Research ArticlePlant biology

Signaling from the Endoplasmic Reticulum Activates Brassinosteroid Signaling and Promotes Acclimation to Stress in Arabidopsis

See allHide authors and affiliations

Science Signaling  28 Sep 2010:
Vol. 3, Issue 141, pp. ra69
DOI: 10.1126/scisignal.2001140


The ability to acclimate to stresses enables plants to grow and develop under adverse environmental conditions. Regulated intramembrane proteolysis (RIP) triggered by endoplasmic reticulum (ER) stress mediates some forms of stress signaling. Brassinosteroids (BRs) have been implicated in plant adaptation to stress, but no mechanisms for activation have been discovered. Here, we reveal a connection between ER stress signaling and BR-mediated growth and stress acclimation. Arabidopsis transcription factors bZIP17 and bZIP28 were translocated from the ER through the Golgi, where they were proteolytically cleaved by site 2 protease and released to translocate into the nucleus. Stresses, including heat and inhibition of protein glycosylation, increased translocation of these two bZIPs to the nucleus. These nuclear-localized bZIPs not only activated ER chaperone genes but also activated BR signaling, which was required for stress acclimation and growth. Thus, these bZIPs link ER stress and BR signaling, which may be a mechanism by which plant growth and stress responses can be integrated.


Brassinosteroid (BR) regulates many growth and developmental processes, such as cell elongation, vascular development, and photomorphogenesis (14). Exogenous BR treatment can also induce plant tolerance to a wide spectrum of biotic and abiotic stresses, such as pathogen attack, low and high temperatures, drought, and high salt (59). Although high temperature has been reported to activate BR signaling (10), it is not known whether endogenous BR signaling normally provides tolerance to stresses, how such stresses activate BR signaling, or which response mechanisms are activated by such BR signaling (11).

Regulated intramembrane proteolysis (RIP) is a conserved stress response mechanism in eukaryotes (12) (fig. S1). In mammals, insects, and fungi, it plays a key role in sensing stresses, such as those resulting from changes in metabolism of lipids (13, 14) and sterols (15, 16), oxygen deficit (1719), and the unfolded protein response (UPR) (20, 21). RIP activates a specific subset of transcription factors that reside as transmembrane proteins within the endoplasmic reticulum (ER) in unstressed cells. In response to stress, these transcription factors are translocated to the Golgi where they are released into the cytosol by proteolytic cleavage by two membrane-embedded proteases (12, 20). One is a subtilisin-like serine protease, which cleaves at one specific site (site 1 protease or S1P), and the other is a metalloprotease (site 2 protease or S2P), which cleaves at a second site within the transmembrane domain. Consequently, the transcription factors are released from the Golgi membrane and translocated to the nucleus where they activate the expression of their target genes (12, 15, 20).

In plants, RIP is implicated in responses to stresses caused by protein misfolding as well as excess salt and heat (2228) (fig. S1). Arabidopsis S1P and S2P homologs have been identified (22, 29, 30). S1P provides tolerance to salt stress by activating bZIP17 (22), a mechanism similar to the ER stress response mediated by the transcription factor ATF6 in mammals (20, 21) (fig. S1). Two other bZIP17 homologs, bZIP28 and bZIP49, have also been identified. bZIP28 mediates tunicamycin (TM)–induced ER stress responses related to UPR (23, 27, 28) and is also activated in response to heat shock (31). The function of bZIP49 is unknown.

Through our investigation of the function of S2P, we discovered a connection among ER stress sensing, BR signaling, and stress acclimation.


S2P mediates bZIP proteolysis and nuclear translocation

We visualized the expression of the S2P gene in transgenic plants expressing the GUS reporter under the control of the S2P promoter. S2P::GUS was expressed mainly in the vasculature of cotyledons, leaves, and roots in a pattern identical to that of S1P::GUS (fig. S2A). With a 35S::S2P-GFP (green fluorescent protein) transgene, we established that S2P localized to the Golgi (Fig. 1A), as did S1P (fig. S2B). We investigated the proteolysis and subcellular localization of bZIP17 and bZIP28 in 35S::GFP-bZIP transgenic wild-type and s2p-null mutant plants (fig. S3, A and B). In roots of wild-type plants, GFP-bZIP17 was observed mainly in the ER as previously demonstrated (22, 27), with little fluorescence in the nucleus (Fig. 1B), although more fluorescence was observed in the nuclei of some cells, particularly in the elongation zone (fig. S2C). However, GFP-bZIP17 relocalized predominantly to the nucleus within 1 hour of heat shock or 4 hours of treatment with TM, an agent that blocks N-linked protein glycosylation in the secretory pathway and induces ER stress (Fig. 1B). In the s2p mutant, GFP-bZIP17 was localized predominantly in the Golgi (Fig. 1C), and heat shock or TM treatment failed to alter its localization (Fig. 1B). GFP-bZIP28 exhibited an S2P-dependent distribution similar to GFP-bZIP17 (fig. S2, D and E).

Fig. 1

Dynamic changes and proteolysis of GFP-bZIP17 under stress. (A) Colocalization of S2P-GFP with Golgi-mCherry in onion epidermal cells after biolistic bombardment. Scale bar, 20 μm. (B) Localization of GFP-bZIP17 in response to heat shock (HS) and ER stress (TM treatment) in wild-type and s2p root tips stably transformed with GFP-bZIP17. Scale bar, 10 μm. (C) Colocalization of GFP-bZIP17 with Golgi-mCherry in s2p root cells stably transformed with GFP-bZIP17 and biolistically bombarded with Golgi-mCherry. Scale bar, 10 μm. (D) Proteolysis of GFP-bZIP17 in response to HS in 7-day-old wild-type and s2p roots. (E) Endo H treatment of protein samples from seedlings of wild-type and s2p roots. GFP-bZIP17 and its truncated form were detected by immunoblotting with antibody against GFP.

To demonstrate whether, in addition to nuclear localization, bZIP17 proteolysis was also S2P-dependent, we monitored the size shift of the GFP-bZIP17 fusion protein before and after heat shock treatments in both wild-type and s2p mutant backgrounds. In the absence of heat shock, we detected three main bands (I, II, and III) of GFP-bZIP in wild-type roots (Fig. 1D). Heat shock induced a mobility shift of band I (the most abundant band before heat shock) to band III (the most abundant band after heat shock), suggesting that band I was the full-length GFP-bZIP17 localized in the ER and band III was the truncated N-terminal form localized in the nucleus (Fig. 1D). Although band III was about the size predicted for S2P-processed GFP-bZIP17 fusion protein (~70 kD), band I was slightly bigger than calculated full-length GFP-bZIP17 fusion protein (~108 kD). We determined that band I was glycosylated GFP-bZIP17, because it was converted to band II when the protein extract was treated with β-N-acetylglucosaminidase H (Endo H) (Fig. 1E). However, in s2p, only band II, the Endo H–resistant form, was observed in the presence or absence of heat shock (Fig. 1, D and E), indicating that GFP-bZIP17 was retained in the Golgi and was not cleaved even under stress conditions. Collectively, these results show that proteolysis, and relocation of bZIPs from ER to the nucleus, occurs through the Golgi and is S2P-dependent. Because of the presence of the Endo H–resistant form in unstressed cells (Fig. 1E) and accumulation of bZIP17 in the Golgi in unstressed s2p cells (Fig. 1, B and C), these results also suggest that transit of bZIPs through the Golgi to the nucleus occurs even in the absence of stress but is greater when plants are subjected to stress.

Stress response is RIP dependent

We found that the s2p mutant had pronounced sensitivity to several abiotic stresses. Formation of the first pair of true leaves in seedlings was inhibited by TM to a greater extent in s2p than in wild type, and s2p was restored to wild-type sensitivity when transformed with S2P genomic DNA (Fig. 2A). Compared to wild-type plants, leaves of the s2p mutant also lost water more rapidly (Fig. 2B) and s2p plants were more sensitive to salt stress (producing larger amounts of anthocyanin) (Fig. 2C).

Fig. 2

s2p is sensitive to various abiotic stresses. (A) Phenotypes of 7-day-old wild type (WT), s2p, and s2p transformed with S2P gene (S2P::S2P/s2p) grown on ½ MS with or without TM (two independent S2P::S2P/s2p transgenic lines were used, #4 and #5). (B) Water loss from detached leaves. Leaves were detached from 4-week-old soil-grown plants, allowed to dry in the growth chamber (23°C, 60% relative humidity), and weighed at indicated intervals. (C) Four-week-old plants were irrigated for 12 days at 4-day intervals with increasing concentrations of NaCl: 50, 100, 200, and 300 mM. Stress sensitivity is indicated by red pigmentation (anthocyanin accumulation). (D) TM sensitivity of s2p was rescued by expressing bZIPΔCs 7 days after sowing. (E) Quantitative reverse transcription PCR (qRT-PCR) analysis of relative gene expression of ER stress marker genes in wild-type and s2p seedlings treated with TM (5 μg/ml) for 2 or 4 hours. Two-sample t tests were performed to compare expression of chaperone genes of s2p with that of wild type at the same time as TM treatment. *P < 0.05. Cont., control. (F) Effects of bZIPΔCs on expression of ER stress marker genes in unstressed seedlings. Two-sample t tests were performed to compare expression of chaperone genes of bZIPΔC transgenic plants with that of wild type. *P < 0.05. (G) Effects of bZIPΔCs on the abundance of the indicated ER stress–related proteins in unstressed seedlings. Chaperones were immunodetected with antibodies against BiP and CRT (53). α-Tubulin (α-TUB) was probed as loading control. Numbers below each panel indicate the relative protein levels quantified by measuring band intensities with ImageJ (, normalized by dividing by the α-tubulin band intensity and presented as the value relative to that of wild type. Experiments were performed three times with similar results. In (A) and (D), percentile represents the percentage of plants with the first pair of leaves on the shoots (±SE from three independent counts). In (B) to (F), error bars indicate SE (n = 3).

To determine whether constitutively active bZIP17 and bZIP28 could overcome the abiotic stress sensitivity of the s2p mutant, we expressed truncated forms of bZIP17 and bZIP28 lacking their C-terminal ER-anchoring transmembrane domains (namely, bZIP17ΔC and bZIP28ΔC) (fig. S4). The truncated bZIPs performed their nuclear functions independently of RIP (22, 23, 27, 31). Either protein complemented the TM response of s2p (Fig. 2D), indicating that they have functionally redundant roles in regulating ER stress responses and that the truncated bZIPs obviate the requirement for S2P-dependent proteolysis.

To further define the role of S2P in ER stress signaling, we examined the expression of genes encoding ER chaperones. ER stress, such as that induced by TM treatment, causes unfolded proteins to accumulate, leading to stimulation of genes encoding chaperones (23, 27, 28, 32). Quantitative real-time polymerase chain reaction (qPCR) showed that the basal expression of some chaperone genes was lower in s2p seedlings and that TM-stimulated induction of some of these genes (especially CRT1, CRT2, and PDIL) was impaired in s2p (Fig. 2E). This defective transcriptional response could explain the sensitivity of s2p to TM. The existence of other RIP-independent ER stress regulatory pathways, such as that mediated by bZIP60 (30) or heterotrimeric G protein signaling (33), could explain why the response of chaperone genes to TM is not completely abolished in s2p. In the s2p mutant containing bZIP28ΔC, basal expression of the chaperone genes was increased in the absence of ER stress (Fig. 2F). Although bZIP17ΔC triggered a weaker induction for most genes analyzed than did bZIP28ΔC when controlled by their native promoters, bZIP17ΔC controlled by the strong 35S promoter enhanced chaperone gene expression to a similar amount as that induced by bZIP28ΔC under its native promoter (Fig. 2F). Consistent with the analysis of transcripts, accumulation of chaperone proteins was lower in s2p, but enhanced by bZIPΔCs in the absence of TM (Fig. 2G). These data support a role for these bZIPs in regulating the ER stress response and suggest that increased basal ER chaperones may be a mechanism for TM tolerance of s2p plants expressing these constitutively active forms of the proteins. These observations are consistent with the model in which ER stress leads to S2P-dependent proteolysis of bZIP proteins (hereafter referred to as S2P-RIP), which in turn activates the production of ER chaperones.

S2P-RIP stimulates BR responses

Although shoot growth appeared normal (Fig. 2A), the s2p mutant exhibited impaired root growth (70% shorter than that of wild type and more branched than that of wild type) when grown on the surface of agar medium (Fig. 3A). When grown in well-watered soil, again s2p root growth was impaired, whereas shoots appeared the same as those of the wild type (Fig. 3, B and C). Root growth was restored when s2p was transformed with the S2P genomic DNA (Fig. 3A). The s1p mutants (22) also had short roots (Fig. 3, D and E). Consistent with bZIP17 and bZIP28 mediating responses of S2P-RIP, expression of bZIP17ΔC or bZIP28ΔC in s2p restored root growth (Fig. 3A).

Fig. 3

Roles of S2P-RIP in plant development and BR response. (A) Complementation of root development of s2p with S2P genomic DNA or bZIPΔC constructs. (B) Shoot phenotypes of 4-week-old wild type and s2p grown in soil under constant light. (C) Root phenotypes of 4-week-old wild type and s2p grown in soil under constant light. (D) Root phenotypes of 1-week-old wild type, s1p-2, and s1p-3 mutants grown vertically on ½ MS plates. (E) Wild type (Col 0), s1p-2, s1p-3, and s2p mutants were grown vertically on ½ MS plates; root lengths were measured at intervals after germination. (F) Effect of 0.5 nM BL on wild-type (Col 0) and s2p root growth. (G) Root growth of seedlings after 5 days on 0.5 or 1 nM BL relative to control without BL. (H) Root growth of seedlings after 7 days on different concentrations of BL, comparing wild type with bzip17 (SALK_104326) and bzip28 (SALK_132285C) mutants. Two-sample t tests were performed to compare root elongation of bzip17 and bzip28 to that of wild type at the same concentration of BL. *P < 0.01 (t test). In (E), (G), and (H), error bars indicate SE (n = 30).

BR signaling activates genes involved in cell wall loosening contributing to increased cell expansion (3436), which is a key aspect of root and shoot development. Disruption of BR signaling causes short root and dwarf shoot phenotypes (4). To better understand the function of S2P in plant development, we tested the sensitivity of root elongation to exogenous BR or in plants with mutations that compromise BR signaling. In wild-type roots, low concentrations of brassinolide (BL) stimulate growth, but at higher concentrations, BL inhibits root growth. To distinguish between wild-type and s2p root response, we used a high concentration of BL that repressed wild-type root growth. Whereas wild-type root growth was repressed by a superoptimal concentration of exogenous BL, s2p roots grew longer in response to this concentration of BL (Fig. 3, F and G). Knockout mutants bzip28 and bzip17 (22, 27, 31) both showed reduced inhibition of root growth by superoptimal BL (Fig. 3H). Expression of bZIP17ΔC or bZIP28ΔC in s2p restored the inhibition of root growth mediated by BL to that of wild type (Fig. 3G).

In addition to assessing how RIP-S2P and bZIP17 and bZIP28 influenced the plant response to exogenous BL, we also evaluated the effect of RIP-S2P on BR target genes and cellular responses. We compared the s2p plants with det2 and bri1-5 mutants, which are impaired in BR synthesis and perception, respectively. The det2 and bri1-5 mutants had short roots (4) (Fig. 4A) and reduced basal expression of the BR target genes of the expansin (EXP) family compared to the root length and EXP expression in wild-type plants (Fig. 4B). The s2p plants also had lower expression of EXP genes in seedlings and smaller root cells than did wild-type plants (Fig. 4C), but s2p seedlings expressing truncated bZIP proteins had increased basal expression of these EXP genes and root cell size that was the same as that of wild type (Fig. 4, B and C).

Fig. 4

Genetic interaction of S2P-RIP with BR mutants. (A) Roots of wild type (WS), bri1-5, s2p bri1-5, det2, and s2p det2 grown vertically on ½ MS plate for 7 days. (B) qRT-PCR analysis of EXP8 and EXPB1 gene expression in response to BL in various genotypes, relative to wild type (Col 0). Error bars indicate SE (n = 3). (C) Length of mature root cortex cells in different genotypes. Error bars indicate SE (n = 150). (D) Roots of s2p and s2p bes1-D after 7 days of growth on ½ MS. (E) Plants of wild type (En 2), bes1-D, and s2p bes1-D grown in constant light in soil for 3 weeks.

We also crossed s2p with constitutively active BR signaling mutant bes1-D. BES1 is a transcription factor activated by dephosphorylation when BR binds to its receptor BRI1 (37, 38). The constitutive signaling mutant bes1-D complemented s2p by repressing lateral root development and partially restoring primary root elongation (Fig. 4D). Furthermore, the elongated leaf petiole and small lamina phenotypes caused by bes1-D were attenuated by the s2p mutation (Fig. 4E). These observations are consistent with a positive interaction of S2P-RIP on BR signaling pathways.

If S2P-RIP positively regulates or enhances BR responses, we may expect to see a positive effect in mutants in which BR signaling is impaired, but not completely absent. We introduced bZIP17ΔC or bZIP28ΔC into bri1-5, a weak bri1 allele with a functional BR receptor that exhibits partial activity because its transfer from the ER to the plasma membrane is impaired (39). Expression of either bZIP17ΔC or bZIP28ΔC in bri1-5 promoted shoot development (Fig. 5, A to C) and hypocotyl elongation in the dark (light inhibits hypocotyl elongation) (Fig. 5D). bZIPΔC expression in bri1-5 also increased expression of the EXPB1 gene in 1-week-old seedlings (Fig. 5E) and EXP8 in 2-week-old seedlings (Fig. 5F). Consistent with the positive effects of bZIP17ΔC and bZIP28ΔC, we also observed that introduction of the s2p mutation into bri1-5 decreased expression of EXP genes (Fig. 5E) and impaired growth (Fig. 5, B to D).

Fig. 5

Genetic interaction of S2P-RIP with bri1-5. (A) The improvement in growth of bri1-5 plants carrying the bZIPΔC constructs is evident in 3-week-old plants. (B) s2p further compromises bri1-5 plant growth and the bZIPΔC constructs rescue growth as seen in the mature plants. (C) bZIPΔCs restore root elongation and plant growth of bri1-5. Error bars indicate SE (n = 20). ND, not detected. (D) Effects of bZIPΔCs on bri1-5 hypocotyl elongation. Seedlings were grown in the dark for 5 days. (E) s2p reduces EXP8 and EXPB1 expression in bri1-5 plants, whereas bZIP17ΔC and bZIP28ΔC differentially affect the expression of these genes in bri1-5 in 1-week-old seedlings. *P < 0.05 (t test). Error bars indicate SE (n = 3). (F) Microarray analysis (as described in Fig. 7) shows that EXP8 expression is significantly enhanced in 2-week-old seedlings.

To investigate the underlying mechanism by which bZIP17 and bZIP28 promote bri1-5 development, we introduced bZIP17ΔC or bZIP28ΔC into bri1-6 and det2. The protein encoded by the bri1-6 allele is located in the plasma membrane and binds BR, but BR signaling is impaired (1, 40). Expression of either bZIP17ΔC or bZIP28ΔC in bri1-6 (Fig. 6A) or det2 (Fig. 6B) had no effect on shoot development, suggesting that BR synthesis and perception are required for the response to S2P-RIP. Furthermore, in det2 plants, neither expression of bZIP17ΔC or bZIP28ΔC nor crossing the det2 plants with s2p plants altered expression of the EXP genes from the reduced expression seen in the det2 plants (Fig. 6C). Introduction of the s2p mutation or expression of bZIP17ΔC or bZIP28ΔC in det2 also failed to alter the reduced hypocotyl elongation of the det2 plants in the dark (Fig. 6D) or alter det2 overall growth (Fig. 6, E and F). Because the effects of altering S2P-RIP occur in bri1-5 but not in det2 or bri1-6, S2P-RIP appears not to function in parallel to BR signaling or to activate downstream BR signaling, but instead may act at the level of BR perception. Changes in BR production leading to increased BR concentrations are unlikely to mediate the effects of expressing bZIP17ΔC or bZIP28ΔC in bri1-5 because bri1 mutants typically have BR concentrations ~20 times as high as that of the wild type (41).

Fig. 6

S2P-RIP–mediated BR signaling requires functional BR receptor and BR biosynthesis. (A) Expression of the bZIP17ΔC transgenes in the bri1-6 background fails to alter plant growth in 3-week-old plants. (B) Expression of the bZIP17ΔC transgenes in the det2 background fails to alter growth in 3-week-old plants. (C) qRT-PCR shows that the S2P-RIP pathway does not alter expression of the EXP8 and EXPB1 genes in the det2 background. Error bars indicate SE (n = 3). (D) The S2P-RIP pathway does not alter hypocotyl elongation of det2 plants. Seedlings were grown in the dark for 5 days. (E) The development of bri1-5, but not det2, was highly impaired by s2p. (F) Expression of the bZIP17ΔC transgenes in the det2 background fails to alter growth in mature plants. (G) Immunoblot analysis of 2-week-old seedlings treated with 1 μM BL for 1 hour shows that expression of the bZIP17ΔC transgenes in the bri1-5 background increases BL-induced BES1 dephosphorylation compared to that in bri1-5 plants. Coomassie blue staining of RbcS served as a loading control. The experiment was repeated three times. (H) Immunoblot analysis of 1-week-old seedlings shows that expression of the 35S:bZIP17ΔC transgenes enhances BES1 dephosphorylation in unstimulated plants. α-Tubulin served as a loading control. The experiment was repeated three times.

Further evidence that S2P-RIP was not regulating the amount of BR comes from expression of BR biosynthesis genes DWF4 and CPD. These genes are highly responsive to changes in the concentration of endogenous BL, are repressed by exogenous BL, and are activated by the BL inhibitor brassinazole (BRZ) in wild-type roots (42). However, basal expression, BL-mediated repression, and BRZ-mediated induction of these genes are normal in roots of s2p (fig. S5, A to C). This indicates, firstly, that S2P-RIP does not affect the amount of endogenous BR in roots and, secondly, that it does not influence the feedback signaling of BR to the DWF4 and CPD biosynthesis pathway. Therefore, we conclude that the reduced expression of EXP genes in s2p is not caused by a reduction in endogenous BR, but is more likely due to impaired signaling to stimulate EXP gene expression and activation of the cell expansion pathway. Consistent with this conclusion, expression of bZIP17ΔC or bZIP28ΔC in s2p or bri1-5 increased expression of EXP genes (Figs. 4B and 5, E and F) but had no effect on DWF4 and CPD genes (fig. S5, D and E). Current evidence suggests that there are different signal transduction pathways from BRI1 to DWF4 and CPD BR biosynthesis genes and to genes involved in growth (13, 9), and we hypothesize that S2P-RIP influences the latter.

If S2P-RIP enhanced BR receptor function in bri1-5, we should be able to detect enhanced BR signaling or the consequences of BR signaling upon expression of bZIP::bZIPΔC transgenes. Indeed, examination of activation of BES1 by dephosphorylation revealed that bZIP17ΔC and bZIP28ΔC partially restored dephosphorylation of BES1 in response to BL treatment in bri1-5 (Fig. 6G). Furthermore, in the absence of treatment with BL, the abundance of dephosphorylated BES1 was increased in the s2p mutant by expression of 35S::bZIP17ΔC (Fig. 6H).

To understand the molecular basis for restoration of bri1-5 development and BR sensitivity provided by constitutively active bZIP17 and bZIP28 and to gain a genome-wide view of the complementation of gene expression of bri1-5 by bZIP17ΔC and bZIP28ΔC, we performed transcriptional microarray analyses with wild-type (WS ecotype), bri1-5, bZIP17::bZIP17ΔC/bri1-5, and bZIP28::bZIP28ΔC/bri1-5 seedlings. We identified 712 genes that were significantly down-regulated (Fig. 7A) and 1320 genes that were significantly up-regulated (Fig. 7B) in bri1-5 mutant when compared with WS. When the bZIPΔCs were introduced into bri1-5 mutant, about 20% (139 of 712) of the down-regulated genes were up-regulated by both bZIPΔ17C and bZIPΔ28C (Fig. 7A and table S1). Similarly, about 30% (391 of 1320) of the up-regulated genes were down-regulated by both bZIPΔCs (Fig. 7B and table S2). Furthermore, the bZIP17ΔC and bZIP28ΔC transgenic plants showed substantial overlap (65%) in the regulation of BR-related genes (of the 817 bri1-5 up- or down-regulated genes that were complemented by at least one bZIPΔC, 530 were complemented by both) (Fig. 7, A and B), again supporting the functionally redundant roles of these bZIPs in regulating BR-responsive genes. Among the genes that were regulated by both bZIPΔCs (table S1 and S2) were many well-characterized BR-regulated genes. For example, IAA14 (an auxin-responsive gene), three auxin-responsive genes of unknown function, and EXPs are all known BR-induced genes, and XTR7, PIF3, PR-3, and ROT3 are BR-repressed genes (Fig. 7, C to E) (34, 35). These genome-wide analyses further support the function of S2P-RIP in regulating the BR signaling pathway.

Fig. 7

Microarray analyses of the complementation of bri1-5 gene expression by bZIPΔCs. (A) Overlap in genes that exhibit reduced expression in bri1-5 mutant and are complemented by introduction of bZIP17::bZIP17ΔC and bZIP28::bZIP28ΔC into the bri1-5 background. See table S1 for a complete list of genes that were reduced in bri1-5 and were complemented by both bZIPΔCs. (B) Overlap in genes that exhibit increased expression in bri1-5 mutant and are complemented by bZIP17::bZIP17ΔC and bZIP28::bZIP28ΔC. See table S2 for a complete list of genes that were increased in bri1-5 and were complemented by both bZIPΔCs. (C) Expression of the bZIPΔC transgenes complements expression of EXP genes in bri1-5 plants. These genes showed reduced expression in bri1-5 plants, and expression was restored by the bZIPΔC transgenes. (D) Expression of the bZIPΔC transgenes complements expression of genes associated with auxin signaling in bri1-5 plants. These genes showed reduced expression in bri1-5 plants, and expression was restored by the bZIPΔC transgenes. (E) Expression of the bZIPΔC transgenes complements expression of four known BR-responsive genes that were increased in bri1-5 plants. All data are derived from three replicate microarrays. Error bars indicate SD.

BR signaling provides stress tolerance

It is reported that increased temperature (29°C) can slowly induce BR signaling (10). We found that TM and heat shock (42°C) treatments quickly induced dephosphorylation of BES1 in wild-type but not in s2p mutant seedlings (Fig. 8, A and B), suggesting that stress-induced activation of BR signaling is an important process for stress responses and is S2P-dependent. Furthermore, we observed that the development of both bri1-5 and det2 was inhibited by TM treatment much more than their corresponding wild types (Fig. 8, C and D). Expression of bZIP17ΔC and bZIP28ΔC in these mutants increased the tolerance of bri1-5 to TM, but not that of det2 (Fig. 8, C and D), again supporting the conclusion that S2P-RIP is acting directly on BR signaling and requires BR production. Furthermore, the constitutive BR signaling mutant bes1-D was similar to wild type in its tolerance to TM, and bes1-D restored TM tolerance to s2p (Fig. 8E). In addition, bes1-D also restored tolerance of s2p leaves to water loss, and bes1-D plants were tolerant to mannitol (osmotic) stress, exhibiting less reduction in root elongation than that of wild-type plants (Fig. 8, F and G). Thus, BR signaling enhanced plant tolerance to multiple types of stress. We found that ER chaperone gene expression was not increased in bes1-D (fig. S6, A and B), suggesting that although various stresses can be sensed by the ER and trigger the S2P-RIP response, the subsequent BR-responsive stress tolerance mechanisms do not necessarily act upon the ER (fig. S7).

Fig. 8

BR signaling provides stress tolerance. (A) s2p exhibits reduced BES1 dephosphorylation in response to TM (5 μg/ml) compared to the response of wild-type 1-week-old seedlings. Numbers below each signal denoting protein abundance relative to that of the wild type at time 0 (calculated as in Fig. 2G). (B) s2p exhibits reduced BES1 dephosphorylation in response to heat shock compared to the response of wild-type 1-week-old seedlings. α-Tubulin served as a loading control in (A) and (B), and experiments were performed at least three times with similar results. (C) Expression of the bZIPΔC transgenes in bri1-5 plants overcomes TM sensitivity. (D) Expression of the bZIPΔC transgenes in det2 plants does not affect TM sensitivity. (E) Introduction of bes1-D into s2p plants improves TM tolerance. En 2 is parental wild type for bes1-D. (F) bes1-D provides tolerance to leaf dehydration for s2p. Leaves were detached and weighed at intervals. Error bars indicate SE (n = 5). (G) bes1-D provides tolerance to osmotic stress. For (C), (D), (E), and (G), pictures were taken 7 days after sowing. In (C) to (E), percentile represents the percentage of plants with a first pair of leaves on the shoots (±SE from three independent counts).

Overall, the positive effect of the S2P-RIP pathway on BR responses was evident not only by genetic interactions of S2P-RIP with bri1-5 but also by the altered BL sensitivity of s2p and bzip mutants (Fig. 3, F to H), the S2P-dependent expression of BR-responsive EXP genes (Fig. 4B), the enhanced BES1 dephosphorylation in 35S::bZIP17ΔC transgenic plants (Fig. 6H), the S2P-dependent growth of det2 seedling roots (Fig. 4A), and the S2P-dependent BES1 dephosphorylation in response to stresses (Fig. 8, A and B).


BRI1 is synthesized in the ER and then translocated through the Golgi to the plasma membrane (43). One way in which BR signaling might be stimulated under stressful conditions is by S2P-RIP–dependent ER chaperone synthesis, which could enhance the correct folding or modification of BRI1 and promote its translocation from the endomembrane to the plasma membrane (3, 39, 40, 4345). Naturally occurring abiotic stresses, such as heat shock, salinity, or osmotic stress, may impair BRI1 maturation and translocation; in the laboratory, agents such as TM or mutations in BRI (for example, bri1-5) may impair BRI1 maturation and translocation. We propose that S2P-RIP can overcome such constraints, enhance BR signaling, and induce acclimation to stress by promoting BRI1 delivery to the plasma membrane and potentially by directly enhancing regulation of some BR target genes (fig. S7). We believe it will be important in the future to investigate the effect of the S2P-RIP pathway on BRI1 synthesis, assembly, and translocation, and its effects on the interaction of BRI1 with other membrane-associated proteins, such as BAK1 and BSKs (4648), which potentially provide specificity to BRI1.

We considered the possibility that bZIP17ΔC and bZIP28ΔC might interact directly with transcription factor BES1 in the nucleus, but obtained no evidence for interaction either in yeast two-hybrid experiments or with bimolecular fluorescence (BiFC) in planta. Elucidating details of the molecular interactions linking ER signaling to BR signaling will be an area for future research.

Previous studies indicated that bZIP28 is involved in TM stress (27), whereas bZIP17 is involved primarily in salt stress (22, 27). We have confirmed and extended these observations to show that both are involved in ER stress and heat stress. These proteins may have overlapping sensitivity to different stresses and partially redundant functions, providing flexibility in perception and response to multiple environmental challenges.

The activation of bZIPs in response to stress, as well as their ability to induce BR signaling (fig. S7), is unexpected because BR is typically viewed as a positive regulator of plant growth, and yet plant growth is often impaired by abiotic stresses, such as heat shock or salt stress. However, in natural environments, plants are constantly subjected to stresses, including extreme fluctuations in temperature and water supply. Differential activation of growth processes under stressful conditions may be required to optimize resource allocation or plant development during acclimation. As examples, moderate water stress can stimulate primary root elongation and the rate of cell production (49), and high temperature can stimulate hypocotyl elongation (10). Integration of stress responses and growth processes through a common BR signaling pathway may provide the means to achieve optimal growth under challenging environmental conditions.

Materials and Methods

Plant materials and growth conditions

The det2 mutant is in Arabidopsis thaliana Columbia ecotype, bes1-D and bri1-6 are in Enkheim-2 (En 2), and bri1-5 is in the WS ecotype. The s2p homozygous mutant (NASC N444004) was identified by PCR with gene-specific and left border (LB) T-DNA primers (table S3). s2p bri1-5, s2p det2, and s2p bes1-D double mutants were generated by crossing single mutants and allowing the F1 to self-fertilize, and then candidate double F2 plants were genotyped through a combination of phenotype and PCR genotype examination. All seeds were surface-sterilized and placed on half-strength Murashige and Skoog (½ MS) medium with 1% (w/v) sucrose or in soil under continuous light at 23°C. For root growth inhibition assay, seeds were placed on culture medium containing different concentrations of BL and then grown vertically. Root length was measured on the fifth or seventh day after germination. For TM sensitivity assay, seeds were placed on culture containing TM (0.02 μg/ml) and then grown horizontally for 7 days. The inhibition of formation of the first pair of true leaves in seedlings was the measure for sensitivity. For transient BL, BRZ, or TM treatments, 7-day-old seedlings or roots were transferred into liquid MS culture containing either mock solution, 5 nM BL, 3 μM BRZ, or TM (5 μg/ml) for different times as indicated.

Plasmid constructs

The coding region of S2P with its own promoter was amplified with genomic DNA template and cloned into pCAMBIA3300 vector and named as S2P::S2P. For bZIP17::bZIP17ΔC and bZIP28::bZIP28ΔC constructs, the truncated coding regions of these bZIPs with their own promoters were amplified from genomic DNA template and cloned into pCAMBIA3300 vector (see pGREEN180-GFP was constructed by introduction of CAMV 2 × 35S promoter and CAMV terminator into pGREEN179 (50) and insertion of enhanced GFP into either the Bam HI–Xba I sites (which creates pGREEN180-GFPC for C-terminal GFP tagging) or the Eco RI–Bam HI sites (which creates pGREEN180-GFPN for N-terminal GFP tagging). N-terminal GFP fusions of bZIPs were constructed by inserting open reading frames of bZIPs into pGREEN180-GFPN. C-terminal fusions of GFP to S1P and S2P were made in pGREEN180-GFPC. Biolistic bombardment into onion epidermal cells or Arabidopsis roots was performed as described previously (51). The plasmid G-RB, expressing mCherry targeted to Golgi (52), was used as a Golgi localization control. For stable transgenic lines, the resulting constructs were introduced into Agrobacterium tumefaciens strain C58, which was used to generate transgenic plants by the floral dip method. Primer sequences for these constructs are provided in table S4.

Protein extraction and Western blot analysis

Proteins were extracted with 2 × SDS sample buffer, boiled for 5 min, and centrifuged for 10 min at 10,000g. The resulting supernatants were transferred into a new microfuge tube. SDS–polyacrylamide gel electrophoresis (SDS-PAGE) was performed in 10% (w/v) polyacrylamide gels. After electrophoresis, proteins were electrophoretically transferred to a nitrocellulose membrane (Amersham) and immunodetected with antibodies against GFP (Invitrogen), immunoglobulin heavy-chain–binding protein (BiP; Santa Cruz Biotechnology), calreticulin (CRT) (53), BES1 (38), or α-tubulin (Sigma).

qPCR and microarray analysis

Total RNA was isolated with an RNeasy kit, treated with RNase-free DNase I (Qiagen) according to the manufacturer’s instructions, and quantified by a NanoDrop (ND-1000) spectrophotometer.

For qPCR analysis, 1 μg of total RNA was reverse-transcribed with the SuperScript III RT kit (Invitrogen). qPCR was performed on a Roche LightCycler 480 qPCR instrument. Transcript abundance was calculated relative to the actin-2 gene as described by Czechowski et al. (54). All qPCR data represent the average of three biological replicate experiments. Primer sequences are provided in table S3.

For microarray analysis, three independent biological replicates for each sample (2-week-old WS, bri1-5, bZIP17::bZIP17ΔC/bri1-5, and bZIP28::bZIP28ΔC/bri1-5 seedlings) were analyzed with Affymetrix Arabidopsis gene chips (ATH1). For each microarray, overall intensity normalization for the entire probe sets was performed with Avadis 4.3 (Strand Life Sciences Pvt Ltd). Using the t test packaged in Mev 4.0 (, we estimated a P value (adjusted Bonferroni correction was applied) with 100 permutations to correct for multiple comparisons; P ≤ 0.01 was considered significant. An additional, arbitrary criterion of a 1.5-fold change cutoff was applied to select genes with up- or down-regulation relative to bri1-5.

Microscopy and imaging

For GUS staining, whole seedlings and excised plant organs and tissues were incubated in 5-bromo-4-chloro-3-indole glucuronide (X-gluc) solution {X-gluc (0.5 mg/ml) in 50 mM tris-NaCl buffer (pH 7.0), 0.5% (v/v) Triton X-100, 0.5 mM K3[Fe(CN)6], 0.5 mM K4[Fe(CN)6], 10 mM Na2EDTA}. X-gluc stock solution (100 mM) was prepared by dissolving 26.1 mg of X-gluc in 0.5 ml of dimethyl sulfoxide just before use. Vacuum infiltration was carried out for 10 min. Tissue was then incubated at 37°C in the dark for 16 hours or until color developed. To improve the contrast, we removed soluble pigments by incubating the stained material in several changes of 70% (v/v) ethanol until the chlorophyll was cleared from the tissue. The stained tissue was examined under bright-field microscopy with an Olympus SZX7 microscope. For fluorescence imaging, 5- or 6-day seedlings were subjected to heat shock (42°C for 60 min) or TM (5 μg/ml for 4 hours) treatments. To measure root hair zone cortex cells, we stained 5-day seedlings for at least 10 min with propidium iodide (0.1 mg/ml) to visualize root cell walls. A Leica TCS SP2 AOBS multiphoton confocal microscope was used with laser excitation of GFP at 488 nm and red fluorescent protein (RFP) or propidium iodide at 561 nm. Emission collection was in the ranges of 495 to 525 nm and 590 to 700 nm, respectively. Confocal images were captured with Leica Confocal Software and manipulated with ImageJ or Photoshop. Root cell length measurements were made with ImageJ.


Acknowledgments: We thank R. Boston (North Carolina State University, Raleigh, NC) for providing the antibody against CRT, Y. Yin (Iowa State University, Ames, IA) for BR mutant seeds and antiserum against BES1, Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH) and Nottingham Arabidopsis Stock Centre (University of Nottingham, Nottingham, UK) for Arabidopsis T-DNA mutants, and G. Ingram (University of Edinburgh, Edinburgh, UK) and D. Nelson (University of Western Australia, Perth, Western Australia, Australia) for critical comments on the manuscript. We acknowledge the Australian Microscopy and Microanalysis Research Facility at the Centre for Microscopy, Characterisation and Analysis, University of Western Australia, a facility funded by The University, State and Commonwealth Governments. Funding: This work was supported by the Australian Research Council (grants FF0457721 and CE0561495) and the Government of Western Australia Centres of Excellence scheme. Author contributions: P.C., W.Z., and S.M.S. initiated the project. P.C. performed all the experiments except the confocal microcopy and microarray, which were performed by J.D.B., and leaf dehydration experiments by G.M.E. and B.J.P. All authors were involved in the interpretation of data and manuscript content. P.C. and S.M.S. wrote the manuscript. Competing interests: The authors declare that they have no competing interests.

Supplementary Materials

Fig. S1. A diagram of regulated intramembrane proteolysis.

Fig. S2. Localization of S1P, S2P, and bZIP28.

Fig. S3. Characterization of s2p mutant.

Fig. S4. Structures of full-length and truncated bZIPs.

Fig. S5. S2P-RIP does not affect BR biosynthesis gene expression.

Fig. S6. BES1 does not induce chaperone gene expression in wild-type or s2p backgrounds.

Fig. S7. Possible interactions of stress-induced S2P-RIP with BR signaling pathway.

Table S1. Complementation of bri1-5 down-regulated genes by bZIPΔC.

Table S2. Complementation of bri1-5 up-regulated genes by bZIPΔCs.

Table S3. Oligonucleotides used for PCR, RT-PCR, and qRT-PCR.

Table S4. Oligonucleotides used for plasmid constructs.

References and Notes

View Abstract

Stay Connected to Science Signaling

Navigate This Article