Introduction

Citrus canker is a severe bacterial disease that impacts the production of citrus because it affects most commercial cultivars of citrus in tropical and subtropical areas. The early symptoms on leaves, stems, and fruits are characterized by tiny, slightly raised blister-like lesions with water-soaked margins, which then develop into necrotic, raised lesions (Brunings et al., 2003). In addition to the severe defoliation and premature fruit dehiscence that usually occur on heavily infected trees, the blemished fruits also have significantly reduced commercial value. The citrus canker is associated with five different bacterial pathotypes, including three pathotypes of Xanthomonas citri subsp. citri (Xcc) (A, A*, and A^w) and two pathotypes of X. fuscans subsp. aurantifolii (B and C) (Gottwald et al., 2002; Sun et al., 2004). The Xcc A-type canker is the most prevalent and affects a wide range of hosts, including Citrus spp. and many closely related rutaceous plants (Graham et al., 2004).

Xcc PthA4 is a transcription activator-like (TAL) effector that is delivered into host plants by a type III secretion system (T3SS). The Xcc strains defective in the pthA4 gene cannot pthA4 alone induces weak canker phenotypes in citrus leaves (Duan et al., 2007). The heterologous expression of T3SS and pthA4 homologous genes in Escherichia coli elicits similar canker symptoms in citrus but not in tobacco (Nicotiana benthianum), soybean (Glycine max), or cotton (Gossypium hirsutum) (Kanamori et al., 1998). Variants of the PthA4 effector are widely present across all five bacterial subspecies that cause citrus canker disease (Swarup et al., 1992; Al-Saadi et al., 2007; Jalan et al., 2013). PthA4 acts as a major virulence determinant for the formation of cankers when translocated into the nuclei of its citrus host (Yang et al., 1995). The superhelical structure of this TAL effector recognizes host sequences through its repeat-variable di-residue (RVD), which specifically recognizes the effector-binding element (EBE) in the gene promoter of host plants, and thus, activates the expression of host target gene (Deng et al., 2012). Notably, PthA4 acts in a protein–protein interaction manner during infection with Xcc (Domingues et al., 2010). In addition to the proteins associated with DNA repair (Domingues et al., 2010; de Souza et al., 2012), PthA4 specifically interacts with citrus CsMAF1 and releases the suppressive effect on RNA Polymerase III to increase the transcription of tRNA and cell proliferation (Soprano et al., 2013). Thus, Xcc PthA4 functions through versatile mechanisms and modulates target genes not only at the transcriptional level but also at the protein level.

PthA4 activates the transcription of the CsLOB1 susceptibility gene in host citrus plants owing to the TAL effector DNA-binding principles and specificity (Moscou et al., 2009; Boch et al., 2009; Hu et al., 2014; Li et al., 2014). Remarkably, gene editing of the coding region or promoter sequence of CsLOB1 has been used to promote the resistance of citrus to infection with Xcc (Jia et al., 2016; Peng et al., 2017). CsLOB1 belongs to the lateral organ boundary domain (LBD) gene family, which encodes a set of plant-specific transcription factors (TFs) that contain cysteine-rich DNA binding motifs (Husbands et al., 2007). Plant LBD genes are involved in the regulation of plant growth, as well as the production of anthocyanins and nitrogen metabolism, and also respond to both hormonal and environmental stimuli (Majer et al., 2010). There are as many as 36 LBD proteins in citrus plants. CsLOB2 and CsLOB3 show the same involvement in inducing the formation of pustules as CsLOB1 when they are induced by Xcc, which expresses an artificial TAL effector (Zhang et al., 2017). In this case, many differentially expressed citrus genes have been identified, including those related to cell wall metabolism. However, the CsLOB2 and CsLOB3 promoters do not contain specific binding elements for PthA4; thus, neither can be induced by PthA4 during infection by Xcc (Zhang et al., 2017).

Since CsLOB1 is a TF, elucidating its downstream target genes is particularly important to understand the molecular events associated with the development of pustule symptoms (Hu et al., 2014; Li et al., 2014). PthA4 and CsLOB1 form a regulatory cascade that activate the citrus genes during Xcc infection (Duan et al., 2018). Our previous research showed that the ectopic expression of the TAL gene avrXa7 in Xcc suppressed the development of cankers but induced a yellowing phenotype around the inoculation site (Sun et al., 2018). Moreover, 138 upregulated citrus plant genes were identified to respond to wild-type (WT) Xcc infection but were relatively downregulated in leaves inoculated with Xcc 29-1 that expresses avrXa7. This study aimed to characterize the target genes of the citrus hosts that are directly regulated by the PthA4-mediated induction of CsLOB1 during Xcc infection, as well as its potential role in canker development.

Results

The expression of Cs9g12620 depends on pthA4 during Xcc infection

Our previous study on a transcriptome analysis reported that 138 genes were upregulated in citrus plants infected with the WT Xcc and downregulated in plants inoculated with Xcc that expressed avrXa7. Those genes were assumed to be related to the major Xcc virulence effector PthA4. PthA4 has 17.5 RVDs in its central domain that correspond to an EBE of 18 nucleotides of the target gene promoter. Based on the TAL effector-DNA binding code, PthA4 can bind an optimal EBE of the 19 bp oligonucleotide -TATAAACCCCCAATACCCT-, in which the first nucleotide “T” at 0 position is strictly conserved (Figure 1A, Figure 1— figure supplement 1). For each of the 138 genes, we downloaded and analyzed the 2000 bp DNA sequence upstream of the translation start site from the public citrus genome database (https://www.citrusgenomedb.org/). Possible PthA4 EBEs were identified in the promoter region of the two genes Cs9g12620 and CsLOB1 (Cs7g27640.1) (Figure 1A).

Figure 1.

Figure 1—figure supplement 1.

The transcript levels of Cs9g12620 and CsLOB1 were studied to determine whether they were associated with PthA4. In view of the transcriptome analysis, their transcript levels were first studied in sweet orange (Citrus sinensis) plants inoculated with the WT Xcc 29-1 and Xcc 29-1 that expressed avrXa7. The inoculation of Xcc 29-1 that expressed avrXa7 did not cause canker symptoms and displayed a yellow phenotype (Figure 1B). In comparison with the uninoculated control, the level of expression of Cs9g12620 was induced remarkably by inoculation with the WT Xcc 29-1 but not Xcc 29-1 that expressed avrXa7 (Figure 1C). Their transcript levels were then studied in C. sinensis plants inoculated with the pthA4 Tn5 insertion mutant Mxac126-80, which was impaired in its ability to cause canker symptoms (Figure 1D). In comparison with the plants inoculated with the WT Xcc 29-1, Cs9g12620 and CsLOB1 were downregulated in the plants infected with Mxac126-80 (Figure 1E). These findings demonstrated that Cs9g12620 was activated by Xcc infection in a PthA4-dependent manner.

Cs9g12620 shows different profiles of expression from its homolog Cs9g12650 owing to the genetic diversity in promoter region

A BLAST search was performed in the citrus genome database using the Cs9g12620 sequence. A highly conserved homolog Cs9g12650 was found closely downstream of Cs9g12620 in chromosome 9 of the C. sinensis chromosome (Figure 2A). Their amino acid sequences are 86% homologous. The existence of Cs9g12620 and Cs9g12650 in the C. sinensis chromosome was verified by PCR amplification with two primer sets anchored to the two genes, respectively (Figure 2B). Their transcription was then assessed by semi-quantitative reverse transcription PCR (RT-PCR) analysis. The results showed that Cs9g12620 was expressed at a high transcript level, while no transcripts of Cs9g12650 were found, even in the plants inoculated with the WT Xcc 29-1 (Figure 2C).

Figure 2.

Figure 2—figure supplement 1.

A sequence analysis of the Cs9g12620 and Cs9g12650 promoter regions revealed that are 71% homologous. A prediction of the promoter in Softberry (http://linux1.softberry.com/berry.phtml) identified the core structure region from the same location in both the Cs9g12620 and Cs9g12650 promoters (Figure 2—figure supplement 1). It should be noted that the predicted PthA4 EBE was located in the promoter core structure. A sequence alignment indicated that the last 22 nucleotides at the 3’ terminus in Cs9g12620 promoter had been lost genetically in the Cs9g12650 promoter (Figure 2—figure supplement 1). Fragments of 463 bp in the Cs9g12620 promoter and 460 bp in the Cs9g12650 promoter that composed the core structure were fused with a luciferase reporter. The transient overexpression of the promoter luciferase fusion indicated that the Cs9g12650 promoter was unable to drive the expression of luciferase in N. benthamiana (Figure 2D). This demonstrated that the Cs9g12650 promoter region lacked promoter activity. To confirm the effect of 22 nucleotides on promoter activity, we subsequently created a Cs9g12620 promoter mutant with a deletion of 22 nucleotides at the 3’ terminus. The mutant completely lost its ability to drive the expression of luciferase, suggesting that the 22 nucleotides are essential for activity of the promoter (Figure 2E).

PthA4 suppresses Cs9g12620 promoter activity independent of the binding action

The PthA4-dependent induction of Cs9g12620 during Xcc infection prompted us to determine whether it is directly regulated by PthA4. The 463 bp Cs9g12620 promoter was used in yeast one-hybrid (Y1H) assays. The EGY48 yeast co-transformed with pGBKT7-pthA4 and pG221 showed weak β-galactosidase activity, which demonstrated that PthA4 functions as a transcriptional activator. The co-transformation of pGBKT7-pthA4 and pG221-PCs9g12620 resulted in deep blue, suggesting that PthA4 binds the Cs9g12620 promoter. The enhanced β-galactosidase activity was quantified by the Miller method (Figure 3A). In an electrophoretic mobility shift assay (EMSA), the application of 0.117, 0.469, 1.875, 7.5, 30 and 120 μg of GST-PthA4 proteins partially hindered the mobility of 25 ng Cs9g12620 promoter DNA. The application of 7.5, 30 and 120 μg GST-PthA4 proteins completely retarded the mobility of 25 ng promoter DNA (Figure 3B). The negative control glutathione-S transferase (GST) tag was unable to bind the Cs9g12620 promoter DNA.

Figure 3.

To examine whether the activity was affected by PthA4, the Cs9g12620 promoter was cloned into the binary vector pCAMBIA1381 fused to a gusA reporter gene. Co-transformation with pthA4 did not induce the Cs9g12620 promoter activity. Histochemical staining revealed that the β-glucuronidase (GUS) activity in the leaves that co-expressed pthA4 suppressed the promoter activity (Figure 3C). For further verification, the levels of expression of gusA and pthA4 were evaluated by real-time quantitative reverse transcription PCR (qRT-PCR). Compared with the expression of the promoter-GUS fusion alone or in combination with the empty vector pHB control, the transcript level of gusA was reduced when the Cs9g12620 promoter-GUS fusion was co-transformed with pthA4 (Figure 3C).

In the EMSA analysis that utilized a synthetic PthA4 EBE (-TCCCACTTTAATATATAAA-) as a probe, 6, 12 and 24 μg of GST-PthA4 protein partially hindered the mobility of oligonucleotide (Figure 3D). This demonstrated that PthA4 truly bound to the predicted EBE located at the promoter core structure (Figure 2—figure supplement 1). The first base pair ‘T’ at 0 position in the sequence was then mutated into ‘A’, ‘C’, and ‘G’ in the Cs9g12620 promoter, which generated the point mutations PCs9g12620-M_A, PCs9g12620-M_C, and PCs9g12620-M_G, respectively. The activities of PCs9g12620-M_A and PCs9g12620-M_C were reduced compared with the WT PCs9g12620. In contrast, the mutation “T” to “G” resulted in a complete loss of promoter activity (Figure 3E). The Y1H and EMSA assays demonstrated that PthA4 could not bind PCs9g12620-M_A, PCs9g12620-M_C, or PCs9g12620-M_G (Figure 3F). However, the co-transformation of PthA4 with PCs9g12620-M_A, PCs9g12620-M_C, and PCs9g12620-M_G retained the ability to suppress the promoter activity (Figure 3G).

CsLOB1 binds multiple sites to induce the Cs9g12620 promoter

CsLOB1 showed transcription activity since β-galactosidase activity was found when it was co-transformed with pG221 into yeast EGY48. The β-galactosidase activity was remarkably enhanced when CsLOB1 was co-transformed with pG221, which harbored PCs9g12620 (Figure 4A). The binding of CsLOB1 to PCs9g12620 was further verified by an EMSA analysis (Figure 4B). Most importantly, the Cs9g12620 promoter activity was significantly induced when CsLOB1 was co-transformed with the PCs9g12620-GUS fusion (Figure 4C).

Figure 4.

Two putative CsLOB1-binding sites (-CGGC-) were found in the Cs9g12620 promoter and located at the -268 and -151 positions (designated LB1 and LB2, respectively) (Figure 4D, Figure 2—figure supplement 1). To examine the role of the two CsLOB1-binding sites, the two binding sites LB1 and LB2 were deleted to generate the mutants PCs9g12620-M_LB1, PCs9g12620-M_LB2, and Cs9g12620-M_LB1/2. Deletion of the LB1 or LB2 site did not block the binding of CsLOB1 to PCs9g12620 in yeast (Figure 4D). In the EMSA analysis, CsLOB1 retarded the mobility of PCs9g12620-M_LB1, PCs9g12620-M_LB2, and PCs9g12620-M_LB1/2 DNA (Figure 4E). Furthermore, the activities of these three promoter mutants were induced by CsLOB1 (Figure 4F). These results demonstrated that CsLOB1 binds to multiple sites in the Cs9g12620 promoter to induce promoter activity. Citrus tristeza virus (CTV)-induced gene silencing was conducted in C. sinensis plants, which generated CsLOB1-silenced plants. The efficiency of infection was evaluated by analyses of the expression of P23 gene harbored in the CTV vector. In comparison with the empty CTV-infected plants, the transcript level of Cs9g12620 was significantly reduced in the CsLOB1 silenced plants (Figure 4G).

PthA4 dynamically regulates the expression of Cs9g12620 during Xcc infection

To examine the exact effect of PthA4 on the Cs9g12620 promoter during Xcc infection, the CsLOB1 promoter GUS fusion (PCsLOB1-GUS) was first co-transformed with PthA4. A comparison with PCsLOB1-GUS expressed alone showed that the expression of PthA4 induced the PCsLOB1 promoter activity (Figure 5A). The co-expression of CsLOB1 with PthA4 had no effect on the PthA4-induced CsLOB1 promoter activity. This demonstrated that CsLOB1 is directly regulated by PthA4. When PthA4 and CsLOB1 were co-expressed with the Cs9g12620 promoter, the expression of PthA4 strikingly suppressed the CsLOB1-induced Cs9g12620 promoter activity (Figure 5B). To verify the suppressive effect by PthA4, A. tumefaciens GV3101 that expressed PthA4-FLAG was arranged at diverse concentrations of OD₆₀₀ values of 0.05, 0.1, 0.2, and 0.4. The results showed that the efficiency of suppression correlated with the high level of expression of PthA4-FLAG (Figure 5C).

Figure 5.

To study the regulatory pattern of pthA4 on the expression of Cs9g12620 during canker development, the levels of expression of pthA4, CsLOB1, and Cs9g12620 were quantitatively assayed in plants inoculated with the WT Xcc 29-1. Even though the levels of expression of pthA4, CsLOB1, and Cs9g12620 were enhanced at 5 and 10 days post-inoculation (dpi) compared with 0 dpi, their pattern of expression differed between 5 and 10 dpi. pthA4 and CsLOB1 showed the highest levels of expression at 10 dpi. In contrast, the level of expression of Cs9g12620 at 10 dpi was lower than that at 5 dpi (Figure 5D). To verify this special pattern of expression, Xcc strain 049, its TAL-free mutant Xcc 049E, and Xcc 049E that expressed pthA4 (Xcc049E/pthA4) were inoculated on C. sinensis plants. The TAL-free mutant 049E caused extremely weak canker symptoms, which were restored by expressing pthA4 in Xcc 049E (Figure 5E). No transcript of pthA4 was detected in the plants inoculated with Xcc 049E. In the plants inoculated with Xcc 049 or Xcc 049E/pthA4, the levels of expression of pthA4 and CsLOB1 were enhanced; moreover, the transcript level at 10 dpi was higher than that at 5 dpi (Figure 5E). In contrast, the level of expression of Cs9g12620 at 10 dpi was distinctively lower than that at 5 dpi (Figure 5E). By comparison with 5 dpi, the cell number of Xcc 29-1 in planta was increased 1.5-fold at 10 dpi (Figure 5F). However, the expressions of pectin esterase and expansin genes associated with canker development were decreased at 10 dpi (Figure 5G). This supported the hypothesis that the overexpression of PthA4 exerts a feedback inhibition on the expression of Cs9g12620, which was consistent with the result that the transient overexpression of PthA4 suppressed the Cs9g12620 promoter activation by CsLOB1. The dynamic regulation of Cs9g12620 is closely related with canker development.

PthA4 interacts with CsLOB1

PthA4 exerted a role in suppressing the Cs9g12620 promoter independently of its binding to EBE, which led us to determine whether PthA4 could physically interact with CsLOB1. In yeast two-hybrid (Y2H) assays, a positive interaction was observed in yeast strain AH109 co-transformed with AD-PthA4 and BD-CsLOB1 (Figure 6A). In a GST pull-down assay, GST-CsLOB1 successfully pulled down MBP-PthA4. In the control lane, the GST tag did not pull down MBP-PthA4 (Figure 6B). The interaction between PthA4 and CsLOB1 in vivo was further verified by split luciferase assays on N. benthamiana (Figure 6C).

Figure 6.

Cs9g12620 is involved in the formation of canker symptoms

The Cs9g12620, as well as CsLOB1 and pthA4, were transiently overexpressed in C. sinensis. The expression of pthA4 induced a canker-like phenotype. The overexpression of Cs9g12620 resulted in yellowing and hypertrophy of the leaves of C. sinensis plants (Figure 7A). The overexpression of CsLOB1 did not cause a distinctive phenotype in C. sinensis plants. Thus, the overexpression of Cs9g12620 elicited a canker-like symptom in C. sinensis that resembled that induced by pthA4, although the symptoms were weaker. Transmission electron microscopy (TEM) was conducted to investigate the changes in C. sinensis leaf structure. In the control leaves infiltrated with empty vector, large air spaces were found between the cells in spongy mesophyll. In comparison, more cells and less air space were found in the spongy mesophyll of the leaves that transiently expressed Cs9g12620 and pthA4 (Figure 7B).

Figure 7.

The contribution of Cs9g12620 to canker development was subsequently examined by citrus tristeza virus (CTV)-induced gene silencing in C. sinensis plants. The efficiency of gene silencing was evaluated in one-month-old leaves at 60 days post-stem inoculation. The infection by CTV was indicated by verification of the transcripts of CTV P23 gene. The transcript level of Cs9g12620 was then verified to be reduced by approximately 70% in the gene-silenced C. sinensis plants (Figure 7C). Inoculation of the WT Xcc 29-1 caused remarkably attenuated canker symptoms in the Cs9g12620-silenced plants. In comparison with the control plants that were infected with empty CTV, there were remarkably fewer symptoms on the Cs9g12620-silenced plants (Figure 7D). The Xcc 29-1 in the Cs9g12620-silenced plants grew more slowly than in the control plants. Significantly fewer cell numbers were isolated from the Cs9g12620-silenced plants than from the control plants at 2, 4, and 6 days post-inoculation (Figure 7E). The expression levels of Cs9g12620 and CsLOB1 genes in Cs9g12620-silenced plants inoculated with Xcc 29-1 showed that the induced transcription levels of both were decreased, and the expression levels of pthA4 were also significantly decreased, down-regulated by 38%, 15% and 16%, respectively (Figure 7F). The TAL-free 049E caused extremely wreak canker symptom, and wildtype 049 and 049E/pthA4 caused severe canker on control plants. On Cs9g12620-silenced plants, wildtype 049 and 049E/pthA4 caused greatly reduced canker symptom, while the 049E/pthA4 caused weak canker symptom alike to TAL-free mutant 049E (Figure 7G). Furthermore, the population of 049 and 049E/pthA4 recovered from Cs9g12620-silenced plants was lower than that in control plants. By contrast, population of 049E was low in either Cs9g12620-silenced or control plants (Figure 7H). This demonstrated expression of pthA4 in 049E did not restore bacterial virulence on Cs9g12620-silenced plants. The level of expression of Cs9g12620 is correlated with the formation of canker symptoms in citrus plants.

Discussion

The ectopic expression of avrXa7 in Xcc 29-1 did not induce canker symptoms on citrus plants (Sun et al., 2018), which was consistent with a previous study that reported that the browning phenotype was elicited by an Xcc L2 strain that expressed avrXa7 (Ishihara et al., 2003). This was hypothesized to be caused by two groups of differentially expressed genes, including (i) resistance genes induced by AvrXa7, and (ii) susceptibility genes induced by PthA4 but suppressed by the expression of AvrXa7. A transcriptome analysis was utilized to identify the different relevant citrus genes and showed that AvrXa7 specifically activated the aspartic protease gene Cs7g06780.1 and linoleate 13S-lipoxygenase gene orange1.1t03773.1 (Sun et al., 2018). In this study, we focused on identification of the genes induced by PthA4 for canker formation. Based on the pattern of expression and analysis of promoter activity, Cs9g12620 was proven to be dynamically regulated by the PthA4-mediated induction of CsLOB1 for the development of canker symptoms. This supported the hypothesis that the PthA4-mediated induction of CsLOB1 serves a critical role in the complicated regulatory network on gene expression during Xcc infection.

The original purpose of this study was to explore the genes that were directly targeted by PthA4 through TAL effector-DNA recognition. Based on a sequence analysis, the two genes Cs9g12620 and CsLOB1 were predicted to contain PthA4 EBEs. An analysis of the pattern of expression revealed that Cs9g12620 and CsLOB1 were induced by Xcc infection in a PthA4-dependent manner. Because the susceptibility gene CsLOB1 has already been proven to be targeted by Xcc PthA4, the binding and regulatory role of PthA4 on the Cs9g12620 promoter was assayed experimentally. Y1H, EMSA, and promoter reporter fusion analyses were used and enabled the conclusion that PthA4 bound to the Cs9g12620 promoter, whereas it suppressed the promoter activity. Since two CsLOB1 putative binding sites were predicted from the Cs9g12620 promoter, we demonstrated that CsLOB1 bound directly to the Cs9g12620 promoter. Most importantly, CsLOB1 could activate the Cs9g12620 promoter. As such, Cs9g12620 is actually a target of CsLOB1.

The ectopic expression of CsLOB1 restores the mutant canker symptoms of Xcc pthA4, which demonstrates that CsLOB1 is the key target of PthA4 and thus, induces pustule symptoms (Duan et al., 2018). The transient overexpression of CsLOB1 did not induce an obvious phenotype in the citrus plant leaves, which was consistent with the findings of previous research (Hu et al., 2014). Instead, the contribution of CsLOB1 to canker symptoms has been verified by the transformation of ‘Duncan’ grapefruit (Citrus x paradisi) plants with a 35S:CsLOB1-glucocorticoid receptor construct, which suggested that the continuous expression of CsLOB1 is essential for the development of canker symptoms (Duan et al., 2018). Duan et al. (2018) reported that a number of citrus genes appear to be associated with CsLOB1, but only the Cs2g20600 promoter probe interacted with CsLOB1. Cs2g20600 encodes a zinc finger C3HC4-type RING finger protein with E3 ubiquitin ligase activity (Duan et al., 2018). In this study, we found that Cs9g12620 was another target that is directly regulated by CsLOB1. The regulatory role of CsLOB1 includes promoting host cell proliferation, regulating citrus cell wall remodeling, and affecting hormone signaling pathways, which is accompanied by the optimal growth of Xcc (Zou et al., 2021). Although it has been reported to specifically bind to -CGGC-in its target promoter region, CsLOB1 retained the ability to bind the Cs9g12620 promoter DNA when the two specific binding sites LB1 and LB2 were deleted. We hypothesized that there are other putative binding sites of CsLOB1 in the Cs9g12620 promoter, and multiple citrus genes are directly targeted by CsLOB1 for the development of cankers during Xcc infection.

PthA4 acts as a sole disease TAL required by Xcc to cause symptoms of canker on citrus plants. It localizes to the nuclei following secretion into host cell by the type III secretion system. A transient overexpression analysis showed that PthA4 exhibited a particular suppressive effect on the CsLOB1-activated Cs9g12620 promoter activity. Since the PthA4 binding site is critical for activity of the Cs9g12620 promoter, the binding of PthA4 suppressed the promoter activity. This prompted us to study whether PthA4 affects CsLOB1 alternatively. Y2H and pull-down assays showed that PthA4 interacted with CsLOB1 at the protein-protein level. This provided a new insight to study the dynamical regulatory role of PthA4 during Xcc infection since we found that PthA4 affects the expression of Cs9g12620 in a dose-dependent manner. When PthA4 was expressed at a high level at 10 dpi, the level of expression of Cs9g12620 was surprisingly lower. This suggested that there is a precise mechanism of the PthA4-CsLOB1-Cs9g12620 regulatory cascade during infection with Xcc.

Although the amino acid sequence of Cs9g12650 was 86.3% homologous with that of Cs9g12620, they differed in their promoter sequences. The promoter of Cs9g12650 lacks 22 bp of nucleotides at its 3’ terminus, which are critical for the activity of its promoter. The activity of Cs9g12620 promoter was completely lost if the 22 bp sequence was deleted by mutation. Therefore, Cs9g12650 is not transcribed in citrus plants. This supported the hypothesis that the null activity of Cs9g12650 promoter was owing to genetic diversity in citrus plants, which leads to different patterns of expression between the two paralogs in citrus. The 22 bp sequence at the Cs9g12620 promoter could be a potential site for gene editing to create citrus plants that are resistant to citrus canker. A similar regulation of expression has been found in CsLOB1; CsLOB2 and CsLOB3 show the same ability to induce pustule formation as CsLOB1, but their promoters do not contain the PthA4 EBE; therefore, neither can be induced by PthA4 (Zhang et al., 2017).

The Cs9g12620 protein contains three distinctive conserved domains, including a signal peptide, a duplicated DUF642 region, and a bacterial exosortase protein domain. The duplicated DUF642 region has been shown to be a carbohydrate-binding domain. The exosortase system is involved in directing transport across the plasma membrane, with additional posttranslational modifications, such as glycosylation (Haft et al., 2012). This system is distributed across many biofilm-producing environmental bacteria and is closely related to the N-acyl amino acid synthases in Proteobacteria (Craig et al., 2011). It was hypothesized that Cs9g12620 could possibly be involved in the transportation of carbohydrates in citrus cells. This was similar to the SWEET genes induced by the effectors that pathogenic bacteria secrete into plants, and they export intracellular sugar into the extracellular space, which can be utilized by pathogenic bacteria (Hu et al., 2014). The binding of carbohydrates and the exact role for canker formation merits further study.

In conclusion, the putative carbohydrate-binding protein gene Cs9g12620 was identified as a target of CsLOB1, which was directly regulated by PthA4-mediated induction of CsLOB1. In addition to the enhanced pattern of expression in response to Xcc infection, Cs9g12620 was dynamically feedback regulated by PthA4 in a dose-dependent manner. At the early infection stage, the presence of PthA4 is responsible for the induction of CsLOB1, which is necessary for the activation of Cs9g12620. At the late infection stage, the over-redundancy of PthA4 exerts a feedback suppression effect on the activation of Cs9g12620 by CsLOB1 (Figure 8). The findings provide strong evidence to understand the molecular complexity of PthA4-mediated induction of CsLOB1 during the development of citrus canker.

Figure 8.

Materials and methods

Plant and bacterial materials

C. sinensis and N. benthamiana were grown in a greenhouse (26/24°C light/dark) under long-day conditions (16/8 h light/dark). The strains and plasmids used in this study are listed in Supplementary file 1a. Xcc strain 29-1 was collected from C. sinensis in Jiangxi Province, China (Ye et al., 2013). The Xcc 29-1/avrXa7 strain is a Xcc 29-1 strain that expresses avrXa7 from Xanthomonas oryzae pv. oryzae, the causal agent of rice bacterial blight (Sun et al., 2018). Mxac 126-80 is a pthA4 mutant with a Tn5 insertion at the 1789 bp position in the middle of RVD coding sequence (Song et al., 2015). Xcc strain 049 was collected from C. sinensis in Chongqing Province, China (Ye et al., 2013). Xcc 049E is a TAL-free strain of Xcc 049, and Xcc 049E/pthA4 is Xcc 049E that expresses the pthA4 gene (Ge et al., 2019). All the Xcc strains were cultured in nutrient broth (NB) or nutrient agar (NA, 1.5% agar) at 28°C (Song et al., 2015). Agrobacterium tumefaciens strain GV3101 was cultured in YEP medium (10 g/L yeast extract, 10 g/L peptone, and 5 g/L sodium chloride, pH 7.0) at 28°C. The yeast strains EGY48 for the Y1H assays and AH109 for the Y2H assays were cultured at 30L in YPD media (10Lg/L yeast extract, 20Lg/L peptone, and 20Lg/L glucose, pH 4.5–5.0).

Inoculation of Xcc in citrus host plants

The cultured Xcc cells were suspended in sterile distilled water to a final concentration of 10⁸ CFU/mL (OD₆₀₀ = 0.3). The bacterial suspension was inoculated onto fully expanded C. sinensis leaves with a needleless syringe (Wu et al., 2019). The canker disease symptoms were recorded at 5 or 10 dpi. To assess the bacterial growth in planta, 0.25 cm² leaf discs from the areas of infiltration were collected to count the cell numbers. The cell numbers were calculated as CFU per cm². All the experiments were repeated three times.

Real-time quantitative reverse transcription PCR (qRT-PCR)

qRT-PCR was performed using a CFX Connect Real Time PCR detection system (Bio-Rad, Shanghai, China) using iTaq Universal SYBR Green Supermix (Bio-Rad). The total RNA was isolated from C. sinensis and N. benthamiana leaves using a Plant RNA Kit (Omega, Shanghai, China). A volume of 2 µg of total RNA was reverse transcribed into single-stranded cDNA using AMV reverse transcriptase (TaKaRa, Dalian, China). The primer sequences used for qRT-PCR are listed in Supplementary file 1b. The C. sinensis CsActin and N. benthamiana NbEF1_α genes were used as internal controls to statistically analyze the levels of relative expression (Wu et al., 2019). All the experiments were repeated three times.

Luciferase assays

To study the interaction in vivo, the PthA4 gene from Xcc was fused with nLUC at the N terminus. The CsLOB1 gene from C. sinensis was fused with cLUC at the C terminus. cLUC-CsLOB1 was co-expressed with nLUC-PthA4 in 4-week-old N. benthamiana leaves. To analyze the promoter activity, the 463 bp Cs9g12620 promoter region was directly introduced into vector pGWB435 by a Gateway cloning strategy to generate PCs9g12620-LUC. Site-directed mutation of the first nucleotide “T” in EBE of PthA4 or deletion of CsLOB1 binding sites LB1 and LB2 were created by overlapping PCR using the primers listed in Supplementary file 1b. The promoter mutants M_A, M_C, M_G, M_LB1, M_LB2, and M_LB1/2 that were generated were fused with luciferase (LUC) in pGWB435. The activities of LUC were assayed after transient overexpression in N. benthamiana. For each assay at 2 dpi, the inoculated leaves were treated with 0.5 mM luciferin, kept in the dark for 1 min to quench the fluorescence, and then used to capture luciferase luminescence images by a cooled charge-coupled device (CCD) imaging apparatus (Roper Scientific, Trenton, NJ, USA). Each experiment had three replicates.

Yeast one-hybrid assays

The pGBKT7-pthA4 and pGBKT7-CsLOB1 constructs were created by inserting the pthA4 and CsLOB1 gene into the pGBKT7 vector, respectively (Supplementary file 1b). A 463 bp DNA fragment of the Cs9g12620 gene promoter was cloned into the pG221 vector, which resulted in pG221-PCs9g12620 (Supplementary file 1a and 1b). To examine the role of PthA4 and CsLOB1 binding sites in interaction, mutants M_A, M_C, M_G, M_LB1, M_LB2, and M_LB1/2 were cloned in pG221 (Supplementary file 1b). Mutants M_A, M_C, and M_G were co-transformed with pGBKT7-pthA4, and mutants M_LB1, M_LB2, and M_LB1/2 were co-transformed with pGBKT7-CsLOB1 into yeast EGY48 using the standard LiAc-mediated yeast transformation (Ye et al., 2004). pG221 that harbored a CYC1 core promoter was used for a transactivation assay in PthA4 and CsLOB1. The yeast transformants were screened on synthetic dropout SD/-Ura and SD/-Ura/-Trp media plates (Ye et al., 2004). The transformed cells were subjected to five freeze–thaw cycles using liquid nitrogen, spotted onto a sterile filter, and soaked with Z buffer that contained 20 µg/mL of X-gal. β-galactosidase activity was scored within 8 hours of incubation. The β-galactosidase activity was quantified using the standard Miller method (Marburg et al., 2016). Each experiment was performed three times.

Electrophoretic mobility shift assays

The binding of both PthA4 and CsLOB1 to the Cs9g12620 promoter DNA (including M_LB1, M_LB2, and M_LB1/2) was detected by an Cy5-labeled probe electrophoretic mobility shift assay (EMSA) as described previously (Fan et al., 2020). The binding of PthA4 to the Cs9g12620 promoter DNA (M_A, M_C, and M_G) was visualized by ethidium bromide staining as part of EMSA. The GST-pthA4 and GST-CsLOB1 constructs were created by separately inserting the pthA4 and CsLOB1 genes into the pET41a(+) vector (Supplementary file 1a). GST-PthA4, GST-CsLOB1, and GST tag were expressed separately in E. coli BL21 (DE3) by induction with 1.0 mM isopropyl-β-D-thiogalactopyranoside (IPTG). The purified GST-PthA4, GST-CsLOB1, and GST tag were subjected to ethidium bromide staining with concentration gradients by a two-fold dilution. The proteins were incubated with 25 ng of promoter DNA in gel shift binding buffer (50 ng/μL Poly (dI•dC), 0.05 M Tris-HCl, 1.0 mM EDTA, 0.15 M KCl, and 1.0 mM dithiothreitol). The GST-PthA4 protein, GST-CsLOB1 protein, and GST tag were diluted into series that consisted of 0.117, 0.469, 1.875, 7.5, 30, and 120 µg; 0.234, 0.938, 3.75, 7.5, 15, and 60 µg; and 1.953, 7.813, 3.75, and 15.625 µg; respectively. After incubation at 28°C for 30 min, the mixtures were subjected to a DNA mobility analysis by electrophoresis in a 6% non-denaturing acrylamide gel. The DNA bands were visualized by fluorescence imaging using a Typhoon Trio Variable Mode Imager (GE Healthcare, Chicago, IL, USA). Analyses of the binding to promoter mutants M_A, M_C, and M_G were conducted with the same strategy but using different GST-PthA4 protein series. Each experiment had three replicates.

GUS activity assays

The Cs9g12620 promoter region was cloned into the binary vector pCAMBIA1381 to fuse with the gusA gene (Supplementary file 1b). The coding sequences of pthA4 and CsLOB1 were individually cloned into the binary vector pHB to obtain pHB-pthA4 and pHB-CsLOB1, respectively. The promoter–GUS fusion was co-transformed with pHB-pthA4 or pHB-CsLOB1 to examine the promoter activity in N. benthamiana. β-Glucuronidase (GUS) activity was assayed by histochemical staining of the activity after 2 dpi (Hu et al., 2014). Moreover, the levels of expression of gusA in the leaf samples were evaluated by qRT-PCR to quantify the promoter activity (Sun et al., 2020). Each experiment had three replicates.

Yeast two-hybrid assays

The coding sequences of pthA4 and CsLOB1 were cloned individually into both the pGADT7 and pGBKT7 vectors. AD-PthA4/BD-CsLOB1 was then transformed into the yeast strain AH109 and screened on synthetic drop-out media that lacked leucine and tryptophan (SD/-Leu-Trp). Single colonies were cultured, serially diluted, and grown on SD/-Leu-Trp and SD/-Ade-Leu-Trp-His media to examine the possible interaction. Additionally, pGADT7-T and pGBKT7-53 were used as positive controls, while pGADT7-T and pGBKT7-lam served as negative controls. Each experiment had three replicates.

GST pull-down assay

PthA4 was cloned into pMAL-c4x to express the MBP-PthA4 fusion protein. CsLOB1 was cloned into pET-41a(+) to express the GST-CsLOB1 fusion protein. MBP-PthA4 and GST-CsLOB1 were expressed separately in E. coli BL21 (DE3) by induction with 1.0 mM IPTG. After the proteins were purified by Glutathione Sepharose 4 Fast Flow (GE Healthcare) and Amylose Resin (New England Biolabs, Ipswich, MA, USA) affinity chromatography, the GST pull-down assays were performed. Each experiment had three replicates.

Western blotting analysis

Western blotting was performed to detect the levels of expression of proteins in N. benthamiana. Leaf samples were harvested at 2 dpi. The leaves were ground in liquid nitrogen and extracted in extraction/washing buffer (Roche, Basel, Switzerland). The proteins that were extracted were eluted with 3×Laemmli loading buffer, resolved by 10% SDS-PAGE, and subjected to western blotting using anti-FLAG (Abmart, Shanghai, China), anti-nLUC (Abmart), or anti-cLUC polyclonal antibody (Abmart). Each experiment had three replicates.

Transient overexpression of the Citrus sinensis genes

The coding sequences of Cs9g12620 were cloned into the binary vector pHB to obtain pHB-Cs9g12620. A. tumefaciens GV3101 cells were transformed with the resulting recombinant plasmids pHB-Cs9g12620, pHB-pthA4, and pHB-CsLOB1. For transient expression in C. sinensis, A. tumefaciens cells that harbored different constructs were cultured and grown overnight in YEP media with shaking. The cells were pelleted by centrifugation at 6,000 rpm for 10 min at 4°C, suspended in infiltration medium (10 mM MgCl₂, 10 mM MES, and 200 mM acetosyringone, pH 5.7) to an OD₆₀₀ of 0.01 and gently injected into the young leaves. The phenotypes were scored at 15 dpi. Each experiment had three replicates.

Transmission electron microscopy

To examine any changes in the leaf tissue, the agroinfiltrated C. sinensis leaves were sampled 15 dpi for observation by TEM. The samples were fixed using 3% glutaraldehyde in 0.1 M potassium phosphate buffer (pH 7.2) and incubated at room temperature for 4 h. The tissues were washed twice with the same buffer and post-fixed in 2% osmium tetroxide in 0.1 M potassium phosphate buffer at room temperature for 4 h (Etxeberria et al., 2009). The samples were then dehydrated in an ethanol series and embedded in Spurr’s resin. Thin sections (80– 100 nm thick) were cut with a diamond knife, collected on 200 mesh copper grids, stained with 2% uranyl acetate, and post-stained with lead citrate. Those sections were then examined under an electron microscope (SU8020; Hitachi, Ltd., Tokyo, Japan) in scanning transmission electron microscope mode. Each experiment had three replicates.

Silencing of Cs9g12620 in the citrus plants

CTV33 is a stable expression vector based on the Citrus tristeza virus (El-Mohtar et al., 2014). A 488 bp DNA fragment of Cs9g12620 and a 264 bp DNA fragment of CsLOB1 were cloned separately into the CTV33 vector to generate the gene silencing vector. The CTV virion was prepared in N. benthamiana and inoculated in C. sinensis using the bark-flap method as previously described (El-Mohtar et al., 2014; Hajeri et al., 2014). Three citrus plants were inoculated for each CTV virion. The plants inoculated with the CTV33 empty vector were used as controls. The infection of CTV was evaluated by the level of expression of the P23 gene harbored in the CTV33 vector. The RNA silencing efficiency of Cs9g12620 and CsLOB1 was evaluated by quantification of the transcript level in the leaves from newly emerging branches. The development of cankers was then assessed by infiltration with Xcc 29-1 inoculum of 10⁷ CFU/mL.

Bioinformatics analysis

Sequence similarity searches of the reference C. sinensis genome database (https://www.citrusgenomedb.org/) were performed using BLAST. Putative promoter sequences were predicted using the online promoter analysis program Neural Network Promoter Prediction (http://www.fruitfly.org/seq_tools/promoter.html). Putative PthA4 EBEs were predicted using DNAMAN 8.0 (Sun et al., 2018). Putative signal peptides were predicted in PrediSi (http://www.predisi.de/) (Hiller et al., 2004). The conserved domains were identified using the NCBI Conserved Domain Database.

Acknowledgements

We thank Gongyou Cheng and Lifang Zou in Shanghai Jiaotong University (Shanghai, China) for their gifts of Xcc strain 049, TAL-free mutant 049E, and 049E/pthA4. We also thank Dingzhong Tang and Haitao Cui for advice on using the cooled CCD imaging apparatus at the Plant Immune Center in Fujian Agriculture and Forestry University (Fuzhou, China). We would like to thank MogoEdit (https://www.mogoedit.com) for its English editing during the preparation of this manuscript.

Additional information

Competing interests

The authors declare that they have no competing interests.

Funding