Alternative splicing expands the repertoire of dominant JAZ repressors of jasmonate signaling


  • Hoo Sun Chung,

    1. Department of Energy-Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA
    2. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA
    Search for more papers by this author
  • Thomas F. Cooke,

    1. Department of Energy-Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA
    2. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA
    Search for more papers by this author
  • Cody L. DePew,

    1. Department of Energy-Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA
    2. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA
    Search for more papers by this author
  • Lalita C. Patel,

    1. Department of Energy-Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA
    Search for more papers by this author
  • Narihito Ogawa,

    1. Department of Biological Engineering, Tokyo Institute of Technology, 4259-B52 Nagatsuta-cho, Midori-ku, Yokohama 226-8501, Japan 4259
    Search for more papers by this author
  • Yuichi Kobayashi,

    1. Department of Biological Engineering, Tokyo Institute of Technology, 4259-B52 Nagatsuta-cho, Midori-ku, Yokohama 226-8501, Japan 4259
    Search for more papers by this author
  • Gregg A. Howe

    Corresponding author
    1. Department of Energy-Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA
    2. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA
    Search for more papers by this author

For correspondence (fax +1517 353 9168; e-mail


Jasmonates (JAs) are fatty acid-derived signaling compounds that control diverse aspects of plant growth, development and immunity. The F-box protein COI1 functions both as a receptor for jasmonoyl-l-isoleucine (JA-Ile) and as the component of an E3-ubiquitin ligase complex (SCFCOI1) that targets JAZ transcriptional regulators for degradation. A key feature of JAZ proteins is the C-terminal Jas motif that mediates the JA-Ile-dependent interaction with COI1. Here, we show that most JAZ genes from evolutionarily diverse plants contain a conserved intron that splits the Jas motif into 20 N-terminal and seven C-terminal (X5PY) amino acid submotifs. In most members of the Arabidopsis JAZ family, alternative splicing events involving retention of this intron generate proteins that are truncated before the X5PY sequence. In vitro pull-down and yeast two-hybrid assays indicate that these splice variants have reduced capacity to form stable complexes with COI1 in the presence of the bioactive stereoisomer of the hormone (3R,7S)-JA-Ile. cDNA overexpression studies showed that some, but not all, truncated splice variants are dominant repressors of JA signaling. We also show that strong constitutive expression of an intron-containing JAZ10 genomic clone is sufficient to repress JA responses. These findings provide evidence for functional differences between JAZ isoforms, and establish a direct link between the alternative splicing of JAZ pre-mRNA and the dominant repression of JA signal output. We propose that production of dominant JAZ repressors by alternative splicing reduces the negative consequences associated with inappropriate or hyperactivation of the JA response pathway.


Jasmonic acid (JA) and its derivatives (collectively referred to as JAs) play a central role in controlling growth, reproduction and defense-related processes in higher plants (Devoto and Turner, 2003; Wasternack, 2007; Howe and Jander, 2008; Browse, 2009). JAs exert their effects by activating large-scale changes in gene expression that depend in part on the transcription factor MYC2 (Boter et al., 2004; Lorenzo et al., 2004; Dombrecht et al., 2007). In resting cells containing low levels of JA, JASMONATE ZIM-domain (JAZ) proteins negatively regulate gene expression by binding to MYC2 and other MYC2-related transcription factors (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007; Melotto et al., 2008; Pauwels et al., 2010). JAZs belong to the TIFY family of proteins that is named for a highly conserved TIFYXG sequence motif located within the ZIM/TIFY domain (Vanholme et al., 2007). This domain functions to recruit transcriptional co-repressors (Pauwels et al., 2010), and also mediates homo- and heteromeric interactions between JAZ proteins (Chini et al., 2009; Chung and Howe, 2009). In response to cues that trigger JA synthesis, JAZ proteins are degraded by the ubiquitin–proteasome pathway, thereby releasing MYC2 from repression. The F-box protein CORONATINE INSENSITIVE 1 (COI1) functions both as a receptor for an amino acid-conjugated form of JA, jasmonoyl-l-isoleucine (JA-Ile), and as an adaptor protein that confers specificity of the SKP1-CUL1-F-box (SCF)-type ubiquitin ligase (SCFCOI1) for JAZ substrates (Xie et al., 1998; Katsir et al., 2008b; Yan et al., 2009). Genetic and biochemical evidence is consistent with a model in which COI1-mediated perception of JA-Ile leads to the formation of COI1/JAZ/JA-Ile complexes, in which JAZs are ubiquitinated and subsequently degraded (Chini et al., 2007; Thines et al., 2007; Chico et al., 2008; Katsir et al., 2008a; Staswick, 2008; Saracco et al., 2009; Wasternack and Kombrink, 2010).

JAZ proteins contain a conserved approximately 27-amino-acid sequence, referred to as the Jas motif, that promotes JA-Ile-dependent interaction with COI1 (reviewed by Chung et al., 2009). Artificially truncated JAZs lacking this sequence are resistant to hormone-induced degradation, and strongly repress JA responses in a dominant fashion (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007; Chung et al., 2008; Shoji et al., 2008). The Jas motif also promotes JAZ interaction with MYC2 in a hormone-independent manner (Chini et al., 2007, 2009; Melotto et al., 2008), and has also been implicated as a nuclear localization signal (NLS) (Grunewald et al., 2009). Basic amino acid residues near the N-terminal end of the Jas motif are required for JAZ interaction with COI1 but not MYC2 (Melotto et al., 2008), suggesting that sequence determinants for JAZ binding to these two proteins are not identical. Further elucidation of the various functions of the Jas motif is central to understanding the molecular mechanism of JA signaling.

The physiological importance of JA as both a growth inhibitor and a positive regulator of stress responses suggests a broader role for the hormone in controlling resource allocation between growth and defense-related processes (Baldwin, 1998; Yan et al., 2007; Zhang and Turner, 2008). Optimization of these antagonistic responses is thought to involve positive and negative feedback loops within the core JA signaling pathway. The rapid activation of JAZ gene expression in JA-stimulated cells (Thines et al., 2007; Chung et al., 2008; Koo et al., 2009) suggests that JA responses may be curtailed through de novo synthesis of JAZ repressors. Unlike most JAZs characterized to date, including JAZ1, JAZ3 and JAZ10, this type of negative feedback loop should, in theory, involve JAZ repressors that are relatively stable in the presence of JA-Ile. Recent studies indicate that alternative splice variants of JAZ10, which are more resistant to JA-induced degradation, may play a role in this form of negative feedback control (Chung and Howe, 2009). Ectopic expression of JAZ10 isoforms with a truncated Jas motif (JAZ10.3) or lacking the entire motif (JAZ10.4) confers dominant insensitivity to JA (Yan et al., 2007; Chung and Howe, 2009). A physiological role for truncated JAZ10 splice variants in attenuating hormone responses in wild-type plants is supported by the finding that plants silenced for JAZ10 expression are hypersensitive to wounding and JA treatment (Yan et al., 2007). Whether this phenomenon is specific for JAZ10, or whether other JAZ splice variants constitute a general post-transcriptional mechanism for fine-tuning the signaling pathway, is currently unknown.

Here, we report that the Jas motif coding region in most JAZ genes from diverse plant species is highly conserved, and includes a homologous intron that divides the motif into segments of approximately 20 N-terminal and seven C-terminal (X5PY) amino acids. We show that: (i) most JAZ genes in Arabidopsis are subject to alternative splicing events, in which retention of this intron generates truncated proteins lacking the X5PY and other C-terminal amino acids; (ii) these truncated JAZs retain the ability to interact with MYC2, but have a reduced capacity to form complexes with COI1 in comparison to their respective full-length isoforms; (iii) overexpression of cDNAs encoding some but not all truncated splice variants results in dominant repression of JA responses; and (iv) high-level constitutive expression of the intron-containing JAZ10 gene is sufficient to generate alternatively spliced JAZ10 transcripts and strong phenotypic repression of JA responses. These findings suggest that alternative splicing of JAZ genes provides a general mechanism to reduce the fitness costs associated with over-stimulation of the signaling pathway.


Alternative splicing of a conserved intron generates transcripts encoding truncated JAZ proteins

We observed that the intron/exon organization of the Jas motif-coding region is conserved in most Arabidopsis JAZ genes (Figure 1a). In nine of the 12 family members, the 27-amino-acid motif is split into 20 N-terminal and seven C-terminal (X5PY) amino acid segments by an intron (the Jas intron) located in phase 2 of the codon specifying the Arg residue (R20) at position 20 (Figure 1b). The 3′-most exon in all nine of these genes encodes the X5PY sequence and additional C-terminal residues of the protein. Interestingly, in seven of the Jas intron-containing genes, the entire coding capacity of the upstream exon (the Jas exon) specifies approximately 20 amino acids that comprise the N-terminal portion of the motif (Figure 1a,b).

Figure 1.

 Bipartite organization of the Jas domain-coding region in Arabidopsis JAZ genes.
(a) Left, phylogenetic tree constructed from the amino acid sequence of 12 full-length JAZs, showing four (A–D) subclades of proteins. Right, intron/exon organization of the corresponding genes. Thick green and blue bars indicate coding regions and non-coding untranslated regions in exons, respectively. The Jas motif coding region is depicted in red. Introns are depicted by a thin black horizontal line or, in the case of the Jas intron, a red line.
(b) Consensus sequence of the Jas motif in 12 Arabidopsis JAZs. The arrow indicates the location of homologous introns. The Jas intron is positioned invariably in phase 2 of the codon specifying the Arg at position 20 of the motif.
(c) Sequence of the Jas exon/intron junction in nine Arabidopsis JAZ genes that contain the intron. Exon and intron sequences are highlighted in black and gray, respectively, together with the predicted amino acid sequence. In the case of JAZ2, JAZ3, JAZ4, JAZ6, JAZ10 and JAZ11, an in-frame stop codon (red) is encountered close to the 5′ end of the Jas intron, if the intron is retained.

Retention of the 5′ splice site of the Jas intron during the pre-mRNA processing of JAZ2, JAZ3, JAZ4, JAZ6, JAZ10 and JAZ11 is predicted to generate a premature stop codon (PTC) immediately or shortly (within six codons) after R20 (Figure 1c). The Arabidopsis Information Resource (TAIR) database contains expressed sequence tag (EST)-supported gene models for JAZ4 (At1g48500.2) and JAZ10 (At5g13220.2 and At5g13220.3) transcripts that are spliced in this manner. Reverse transcriptase-mediated PCR (RT-PCR) performed with RNA from seedlings that had been treated with methyl-JA (MeJA) to induce JAZ expression identified fully spliced (i.e. lacking the Jas intron) mRNAs for all nine genes that contain the Jas intron (Figure 2). RT-PCR experiments performed with the same RNA and gene-specific primers that anneal within the Jas intron amplified alternatively spliced forms of JAZ2, JAZ3, JAZ4, JAZ6, JAZ9, JAZ10 and JAZ12 mRNA (Figure 2). Sequencing of these PCR products showed that the transcripts contain the 5′ end of the Jas intron, but not other intron sequences, thereby excluding the possibility of genomic DNA contamination in the PCR reactions. Thus, most JAZ genes (except JAZ1, JAZ7 and JAZ8) expressed in JA-treated Arabidopsis seedlings produce alternatively spliced mRNAs predicted to encode truncated proteins that lack the X5PY sequence and additional C-terminal amino acids. We henceforth refer to these splice variants as ΔPY JAZs.

Figure 2.

 RT-PCR-based detection of Jas intron retention in various JAZ transcripts.
(a) Primer design strategy showing JAZ2 as an example. Structural elements of the gene include the Jas motif coding region (black box) and Jas intron (black line). Light gray and dark gray bars indicate coding regions and non-coding untranslated regions in exons, respectively. Arrows indicate forward (FP) and reverse (RP) primers used for RT-PCR assays designed to amplify transcripts in which the Jas intron is spliced out or retained. The RP1 primer spans the exon–exon junction, and thus hybridizes to transcripts in which the Jas intron is removed. The RP2 primer hybridizes to a gene-specific sequence near the 5′ end of the Jas intron, thereby amplifying transcripts in which the 5′ splice site of the intron is retained. The primer sequences are listed in Table S2.
(b) RNA extracted from methyl-jasmonate (MeJA)-treated Arabidopsis seedlings was used as a template for RT-PCR (see Experimental procedures). Gene-specific FP-RP1 primer sets amplified fully-spliced transcripts (upper panel), whereas FP-RP2 primer sets amplified transcripts in which the 5′ end of the Jas intron was retained (lower panel). PCR-amplified products were separated by agarose gel electrophoresis and visualized by staining with ethidium bromide. For those RT-PCR assays yielding more than one main product, asterisks denote the transcript containing the Jas intron but not the other introns, as determined by DNA sequencing of cloned products.

To gain additional insight into the functional significance of the Jas intron, we sought to determine whether JAZ genes from phylogenetically diverse plant species contain this sequence. Analysis of select genome sequence databases (see Experimental procedures) showed that most (approximately 76%) JAZ-related genes in rice (Oryza sativa), grapevine (Vitis vinifera), poplar (Populus trichocarpa), Brachypodium distachyon, moss (Physcomitrella patens) and the lycophyte Selaginella moellendorffii contain a homologous intron at this position within the gene (i.e. phase 2 of the R20 codon; Tables 1 and S1). Three EST-supported gene models (Bradi3g10820.2, Bradi4g31240.2 and Bradi5g24410.4; Table S1) for transcripts encoding ΔPY JAZs in B. distachyon provide evidence that alternative splicing events involving the Jas intron are conserved between monocots and dicots. As is the case for Arabidopsis JAZ2, JAZ6, JAZ10 and JAZ11 (Figure 1c), hypothetical retention of the 5′ donor splice site of the Jas intron in 59% of the non-Arabidopsis JAZ-related genes generates a PTC immediately after the R20 codon in which the intron resides (Tables 1 and S1). The high proportion of predicted transcripts with a PTC at this position appears to be a consequence of the compositional bias of 5′ splice sites, which in plants have a consensus sequence AG/GTAA (stop codon in bold) (Reddy, 2007).

Table 1.   The Jas intron is highly conserved in JAZ-related genes from diverse plant species
SpeciesNo. JAZ genesaNo. JAZ genes with the Jas intronArg20-PTCb
  1. aThe number of JAZ genes in Arabidopsis and rice (Oryza sativa) or, for all other species, the number of group-II TIFY genes (which includes JAZ and PPD genes) identified in genome databases.

  2. bThe number of genes in which a premature stop codon (PTC) occurs immediately after the codon in which the Jas intron resides. This codon typically (but not invariably) codes for Arg (Table S1).

Physcomitrella patens973
Selaginella moellendorffii997
Arabidopsis thaliana1294
Brachypodium distachyon1174
Oryza sativa1595
Populus trichocarpa14126
Vitis vinifera975

ΔPY JAZ proteins differentially associate with COI1

We used in vitro pull-down (PD) assays to study the hormone-dependent interaction of Arabidopsis ΔPY isoforms with COI1. This assay was first tested with splice variants of JAZ10, which were previously shown to differentially associate with COI1 in yeast cells grown in the presence of coronatine, a potent agonist of the JA-Ile receptor (Katsir et al., 2008b; Chung and Howe, 2009; Yan et al., 2009). A biologically active form of JA-Ile, (3R,7S)-JA-Ile (Fonseca et al., 2009), strongly stimulated the ability of recombinant JAZ10.1-His to recover epitope-tagged COI1 (AtCOI1-Myc) from crude leaf extracts (Figure 3a). In comparison with the full-length JAZ10.1 isoform, the ΔPY splice variant JAZ10.3 recovered only minor quantities of COI1-Myc in the presence of the highest concentration of JA-Ile tested (500 nm). (3R,7S)-JA-Ile did not stimulate recovery of COI1 by JAZ10.4, which lacks the entire Jas motif as a consequence of an alternative splicing event within the third exon (Chung and Howe, 2009).

Figure 3.

 ΔPY JAZ isoforms differentially associate with COI1 in the presence of receptor ligands.
(a) Hormone-dependent interaction of JAZ10 splice variants with COI1. Pull-down assays were performed using recombinant JAZ10 splice variants and crude leaf extracts from 35S-COI1-Myc leaves as a source of COI1. JAZ10.4 lacks the entire Jas motif. Reaction mixtures were supplemented with the indicated concentration of (3R,7S)-JA-Ile, or an equivalent volume of assay buffer (0). Protein bound to JAZ10-His was separated by SDS-PAGE and analyzed by immunoblotting with anti-Myc antibody for the presence of COI1-Myc. As a loading control, the immunoblotted membrane was stained with Coomassie blue to detect JAZ-His.
(b) Effect of alternative splicing of the Jas intron on the sequence of the Jas motif in full-length (JAZ10.1, JAZ2.1 and JAZ3.1) and ΔPY (JAZ10.3, JAZ2.2 and JAZ3.4) isoforms of JAZ10, JAZ2 and JAZ3.
(c, d) Hormone-dependent interaction of full-length (JAZ2.1 and JAZ3.1) and ΔPY (JAZ2.2 and JAZ3.4) splice variants with COI1. Experiments were performed as described in panel (a).
(e) JA-Ile-dependent interaction of JAZ2 (upper) and JAZ3 (lower) splice variants with COI1 in the yeast two-hybrid (Y2H) system. Yeast strains expressing both COI1 (as a DNA binding domain fusion) and the indicated JAZ splice variant (as an activation domain fusion) were plated on media containing X-gal and the indicated concentration (μm) of (3R,7S)-JA-Ile. LacZ-mediated blue color formation is indicative of the strength of the COI1–JAZ interaction.
(f) Coronatine-dependent interaction of JAZ2 (upper) and JAZ3 (lower) splice variants with COI1 in the Y2H system. Experiments were performed as described in panel (e), except that coronatine (COR) was included at the indicated concentration (μm) in the yeast medium.

Genes belonging to three (A, B and C) of the four JAZ phylogenetic clades in Arabidopsis contain the Jas intron (Figure 1a). To determine whether the weak COI1 interaction observed with JAZ10.3 (clade C) is a general property of ΔPY splice variants, we performed PD assays with representative members of clades A (JAZ3) and B (JAZ2). In accordance with established gene models in TAIR, we refer to proteins produced from transcripts encoding full-length JAZ2 and JAZ3 as JAZ2.1 and JAZ3.1, respectively, and the corresponding ΔPY isoforms as JAZ2.2 and JAZ3.4 (Figure 3b). The capacity of JAZ2.1-His and JAZ2.2-His to recover COI1 from leaf extracts in the presence of (3R,7S)-JA-Ile was very similar to that of JAZ10.1 and JAZ10.3, respectively. For example, the quantity of COI1 bound by JAZ2.1-His in the presence of 10 nm JA-Ile was comparable with that bound by JAZ2.2 in the presence of 500 nm JA-Ile (Figure 3c). (3R,7S)-JA-Ile stimulated a strong interaction of JAZ3.1-His with COI1, as previously reported (Melotto et al., 2008; Fonseca et al., 2009), and also acted in a dose-dependent manner to promote COI1 binding to JAZ3.4-His (Figure 3d). The quantity of COI1-Myc bound to JAZ3.4 at a given hormone concentration, however, was reproducibly less than that recovered by JAZ3.1-His.

We used the yeast two-hybrid (Y2H) system as a second approach to assess the ability of ΔPY splice variants to interact with COI1. Dose-dependent COI1-JAZ2.1 and COI1-JAZ3.1 interactions were observed in yeast cells grown in the presence of increasing concentrations of (3R,7S)-JA-Ile (Figure 3e). These interactions were not observed in yeast cells grown in the presence of 100 μm of the trans stereoisomer (3R,7R)-JA-Ile (Figure S1), which is a largely inactive form of the hormone (Fonseca et al., 2009). Remarkably, bioactive (3R,7S)-JA-Ile failed to promote a detectable interaction of COI1 with JAZ2.2 and JAZ3.4 in yeast (Figure 3e). Coronatine strongly stimulated COI1 binding to JAZ2.1 and JAZ3.1 in yeast, and was significantly more active in these assays than our synthetic preparation of (3R,7S)-JA-Ile (Figure 3f). COI1 interacted weakly with JAZ2.2 in the presence of high concentrations of coronatine (100 μm). COI1 interacted more strongly with JAZ3.1 than with JAZ3.4 in the presence of low concentrations of coronatine (e.g. 5–10 μm), whereas discrimination between these two splice variants by COI1 was not observed at higher concentrations (e.g. 100 μm) of coronatine (Figure 3f). Western blot analysis showed that all JAZ proteins were expressed in yeast (Figure S2). These findings show that ΔPY splice variants of JAZ2 and JAZ3, like JAZ10, have a reduced capacity to form complexes with COI1 in the presence of the receptor-active form JA-Ile and the potent agonist coronatine.

We hypothesized that the reduced ability of ΔPY JAZs to associate with COI1 may reflect a role for the highly conserved PY sequence in promoting COI1–JAZ interactions. To test this idea, we substituted the PY sequence in JAZ2.1 and JAZ10.1 with Ala residues, and tested the resulting proteins in PD assays for their ability to bind COI1. The results showed that the PY → AA mutation did not diminish the ability of JAZ2.1-His and JAZ10.1-His to recover COI1-Myc from crude leaf of extracts in the presence of JA-Ile or coronatine (Figure 4), indicating that the PY motif is not required for these COI1–JAZ interactions.

Figure 4.

 The PY submotif of JAZ2.1 and JAZ10.1 is not required for COI1 interaction.
(a, b) Site-directed mutagenesis was used to change the PY diamino acid sequence in the Jas motif of JAZ2.1 (amino acid position 227–228) and JAZ10.1 (amino acid position 191–192) to AA. The resulting wild-type (WT) and mutant (PY → AA) proteins were expressed in Escherichia coli as MBP-His fusions. Pull-down reactions containing purified JAZs and protein extract from leaves of the 35S-AtCOI1-Myc transgenic line were supplemented (+) with 2.5 μm (3R,7S)-JA-Ile (a) or 0.1 μm coronatine (b). Control reactions contained an equivalent quantity of binding buffer (–). Reactions were incubated and processed as described in the legend for Figure 3a. The Coomassie blue stained gel shows the input of MBP-JAZ-His fusion protein.

Attenuation of JA signaling by JAZ splice variants

The reduced capacity of JA-Ile to stimulate COI1 interaction with JAZ2.2 and JAZ3.4 led us to hypothesize that these splice variants may act to repress JA responses in vivo. To address this question, we overexpressed cDNAs encoding full-length and ΔPY variants of JAZ2 and JAZ3 in Arabidopsis, and tested the resulting transgenic lines for altered responsiveness to exogenous methyl-JA (MeJA) using a root growth-inhibition assay. The overexpression of cDNAs encoding JAZ2.1, JAZ3.1 and JAZ3.4 did not obviously affect the MeJA-induced inhibition of root growth in progeny of 10 independent lines tested for each construct (data not shown). In the case of JAZ2.2, however, seven of 15 independent T1 lines expressing the 35S-JAZ2.2 transgene produced progeny that were significantly less sensitive than wild-type seedlings to MeJA (Figure 5). Root-length assays showed that the sensitivity of 35S-JAZ2.2 homozygous seedlings to MeJA is similar to, or slightly less than, that of 35S-JAZ10.3 seedlings, which overexpress the ΔPY variant of JAZ10 (Figure S3). 35S-JAZ2.2 homozygous lines did not display obvious defects in fertility (data not shown). Y2H assays showed that JAZ2.2, like the full-length JAZ2.1 protein (Chini et al., 2009), retains the ability to interact with MYC2 (Figure 5c). These results suggest that JAZ2.2 may exert its dominant effect through the inhibition of MYC2 or related transcription factors.

Figure 5.

 Ectopic expression of JAZ2.2 attenuates jasmonate responses.
(a) Photograph of wild-type (WT), 35S-JAZ2.1 and 35S-JAZ2.2 seedlings grown for 9 days on MS medium containing 50 μm methyl jasmonate (MeJA). Seedlings from two independent lines for each construct are shown.
(b) Quantification of MeJA-induced root growth inhibition of seedlings shown in (a). Data show means ± SDs (n = 15 seedlings per genotype).
(c) JAZ2.1 and JAZ2.2 interact with MYC2 in yeast. Yeast strains co-transformed with plasmids encoding MYC2 (as an activation domain fusion; AD) and the indicated JAZ2 splice variant (as a DNA binding domain fusion; BD) were plated on media containing X-gal. LacZ-mediated blue color formation is indicative of the JAZ–MYC2 interaction. Yeast strains expressing BD-JAZ and an empty AD vector (AD-EV) control did not exhibit visible blue color formation.

Previous functional analysis of JAZ splice variants relied on the overexpression of individual cDNAs encoding truncated JAZ isoforms (Chung and Howe, 2009; Yan et al., 2007; Figure 5). To further address the functional significance of JAZ alternative splicing in JA signaling, we transformed Arabidopsis with a full-length genomic JAZ10 sequence under the control of the CaMV 35S promoter. Fifteen of 20 independent T1 lines expressing the 35S-JAZ10G transgene produced progeny that were strongly insensitive to root growth inhibition by MeJA (Figure 6a). Root length measurements performed with 9-day-old seedlings showed that the strength of the JA-insensitive phenotype in three representative T2 lines was greater than that of 35S-JAZ10.3 seedlings, and was comparable with 35S-JAZ10.4 seedlings that express the COI1-non-interacting JAZ10.4 isoform (Figure 6b). In contrast to the male-sterile phenotype of 35S-JAZ10.4 plants (Chung and Howe, 2009), reproductive defects were not observed in any of the 50 independent 35S-JAZ10G lines tested, including homozygous lines (data not shown). RT-PCR experiments confirmed that JAZ10.1, JAZ10.3 and JAZ10.4 transcripts derived from the transgene are expressed in 35S-JAZ10G seedlings (Figure 6c). These findings establish a causal link between the strong expression of JAZ10, the alternative splicing of JAZ10 pre-mRNA and the functional repression of JA responses.

Figure 6.

 Overexpression of a JAZ10 genomic clone attenuates jasmonate signaling.
(a) Photograph of wild-type (WT) and transgenic seedlings of the indicated genotype grown for 9 days on MS medium containing 50 μm methyl jasmonate (MeJA). 10.1, 10.3 and 10.4 refer to transgenic lines overexpressing cDNAs for JAZ10.1, JAZ10.3 and JAZ10.4, respectively (Chung and Howe, 2009), which were grown together with three independent 35S-JAZ10G lines (10G-2, 10G-4 and 10G-5).
(b) Quantification of MeJA-induced root growth inhibition of seedlings shown in (a). Data show means ± SDs of at least 11 seedlings per genotype.
(c) Semi-quantitative RT-PCR analysis of JAZ10 transcripts in 2-week-old 35S-JAZ10G seedlings (line 2) grown on MS medium. The expression of transcripts derived from the 35S-JAZ10G transgene was assessed using primer sets (Table S2) that amplify all three JAZ10 transcripts (top panel), JAZ10.3 only (second panel) or JAZ10.4 only (third panel). PCR reactions contained a forward primer that hybridizes specifically to the 35S-JAZ10G transgene, and were limited to 22 cycles. The experiment was repeated with RNA isolated from two independent sets of 35S-JAZ10G-2 plants (P3 and P4). A primer set that amplifies an actin transcript (lower panel) was used as a control.


A large proportion of intron-containing plant genes are subject to alternative splicing (Campbell et al., 2006; Wang and Brendel, 2006; Reddy, 2007; Barbazuk et al., 2008; Filichkin et al., 2010). Examples of evolutionarily conserved splicing events that have functional significance in plant biology, however, are scarce (Ner-Gaon et al., 2007; Filichkin et al., 2010). Here, we analyzed splicing events affecting the JAZ family of proteins that, together with COI1 and MYC2, comprise the core JA signaling pathway in vegetative tissues. Focus was placed on the analysis of splicing events that alter the protein’s C-terminal Jas motif, which mediates interaction with COI1, MYC2 and possibly JA-Ile as well. We show that the intron/exon organization of the Jas motif-coding region in most JAZ genes consists of two exons separated by an intron (the Jas intron), which is found in most JAZ-related genes from diverse plant species. The presence of the Jas intron in JAZ orthologs from P. patens indicates that the sequence arose early in the evolution of the JAZ gene family, and has been retained since P. patens diverged from higher plants approximately 400 Myr (Rensing et al., 2008). Evolutionary conservation of this gene architecture implies an important biological function for the Jas intron; our results provide evidence that alternative splicing events affecting this intron expand the functional repertoire of JAZ proteins to modulate JA signaling. These findings provide a counter example to the view that alternative splicing has only a limited role in expanding the diversity of plant proteomes (Severing et al., 2009).

The approximately 20 N-terminal amino acids of the Jas motif in many JAZ proteins are encoded by a dedicated exon, which we refer to as the Jas exon. The modular and conserved nature of this coding sequence implies an important role in JAZ function. It is intriguing that the amino acid sequence specified by the Jas exon resembles the N-terminal portion of the CCT (CO, CO-like, TOC1) motif found in proteins that control plant circadian rhythm and other responses to environmental cues (Chung et al., 2009). This sequence similarity accounts for the annotation of the Jas motif as a CCT-like domain in Pfam ( and InterPro ( databases. An evolutionary link between CO/COL/TOC1 and JAZ proteins is supported by the fact that ZIM and ZIM-like members of the TIFY family contain a CCT domain.

Alternative splicing events involving the retention of the 5′ splice site of the Jas intron generate transcripts in which a PTC effectively removes the X5PY sequence and other amino acids encoded by the 3′ exon. PTC-containing (PTC+) transcripts may be targeted for destruction by the nonsense-mediated mRNA decay (NMD) pathway (Reddy, 2007). Among the features that trigger transcripts for NMD are PTCs located >50–55 nucleotides upstream of an exon–exon junction and abnormally long 3′ untranslated regions (Chang et al., 2007; Hori and Watanabe, 2007). Although additional work is needed to determine the role of the NMD pathway in regulating JAZ mRNA accumulation, RT-PCR experiments showed that several PTC+JAZ transcripts are expressed in Arabidopsis. At least some of these transcripts produce JAZ isoforms that, when ectopically expressed, repress JA responses (Chung and Howe, 2009; Yan et al., 2007; this study). Given the widespread occurrence of the Jas intron in the plant kingdom and the high frequency of intron retention in plants (Ner-Gaon et al., 2004; Wang and Brendel, 2006; Filichkin et al., 2010), it seems likely that JAZ genes in other plants are spliced in a similar manner. Evidence for cross-species conservation ΔPY splice variants comes from three sequence-supported JAZ gene models in B. distachyon (Table S1).

The JA-resistant phenotype of 35S-JAZ10G plants establishes a direct link between the alternative splicing of JAZ pre-mRNA and the dominant attenuation of the JA signal output. High-level constitutive expression of JAZ10 from the 35S promoter is sufficient to generate a pool of pre-mRNA that is subject to alternative splicing, and the production of the dominant JAZ10.3 and JAZ10.4 isoforms. In wild-type plants, the transcription of JAZ genes is rapidly and strongly induced in response to cues that trigger the synthesis of bioactive JAs (Thines et al., 2007; Yan et al., 2007; Chung et al., 2008; Koo et al., 2009). Alternative splicing of the resulting JAZ pre-mRNA, together with the enhanced stability of truncated splice variants, could provide a mechanism to accumulate dominant JAZ repressors during the early phase of JA-mediated stress responses. This model is consistent with studies showing the coordinate accumulation of JAZ10.1 and JAZ10.3 transcripts in wounded Arabidopsis leaves (Yan et al., 2007). It is also possible that alternative splicing of JAZ pre-mRNA is controlled in a cell- or tissue-type specific manner by trans-acting factors of the spliceosome machinery. The fact that roots of 35S-JAZ10.4 and 35S-JAZ10G plants are both strongly insensitive to JA-induced growth inhibition, whereas only 35S-JAZ10.4 plants are male sterile, raises the possibility that JAZ10.4 is not normally produced in male reproductive tissues. It is also possible that JAZ10.4 transcript levels are higher in 35S-JAZ10.4 plants than in 35S-JAZ10G plants.

Our results indicate that repression of JA responses by ΔPY JAZs results in part from a decreased interaction with COI1, and, as a consequence, an increased stability of JAZ in the presence of JA-Ile. This finding implies a role for the X5PY sequence or additional amino acids at the protein C terminus in promoting COI1-JAZ complex formation. Site-directed mutagenesis experiments excluded the possibility that the highly conserved PY is required for this aspect of JAZ2 and JAZ10 function. We also found that coronatine promotes the robust binding of COI1 to JAZ3.4, which lacks the PY sequence, in PD and Y2H assays. Ectopic expression of ΔPY isoforms of JAZ10 (JAZ10.3) and JAZ2 (JAZ2.2) conferred an obvious JA-insensitive phenotype, whereas overexpression of JAZ3.4 did not, indicating that some but not all ΔPY isoforms exert dominant effects. However, because the ligand-dependent interaction of COI1 with JAZ3.1 was stronger than that with JAZ3.4, it is possible that JAZ3.4 affects JA responses in a manner that was not detected in our analysis.

In considering the potential functional differences between the dominant JAZ2.2/JAZ10.3 proteins and JAZ3.4, it is worth noting that the latter isoform contains five amino acids C-terminal to the KRKD/ER sequence, whereas JAZ2.2/JAZ10.3 and most other predicted ΔPY splice variants are truncated immediately after KRKD/ER (Table 1). We speculate that the C-terminal extension of JAZ3.4 may contribute indirectly to the formation of COI1-JAZ complexes by stabilizing the structure of the KRKD/ER region, which is predicted to have an α-helical character (Chung et al., 2009). The most straightforward interpretation of our results, together with other structure–function studies (Melotto et al., 2008), is that the primary sequence determinant for COI1 binding resides in the N-terminal region of the Jas motif (i.e. residues encoded by the Jas exon), and that the C-terminal end of the Jas motif (i.e. X5PY) contributes indirectly to ligand binding by stabilizing the COI1-ligand-JAZ ternary complex.

The X5PY sequence may also function as part of an NLS for targeting to or within the nucleus (Grunewald et al., 2009). A similar sequence motif (RX2-5PY) is part of an NLS in yeast and human proteins (Lee et al., 2006). Ectopically expressed JAZ proteins lacking the X5PY submotif clearly retain the ability to modulate JA signaling, presumably by acting in the nucleus (Chini et al., 2007; Chung and Howe, 2009; Thines et al., 2007; Yan et al., 2007; this study). Altered localization of ΔPY JAZs within the nucleus may reduce their accessibility to COI1. Alternatively, the manner in which ΔPY JAZs bind MYC2 or other proteins may hinder their interaction with COI1.

Our results are consistent with the view that alternative splicing plays a role in optimizing plant adaptation to stress (Reddy, 2007; Barbazuk et al., 2008; Filichkin et al., 2010). In contrast to our understanding of how stress-regulated genes are activated in plants, relatively little is known about how these responses are restrained once they are initiated (Kazan, 2006). In wounded Arabidopsis leaves, downregulation of response genes during a time period when JA-Ile levels remain high suggests that stressed tissues become desensitized to the hormone (Koo and Howe, 2009; Koo et al., 2009). Epimerization of receptor-active (3R,7S)-JA-Ile to the inactive (3R,7R)-JA-Ile isomer was proposed as a mechanism to inactivate signaling (Fonseca et al., 2009). Measurement of JA-Ile stereoisomers in wounded tomato leaves, however, indicates that endogenous pools of (3R,7S)-JA-Ile are not readily epimerized in planta (Suza et al., 2010).

Our data support an alternative hypothesis in which stable ΔPY JAZ isoforms act to restrain the transcription of JA-response genes in the presence of (3R,7S)-JA-Ile. Dominant isoforms such as JAZ2.2 and JAZ10.3 retain the ability to interact with MYC2, presumably through the truncated Jas motif, whereas JAZ10.4 appears to bind MYC2 through a region outside the Jas motif (Chung and Howe, 2009). These endogenous repressors contain an intact ZIM/TIFY domain, and thus may actively repress gene expression through the NINJA/TOPLESS transcriptional co-repressor complex (Pauwels et al., 2010). Fitness costs resulting from inappropriate or hyperactivation of JA responses (Staswick et al., 1992; Baldwin, 1998; Yan et al., 2007; Zhang and Turner, 2008; Moreno et al., 2009) may have provided selective pressure to evolve dominant JAZ isoforms through alternative splicing events within the Jas intron. Other endogenous JAZ repressors (e.g. JAZ10.4) are generated by alternative splicing events outside the Jas intron (Chung and Howe, 2009). Thus, multiple mechanisms may contribute to expanding the functional repertoire of JAZ proteins to fine-tune JA responses.

Experimental procedures

Biological materials

Growth conditions for Arabidopsis thaliana (ecotype Col-0), Agrobacterium tumefaciens-mediated transformation, and root growth inhibition assays were conducted as previously described (Chung and Howe, 2009).

Identification of JAZ gene sequences

Rice JAZ genes were previously described (Ye et al., 2009). Genome sequences for poplar (Tuskan et al., 2006), Brachypodium (International Brachypodium Initiative, 2010), Physcomitrella (Rensing et al., 2008), Selaginella ( and grapevine (Jaillon et al., 2007) were searched with BLASTP and TBLASTN for matches to Arabidopsis JAZ1. Retrieved sequences were used to re-query the database for that species. Following the removal of redundant sequences, protein sequences were manually annotated for the presence of the ZIM/TIFY and Jas motifs. No attempt was made to discriminate between genes encoding JAZ and PEAPOD (PPD) proteins, both of which contain ZIM/TIFY and Jas-like motifs. The intron/exon structure of all identified genes was assessed with the genome browser in the respective database, as well as with the ClustalW-based alignment of the genomic DNA sequence with the predicted coding sequence of the gene. The conserved intron/exon organization of the Jas motif coding region was verified with software that predicts common introns within orthologous genes ( (Wilkerson et al., 2009), and with comparative genome analysis tools available at The consensus sequence for the Jas motif in Arabidopsis JAZs was determined with the web-based application WebLogo (Crooks et al., 2004).

Molecular biology procedures

RT-PCR was used to identify Arabidopsis JAZ transcripts that either contain or do not contain the Jas intron. RT-PCR reactions were performed with RNA prepared from 12-day-old wild-type seedlings treated with 100 μm MeJA for 2 h (Chung et al., 2008), Taq DNA polymerase (Invitrogen,, and the transcript-specific primer sets listed in Table S2. PCR products were cloned into pGEM-T Easy (Promega, for sequencing with T7 and SP6 primers. A PCR product containing the full-length JAZ10 genomic sequence was amplified from Arabidopsis genomic DNA using Pfu Turbo DNA polymerase (Stratagene, now part of Agilent, and the primers listed in Table S2. The resulting approximately 1.9-kb product was subcloned into pGEM-T Easy (Promega), followed by recloning into the BamHI site of a modified pBI121 binary vector (Schilmiller et al., 2007). 35S-JAZ10G transgenic plants harboring this construct were generated as described above. PCR-based site-directed mutagenesis of the PY submotif in JAZ2.1 and JAZ10.1 was performed as previously described (Chung and Howe, 2009). Primers for these reactions are listed in Table S2.

COI1–JAZ interaction assays

Appropriate restriction sites were added to JAZ cDNAs (Chung and Howe, 2009) by PCR amplification with the primer sets listed in Table S2. PCR products were cloned into the corresponding site of pRMG-nMAL to produce plasmids encoding MBP-JAZ-His6 fusion proteins (referred to as JAZ-His). JAZ fusion proteins were expressed in Escherichia coli and purified by Ni-affinity chromatography, as described previously (Thines et al., 2007; Katsir et al., 2008b), with the following minor modifications. The lysis buffer consisted of 50 mm Na-phosphate, pH 7.8, 500 mm NaCl, 0.1% Tween-20, 0.1 mm phenylmethyl sulfonyl fluoride and 15 mm imidazole. The wash buffer was identical to the lysis buffer, but contained 25 mm imidazole. Leaf extract from a transgenic Arabidopsis line expressing AtCOI1-9xMyc (Melotto et al., 2008) was used as the source of COI1. Soluble protein was extracted from rosette leaves (ground in liquid nitrogen) in binding buffer (50 mm Na-phosphate, pH 7.7, 100 mm NaCl, 10% glycerol, 0.1% Tween-20, 25 mm imidazole, 20 mmβ-mercaptoethanol and Roche Complete Mini protease inhibitor tablet-EDTA free) and clarified by centrifugation at 15 000 g at 4°C for 10 min. Each PD assay contained 1 mg of leaf protein and 25 μg of recombinant JAZ–His in a total volume of 200 μl. Reactions were incubated for 30 min at 4°C in the absence or presence of JA-Ile or coronatine. Following the addition of 60 μl Ni-NTA Resin (Qiagen,, the reaction was incubated for an additional 30 min at 4°C. Ni-NTA resin was recovered by centrifugation on a spin column (Bio-Rad, and washed three times with 250 μl of binding buffer. The affinity resin was eluted with 30 μl of a solution containing 350 mm imidazole. The eluted protein was separated by SDS-PAGE on a 10% gel, transferred to polyvinylidene fluoride (PVDF) membrane, and probed with an anti-c-Myc antibody (Roche, For Y2H analysis, JAZ cDNAs were amplified using the primer sets listed in Table S2, and were subsequently cloned into pB42AD. Y2H assays were performed as described previously (Chung and Howe, 2009). Stereoisomers of JA-Ile were chemically synthesized as described by Ogawa and Kobayashi (2008). Coronatine was purchased from Sigma-Aldrich (


We are grateful to Sheng Yang He (Michigan State University) for providing the Arabidopsis line expressing AtCOI1-9xMyc and Todd Mockler (Oregon State University) for sharing unpublished data on alternatively spliced JAZ transcripts in Arabidopsis and Brachypodium. We thank Abe Koo for the mass spectrometry analysis of the JA-Ile isomers, and Laura Sheard and Ning Zheng (University of Washington) for helpful discussions during the course of the research. This work was supported by the National Institutes of Health (grant R01GM57795) and the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, US Department of Energy (grant DE–FG02–91ER20021).