A Novel ‘Candidatus Liberibacter asiaticus’-Encoded Sec-Dependent Secretory Protein Suppresses Programmed Cell Death in Nicotiana benthamiana

‘Candidatus Liberibacter asiaticus’ (CLas) is one of the causal agents of citrus Huanglongbing (HLB), a bacterial disease of citrus trees that greatly reduces fruit yield and quality. CLas strains produce an array of currently uncharacterized Sec-dependent secretory proteins. In this study, the conserved chromosomally encoded protein CLIBASIA_03875 was identified as a novel Sec-dependent secreted protein. We show that CLIBASIA_03875 contains a putative Sec- secretion signal peptide (SP), a 29 amino acid residue located at the N-terminus, with a mature protein (m3875) of 22 amino acids found to localize in multiple subcellular components of the leaf epidermal cells of Nicotiana benthamiana. When overexpressed via a Potato virus X (PVX)-based expression vector in N. benthamiana, m3875 suppressed programmed cell death (PCD) and the H2O2 accumulation triggered by the pro-apoptotic mouse protein BAX and the Phytophthora infestans elicitin INF1. Overexpression also resulted in a phenotype of dwarfing, leaf deformation and mosaics, suggesting that m3875 has roles in plant immune response, growth, and development. Substitution mutagenesis of the charged amino acid (D7, R9, R11, and K22) with alanine within m3875 did not recover the phenotypes for PCD and normal growth. In addition, the transiently overexpressed m3875 regulated the transcriptional levels of N. benthamiana orthologs of CNGCs (cyclic nucleotide-gated channels), BI-1 (Bax-inhibitor 1), and WRKY33 that are involved in plant defense mechanisms. To our knowledge, m3875 is the first PCD suppressor identified from CLas. Studying the function of this protein provides insight as to how CLas attenuates the host immune responses to proliferate and cause Huanglongbing disease in citrus plants.


Introduction
Upon invasion of the host, phytopathogens produce numerous compounds (e.g., metabolites) to facilitate their survival in the host. These compounds are termed pathogen-associated molecular patterns (PAMPs), some of which are detected by host pattern recognition receptors (PRRs) localized in the cell membrane. The resulting interaction between the PAMPs and PRRs promotes (1) transcriptional reprogramming, (2) protein phosphorylation, (3) the activation of ion channels, (4) the production of reactive oxygen intermediates, (5) cell wall reinforcement, and (6) the accumulation of (Supplementary Figure S1) but was absent from 'Candidatus Liberibacter africanus' (CLaf) and 'Candidatus Liberibacter americanus' (CLam), the other two causal agents of HLB [14,30]. Therefore, CLIBASIA_03875 may be a conserved protein of the CLas strains. Diagrams of the prokaryotic expression cassettes for the phoA gene. pET-phoA harboring the full length phoA gene was used as a positive control, while pET-mphoA was a negative control that harbors mphoA lacking its native SP-encoding sequence. (C) 3875SP directed the extracytoplasmic export of the mPhoA moiety. After 6 h of incubation at 37 °C on LB media containing BCIP (90 μg/mL), IPTG (100 mM) and Na2HPO4 (75 mM), the Escherichia coli cells expressing mphoA remained colorless, but the cells carrying pET-3875SP-mphoA clearly turned blue, like those that harbored pET-phoA.

CLIBASIA_03875 Was a Sec-Dependent Secretory Protein, and Its Mature Form Was Present in Multiple Subcellular Components of N. benthamiana Cells
A bioinformatics analysis using PrediSi [31] revealed that CLIBASIA_03875 contained a putative signal protein (SP) (3875SP) corresponding to its N-terminal 29 amino acids ( Figure 1A), which is basically consistent with a prediction based on SignalP [32] in a recent study [33]. To validate the extracytoplasmic transport signal function of 3875SP, its coding sequence was inserted into the pET-mphoA vector ( Figure 1B) and was subjected to an Escherichia coli (E. coli) phoA gene fusion assay [34]. The cells expressing the fusion protein 3875SP-mphoA, like the positive control cells that harbored pET-phoA, turned dark blue after 6 h of incubation on indicator LB agar, while the E. coli cells that harbored pET-mphoA (the negative control) remained colorless even after 24 h of incubation ( Figure  1C). Collectively, the in silico prediction and experimental data provide evidence that 3875SP directs the extracytoplasmic export of the mPhoA moiety in E. coli, indicating that CLIBASIA_03875 was a novel Sec-dependent secretory protein of CLas.
Sec-dependent secretory proteins simultaneously undergo SP cleavage during extracytoplasmic translocation, yielding a mature protein [35]. Accordingly, after the removal of 3875SP, CLIBASIA_03875 would transition to its mature form, m3875. Since CLas is an intracellular pathogen [14], m3875 would be released directly into the host cells. To determine the distribution of m3875 in plant cells, a fusion protein m3875-GFP (Figure 2A) was expressed in Nicotiana benthamiana via Agrobacterium-mediated transient expression [36]. Using confocal microscopy, we observed that the green fluorescence of m3875-GFP, resembling that of free green fluorescent protein (GFP), was present in the nucleus, cytoplasm and cytoplasmic membrane ( Figure 2B), indicating the multiple subcellular localization of m3875 in plant cells.

m3875 Suppressed Bax-and INF1-Triggered PCD in Nicotiana benthamiana
PCD functions in a variety of cellular processes, including plant immunity. A specialized form of PCD termed HR occurs in response to a pathogen and triggers rapid cell death to restrict the biotrophic/hemibiotrophic pathogens at the site of infection [37,38].
The mouse BAX and the PAMP INF1 of Phytophthora infestans are two well-known HR PCD inducers [39][40][41]. Based on a PVX-based expression system, both BAX and INF1 have been extensively utilized to identify the PCD suppressor from a range of phytopathogens in N. benthamiana [42][43][44][45][46]. With the same strategy, we found that m3875 completely suppressed both BAX-and INF1triggered PCD ( Figure 3A). The histochemical detection of H2O2, a critical reactive oxygen species (ROS) that contributes to plant cell death [47], revealed that BAX-and INF1-induced H2O2 accumulation were significantly inhibited by m3875 ( Figure 3B). In contrast, the negative controls (infiltration buffer and GFP) did not suppress either cell death or the H2O2 accumulation triggered by BAX or INF1 ( Figure 3A,B).
To identify the amino acids of m3875 involved in PCD suppression, the charged amino acids (D 7 , R 9 , R 11 and K 22 ) within the protein were individually or simultaneously substituted with the neutral amino acid alanine, resulting in mutants A7, A9, A11, A7-11 and A22 ( Figure 4). As a result, all the mutants exhibited only a slightly reduced ability to suppress BAX-and INF1-mediated PCD ( Figure  4 and Table 1), indicating the mild effects of D 7 , R 9 , R 11 and K 22 on PCD suppression. It was notable that the activity of A7-11, the mutant that harbored the simultaneous substitution of D 7 , R 9 and R 11 with alanine residues, remained nearly unchanged at suppressing PCD, comparable to that of the wild-type (WT) m3875 ( Figure 4 and Table 1). Given the relatively small size (22 amino acids) of m3875, the data suggested the robustness of m3875 in the role of PCD suppression.

m3875 Interfered with the Development of Nicotiana benthamiana
To further explore the biological significance of m3875 in planta, A. tumefaciens cells with pPVX-m3875 were infiltrated into the N. benthamiana seedlings ( Figure 4A), and A. tumefaciens cells harboring empty PVX-releasing pGR107 were included as a parallel control. Both the PVX-and PVX-m3875-infected plants initially induced systemic symptoms of leaf crinkling and veinal chlorosis at 5-6 days post inoculation (dpi). Continuous observations up to 15 dpi showed that the PVX-infected plants gradually recovered and were comparable with the healthy N. benthamiana. However, the plants infected with PVX-m3875 developed severe symptoms, including dwarfing, leaf deformation and mosaic ( Figure 5A). Northern blot analysis detected a similar viral load of PVX and PVX-m3875 in the upper leaves of N. benthamiana ( Figure 5B), indicating that m3875 had little impact on the multiplication of PVX, despite functioning as a PCD suppressor. Taken together, these data indicated that the severe symptoms displayed by PVX-m3875-infected plants were most likely caused by the heterologously expressed m3875, rather than the increased replication of PVX, suggesting that m3875 interfered with plant development.
Since amino acid changes had little effect on m3875-mediated PCD suppression (Figure 4), we questioned if the mutations disrupted the interference of m3875 in plant development. Using the same strategy as above, A. tumefaciens cells harboring pPVX-A7, pPVX-A9, pPVX-A11, pPVX-A7-9-11 or pPVX-A22 were infiltrated into N. benthamiana. All the mutants, including A7-11, induced symptoms similar to that of the WT m3875 ( Figure 5A) and had little effect on the multiplication of PVX ( Figure 5B), providing additional evidence for the robustness of m3875 in its associated roles.

m3875 Regulated Transcription of the Defense-Related Genes in Nicotiana benthamiana
Since m3875 suppressed the HR PCD formation, we then examined the transcriptional levels of the defense-related genes in N. benthamiana that transiently expressed m3875 and eGFP. Calcium is a universal secondary messenger that has been implicated in plant innate immunity [48]. In plants, Ca 2+ transport across the plasma membranes is mediated by several families of ion channels, including cyclic nucleotide-gated channels (CNGCs) [49,50]. The plant CNGCs family is classified into five groups (I, II, III, IVa, and IVb) [51] with the members of group-IVb determined to be positive regulators of HR formation [52,53]. Thus, the transcripts of all four CNGC group-IVb members (NbCNGC23-26) in N. benthamiana were studied [54]. The expression of all the genes tested were significantly reduced in the N. benthamiana leaves transiently expressing m3875 compared with those expressing eGFP, further supporting the hypothesis that m3875 acts to suppress the host immune response ( Figure 6A).
We next focused on Bax-inhibitor 1 (BI-1), a common cell death suppressor in plants, as well as animals [55]. In N. benthamiana, there are two BI-1 homologs (designated NbBI-1 and NbBI-2), which share ~84% identity [56]. Interestingly, m3875 dramatically enhanced the expression of NbBI-2 but had only a minor effect on NbBI-1 ( Figure 6B). In addition, the transcriptional level of NbWRKY9, like that of NbBI-2, was also significantly upregulated upon m3875 overexpression ( Figure 6C). Consistent with a close phylogenetic relationship between NbWRKY9 and AtWRKY33 [57], BLAST searches of the Arabidopsis database showed that NbWRKY9 shared the highest identity with AtWRKY33 (Genbank No. NP_181381.2), suggesting that NbWRKY9 is an AtWRKY33 ortholog in N. benthamiana. AtWRKY33 is a well-documented WRKY transcription factor that negatively regulates defense responses mediated by salicylic acid (SA), a critical signaling molecule responsible for triggering PCD [58]. In particular, the Arabidopsis thaliana plants overexpressing AtWRKY33 confer high susceptibility to Pseudomonas syringae, a biotrophic bacterium [59]. Total RNA was extracted from the leaves that transiently expressed m3875 or GFP at 48 h post inoculation (hpi) and used to measure the relative expression levels of NbCNGC23, NbCNGC24, NbCNGC25, NbCNGC26 (A), NbBI-1, NbBI-2 (B), and NbWRKY9 (C) using RT-qPCR. Three biological replicates (each consisting of three technical replicates) were performed. Expression levels of the indicated genes were individually normalized to the internal reference Nbactin. Bars represent the mean ± standard error, and asterisks denote significant differences according to a Student's t test (** p <0.01 and * p <0.05).

Discussion
Bacterial pathogens often secrete proteins (also called effector proteins) that contribute to disease pathogenesis [60]. As a Gram-negative bacterium, CLas does not possess secretion systems T3SS, T4SS and T6SS, but it has the Sec secretion system [26]. To date, a total of 86 CLas-encoded proteins have been determined as Sec-dependent secretory proteins [27]. Although evidence implicates these secreted proteins as being involved in CLas pathogenesis [22,27,34,61,62], information about the contributions of these proteins to the virulence of CLas is limited. In this study, we identified a new Sec-dependent secreted protein (CLIBASIA_03875) from CLas. The transient overexpression of m3875, the mature form of CLIBASIA_03875, completely suppressed INF1-and BAX-triggered HR PCD, as well as induced dwarfing, leaf deformation and mosaics in N. benthamiana. This is reminiscent of a recent study in which CLIBASIA_03875 was predicted to encode an effector protein and expressed relatively high mRNA levels in both the leaf and root tissues of Duncan and Cleopatra, two citrus species that differ in their tolerance to HLB [33]. This study, along with our current data, strongly indicates the critical role of CLIBASIA 03875 in CLas pathogenesis.
The plant HR PCD is initiated as a strong immune response to block the infection of biotrophic/hemibiotrophic pathogens [63]. Conversely, to enable a successful infection, the biotrophs and hemibiotrophs often evolve suites of PCD-suppressing effectors to impede the plant immune response. For example, the majority of the type III secretion proteins produced by the biotrophic bacterium P. syringae pv. tomato DC3000 are devoted to suppressing the host HR PCD [64][65][66]. In addition, hemibiotrophic Phytophthora species usually express a significant number of effectors to inhibit PCD before switching to a necrotrophic infection phase, thereby facilitating the initial infection in living host cells [45,67]. In this study, the secreted protein m3875 was identified to be a PCD suppressor, suggesting that the obligate biotroph CLas utilizes the tactic frequently employed by the biotrophic/hemibiotrophic pathogens to prevent the plant HR PCD.
The transient overexpression of m3875 in N. benthamiana regulates the transcriptional levels of a range of defense-related genes, including NbCNGC23-26, NbBI-2 and Nbwrky9. NbCNGC23-26 was of particular interest, since this gene is essential for conducting Ca 2+ into the cytosol. Ca 2+ is a key regulator that contributes to triggering the HR PCD and ensures its timely and controlled execution. An iIncrease in Ca 2+ influx is therefore an initial protective response following pathogen detection [37,48]. Previous studies have delineated that the effectors typically suppress PCD via targeting pathogen perception, ROS signaling networks, or mitogen-activated protein kinase (MAPK) cascades [7,68]. However, there is little information regarding the mechanism by which PCD-suppressing effectors can interfere with Ca 2+ influx. In this study, we detected that m3875 caused a significant reduction in the expression of NbCNGC23-26, which are all CNGCs group-IVb members of N. benthamiana. It has long been known that the CNGCs group-IVb members of A. thaliana (AtCNGC2 and AtCNGC4) are positive regulators of HR [52,53]. Taken together, we present a novel PCDsuppressing pathway in which m3875 indirectly reduces calcium influx by reducing the expression of NbCNGC23-26 to ultimately suppress PCD.
In contrast to NbCNGC23-26, NbWRKY9 and NbBI-2 were significantly upregulated during the transient expression of m3875. NbWRKY9 is required to activate the RBOH (respiratory burst oxidase homologue)-dependent ROS burst in N. benthamiana [57]. The ectopic overexpression of AtWRKY33, the closest Arabidopsis WRKY to NbWRKY9, in A. thaliana has been shown to improve disease resistance to the necrotrophic fungi Botrytis cinerea and Alternaria brassicicola [58]. However, the AtWRKY33-overexpressing plants exhibited increased susceptibility to the biotrophic bacterium P. syringae [58]. Similarly, the overexpression of HvBI-1 in barley increases resistance to the necrotrophic fungus Fusarium graminearum [69] but confers increased susceptibility to the biotrophic fungus Blumeria graminis f. sp. hordei [69][70][71][72]. These observations suggest that both WRKY33 and BI-1 act as so-called susceptibility (S) factors for biotrophs [73]. In addition, BI-1 is a well-known PCD suppressor [55], although BI-1 acts to trigger cell death when transiently overexpressed under the control of a double 35S promoter and a translation enhancer TMV Omega leader sequence [56]. Therefore, the m3875-enhanced expression of NbWRKY9 and NbBI-2 may reduce the host innate immune response against biotrophs, indicating a potential strategy for CLas infection. Notably, m3875 did not alter the expression of NbBI-1, an additional BI-1 homolog in N. benthamiana known to interact with ATG6 to regulate autophagy and PCD [56], suggesting that NbBI-1 and NbBI-2 have different roles in the susceptibility to the biotrophic pathogen.
The putative m3875 is a small protein comprising only 22 amino acids. However, site-directed mutagenesis of the charged amino acid (D 7 , R 9 , R 11 or K 22 ) or even the simultaneous substitution of D7, R9, and R11 with alanine residues within m3875 had little effect on its functions in PCD suppression and plant development. While there have been repeated observations that proteins exist with robustness against site mutations, without conferring loss of function [74,75], the current data are still surprising considering the small size of m3875. Previous studies have elucidated that this mutational robustness promotes the evolution of protein [76,77], resulting in more robust proteins with a greater capacity to acquire new or improved functions over large evolutionary time-scales [78]. In view of the positive correlation between mutational robustness and evolvability [79,80], it is plausible that the robustness of m3875 would facilitate its evolvability, thereby offering benefits to CLas for its survival and colonization in host plants.
Collectively, in this study, the mature form of a new CLas-encoded Sec-dependent secretory protein was shown to inhibit HR PCD formation in N. benthamiana, and the functional mechanisms behind this interaction were preliminarily investigated. It is known that CLas infects all commercial citrus species and scion cultivars regardless of their rootstock [17,18]. Unfortunately, the underlying mechanisms regarding how this bacterium overwhelms plant immune responses remain elusive. In this study, the identification of m3875 as the PCD suppressor suggests a novel strategy of CLas to antagonize the host innate immunity, which may extend the current understanding of CLas virulence mechanisms, as well as host plant susceptibility, and therefore, merits further investigation.

Plants, Microbial Strains, and Growth Conditions
The N. benthamiana plants were grown in a greenhouse at room temperature with a photoperiod of 18 h light/6 h darkness. Escherichia coli strains DH5α and BL21 were grown on Luria-Bertani (LB) medium at 37 °C, while Agrobacterium tumefaciens strains EHA105 and GV3101 were on Yeast Extract Broth (YEB) supplemented with 50 μg/mL rifampicin, 50 μg/mL kanamycin, and 2 μg/mL tetracycline as needed.

In Silico Analysis of the Signal Peptide
The putative SP of CLIBASIA_03875 was predicted using the online algorithms of PrediSi [31] and SignalP version 3.0 [32] with default settings for Gram-negative bacteria.

Alkaline Phosphatase (PhoA) Assay
A PhoA assay system [34] was utilized to verify the putative SP of CLIBASIA_03875. Briefly, the gene encoding CLIBASIA_03875 was cloned with the primer pair 3875F/3875R (Supplementary Table  S1) and inserted into vector pMD18-T (TaKaRa, Kusatsu, Japan) to generate pMD-3875. Using pMD-3875 as a template, PCR was performed with the primers 3875SP-F/3875SP-R (Supplementary Table  S1). The PCR amplified the DNA sequence of the CLIBASIA_03875 N-terminal 35 amino acids, which contained the putative SP, followed by six more downstream residues. The resulting PCR product was ligated into NdeI-HindIII-digested pET-mphoA [34], generating a construct pET-3875SP-mphoA containing an in-frame gene fusion between the SP and mphoA genes. The pET-3875SP-mphoA was then introduced into E. coli BL21 chemically competent cells, followed by the detection of the PhoA activity of the transformants on indicator LB agar with 90 μg/mL BCIP, 100 mM IPTG and 75 mM Na2HPO4. The E. coli BL21 cells with the pET-phoA plasmid were used as a positive control, while the negative control was the BL21 cells that harbored pET-mphoA. After 6-10 h incubation at 37 °C, the change to blue color of the transformants indicates PhoA activity, while the colonies that remained white were deemed to lack PhoA.

Subcellular Localization of m3875 in Plant Cells
The sequence of m3875 was amplified using the primers m3875gfp-F/m3875gfp-R (Supplementary Table S1) and ligated into KpnI-XhoI-digested pCAMBIA1300-35S-GFP. The resulting construct pm3875-GFP could express the C-terminal GFP fusion protein m3875-GFP. This construct was then transformed into A. tumefaciens EHA105, followed by the agroinfiltration of 4week-old N. benthamiana as previously described [34]. GFP fluorescence of the infiltrated leaves was visualized at 60 h post inoculation (hpi) with a LSM700 confocal microscope (Zeiss, Munich, Germany).
The PCD suppression assay was performed as previously described [45]. Briefly, A. tumefaciens cells harboring pPVX-m3875, pPVX-A7, pPVX-A9, pPVX-A11, pPVXA7-11, or pPVX-A22 were first infiltrated into the newly expanded leaves of six N. benthamiana plants at the six-to seven-leaf stage. A negative control of A. tumefaciens cells harboring pPVX-GFP was used. At 24 hpi, the initially infiltrated sites were challenged with A. tumefaciens cells harboring the pGR-Bax or pGR-INF1. After 72 hpi, one third of the infiltrated leaves were detached from the plants to detect H2O2 accumulation with 3,3′-diaminobenzidine (DAB) staining as previously described [81], while the remaining leaves were observed for up to 5 days dpi to record the symptoms of cell death. The experiment was repeated three times.
To further determine the virulence of m3875, A. tumefaciens cells harboring pPVX-m3875 or the alanine-substitution m3875 mutants were infiltrated into six N. benthamiana seedlings at the three-to four-leaf stage, as previously described [28]. A. tumefaciens cells harboring empty pGR107 were used as a control. At 15 dpi, the systemic symptoms of the infiltrated plants were recorded, and the systemically infected leaves at a similar developmental stage were harvested to prepare the total RNA. A Northern blot analysis was subsequently performed to detect viral RNAs using a DIG-labeling RNA probe complementary to the PVX coat protein-encoding gene (GenBank No. NC_011620) as previously described [34]. The intensities of the RNA bands were quantified using Image J Version 1.45S (NIH, Bethesda, MD, USA). The level of accumulation of each construct was quantified as the ratio of the intensity of the viral gRNA to that of the corresponding 28S rRNA and normalized to that of PVX. Each construct was evaluated in three independent experiments.

RT-qPCR Analysis
A. tumefaciens cells harboring pPVX-m3875 or pPVX-eGFP were infiltrated into the fourth and fifth leaves of six N. benthamiana plants at the six-leaf stage. At 48 hpi, the infiltrated leaf tissues were collected and pooled for total RNA extraction using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. Immediately, total RNA was reverse-transcribed into first strand cDNA using the PrimeScript TM RT reagent Kit (TaKaRa, Kusatsu, Japan) with oligo (dT) as the primer. qPCR was performed on an ABI StepOne plus Real-time PCR instrument (Applied Biosystems, Foster City, CA, USA) with a 20 μL reaction system containing 5 μL of 2 X SYBR Premix Ex Taq (TaKaRa), 2 μmol/L of each gene-specific primers (listed in Supplementary Table S1), 0.5 μL of the cDNA sample, and 0.2 μL of Rox Reference Dye II (TaKaRa, Kusatsu, Japan). The reactions were performed using the following program: 95 °C for 5 min, 40 cycles of 95 °C for 15 s, 55 °C for 30 s, and 72 °C for 30 s. The N. benthamiana actin gene (Genbank No. JQ256516.1) was used as an internal reference. The experiment was performed with three biological replicates (each consisting of three technical replicates). Finally, the 2 -ΔΔCt method [82] was utilized to calculate the relative gene expression values, which were subsequently transformed to fold-change and plotted in the figures. Statistical analyses of all data were conducted using the Student's t-test (SPSS 10.0, Chicago, IL, USA).