SEARCH

SEARCH BY CITATION

Keywords:

  • Jasmonate;
  • JAZ;
  • ZIM domain;
  • Jas motif;
  • homo- and heteromerization;
  • MYC2

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Discovery of the jasmonate ZIM-domain (JAZ) repressors defined the core jasmonate (JA) signalling module as COI1–JAZ–MYC2, and allowed a full view of the JA signalling pathway from hormone perception to transcriptional reprogramming. JAZ proteins are repressors of MYC2 and targets of SCFCOI1, which is the likely jasmonate receptor. Upon hormone perception, JAZ repressors are degraded by the proteasome releasing MYC2 and allowing the activation of JA responses. All members of the JAZ family share two conserved domains, the Jas motif, required for JAZ interactions with MYC2 and COI1, and the ZIM domain, the function of which is so far unknown. Here, we show that the ZIM domain acts as a protein–protein interaction domain mediating homo- and heteromeric interactions between JAZ proteins. These JAZ–JAZ interactions are independent of the presence of the hormone. The observation that only a few members of the JAZ family form homo- and heteromers may suggest the relevance of these proteins in the regulation of JA signalling. Interestingly, the JAZ3ΔJas protein interacts with several JAZ proteins, providing new clues to understanding the dominant JA insensitivity promoted by truncated JAZΔJas proteins. We also provide evidence that the Jas motif mediates the hormone-dependent interaction between Arabidopsis JAZ3 and COI1, and further confirm that the Jas motif is required and sufficient for Arabidopsis JAZ3–MYC2 interaction. Finally, we show that interaction with MYC2 is a common feature of the JAZ family, as most JAZ proteins can bind MYC2 in pull-down and yeast two-hybrid assays.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The fitness and survival of plants in nature relies on their ability to quickly perceive and respond to external challenges. This ability largely depends on complex regulatory networks that modulate the energetic balance between growth and defence, primarily through the transcriptional reprogramming of the cell in response to external stimuli (Schenk et al., 2000; Jiao et al., 2007; Wise et al., 2007; Baena-Gonzalez and Sheen, 2008; McClung, 2008; McSteen and Zhao, 2008). Among the phytohormones that participate in these regulatory networks, jasmonates (JAs) have been investigated for several decades, and are probably the best-characterized oxylipins. Their signalling pathways modulate plant defence responses against biotic and abiotic challenges (Farmer et al., 2003; Wasternack, 2007; Balbi and Devoto, 2008; Browse and Howe, 2008; Chico et al., 2008). JAs also regulate developmental processes such as root elongation, tuberization, fruit ripening, tendril coiling, cell-cycle progression and senescence (Pozo et al., 2004; Glazebrook, 2005; Lorenzo and Solano, 2005; Wasternack, 2007; Balbi and Devoto, 2008; Pauwels et al., 2008; Zhang and Turner, 2008). The recent discovery of the jasmonate ZIM-domain (JAZ) family of repressors has evidenced the parallelism between the auxin and the JA signalling pathways, and has defined the core JA signalling module as COI1–JAZ–MYC2 (Chini et al., 2007; Thines et al., 2007; Chico et al., 2008; Katsir et al., 2008a; Staswick, 2008). (−)-JA-Ile has recently been shown to be the endogenous bioactive hormone (Thines et al., 2007; Staswick, 2008). This molecule and its bacterial mimic coronatine (COR) promote the hormone-dependent binding of COI1, the F-box component of an Skp1-Cullin-F-box (SCF) E3-ubiquitin ligase (Feys et al., 1994; Xie et al., 1998), to several JAZ proteins from different plant species (Thines et al., 2007; Katsir et al., 2008b; Melotto et al., 2008). Moreover, binding of radiolabelled COR by tomato cellular extracts requires COI1, and can be competed by JA-Ile, suggesting that COI1 is the JA-Ile and COR receptor (Katsir et al., 2008b). Upon hormone perception, SCFCOI1-dependent proteasome degradation of several JAZ proteins (JAZ1, JAZ3 and JAZ6) has been described (Chini et al., 2007; Thines et al., 2007; Shoji et al., 2008), triggering a quick JA-dependent transcriptional reprogramming (Mandaokar et al., 2003; Reymond et al., 2004; Devoto et al., 2005; Uppalapati et al., 2005; Chini et al., 2007; Chico et al., 2008). Among the limited number of transcription factors (TFs) involved in JA signalling, MYC2 is probably the best characterized because of its implication in numerous JA-dependent responses (Boter et al., 2004; Lorenzo et al., 2004; Dombrecht et al., 2007; Pozo et al., 2008). MYC2 was found to be a direct target of JAI3/JAZ3, and proteasome degradation of this repressor is required to trigger MYC2 activity (Chini et al., 2007). Recently, Arabidopsis JAZ1 and JAZ9 were also shown to bind MYC2, and, by analogy with JAZ3, were proposed to repress its transcriptional activity (Melotto et al., 2008).

The JAZ proteins share two conserved regions: the ZIM domain, located in the central part of the protein, and the Jas domain, at the C terminus (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007). The jai3-1 mutant as well as transgenic plants overexpressing truncated versions of JAZ1, JAZ3 and JAZ10 proteins lacking the Jas motif (JAZ1ΔJas, JAZ3ΔJas and JAZ10ΔJas) exhibit a dominant JA-insensitive phenotype (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007; Shoji et al., 2008). Furthermore, JAZ1ΔJas, JAZ3ΔJas and JAZ6ΔJas proteins are not degraded upon hormone perception (Chini et al., 2007; Thines et al., 2007; Shoji et al., 2008). Consistently, the hormone-dependent COI1–JAZ interaction occurs through the Jas domain, and requires two positively charged conserved amino acids within this domain (Katsir et al., 2008b; Melotto et al., 2008). Interestingly, the Jas motif is also necessary for the hormone-independent binding of JAZ proteins to MYC2, but this interaction does not require these two positively charged conserved amino acids of the Jas domain (Chini et al., 2007; Katsir et al., 2008b; Melotto et al., 2008). Therefore, different surfaces of the Jas domain were proposed to modulate specific interactions with COI1 and MYC2 (Melotto et al., 2008).

The ZIM domain contains a highly conserved sequence (the TIFY motif), and, in addition to JAZ, is present in PPD, ZIM and ZIM-LIKE proteins, which are collectively grouped in the TIFY superfamily (Shikata et al., 2004; White, 2006; Vanholme et al., 2007). The importance of the ZIM domain in the evolution of land plants is underscored by the presence of this domain in several phylogenetically distant species of land plants, but not in green algae (Chico et al., 2008; Katsir et al., 2008a). However, in spite of its importance, the biochemical and physiological function of the ZIM domain remains unknown so far.

Here, we show that the JAZ proteins can form homo- and heteromeric interactions. The ZIM domain mediates these interactions in a hormone-independent manner. The ability of a few members of the JAZ family to homo- and heteromerize could reflect their relevance in JA signalling regulation. Interestingly, the truncated JAZ3ΔJas protein, retaining the ZIM domain and lacking the Jas motif, interacts with several JAZ proteins, even more than the full-length JAZ3 protein, providing new insights to explain the dominant JA-insensitive phenotype promoted by the JAZΔJas proteins. We also provide additional evidence that the Jas motif mediates the hormone-dependent Arabidopsis JAZ3–COI1 interaction, as well as the hormone-independent binding of JAZ3 to MYC2. Finally, we demonstrate that MYC2 interaction is a common feature of the JAZ family, as most JAZ proteins can bind MYC2.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The Jas motif is sufficient for the interaction of JAZ3 with both COI1 and MYC2

It has been previously shown that the hormone-dependent interaction of some JAZ proteins with COI1 involves the Jas motif (Katsir et al., 2008b; Melotto et al., 2008). To test if the Jas domain of Arabidopsis JAZ3 protein is responsible for the hormone-dependent interaction with COI1, recombinant JAZ3 and JAZ3 derivatives fused to MBP were used in pull-down (PD) experiments with cell extracts containing COI1-flag (from 35S:COI1-flag transgenic plants) (Figure S1). As shown in Figure 1a, full-length MBP–JAZ3 and MBP–JAZ3ΔZIM proteins, but not MBP–JAZ3ΔJas, were able to pull-down COI1-flag from the extracts in the presence of COR (a hormone agonist), indicating that the Jas motif is required for the hormone-dependent interaction between JAZ3 and COI1. Yeast two-hybrid analysis consistently confirmed these results (Figure 1b).

image

Figure 1.  The Jas motif of JAZ3 is necessary and sufficient to mediate the JAZ3–COI1 interaction in the presence of coronatine (COR). (a) Interaction of COI1-flag with MBP-JAZ3 and truncated versions. Immunoblot (anti-flag antibody) of recovered COI1-flag (from 35S:COI1-flag plant extracts) after pull-down reactions using recombinant MBP fusions of JAZ3, full length and truncated versions (JAZ3ΔJas and JAZ3ΔZIM), without COR (−), or in the presence of 50 μm COR (+). Recombinant MBP was used as the control. Crude protein extracts from 35S:COI1-flag transgenics (COI1-flag ext) are also shown. Lower panel shows the Coomassie staining of the input quantity of recombinant proteins used. (b) Yeast cells co-transformed with pGBKT7-COI1 (bait) and indicated pGADT7-JAZ derivatives (prey) were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2), as a transformation control, or on selective media (−4) (lacking Ade, His, Leu and Trp), to test protein interactions in the presence or absence of 30 μm COR. pGBKT7-COI1 co-transformation with pGADT7 vector (control) was included.

Download figure to PowerPoint

Interestingly, yeast two-hybrid assays have previously shown that JAZ3 interacts with MYC2, and that this interaction also involves the Jas motif of JAZ3 and the N-terminal half of MYC2 (Chini et al., 2007; Melotto et al., 2008). To further confirm this interaction, we first obtained transgenic plants expressing a GFP-tagged MYC2 protein (MYC2–GFP) in a myc2 mutant background (jin1-2). Microscope visualization of GFP confirmed that the fusion protein was nuclear and fully functional, complementing the JA insensitivity of jin1-2 (Figure 2a, b). We next performed PD experiments using Arabidopsis cell extracts from the 35S:MYC2–GFP transgenic plants and recombinant MBP–JAZ3 or truncated derivatives lacking either the ZIM domain (MBP–JAZ3ΔZIM) or the Jas motif (MBP–JAZ3ΔJas). As shown in Figure 2c, MYC2–GFP protein was pulled-down by MBP–JAZ3 or MBP–JAZ3ΔZIM, independently of the presence or absence of COR. In contrast, JAZ3ΔJas failed to recover MYC2–GFP protein, confirming that JAZ3 interaction with MYC2 occurs through the Jas motif of JAZ3. To further define a shorter domain in the JAZ proteins required for MYC2 interaction, we performed yeast two-hybrid assays with additional JAZ3 derivatives. As shown in Figure 2d, in addition to JAZ3 and JAZ3ΔZIM proteins (previously shown to interact with MYC2; Chini et al., 2007), the Jas motif alone (Jas) was sufficient for interaction with MYC2. Interestingly, however, this protein derivative (Jas) did not interact with COI1, suggesting that the interaction domain may be larger. Moreover, interaction with COI1 is hormone dependent, and the interaction with MYC2 is hormone independent. Therefore, the interaction of JAZ3 with both COI1 and MYC2 occurs through the same domain in JAZ3, the Jas motif, but via partially different regions, and with a different requirement of the hormone.

image

Figure 2.  The Jas motif of JAZ3 is necessary and sufficient to mediate the JAZ3–MYC2 interaction independently of the presence of the hormone. (a) Microscope visualization of nuclear-localized MYC2-GFP fusion protein in the roots of 35S:MYC2-GFP transgenic Arabidopsis plants. Scale bar: 0.1 mm. (b) Root-growth inhibition assays of 10-day-old Arabidopsis wild-type Col-0 seedlings, jin1-2 mutants and 35S:MYC2-GFP transgenic plants in a jin1-2 mutant background grown in Johnson’s media supplemented with 50 μm jasmonate (JA). The MYC2-GFP protein is fully functional as it complements the JA-insensitivity of jin1-2. Scale bars: 1 cm. (c) Interaction between MYC2-GFP and MBP-JAZ3 full-length or truncated protein versions. Immunoblot (anti-GFP antibody) of recovered MYC2-GFP (from 35S:MYC2-GFP;jin1-2 plant extracts) after pull-down reactions using recombinant MBP-JAZ3 protein and truncated versions, without COR (−) or in the presence of 50 μm COR (+). The asterisk indicates the MYC2-GFP protein. Recombinant MBP was used as the control. Crude protein extracts from the wild type (WT ext) and from 35S:MYC2-GFP;jin1-2 plant extracts (MYC2-GFP ext) are also shown for comparison. (d) Yeast cells co-transformed with pGBKT7-MYC2 (prey) and the indicated pGADT7-JAZ3 derivatives (bait) were selected and subsequently grown on yeast synthetic drop-out, lacking Leu and Trp (−2) as a transformation control, or on selective media (−4) (lacking Ade, His, Leu and Trp), to test protein interactions. All pGBKT7-JAZ3 derivatives (bait) co-transformed with pGADT7 empty vector (control) were included as controls.

Download figure to PowerPoint

Interactions of JAZ repressors with MYC2

In addition to JAZ3, JAZ1 and JAZ9 have recently been shown to interact with MYC2 in yeast two-hybrid assays (Melotto et al., 2008). In order to test whether interaction with MYC2 is a general feature of the JAZ repressors, we assess the capacity of MYC2 to interact with all members of the Arabidopsis JAZ family using PD and yeast two-hybrid experiments. As shown in Figure 3a, MYC2–GFP was recovered by all MBP–JAZ proteins, except for JAZ7, in PD experiments. The interaction was strong in the case of JAZ1, JAZ2, JAZ3, JAZ5, JAZ9, JAZ10 and JAZ11, and was weak in the case of JAZ4, JAZ6, JAZ8 and JAZ12. Consistently, most JAZ proteins, with the exceptions of JAZ4 and JAZ7, interacted with MYC2 in yeast two-hybrid assays, also with different affinities (Figure 3b). We conclude that most JAZ proteins can interact with MYC2.

image

Figure 3.  Jasmonate ZIM-domain (JAZ) repressors interact with MYC2. (a) Immunoblot (anti-GFP antibody) of recovered MYC2-GFP (from 35S:MYC2-GFP jin1-2 plant extracts) after pull-down reactions using the 12 MBP–JAZ recombinant proteins. Recombinant MBP was used as the control. Crude protein extracts from the wild type (WT ext) and 35S:MYC2-GFP; jin1-2 plant extracts (MYC2-GFP ext) are also included. The lower panel shows Coomassie staining of the input quantity of recombinant proteins. (b,c) Yeast cells co-transformed with pGBKT7-JAZ (bait) and pGADT7-MYC2 (prey) (b), or pDEST32–JAZ (bait) and pDEST22–MYC2 (prey) (c), were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2), as a transformation control, or on selective media (−4) (lacking Ade, His, Leu and Trp), to test protein interactions. pGADT7–MYC2 co-transformation with pGBKT7 vector (control) and pDEST22–MYC2 co-transformation with pDEST32 vector (control) were included.

Download figure to PowerPoint

The ZIM domain mediates homomeric interactions between JAZ proteins

As bHLH proteins act as dimmers (Ellenberger et al., 1994; Ma et al., 1994; Brownlie et al., 1997), we reasoned that an MYC2 repressor such as JAZ3 could also function as a dimmer. Supporting this hypothesis, yeast two-hybrid screening for the interactors of JAZ2 rendered a JAZ2 derivative missing the N-terminal 33 amino acids (data not shown). To further test this hypothesis we checked JAZ3 self-interaction by PD and yeast two-hybrid experiments. As shown in Figure 4a, PD assays using plant extracts expressing JAZ3–GFP and recombinant MBP–JAZ3 revealed that this protein can form homomers independently of the presence of the hormone (Figure 4a, b). To identify the domain responsible for dimerization, we used JAZ3 derivatives in PD experiments, defining the ZIM domain as the motif necessary and sufficient for homomeric interactions (Figure 4c). Yeast two-hybrid assays further confirmed these results (Figures 4d and S2), leading us to conclude that the ZIM domain in JAZ3 is responsible for its homomeric interaction.

image

Figure 4.  The ZIM domain is necessary and sufficient for homomerization. (a,b) Interaction between JAZ3–GFP and MBP–JAZ3 protein. Immunoblot (anti-GFP antibody) of recovered JAZ3–GFP (from 35S:JAZ3–GFP plant extracts) after pull-down (PD) reactions using recombinant MBP–JAZ3 and MBP proteins, in the absence (−) or presence (+) of 50 μm COR (b). Incubation of MBP–JAZ3 with wild-type extract (WT ext) was used as a control for spurious bands (a). Crude protein extracts from the wild type and 35S:JAZ3–GFP plant extracts (JAZ3–GFP ext) are also shown as controls. (c) Immunoblot (anti-GFP antibody) of recovered JAZ3ΔJas–GFP (from 35S:JAZ3–GFPΔJas plant extracts) after pull-down reactions using recombinant MBP–JAZ3, MBP–JAZ3 derivative proteins and MBP. Wild-type extracts (WT ext) were used in PD reactions as a control. Crude protein extracts from the wild type and 35S:JAZ3ΔJas–GFP plant extracts (JAZ3ΔJas–GFP ext) are also shown as controls. The lower panels in (a), (b) and (c) show the Coomassie staining of the input quantity of recombinant proteins. (d) Yeast cells co-transformed with pGBKT7–JAZ3 or derivatives (bait) and pGADT7–JAZ3 or derivatives (prey) were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2), as a transformation control (shown in Figure S2), or on selective media (−4) (lacking Ade, His, Leu and Trp), to assess the protein domains responsible for homomerization. The MYC2 interaction with JAZ3 derivatives was also tested. pGBKT7–JAZ3 constructs co-transformed with pGADT7 vector (control), and pGADT7–JAZ3 derivatives co-transformed with pGBKT7 vector (controls), were included.

Download figure to PowerPoint

We next tested homomeric interactions among all JAZ proteins in two-hybrid assays. As shown in Figure 5, the homomerization of full-length proteins was only observed for JAZ1, JAZ3, JAZ4 and JAZ9, suggesting that not all JAZ proteins can form homodimers. These assays were carried out simultaneously with those described in Figure 3b, where most of these constructs produced a positive interaction with MYC2, therefore supporting that all JAZ proteins are being properly expressed and are functional in yeast (also see Figure S3 for protein expression).

image

Figure 5.  Homomeric interactions among jasmonate ZIM-domain (JAZ) proteins. (a) Yeast cells co-transformed with pGBKT7–JAZ (bait) and pGADT7–JAZ proteins (prey) were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2), as a transformation control, or on selective media (−4) (lacking Ade, His, Leu and Trp), to test homomerization. (b) A 1:10 dilution of the same yeast cells shown in (a) were also tested in −2 and −4 media to assess the strength of homomerization. (c) pGBKT7–JAZ (bait) constructs co-transformed with pGADT7 empty vector were included to exclude JAZ binding domain (BD) auto-activation.

Download figure to PowerPoint

Heteromeric interactions among JAZ proteins

We also checked the possibility that JAZ proteins could form heteromers by testing all 144 possible combinations between the 12 JAZ repressors in yeast two-hybrid assays. As shown in Figure 6a, the same proteins that produced homomeric interactions (JAZ1, JAZ3, JAZ4 and JAZ9) also rendered heteromeric interactions. Thus, JAZ3, JAZ4 and JAZ9 produced the corresponding heteromers in reciprocal transformations (i.e. with the construct either as prey or as bait; Figures 4, S3 and S4). Interactions between JAZ1 and JAZ8 were also observed in reciprocal transformations, whereas interactions between JAZ1 and JAZ4 or JAZ9 were only observed with JAZ1 as the bait (Figures 6a and S5). All other combinations tested between JAZ proteins failed to establish heterodimers; however, these data in yeast do not preclude the scenario that additional JAZ proteins could heteromerize in vivo.

image

Figure 6.  Heteromeric interactions among jasmonate ZIM-domain (JAZ) proteins. (a) Yeast cells co-transformed with all 12 pGBKT7–JAZ proteins (bait) and all 12 pGADT7–JAZ proteins (prey) were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2), as a transformation control (shown in Figure S4), or on selective media (−4) (lacking Ade, His, Leu and Trp), to test heteromerizations. pGBKT7–JAZ (bait) constructs co-transformed with pGADT7 empty vector were included to exclude JAZ binding domain (BD) auto-activation. (b) Phylogenetic tree of the Arabidopsis JAZ proteins. Phenogram representation of the neighbour-joining for the 12 full length JAZ and the two PPD sequences. Sequence alignment was generated using DiAlign (Genomatix) and the tree was created by Phylodendron (University of Indiana). Branch lengths are proportional to the estimated evolutionary distance. Bootstrap values are included. JAZ proteins can be tentatively grouped in four clades.

Download figure to PowerPoint

To assess if the ZIM domain is also responsible for heteromerization, we tested the interaction between each JAZ protein and JAZ3ΔJas. Interestingly, in addition to the interactions promoted by the full-length JAZ3, the truncated derivative also rendered a strong interaction with JAZ1 (Figure 7a, b). These results indicate that heteromeric interactions also involve the ZIM domain. Although the interactions between JAZ1 and JAZ3ΔJas were only observed with JAZ1 as the bait, these data also suggest that the lack of the Jas motif could broaden the range of JAZ partners interacting with a particular JAZΔJas, compared with the corresponding full-length JAZ protein.

image

Figure 7.  The ZIM domain is necessary for heteromeric interactions. (a, b) Yeast cells co-transformed with all 12 pGBKT7–JAZ proteins (bait) and pGADT7–JAZ3 protein (prey) (a) or pGADT7–JAZ3ΔJas protein (prey) (b), were selected and subsequently grown on yeast synthetic drop-out lacking Leu and Trp (−2) as a transformation control, or on selective media (−4) (lacking Ade, His, Leu and Trp), to test protein interaction. (c) Interaction between JAZ3ΔJas-GFP and MBP–JAZ proteins. Immunoblot (anti-GFP antibody) of recovered JAZ3ΔJas–GFP (from 35S:JAZ3ΔJas–GFP plant extracts) after pull-down reactions using the 12 MBP–JAZ and MBP recombinant proteins. A pull-down with wild-type extracts (WT ext) was used as the control. Crude protein extracts from the wild type and 35S:JAZ3ΔJas–GFP plant extracts (JAZ3ΔJas–GFP ext) were also included. The lower panels show the Coomassie staining of the input quantity of recombinant proteins.

Download figure to PowerPoint

Additional evidence of heteromeric interactions, and of the involvement of the ZIM domain, was obtained by testing the interaction of all 12 MBP–JAZ fusion proteins with JAZ3ΔJas–GFP in PD experiments using cell extracts from JAZ3ΔJas–GFP transgenic plants (Figure 7c). Consistent with the yeast two-hybrid results, JAZ1, JAZ3, JAZ4 and JAZ9 were the proteins that most strongly pulled-down JAZ3ΔJas–GFP from the extracts. JAZ2, JAZ5, JAZ6 and JAZ10 only showed a weak interaction with JAZ3ΔJas–GFP, and no interaction over the background was observed for JAZ7, JAZ8, JAZ11 and JAZ12. These results further support the existence of heteromeric interactions among JAZ proteins involving the ZIM domain.

Homo- and heteromeric interactions in vivo

To demonstrate that interactions among JAZ proteins can also occur in vivo, we checked homo- and heteromeric interactions by bimolecular fluorescence complementation (BiFC), and by co-immunoprecipitation (co-IP) assays. As shown in Figure 8a, fusion proteins of JAZ3ΔJas to the N-terminal and C-terminal halves of YFP (JAZ3ΔJas-NY and CY-JAZ3ΔJas, respectively) complemented the YFP, and produced fluorescent nuclei at the YFP wavelength (also see Figure S6). Moreover, JAZ3ΔJas-GFP and JAZ9-GFP, transiently expressed proteins in Nicotiana benthamiana, could be detected after co-IP experiments of JAZ9-HA using the α-HA matrix (Figure 8b). These results indicate that both homomeric and heteromeric interactions can occur in planta, and further support that ZIM mediates dimerization between JAZ proteins.

image

Figure 8.  Homo- and heteromeric interactions in vivo. (a) Bimolecular fluorescence complementation by JAZ3ΔJas homodimerization. Confocal images of the YFP detected following Agrobacterium-mediated transient expression in Nicotiana benthamiana of JAZ3ΔJas-NY and CY-JAZ3ΔJas (upper panels), compared with the negative control (lower panels). Nuclei were stained by 4′,6-diamidino-2-phenylindole (DAPI), and the YFP/DAPI merged images highlight the YFP fluorescence localized in the nuclei. Scale bars: 50 μm. Abbreviations: NY, N-terminal half of YFP; CY, C-terminal half of YFP. (b) Immunoblot of inmunoprecipitated proteins (JAZ9–HA and HA; upper panel), co-immunoprecipipated proteins (JAZ3ΔJas–GFP, JAZ9–GFP and GFP; medium panel) transiently expressed in N. benthamiana. Immunoprecipitation was performed using anti-HA matrix, and co-immunoprecipitated proteins were detected using anti-GFP antibody. The expression levels of the input GFP-fused proteins were assessed by anti-GFP of crude extracts (lower panel).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

JAZ repressors share two conserved domains, Jas and ZIM, the functions of which are not yet fully understood (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007). Previous results suggest that both domains could be protein–protein interaction surfaces that mediate association with their targets (i.e. MYC2) and/or regulators (i.e. SCFCOI1). Thus, hormone-dependent interaction of some JAZ proteins with COI1 involves the Jas motif (Katsir et al., 2008b; Melotto et al., 2008). Moreover, two positively charged amino acids within this motif are required for this interaction (Melotto et al., 2008). Our demonstration that the Jas motif in Arabidopsis JAZ3 is required and sufficient for the hormone-dependent interaction with COI1 further supports these observations. Moreover, it is fully consistent with the stabilization of the JAZ3ΔJas–GFP protein against proteasome degradation (Chini et al., 2007), as the lack of the Jas motif would prevent interaction with COI1, and therefore, the COI1-mediated ubiquitination required for proteasome degradation upon hormone perception.

Paradoxically, the Jas motif is also required and sufficient for JAZ3 interaction with MYC2 (Chini et al., 2007; Melotto et al., 2008; present work). Moreover, this seems to be a general feature of the JAZ family, as we found that most JAZ proteins interact with MYC2 with different affinities in PD or yeast two-hybrid assays. This interaction, however, does not require the two positively charged residues involved in JAZ–COI1 binding (Melotto et al., 2008). Moreover, only the Jas motif (Jas in Figure 2) is sufficient for an interaction with MYC2, but not with COI1. These observations suggest that the interacting regions of the Jas motif with COI1 and MYC2 are not exactly the same, or that the amino acids within this motif establishing the connection with COI1 and MYC2 are different. Moreover, JAZ–MYC2 binding is independent on the hormone. There is, therefore, a growing body of evidence indicating that the Jas motif is responsible for the interaction with both COI1 and MYC2, probably through different surfaces or amino acid contacts. Whether the presence of the hormone determines the switch of JAZ protein interaction from MYC2 to COI1 requires further experimental support.

Concerning the ZIM function, in spite of its annotation in public databases as a putative novel DNA-binding domain (ZIM, Pfam PF06200; TIFY, PROSITE profile PS51320), the different types of experiments aimed to test this hypothesis failed to provide evidence of ZIM–DNA interaction (not shown). In contrast, however, yeast two-hybrid screening for partners of JAZ2 rendered a JAZ2-derivative lacking the N-terminal 33 amino acids, but containing both the ZIM and Jas domains. Additional PD and yeast two-hybrid experiments using truncated derivatives of JAZ3 reduced the region necessary and sufficient for JAZ3 homomeric interactions to the ZIM domain. Additional JAZ proteins can also form homodimers, suggesting that self-assembling interactions could be a common feature of ZIM-containing proteins (PPDs and ZIM/ZIM-LIKE as well as JAZ proteins). Moreover, these assays also demonstrated that heteromeric interactions among JAZ repressors are possible, and also involve the ZIM domain. Thus, using cell extracts from transgenic plants expressing JAZ3ΔJas–GFP, this protein could be pulled-down by most JAZ proteins. In addition to JAZ3 itself, the interaction was very strong for JAZ1, JAZ4 and JAZ9.

Consistently, the strong interactors/partners in PD assays were coincident with the proteins homomerizing in yeast two-hybrid experiments, suggesting that these interactions may occur in intact cells. BiFC and co-IP experiments further demonstrated that homo- and heteromeric interactions can occur in vivo, and that ZIM is the domain involved in dimerization. It is noteworthy that most JAZ proteins that are capable of establishing homomeric interactions belong to clade A (i.e. JAZ3, JAZ4 and JAZ9; Figure 6b), thereby evidencing the importance of amino acid sequence conservation for homomerization. However, these results do not preclude additional protein interactions in vivo or in vitro, and therefore the determination of the real homo- and heteromeric complexes occurring in vivo will require further experimental work.

An important question still unanswered is how the JAZΔJas proteins lacking the Jas motif, required for interaction with both MYC2 and COI1, can cause JA insensitivity. Two alternative hypotheses have been proposed so far. The first hypothesis attributes the dominant JA insensitivity to the continuous repression of JA-related transcription factors by stabilized JAZΔJas proteins (Thines et al., 2007; Melotto et al., 2008). Although this hypothesis is consistent with the requirement of the Jas motif for COI1 interaction, it is unlikely, as this domain is also required for JAZ interaction with MYC2. The second hypothesis proposed that JAZ3ΔJas protein could interfere with COI1 activity, slowing the turnover of JAZs, and, therefore, causing the JA insensitivity (Chini et al., 2007). This hypothesis is based on the stabilization of full-length JAZ proteins in the presence of JAZ3ΔJas in vivo and in vitro, and on a hormone-independent COI1–JAZ3ΔJas interaction detected in vitro (Chini et al., 2007). However, this hormone-independent interaction could not be detected in the PD experiments or yeast two-hybrid assays described within this work, therefore indicating that the interaction must be very weak and is unlikely to interfere with COI1 activity. Moreover, in light of the results reported here, the stabilization of JAZ proteins in the presence of JAZ3ΔJas could be reinterpreted as a consequence of their homo- and heteromeric interactions. Remarkably, JAZ3ΔJas can form heterodimers with JAZ1 and JAZ9, the two proteins whose degradation was previously reported to be inhibited by JAZ3ΔJas (Chini et al., 2007). In light of the results reported here, we can therefore reconcile both hypotheses proposing that heteromeric JAZ–JAZΔJas complexes could interact with MYC2 (and probably other JA-related TFs), but are inefficiently recognized by COI1. In this case, the JAZ3ΔJas dominant effect would be indirectly established by dimerization, instead of by direct COI1 interaction.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant materials and growth conditions

Arabidopsis thaliana Col-0 is the genetic background of wild-type and transgenic lines used throughout the work. Plant growth conditions in vitro (in Johnson’s media) were as previously described (Lorenzo et al., 2003).

The generation of transgenic lines overexpressing JAZ3–GFP and JAZ3ΔJas–GFP are described in Chini et al. (2007). To generate transgenic plants expressing MYC2–GFP in the jin1-2 background the full-length MYC2 coding sequence was amplified with high-fidelity Taq-polimerase (Roche, http://www.roche.com), using Gateway-compatible primers (see Table S1 for primers sequences). PCR products were cloned into pDONR207 with a Gateway BP II kit (Invitrogen, http://www.invitrogen.com) and sequenced. A Gateway LR II kit (Invitrogen) and the destination vector pGWB5 (Mita et al., 1995) was used to generate the 35S:MYC2GFP construct, which was then transferred to Agrobacterium tumefaciens C58C1 carrying the pGV2260 by heat shock, and Arabidopsis jin1-2 plants were then transformed by floral dipping (Clough and Bent, 1998). GFP was visualized by fluorescence microscopy using DMR and confocal microscopes (Leica, http://www.leica.com). Photographs of cells expressing the GFP were taken as previously described (Lorenzo et al., 2004).

Transgenic plants expressing COI1-flag in Col-0 (kindly provided by the laboratory of X.-W. Deng; Feng et al., 2003) and wild-type plants were grown in the same conditions, collected after 12 days, immediately frozen, and were then used for pull-down assays.

Recombinant proteins

Coding sequences for full-length JAZ proteins and for truncated versions of JAZ3 were PCR amplified from plasmid templates provided by TAIR (Table S2) or cDNA, maintaining both the frame and the stop codons (see Table S1 for primer sequences). Using the Gateway system (Invitrogen), these amplicons were cloned into pDONR201/pDONR207 and recombined in pDEST-TH1 (Hammarstrom et al., 2002) to obtain N-terminal MBP-fusions. All constructs were verified by sequencing prior to protein expression. Recombinant MBP-fusion proteins were expressed in Escherichia coli BL21 cells, and were then purified in amylose resin columns (New England Biolabs, http://www.neb.com), following the method described by Chini et al. (2007). Protein purity was assessed by Coomassie gel staining, and quantification was performed in gels by comparison with known concentrations of BSA.

Protein extracts and pull-down assays

Arabidopsis thaliana wild-type seedlings, lines expressing 35S:COI1-flag (Feng et al., 2003), 35S:JAI3-GFP or 35S:JAI3ΔJas-GFP (Chini et al., 2007), and jin1-2 plants expressing 35S:MYC2-GFP were ground in liquid nitrogen and homogenized in extraction buffer containing 50 mm TrisHCl, pH 7.4, 100 mm NaCl, 10% glycerol, 0.1% Tween-20, 1 mm DTT, 1 mm phenylmethylsulphonyl fluoride (PMSF), complete protease inhibitor cocktail (Roche) and 50 μm MG132 (Sigma-Aldrich, http://www.sigmaaldrich.com). After two rounds of 15 min of centrifugation at 16 000 g, at 4°C, the supernatant was collected and the total protein quantified by the Bradford method.

For pull-down experiments, 6 μg of resin-bound MBP fusion protein was added to 1.8 mg of total protein extract, and, when indicated, supplemented with coronatine (COR) (Sigma-Aldrich), and then incubated for 1 h at 4°C under rotation. Samples were resuspended in 30 μL of extraction buffer and supplemented with 2 mm CaCl2 and 1 μL factor Xa (New England Biolabs), to digest MBP-fused proteins (3 h at room temperature (23°C), RT), so as to facilitate COI1-flag detection. Samples were boiled with loading buffer and run on 8% SDS–PAGE gels. Proteins were transferred onto polyvinylidene fluoride (PVDF) membranes and incubated with monoclonal anti-FLAG antibody (Sigma-Aldrich), or were transferred to nitrocellulose membranes and incubated with anti-GFP monoclonal antibody (Roche).

A 3-μl aliquot was taken from each sample to check the quantity of MBP-fused protein used in each PD sample. These samples were loaded into SDS–PAGE gels and stained with Coomassie.

Co-immunoprecipitation

JAZ3ΔJas and JAZ9 proteins were fused to a GFP tag by cloning in the pGWB5 vector, and JAZ7 and JAZ9 were fused to an HA tag by cloning in pGWB14 (Mita et al., 1995). Empty pGWB vectors were used for the expression of GFP, and HA proteins were used as the controls. N. benthamiana leaves were infiltrated with Agrobacterium harbouring these constructs, and were collected after 2 days. For each sample, 0.6 g of agroinfiltrated leaves were homogenized in 2 mL of co-IP buffer containing 50 mm Tris–HCl, pH 7.5, 100 mm NaCl, 2 mm DTT, 0.1% Tween-20, 1 mm PMSF, 50 μm MG132 and complete protease inhibitor cocktail (Roche), and were centrifuged twice at 16 000 g at 4°C. The supernatant was incubated for 2 h (4°C, with rotation) with the anti-HA affinity matrix (Roche), and was washed three times with 1 mL of IP buffer. After denaturalization in Laemmli SDS–PAGE loading buffer, samples were loaded into 10 or 12% SDS–PAGE gels, transferred to PVDF membranes (Millipore, http://www.millipore.com) and incubated with anti-HA-HR-peroxidase (Roche) and anti-GFP-HRP antibodies (Miltenyi Biotec, http://www.miltenyibiotec.com). A 15-μL aliquot of total protein extract was also used for immunoblot with the same antibodies to evaluate the expression of recombinant proteins in each sample.

Bimolecular fluorescence complementation

Truncated versions of JAZ3 were PCR amplified, with or without the stop codon, as previously described (Chini et al., 2007). Using the Gateway system (Invitrogen), these amplicons were cloned into pDONR201 and subsequently to the destination vector series pBiCF (To et al., 2006). The four resulting plasmids (pBiFC1–JAZ3ΔJas = JAZ3ΔJas–NYFP; pBiFC2–JAZ3ΔJas = NYFP–JAZ3ΔJas; pBiFC3–JAZ3ΔJas = CYFP–JAZ3ΔJas; pBiFC4–JAZ3ΔJas = JAZ3ΔJas–CYFP) were transformed in A. tumefaciens. The four possible combinations of CYFP and NYFP JAZ3ΔJas constructs were employed to infect N. benthamiana plants, as previously described (Chini et al., 2007). Leaves of infected plants were analysed under a TCS SP5 Leica Microsystems confocal laser microscope. 4′,6-Diamidino-2-phenylindole (DAPI) solution was injected into plant leaves 2 min before imaging. DAPI and YFP fluorescence were analysed at the same time via excitation at a wavelength of 504 and 514 nm, respectively, and emission scanning at a wavelength between 418 and 479 nm (DAPI) and between 520 and 615 nm (YFP).

Yeast two-hybrid assays

The pDONR constructs described above were used in Gateway LR (Invitrogen) reactions, in combination with the destination vector pGADT7gateway (Gal4 activation domain, AD) and pGBKT7gateway (Gal4 DNA binding domain, BD), in which the gateway cassette was cloned within the EcoRI and NdeI restriction sites, to generate all vectors described in the text. All these constructs were checked by sequencing. All JAZ cDNA and MYC2 in pDNOR were also cloned in the destination low-copy yeast expression vector pDEST22 (Gal4 AD) and pDEST32 (Gal4 BD), and were then checked by sequencing.

To assess protein interactions, the corresponding plasmids were co-transformed into Saccharomyces cerevisiae AH 109 cells following standard heat-shock protocols (Chini et al., 2007). Successfully transformed colonies were identified on yeast synthetic drop-out lacking Leu and Trp. At 3 days after transformation, yeast colonies were grown in selective lacking Leu and Trp (–2) liquid media for 6 or 7 h, and the cell density was adjusted to 3 × 107 cells mL−1 (OD600 = 1). A 4-μl sample of the cell suspensions was plated out on yeast synthetic drop-out lacking Ade, His, Leu and Trp to test protein interaction (supplemented with 30 μm COR, where indicated). Plates were incubated at 28°C for 2–4 days. The empty vector pGADT7gateway was also co-transformed with pGBKT7 constructs as a negative control.

The expression of recombinant proteins in yeast was assessed by western blot. Total yeast proteins were purified by growing yeast colonies in 10 mL of selective –2 liquid media until reaching a cell density of 2 × 107 cells mL−1 (OD600 = 0.7). Yeast cells were disrupted by FastPrep extraction (at a speed of 5.5 meters for 30 sec) in extraction buffer containing 125 mm Tris–HCl, pH 6.8, 4% SDS, 22% glycerol, 0.005% bromophenol blue, 140 mmβ-mercaptoethanol and complete protease inhibitor cocktail (Roche). Samples were boiled and run on 10% SDS–PAGE gels. Proteins were transferred onto PVDF membranes and incubated with anti-HA-peroxidase (Roche), following the manufacturer’s instructions, or anti-MYC-peroxidase (Boehringer Mannheim, now part of Roche) antibody to test the protein expressions of constructs expressed by the pGADT7 and pGBKT7 plasmid, respectively.

The β-galactosidase activity in yeast was measured with liquid cultures, as described by the manufacturer (Clontech, http://www.clontech.com), using o-nitrophenyl-β-d-galactopyranoside (ONPG) as a substrate.

Phylogenetic tree

Sequence analyses and phylogenetic trees were carried out as previously described (Chini and Loake, 2005). Briefly, alignments of protein sequences were generated using DiAlign (Genomatix, http://www.genomatix.de) and ClustalW (EBI, http://www.ebi.ac.uk). The phenogram representation of the neighbour-joining tree of the JAZ family was created by ClustalW [1.75 (http://sci.cnb.uam.es/Services/MolBio/clustalw)].

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank X.-W. Deng, who kindly provided seeds of the transgenic plants expressing COI1-flag in the Col-0 background. The plasmids reported in Table S2 were obtained from the Arabidopsis Biological Resource Centre (ABRC, http://www.biosci.ohio-state.edu/~plantbio/Facilities/abrc/abrchome.htm). Vectors pGBKT7gateway and pGADT7gateway were developed by M. Boter. We are grateful to Romel Gonzalez for technical assistance, and to Sylvia Gutiérrez for assistance with the confocal imaging.

This work was financed by grants to RS from the Spanish Ministerio de Ciencia y Tecnología (BIO2004-02502, BIO2007-66935, GEN2003-20218-C02-02 and CSD2007-00057-B) and from the Comunidad de Madrid (GR/SAL/0674/2004).

AC was supported by a ‘Juan de la Cierva’ fellowship from the Spanish Ministerio de Educación y Ciencia and an ‘EMBO long-term’ fellowship. SF has been supported by postdoctoral fellowships from the Portuguese Foundation for Science and Technology (BPD/21045/2004) and from the Spanish Ministerio de Educación y Ciencia (JAEDoc015). JMC was supported by a postdoctoral contract from the Spanish Ministerio de Educación y Ciencia (CSD2007-00057-B). PFC was supported by a predoctoral fellowship from Comunidad de Madrid (FPI-CAM).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Figure S1. JAZ3 construct derivates. Schematic representation of the wild-type JAZ3 and JAZ3 construct derivates employed in PD and Y2H experiments. Primer sequences employed to generate these constructs are reported in Supplementary Table 1.

Figure S2. Positive control of yeast two-hybrid co-transformation shown in Figure 4d. Yeast cells co-transformed with pGBKT7-JAZ3 derivates (bait) and indicated pGADT7-JAZ3 derivates (prey) were selected and subsequently grown on yeast synthetic dropout lacking Leu and Trp (−2) as transformation controls. Protein interaction was assessed in selective – 4 media as reported in Figure 4d.

Figure S3. Positive control of yeast two-hybrid co-transformation, shown in Figure 6a Yeast cells co-transformed with pGBKT7-JAZ (bait) and pGADT7-JAZ (prey) were selected and subsequently grown on yeast synthetic dropout lacking Leu and Trp (−2) as transformation controls. Protein interaction was assessed in selective – 4 media as reported in Figure 6a.

Figure S4. Expression of recombinant yeast proteins. Total protein extracts from yeast cells were used for immunoblot with anti-HA (a) and anti-MYC (b) antibodies to evaluate expression levels of recombinant proteins expressed by pGADT7 (prey) and pGBKT7 (bait), respectively.

Figure S5. Quantification of β-galactosidase activity in selected yeast strains testing heteromeric jasmonate ZIM-domain (JAZ) interaction. Quantification of β-galactosidase activity in yeast strains cotransformed with JAZ1/JAZ4 or JAZ1/JAZ9. Data show the mean of three replicates normalized to backgroung level of each bait construct.

Figure S6.In planta homomeric interaction of JAZ3&Dgr;Jas. (a) YFP confocal image following transient expression of JAZ3&Dgr;Jas-NY and CY-JAZ3&Dgr;Jas. (b) Scan of the fluorescence emission spectrum (from 520 to 610 nm) of the positive nucleus was carried out to confirm the emission specificity. The resulting fluorescence emission spectrum shows a sharp peak at the specific YFP emission wavelength.

Table S1. Oligonucleotides used as described in Experimental procedures.

Table S2. Plasmid templates used for amplification of jasmonate ZIM-domain (JAZ) cDNAs.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TPJ_3852_sm_figs.pdf240KSupporting info item
TPJ_3852_sm_tables.pdf24KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.