AVG showed dramatic symptoms of cell wall defects such as blebbing (Fig. 2B), while this was practically never seen when isoxaben alone was applied. These observations demonstrate that cell expansion can proceed despite cell wall defects. The reduction of elongation triggered by blocking cellulose biosynthesis is an active, AVG-sensitive process rather than a passive biomechanical consequence of weakened walls.
The results of the inhibitor treatments were surprising because they suggested that ethylene perception is not required for this response while AVG blocks it. In a canonical ethylene-dependent pathway, inhibiting any step between ACC biosynthesis and ethylene perception should block the response. An alternative ethylene receptor blocker, norbornadiene, did not restore elongation either (Supplemental Fig. S2A). To assess the efficiency of silver treatment, we treated seedlings with ACC or with ethephon, a compound that hydro- Figure 2. Cell wall stress reduces root elongation via an AVG-sensitive, ethylene-independent pathway. A, Inhibition of ACC biosynthesis but not ethylene action restores isoxabeninduced root growth inhibition. *** and n.s. indicate LEH difference significant (P , 0.0001) or not significant (P . 0.05; two-tailed t test). B, Elongation and differentiation zone of roots treated for 4 h with isoxaben, AVG, or both (environmental scanning electron microscopy images). Bar = 200 mm. Note the severe bulging of roots treated with both compounds. C, The effect of ACC on root elongation is partially ethylene independent. a and b indicate results significantly different (P , 0.0001) from control or Ag + -only treated roots, respectively; n.s. indicates results not significantly different (P . 0.05; two-tailed t test). D, The ethylene-insensitive mutant ein3 eil1 responds to ACC and isoxaben. lyzes to ethylene above pH 3.5 (such as in plant cell cytoplasm). Both compounds reduced the LEH, but only the response to ethephon-released ethylene could be completely reversed with 10 mM silver thiosulfate (Fig. 2C). In some cases, the effect of ethephon has been shown to be independent of ethylene generation (Lawton et al., 1994). However, a control experiment showed that buffered phosphoric acid plus hydrochloric acid (the other hydrolysis products of ethephon) at the equivalent concentration did not affect LEH (Supplemental Fig. S2B).
A "side effect" of silver treatment, the promotion of indole-3-acetic acid efflux, has recently been described (Strader et al., 2009). Therefore, we sought additional confirmation of ethylene-independent short-term effects of ACC on root elongation. Externally applied ACC reduced root elongation in the same time scale as isoxaben, with a maximal response above approximately 100 nM (Le et al., 2001;Supplemental Fig. S1). The ein3 eil1 mutant is completely ethylene insensitive; hypocotyls of dark-grown ein3 eil1 seedlings show no response to ACC (Alonso et al., 2003;Supplemental Fig. S3), and long-term root growth (4-5 d) is insensitive to both ACC and ethephon (data not shown). While the LEH in wild-type seedlings (ecotype Columbia-0) was reduced by ethylene ( Fig. 2D; here applied as gas generated from ethephon without direct contact [Zhang and Wen, 2010]), ein3 eil1 seedlings were insensitive. In contrast, isoxaben and ACC significantly reduced the LEH both in the mutant and the wild type. ACC thus appears to have a short-term influence on root cell elongation that is independent of the canonical ethylene signaling pathway.
To further analyze the AVG-silver discrepancy, we also applied aminoxyacetic acid (AOA), another inhibitor of ACC synthase (ACS; Yu et al., 1979), and a-aminoisobutyric acid (AIB), an inhibitor of ACC oxidase (Satoh and Esashi, 1983). AIB or AOA had no effect on LEH, but all three inhibitors fully restored LEH in isoxaben-treated roots (Fig. 3A, light gray bars). In batches of seedlings with lower initial LEH, AVG increased elongation relative to the control. Chemical inhibitors of ACC biosynthesis are widely used because ACS are encoded by a multigene family and complete elimination of all nine members causes embryonic lethality (Tsuchisaka et al., 2009). However, AVG and AOA not only inhibit ACS but all pyridoxal phosphaterequiring enzymes, including Trp aminotransferase in the auxin biosynthetic pathway (Soeno et al., 2010). To test their specificity in our assay, we applied ACC together with either inhibitor. If AVG and AOA act upstream of ACC by inhibiting ACS activity, root elongation should still be reduced by ACC. Instead, both compounds prevented most of the ACC-induced reduction of LEH (Fig. 3A, dark gray bars). This indicates that in our experimental system, AVG and AOA inhibit processes that are downstream or independent of ACC biosynthesis so are not suitable inhibitors. Conversely, AIB should act downstream of ACC in the ethylene biosynthesis pathway but did not affect the ACC response (Fig. 3A). It is possible that AIB inhibits isoxaben-induced processes that are upstream or independent of ACC. Another not mutually exclusive explanation is that the 10 mM dose of ACC used here is too high for ethylene generation to be efficiently inhibited by AIB, a very weak competitive inhibitor of ACC oxidase. To reverse most of the root response to 1 mM external ACC, the highest tested dose of 1 mM AIB was required (Supplemental Fig. S1). Our finding that 10 mM AIB reverses the isoxaben response suggests that only small amounts of ACC are generated and/or that the inhibitor acts on a different target.
Recently, novel inhibitors of ACS have been identified from a chemical screen (Lin et al., 2010). To obtain additional evidence that ACC biosynthesis is required to block root elongation in response to isoxaben, we tested the compound that was most active as a suppressor of the constitutive triple response phenotype of the eto1-4 mutant (Lin et al., 2010), 2-anilino-7-(4-methoxyphenyl)-7,8-dihydro-5(6H)-quinazolinone (7303). At 5 mM, this compound did not affect elongation on its own but completely inhibited the acute response to isoxaben (Fig.  3A). In contrast, externally added ACC reduced elongation as much as in the untreated control. Since this compound is an uncompetitive inhibitor of ACS that is structurally and mechanistically unrelated to the more established AVG, these results independently corroborate ACC biosynthesis as a necessary component of cell wall integrity signaling.
Other cell wall stress inducers, like the cell wallbinding dye Congo red, reduced root elongation on a similar time scale, and the two ACS inhibitors restored most or all of this inhibition (Fig. 3B). Stress factors other than cell wall defects are also known to reduce plant growth in general and/or root elongation in particular, including microbial PAMPs or elicitors (Gomez-Gomez et al., 1999). We tested the short-term effect of flagellin-22, which is known to trigger defense responses in roots (Millet et al., 2010). Flagellin reduced the LEH within 5 h to about the same level as isoxaben at 3 h. This reduction, too, was sensitive to AIB and 7303 (Fig. 3B). These results show that root elongation is rapidly controlled via ACC biosynthesis in response to a wide range of stress triggers.
Regulation of ACS activity occurs at both the transcriptional and posttranscriptional levels (Argueso et al., 2007). To test whether ACS expression was induced in the elongation zone in response to isoxaben, we analyzed all available pACS::GUS lines. ACS isoforms ACS2, -4, -6, and -8 showed strong (and ACS5 showed weak) constitutive expression, while ACS9 was not detected. In contrast, ACS11 was induced in the root elongation zone within a few hours of isoxaben treatment (Supplemental Fig. S4).
A complex interplay between hormones regulates root growth (Benková and Hejátko, 2009). Several recent studies have shown that auxin biosynthesis and transport are required to mediate ethylene responses in roots (Rů zicka et al., 2007;Stepanova et al., 2007;Swarup et al., 2007): ACC/ethylene activates auxin synthesis via a root tip-specific pathway involving anthranilate synthase and Tyr aminotransferase (Stepanova et al., 2005(Stepanova et al., , 2008. In addition, basipetal auxin transport via AUX1 and EIR1/PIN2 (but not acropetal transport via PIN1 and PIN4) is required. We have analyzed the isoxaben-induced LEH reduction in these mutant backgrounds. pin4-3 responded like the wild type; the aux1-T, eir1-1, and tir1-1 mutants showed a response that was slightly smaller in amplitude than the wild type, but the reduction in these cases was from a considerably bigger cell size down to the level of untreated wild-type roots ( Fig. 4A; Supplemental Fig. S5). To phenocopy the effect of strong auxin-resistant mutants, we used a synthetic antagonist of TIR1 receptor function, a-(phenylethyl-2-oxo)indole acetic acid (PEO-IAA; Hayashi et al., 2008) as well as its inactive 5-methyl derivative. The antagonist PEO-IAA, but not 5-methyl-PEO-IAA, completely abrogated the effect of isoxaben or ACC on LEH ( Fig. 4B;  Supplemental Fig. S6). Therefore, reduced elongation triggered by stress/ACC requires auxin signaling but is only partially dependent on basipetal auxin transport via AUX1 and PIN2.
De Cnodder et al. (2005) have shown that the effect of ACC or ethylene on root elongation is mediated by extracellular events that affect cell wall cross-linking. Specifically, the production of ROS and cross-linking of Hyp-rich glycoproteins was linked with the reduction of root elongation. As was previously shown for the response to ACC (De Cnodder et al., 2005), root elongation in the presence of isoxaben could be completely restored by diphenylene iodonium (Fig. 4C), an inhibitor of flavin-containing enzymes, including NADPH oxidases.

DISCUSSION
Growing primary roots are vulnerable to cell wall damage; the rapid elongation phase increases the cell surface area by an order of magnitude. This requires massive cell wall rearrangement and greatly increased polysaccharide biosynthesis. We have shown that in- hibition of cellulose biosynthesis or interference with cell wall assembly rapidly reduces elongation. Understanding the molecular mechanisms of cell wall integrity control poses a dilemma: cell wall damage, be it genetically or pharmacologically induced, will ultimately lead to structural problems that are incompatible with growth. It is conceivable that growing cells with too weak walls simply burst and die. Growth arrest, as observed in isoxaben-treated seedlings, might be due to osmotic stress, plasma membrane stretch, and/or general stress, rather than a dedicated system of cell wall integrity surveillance, as in yeast and other fungi. However, two major lines of evidence have supported the existence of such a monitoring system so far. First, RLKs without obvious cell wall biosynthetic capacity have been found (e.g. FEI1 and FEI2, FERONIA, THESEUS, and HERCULES1; Xu et al., 2008;Guo et al., 2009) whose absence causes cell wall defects similar to mutants in cell wall biosynthetic genes. This suggests control of wall biogenesis by an integrated signaling pathway. Second, if a signaling process lies between cell wall defects and growth reduction or other compensatory responses, it should be possible to restore growth by disrupting signal transduction without restoring the cell wall defect. The identification of the receptor kinase THE1 (Hématy et al., 2007) provided clear genetic evidence that this is the case in hypocotyls. Dwarfism and reduced elongation of etiolated hypocotyls in several moderately cellulose-deficient mutants is much less pronounced in the absence of THE1. In this study, we have demonstrated that an active signaling process reduces root elongation when cell wall biosynthesis is impaired. Disrupting this signaling process at any one of three different steps (ACC biosynthesis, auxin signaling, and superoxide production) fully restores elongation in the short term despite clearly visible cell wall damage. We have thus confirmed the existence of a cell wall integrity signaling pathway for roots. Interestingly, neither of the above-mentioned RLK mutants with a cell wall integrity phenotype (fei1 fei2 and the1) was affected in the rapid response to isoxaben in our assays (Supplemental Fig. S5; data not shown). The original sensor(s) that communicate(s) deficient cell wall structure or biosynthesis to the cytoplasm thus also remains to be identified. The cell wall integrity pathway merges with a general rapid stress response pathway that requires ACC biosynthesis but not ethylene perception to control root elongation.
Ethylene reduces root elongation via auxin biosynthesis in the root tip (Stepanova et al., 2005(Stepanova et al., , 2008 and basipetal auxin transport (Rů zicka et al., 2007;Stepanova et al., 2007;Swarup et al., 2007). Farther down in the pathway, NADPH oxidase-dependent superoxide production and cell wall protein maturation and cross-linking are required (De Cnodder et al., 2005). Our experiments first seemed to show that the cell wall damage response (as well as the response to microbial PAMPs) follows this established pathway, but surprisingly, ethylene perception was not required for this response. As summarized in Figure 5, in the rapid control of root cell elongation, both ACC and ethylene can act via what appears to be the same pathway. For long-term growth responses, such as those shown in Supplemental Figure S3, conversion of ACC to ethylene is required. We have also found that the expression of some rapidly isoxaben-induced genes requires ethylene perception (data not shown). The "shortcut" pathway from ACC may be limited to short-term responses including reduced elongation. We have not formally established that auxin is upstream of ROS production in this context. While this order is best supported by the literature (Joo et al., 2001;Schopfer et al., 2002), auxin-ROS interactions are not one way, and both have independent functions in root growth control (Potters et al., 2007;Tsukagoshi et al., 2010). The exact process inhibited by diphenylene iodonium-ROS signaling or direct oxidative remodeling of cell walls-remains open.
One puzzling fact remains. Externally added ACC acts on root elongation both as ACC itself and after conversion to ethylene, as shown by the partial sensitivity to silver. Why, then, is the short-term response to cell wall damage or PAMPs completely ethylene independent? If ACC biosynthesis is activated, how could it not be converted to ethylene and thus act on elongation? One explanation could be that either ethylene formation or ethylene signaling is (temporarily) suppressed. An alternative, perhaps more likely, working hypothesis is that ACC concentrations induced by cell wall or biotic stress are initially low and that ACC binds to a dedicated receptor or binding protein with higher affinity than to ACC oxidase, triggering a reduction in elongation. As ACC accumulates, this system is saturated and ACC oxidation to ethylene begins. Externally added ACC would practically always exceed the capacity of this system and exert its effects at least partially via ethylene. Low concentrations of AIB efficiently inhibit the postulated ACC- binding protein, as proposed by Xu et al. (2008), while millimolar concentrations are required to block ACC oxidase. It remains unclear exactly how ACC biosynthesis is controlled in the cell wall damage response. Although expression of ACS11 is induced rapidly by isoxaben, several other isoforms are constitutively expressed. Posttranslational mechanisms play an important role in the regulation of ACS stability (Argueso et al., 2007) and might explain the discrepancy.
The idea that in some developmental pathways ACC might act as a signal in its own right, independently of ethylene receptors or the canonical downstream pathway, has been proposed before (Xu et al., 2008;Tsuchisaka et al., 2009). Multiple knockouts of ACS genes have increasingly severe developmental defects (Tsuchisaka et al., 2009), while completely ethylene-insensitive mutants like ein2 and ein3 eil1 are relatively healthy (Alonso et al., 1999(Alonso et al., , 2003. Currently, we do not know to what degree the responses to ACC and to ethylene overlap. ACC could trigger distinct responses or "just" be a shortcut to ethylene responses. The advantage of an ACC shortcut to at least some ethylene responses could be that they can be triggered earlier in a cell-autonomous way before threshold concentrations of the easily dissipating gaseous ethylene are reached. We note that for the shortterm response to isoxaben or ACC, auxin signaling but not auxin transport is required ( Fig. 4B; Supplemental  Fig. S5), while the long-term growth response requires both (Rů zicka et al., 2007;Strader et al., 2010). This suggests that for auxin, too, only cell-autonomous functions may be required in the hypothetical ACC shortcut. Which of the established later elements of the ethylene pathway (if any) are required to transmit this rapid ACC-dependent signal remains the subject of further research. CONCLUSION We have established that a rapid reduction in root cell elongation is an acute response to perturbation of cell wall integrity. We have shown that this is a response mediated by an active signaling pathway rather than a simple consequence of cell wall failure. For at least a few hours, elongation can proceed despite cell wall damage if signaling is blocked. This means that the short-term LEH assay can be used as a tool to separate cell wall signaling from broader structural damage induced by isoxaben. We are currently analyzing a range of mutants in candidate cell wall receptors and other signaling proteins for loss of shortterm responsiveness in this assay.
It is intriguing that ACC biosynthesis but not canonical ethylene signaling is required for this stressinduced morphogenic response (Potters et al., 2007). Having established the root elongation zone as a target of an ACC shortcut, it would be very interesting to dissect the transcriptional response to ACC versus ethylene in roots of the wild type and ethylene-insensitive mutants. The relative contribution of canonical and ethylene-independent pathways remains to be established, and unique genetic components of ACC signaling have yet to be identified.

Root Treatments and LEH Measurements
At least 20 4-d-old seedlings were carefully transferred onto microscopic slides with a "cushion" of half-strength Murashige and Skoog agar containing the treatment. The slides were kept in a petri dish with a wet filter paper, sealed with surgical tape, and returned to the growth chamber typically for 3 h. The following final concentrations of reagents were used: 150 nM isoxaben, AOA, AVG, and AIB all at 10 mM, 50 mM flg-22 peptide (Eurogentech), 5 mg L 21 Congo red, 10 mM silver thiosulfate (a 20 mM stock was freshly prepared by mixing 1 volume of 100 mM silver nitrate with 4 volumes of 100 mM sodium thiosulfate), 10 mM compound 9127303 ("7303"; Hit2Lead; Chembridge Corp.), 200 mM ethephon (from 5 mM stock mixed with an equal volume of 15 mM HEPES/KOH, pH 6.5) for direct contact or 200 mL of 5 mM ethephon added to a filter paper wetted with 1 mM K 2 HPO 4 , pH 9, for ethylene gas generation (resulting in a maximal concentration of 500 mL L 21 ethylene gas), 25 mM PEO-IAA and 5-methyl PEO (a kind gift of Ken-ichiro Hayashi), and 10 mM diphenylene iodonium. All chemicals were from Sigma-Aldrich unless noted otherwise.
After 3 h of treatment, roots were analyzed directly (i.e. without a coverslip) with a microscope (DMR; Leica) fitted with a 203 objective (HL PL Fluotar; Leica; numerical aperture = 0.50). Images were taken with a SPOT Xplorer 4Mp camera (Diagnostic Instruments), and the LEH (Le et al., 2001) was measured with the SPOT software or ImageJ. On each root, one to three cells could be measured with confidence, resulting in 30 to 50 measurements per treatment. Each result is representative of at least three independent experiments.

Environmental Scanning Electron Microscopy
The microscope images were taken using a FEI Quanta 200 scanning electron microscope (Philips) in environmental scanning electron microscopy mode (gaseous secondary electron detector, 30 kV, 3.5-5.5 Torr, sample cooled to 5°C). The apical 5 to 6 mm of root tips was detached with a razor blade, transferred to a drop of water on the sample stub, and inserted into the chamber. Images were taken just before all water had evaporated from the sample stub.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. AIB acts as competitor of the ACC effect on root elongation.
Supplemental Figure S5. LEH response to isoxaben treatment in auxin transport or signaling mutants and in fei1 fei2.
Supplemental Figure S6. The auxin response inhibitor PEO-IAA, but not its inactive 5-methyl derivative, restores elongation in the presence of ACC.