Citrus MAF1, a repressor of RNA polymerase III, binds the Xanthomonas citri canker elicitor PthA4 and suppresses citrus canker development

Abstract

Transcription activator-like (TAL) effectors from Xanthomonas species pathogens act as transcription factors in plant cells; however, how TAL effectors activate host transcription is unknown. We found previously that TAL effectors of the citrus canker pathogen Xanthomonas citri, known as PthAs, bind the carboxyl-terminal domain of the sweet orange (Citrus sinensis) RNA polymerase II (Pol II) and inhibit the activity of CsCYP, a cyclophilin associated with the carboxyl-terminal domain of the citrus RNA Pol II that functions as a negative regulator of cell growth. Here, we show that PthA4 specifically interacted with the sweet orange MAF1 (CsMAF1) protein, an RNA polymerase III (Pol III) repressor that controls ribosome biogenesis and cell growth in yeast (Saccharomyces cerevisiae) and human. CsMAF1 bound the human RNA Pol III and rescued the yeast maf1 mutant by repressing tRNA(His) transcription. The expression of PthA4 in the maf1 mutant slightly restored tRNA(His) synthesis, indicating that PthA4 counteracts CsMAF1 activity. In addition, we show that sweet orange RNA interference plants with reduced CsMAF1 levels displayed a dramatic increase in tRNA transcription and a marked phenotype of cell proliferation during canker formation. Conversely, CsMAF1 overexpression was detrimental to seedling growth, inhibited tRNA synthesis, and attenuated canker development. Furthermore, we found that PthA4 is required to elicit cankers in sweet orange leaves and that depletion of CsMAF1 in X. citri-infected tissues correlates with the development of hyperplastic lesions and the presence of PthA4. Considering that CsMAF1 and CsCYP function as canker suppressors in sweet orange, our data indicate that TAL effectors from X. citri target negative regulators of RNA Pol II and Pol III to coordinately increase the transcription of host genes involved in ribosome biogenesis and cell proliferation.

Figure 1.

Specific interaction between CsMAF1 and PthA4. A, Yeast two-hybrid assays showing that CsMAF1 specifically interacts with PthA4. Full-length CsMAF1 fused to the yeast GAL4-AD domain or the control plasmid (GAL4AD) was moved into yeast cells carrying one of the four PthA variants (PthA1–PthA4) fused to the GAL4-BD domain (GAL4BD-PthA constructs). The PthA proteins, which lack the first 128 residues and carry the entire repetitive DNA-binding domain (RD), invariable LRR region (LRR), C terminus with the nuclear localization signals (NLS), as well as the acidic activation domain (AAD), are schematically represented. B and C, No interaction was observed between CsMAF1 and RD4 (B), whereas weak interactions were detected with the LRR and the C-terminal domain plus 5.5 repeat units (5.5R) of the DNA-binding domain (C). D, Western-blot detection of the eluted fraction of a GST-pulldown assay using the purified 6xHis-PthA4 as prey and immobilized GST or GST-CsMAF1 as bait, confirming the interaction between PthA4 and CsMAF1. Arrowheads indicate bands corresponding to the GST-CsMAF1 fusion protein (approximately 52 kD) and GST alone (approximately 26 kD) detected by the GST antiserum. PthA4 (approximately 122 kD) was detected by the anti-PthA serum. E, Coimmunoprecipitation (Co-IP) assay of PthA4 and CsMAF1. Cell extracts of citrus epicotyls transiently expressing PthA4 (input) were incubated with the anti-CsMAF1 or preimmune serum (control) and protein A-Sepharose. After washing the beads, bound proteins were resolved on a 10% SDS-polyacrylamide gel and probed with the indicated antibodies. PthA4 was detected in the coimmunoprecipitation performed with the anti-CsMAF1 but not with the preimmune serum.

Figure 2.

CsMAF1 is a homolog of the human and yeast MAF1 proteins. A, Protein sequence alignment of CsMAF1 with the human (HsMAF1) and yeast (ScMAF1) proteins performed by ClustalW. Identical and similar residues are shaded in black and gray, respectively. The two MAF1 signature sequences, which are conserved in CsMAF1, and the nuclear localization signals (NLS) are indicated. Asterisks indicate Thr and Ser residues that are phosphorylated in the human and/or yeast MAF1 proteins. The sequence comprising amino acid residues 60 to 202 of the yeast protein, as well as some of the C-terminal residues of all three proteins, were omitted to improve the alignment. B, Immunoblot detection of recombinant CsMAF1 (RP) and CsMAF1 from cell extracts of sweet orange leaves (Citrus). Arrowheads indicate two major forms of CsMAF1 in citrus cell extracts. C, SDS-PAGE showing that CsMAF1 phosphorylated by PKA (+) migrates slower than nonphosphorylated (−) CsMAF1 (top panel); the bottom panel shows the radioactive detection of the phosphorylation reaction with [³²P]ATP. Arrows indicate the CsMAF1 protein, whereas M indicates the molecular weight marker. D, Phylogenetic analysis of MAF1 proteins from plants (monocots and dicots), animals, yeast, and fungi, showing that CsMAF1 clusters with uncharacterized MAFs from dicot plants. E, GST-pulldown assay showing that GST-CsMAF1, but not GST, binds to the human RNA Pol III from lysed Hek 293 cells. Human RNA Pol III was detected by the monoclonal antibody raised against the human RNA Pol III 39-kD subunit, whereas CsMAF1 was detected by an anti-GST serum (arrowheads).

Figure 3.

CsMAF1 inhibits tRNA^His synthesis in the yeast maf1 mutant. A, The sweet orange MAF1 gene was expressed in the yeast maf1 mutant as a GST fusion (approximately 52 kD) under the control of a copper-induced promoter. After induction with copper sulfate (Cu⁺²), CsMAF1 was purified and detected with the anti-CsMAF1 serum. B, Northern analysis of tRNA^His in the wild-type yeast (WT), the maf1 deletion mutant (maf1⁻), and maf1⁻ cells complemented with CsMAF1 grown in the presence or absence of copper sulfate (top panel). The expression of CsMAF1 in the maf1 mutant significantly reduces tRNA^His synthesis, which is elevated in the mutant, relative to the wild-type cells. However, upon copper treatment, the tRNA^His levels in the mutant are substantially lower, indicating that CsMAF1 complements the function of the yeast MAF1 protein. The 25S rRNA was used as a loading control (bottom panel). C, Northern analysis showing that CsMAF1 represses tRNA^His transcription in the maf1 mutant, whereas its coexpression with PthA4 slightly restores tRNA^His synthesis, suggesting that PthA4 inhibits CsMAF1 activity (top panel). The 25S rRNA was used as a loading control (bottom panel).

Figure 4.

Silencing of CsMAF1 enhances tRNA synthesis and canker development. A, Transgenic plants carrying the CsMAF1 hairpin construct were analyzed by western blot using the anti-CsMAF1 serum (anti-CsMAF1). The CsMAF1 RNAi plants (4, 5, and 6) showed reduced levels of CsMAF1 protein relative to an untransformed control (C). Protein loads were controlled by Coomassie Brilliant Blue staining of the bands corresponding to the large subunit of ribulose-1,5-bisphosphate carboxylase (RBCL). B, Plot of the relative band area of the corresponding top (slow-migrating) and bottom (fast-migrating) bands depicted in A. Measurements of the band areas (in pixels) were performed using ImageJ software. C, Quantitative reverse transcription-PCR analysis of the relative abundance of tRNA^His, tRNA^Leu, and tRNA^Thr in the CsMAF1 RNAi lines (R4, R5, and R6) and in a control plant (C). The silenced lines showed a significant increase in tRNA expression relative to the control. Asterisks indicate that the differences between the control and transgenic means are statistically significant at P = 0.05. D, Sweet orange leaves of control and three CsMAF1 RNAi plants challenged with X. citri (approximately 10⁴ cells), showing that canker lesions were substantially enhanced in the RNAi compared with control plants. Photographs at 10× magnification were taken 14 d after bacterial infiltration.

Figure 5.

Overexpression of CsMAF1 correlates with seedling growth arrest and the inhibition of canker formation. A, Sweet orange epicotyls transformed with the 35S::CsMAF1 construct showing higher levels of CsMAF1 in more stunted and senescing shoots. RBCL, Ribulose-1,5-bisphosphate carboxylase. B, CsMAF1 levels in young leaves of the transgenic lines recovered after selection for the 35S::CsMAF1 construct. Only line 34 shows a relatively higher level of CsMAF1. C, Control plant. C, Plot of the relative band area of the corresponding top (slow-migrating) and bottom (fast-migrating) bands depicted in B, measured by ImageJ software. D, Quantitative reverse transcription-PCR analysis of the relative abundance of tRNA^His, tRNA^Leu, and tRNA^Thr in line 34 and a control plant, showing that line 34 accumulates less tRNA^His and tRNA^Leu than the control plant. Asterisks indicate that the differences between the control and transgenic means are statistically significant at P = 0.05. E, Sweet orange leaves of line 34 and a control plant infected with X. citri (approximately 10⁴ and 10⁵ cells) showing that canker lesions are less severe in line 34 than in the control. Photographs at 10× magnification were taken 20 d after bacterial infiltration.

Figure 6.

CsMAF1 depletion in canker lesions correlates with the presence of PthA. A, Sweet orange cv Pera infiltrated with a water suspension of X. citri (approximately 10⁵ cells) showing canker lesions that developed 10 d after bacterial inoculation. Soluble proteins extracted from the indicated leaf sectors (dotted lines) were resolved on a 10% SDS-polyacrylamide gel and probed with anti-CsMAF1. Protein loads were controlled by probing the samples with anti-CsCYP, since the large subunit of ribulose-1,5-bisphosphate carboxylase shows considerable down-regulation during canker formation (Cernadas et al., 2008). CsMAF1 levels are significantly reduced in the hyperplastic lesions relative to CsCYP. B, Leaf sectors of sweet orange cv Hamlin infiltrated with the wild-type (wt) X. citri or the pthA4 deletion mutant (approximately 10⁵ cells) showing that PthA4 is required to elicit cankers in sweet orange. C, Detailed view (10× magnification) of sweet orange cv Pera and Valencia leaves challenged with wild-type X. citri or the pthA4 deletion mutant (approximately 10⁵ cells) showing the absence of canker symptoms in the leaf sectors infiltrated with the pthA4 deletion mutant 10 d after bacterial inoculation. D, Soluble proteins from cv Pera leaf sectors infiltrated with wild-type X. citri or the pthA4 deletion mutant (approximately 10⁵ cells), or from noninfiltrated leaf sectors (control), were extracted 5 and 10 d after bacterial inoculation (left and right panels, respectively) and analyzed by western blot with the anti-CsMAF1 or anti-CsCYP antiserum. The blots show that depletion of CsMAF1, but not CsCYP, correlates with the development of cell hypertrophy and hyperplasia and with the presence of PthA4.