Multiple Classes of Immune-Related Proteases Associated with the Cell Death Response in Pepper Plants

Proteases regulate a large number of biological processes in plants, such as metabolism, physiology, growth, and defense. In this study, we carried out virus-induced gene silencing assays with pepper cDNA clones to elucidate the biological roles of protease superfamilies. A total of 153 representative protease genes from pepper cDNA were selected and cloned into a Tobacco rattle virus-ligation independent cloning vector in a loss-of-function study. Silencing of 61 proteases resulted in altered phenotypes, such as the inhibition of shoot growth, abnormal leaf shape, leaf color change, and lethality. Furthermore, the silencing experiments revealed that multiple proteases play a role in cell death and immune response against avirulent and virulent pathogens. Among these 153 proteases, 34 modulated the hypersensitive cell death response caused by infection with an avirulent pathogen, and 16 proteases affected disease symptom development caused by a virulent pathogen. Specifically, we provide experimental evidence for the roles of multiple protease genes in plant development and immune defense following pathogen infection. With these results, we created a broad sketch of each protease function. This information will provide basic information for further understanding the roles of the protease superfamily in plant growth, development, and defense.


Introduction
Proteases catalyze the hydrolytic cleavage of peptide bonds, which are present in all living organisms and play crucial roles in many biological processes [1]. Higher plants are autotrophic organisms that can synthesize all of their organic molecular components from inorganic nutrients without digestion of heterotrophic protein. However, hundreds of proteases are encoded by the plant genome, suggesting that these enzymes have essential roles in various plant processes including responses to developmental and environmental cues, metabolism, and immunity [2][3][4].
Proteases are classified into five families according to their catalytic activity, namely the cysteine, serine, threonine, metalloand aspartic proteases based on their nucleophile and oxyanion stabilizer [5]. Depending on the catalytic mechanism, serine, cysteine, and threonine proteases use a portion of their amino acids as the catalytic site for the nucleophile, whereas metallo-and aspartic proteases use an activated water molecule as the nucleophile [6].
Several members of the serine, cysteine, and threonine protease families have reported regulatory roles in the development and morphogenesis of different stages of the plant life cycle. For instance, defective kernel 1 (DEK1), a cysteine protease, is involved in an epidermal cell fate stage, and a dek1 mutant in maize (Zea mays) causes a lethal defect in which kernels lack an aleurone layer [7,8]. Two other cysteine proteases, ubiquitin Cterminal hydrolase1 (UCH1) and 2 (UCH2), regulate branching in Arabidopsis thaliana. The uch1/uch2 mutant strain exhibits reduced branched primary inflorescence under short day conditions [9]. In addition, the serine protease abnormal leaf shape 1 (ALE1) function in embryo cuticle deposition as evidenced by the lack of cuticle in embryos of ale1 mutants [10].
Along with these roles in development, many aspartic and metalloproteases have been associated with immune defense in plants. Constitutive disease resistance 1 (CDR1) from Arabidopsis, an aspartic protease localized in the apoplast of plant cells, induces local and systemic defense responses. Overexpression of CDR1 caused dwarfing and resistance to the virulent pathogen Pseudomonas syringae and CDR1 deficiency resulted in increased susceptibility to infection by this pathogen [2,11]. Aspartic proteases (APs) from Solanum tuberosum exhibit antimicrobial activity induced by abiotic and biotic stress by interacting directly with microbial cell surface proteins followed by membrane permeabilization [12]. Furthermore, APs from barley, which are specifically expressed in nuclear cells during degeneration, is reported to be involved in programmed cell death [13]. Moreover, matrix metalloproteases 2 isolated from Glycine max is induced by biotic stresses such as fungal or bacterial pathogens [14]. Altogether these reports suggest that proteases may play roles in the immune response against pathogenic infection in plants.
Functional characterization of plant protease gene family members has been conducted in a fragmentary scale to uncover a specific function for each protease. To understand the broad role of proteases in plants, we isolated the protease superfamily from pepper (Capsicum annuum) expressed sequenced tags (ESTs, http:// genepool.kribb.re.kr/pepper/) on the basis of MEROPS (http:// merops.sanger.ac.uk/) classification [15,16]. We analyzed multiple pepper ESTs to gain a better understanding of proteases involved in plant development and immune defense. In an effort to identify protease function, a Tobacco rattle virus (TRV)-based virus-induced gene silencing (VIGS) technique was adapted as an easy and powerful method for silencing uncharacterized genes on a large scale [17]. To date, VIGS has been used successfully in studies of Nicotiana benthamiana, Solanum lycopersicum, Solanum bulbocastanum, C. annuum, Arabidopsis, and Papaver somniferum among other plants [18][19][20][21][22][23].
In this study, we present the phenotypic profile of N. benthamiana plants following silencing of multiple proteases. We further characterized some selected proteases with roles in pathogeninduced plant cell death. This study of 153 different protease genes from pepper plant could provide an overall sketch of the biological roles of plant proteases in growth, development, and defense against pathogens.

Plant Growth and Agroinfiltration
N. benthamiana plants were grown in a 4.09 inch soil pot. The pots were grown in a growth chamber at 22uC under a cycle of 16 h light and 8 h dark. TRV-based VIGS on N. benthamiana was performed as described by Dong et al. [24]. The phenotypic changes of protease-silenced plants were observed at 4 or 5 weeks after infiltration.

Cloning of Pepper EST for VIGS
Pepper ESTs were amplified with: 59-GACGACAAGACCCT (adaptor sequence) -GTAATACGACTCACTATAGGGC (pBluescript SK-specific sequence: ESTs were packaged into pBluescript SK-; Stratagene) -39 and 59-GAGGAGAA-GAGCCCT (adaptor sequence) -CGCTCTAGAACTAGTG-GATCC (pBluescript SK-specific sequence)-39. The PCR products were purified with DNA Clean and Concentrator TM (Zymo Research) to remove primers and nonspecific PCR products. A total 15 fmol of purified PCR product was treated with T4 DNA polymerase (Novagen) in 106reaction buffer containing 5 mM dATP at 22uC for 30 min and 70uC for 20 min for inactivation of T4 DNA polymerase. The TRV-LIC vector was digested with PstI and treated with T4 DNA polymerase similarly with dTTP instead of dATP [24]. A total 15 fmol of T4 DNA polymerase treated PCR products and TRV-LIC vector were mixed and incubated at 65uC for 1 min and slowly cooled down to room temperature. Then, the mixture was transformed into E. coli DH5a competent cells. Transformants were tested by PCR amplification using primers 59-TGTTACT-CAAGGAAGCACGATGAGCT-39 and 59-CAGGCACG-GATCTACTTAAAGAACGTAG-39.PCR products with EST insertions were confirmed by DNA sequencing (NICEM, http:// nicem.snu.ac.kr/).

RNA Extraction and qRT-PCR Analysis
Total RNA was extracted from leaves of N. benthamiana plant using TRI Reagent (Molecular Research Center) according to the manufacturer's instruction. First-strand cDNA was synthesized using 5 mg of total RNA in a mixture with anchor primer (oligo-dT) and SUPERSCRIPTH II Reverse-Transcriptase (Invitrogen) according to the manufacturer's protocol. RT-PCR products were used for quantitative RT-PCR to monitor gene expression levels in protease-silenced plants. Triplicate samples from 3 independently silenced plants of cDNA were analyzed using Rotor-Gene 6000 apparatus (QIAGEN) with SYBR Green (Invitrogen) according to manufacturer's instructions. All calculations and statistical analyses were demonstrated as described by the manufacturer. Primers for qRT-PCR of specific proteases were presented in Table S3. To normalize the expression levels, transcript level of NbActin gene was used as a control.

Pathogen Preparation and Inoculation
The bacterial strains used for this study were P. syringae pv. tomato T1 which causes hypersensitive response (HR) cell death and P. syringae pv. tabaci which causes wild fire disease on N. benthamiana [29,30]. Both bacterial pathogens were grown in yeast extract peptone (YEP) medium and resuspended in 10 mM sterile MgCl 2 solution. N. benthamiana leaves were pressure infiltrated using a needless syringe with bacterial concentration of OD 600 = 0.2 (P. syringae pv. tomato T1) and OD 600 = 0.005 (P. syringae pv. tabaci).
Changes of the HR cell death and symptom development in bacteria-infiltrated plants were observed for 3 and 7 days, respectively [31].
P. syringae pv. tabaci cultures (OD 600 = 0.005, approx. 1610 8 CFU ml 21 ) were resuspended in 10 mM MgCl 2 and infiltrated into the protease-silenced leaves using a 1 ml needless syringe. The growth of bacterial in the leaf was measured on four replicate plants. Two leaf discs (1 cm in diameter) per plant were collected from the P. syringae pv. tabaci infiltrated region of each proteasesilenced and control plants. Numbers of the bacteria were measured by grinding the leaf discs in 10 mM MgCl 2 and plating serial dilutions LB agar plants containing 100 ug ml 21 rifampicin on a daily basis for 2 dpi.

Staining with Trypan Blue
To detect the cell death, leaves were stained with lactophenol/ trypan blue (10 ml glycerol, 10 ml lactic acid, 10 g phenol and 10 mg trypan blue dissolved in 10 ml distilled water). Pathogeninfected leaves were vacuum-infiltrated two-times for 5 min in staining solution and incubated for overnight. Stained leaves were de-stained in 2.5 g ml 21 chloral hydrate (TOKYO KASEI). The de-stained leaves were mounted in 50% glycerol and photographed under a digital microscope (DIMIS M, Siwon Optical Technology, http://www.dimis.co.kr/) [32].

Measurement of Ion Leakage
One and two days after infiltration with P. syringae pv. tomato T1, two leaf discs (1 cm in diameter) were floated in 5 ml of distilled water for 4 h at room temperature. Electrical conductivity was measured using a conductivity meter (EC215, HANNA instruments).

Isolation of Putative Protease Genes in Pepper ESTs
The pepper EST database, which is a useful collection for functional genomics, consists of 122,582 ESTs including 22,011 unigenes, 11,225 consensus sequences and 5,585 full-length cDNAs have been published [15]. From the pepper EST database, putative protease genes (totaling 939 entries, data not shown) were selected computationally using Hmmpfam and Hmmsmart for protease domain prediction based on classifications from MER-OPS. This protease database, displays hierarchical classifications and homologous proteases grouped into species, families, and clans based on evolutionary relationships. MEROPS can also be used to obtain information about the substrate and inhibitors of particular protease [16].
To select representative protease genes, sequence alignments using TBLASTX, BLASTX, and BLASTP were performed to avoid silencing multiple genes with homology. Ultimately, 153 pepper cDNAs were selected as representatives of protease genes for VIGS. The 153 pepper cDNA is presented with its EST ID, MEROPS ID, e-value, and the accession numbers of corresponding from N. benthamiana, S. lycopersicum, S. phureja, Arabidopsis genome database and other organisms are summarized in Table S1 [25][26][27][28]. The selected protease genes consisted of 15 aspartic, 27 cysteine, 28 metalloproteases, 72 serine, and 11 threonine proteases as classified by MEROPS. All of the selected cDNAs were cloned into the TRV-LIC vector for the VIGS assay [24].

Phenotypic Analysis of Protease-silenced Plants
VIGS experiments were performed in N. benthamiana plants with the selected 153 cDNAs. Since pepper and N. benthamiana belong to the same solanaceae family, coding sequences from both plants exhibit a high degree of similarity. Previous studies have also demonstrated successful heterologous silencing of tomato genes in N. benthamiana [24]. Therefore, it is not surprising that several genes from pepper have been characterized in N. benthamiana using the VIGS assay [33,30,32]. The pepper ESTs investigated in this study are listed with their corresponding N. benthamiana proteases in Table 1.
The developmental phenotypes of protease-silenced plants were observed 3 or 4 weeks after VIGS and representative examples were shown ( Fig. 1 and Table S2). These phenotypes were grouped into six classes, including no difference, inhibition of shoot growth, inhibition of shoot growth with abnormal leaf shape, lethality, leaf color change, and abnormal leaf shape (Fig. 1). The largest class, of which proliferation-association protein 1 silencing is an example (Ncn881, Fig. 1), showed no phenotypic differences compared to the control plant and was composed of 92 proteases. The second largest class contained 35 proteases and gene silencing resulted in severe growth retardation relative to control plants ( Fig.  S1A), shown in the silencing phenotype of prolyl aminopeptidase (Ncn1998, Fig. 1). The third largest class consisted of 11 proteases and the silencing phenotype was characterized by severe stunting with crinkled leaves (Fig. S1B). This is evidenced by an ortholog of At5g22030 (Arabidopsis ubiquitin-specific protease 8) (Ncn7292, Fig. 1). The fourth largest class, composed with 6 proteases (Fig. S1C), was characterized by lethality as shown by the silencing phenotype of proteasome subunit beta 3 (Ncn2134, Fig. 1). The fifth class contained 6 proteases ( Fig. S1D) and silencing resulted in photobleaching and yellowing of leaves as shown by chloroplast (stromal) processing peptidase silencing (Ncn4707, Fig. 1). This altered phenotype showed similar developmental phenotypes to those in previous studies of downregulation of the gene product [34,35]. The smallest class was made up of 3 proteases (Fig. S1E) that, when silenced, lead to disruption of leaf development resulting crinkled leaves as shown by silencing an ortholog of At1g49050 (aspartyl protease family) proteintype peptidase (Ncn9390, Fig. 1).
Nearly half of the members exhibited altered phenotypes in aspartic, cysteine and metalloproteases family (aspartic: 9 out of 15, cysteine: 14 out of 27, metallo-: 13 out of 28) (Table S2). However, the serine protease family was the least affected group with only 16 out of 72 members displaying a phenotypic change compared to that of control plants.
Interestingly, 9 out of 11 threonine protease family members produced altered developmental phenotypes, such as severe stunting, crinkled leaves, and lethality. Consistant with these results, previous studies have demonstrated that threonine proteases play a crucial role in the 26S proteasome and regulation of all aspects of plant development [36,37]. These data suggest that threonine proteases have an essential role in plant growth and development.

Temporal Change in Hypersensitive Response (HR) Cell Death in Protease-silenced Plants
To investigate the role of proteases in immune defense, an avirulent pathogen P. syringae pv.tomato T1 capable of inducing HR cell death in N. bethamiana was introduced into protease-silenced plants [38]. Three weeks after silencing, plants were inoculated with a suspension of P. syringae pv. tomato T1 at a density of OD 600 = 0.2. HR cell death was observed at days 1 and 2 postinoculation (1 dpi and 2 dpi, respectively) in TRV-DGFP control plants. After 1 dpi, 26 protease-silenced plants showed temporally advanced HR cell death compared to that of control plants while the control plant showed HR cell death at 2 dpi. This group consisted of 11 cysteine, 7 metallo-, 5 serine, 2 threonine, and 1 aspartic protease family members ( Fig. 2A and Fig. S2A). Cysteine protease deficiency produced the greatest effect on HR cell death in response to avirulent pathogen. This result is consistent with previous reports that demonstrated a crucial role for cysteine proteases in the plant-pathogen/pest interaction [39,40]. After 2 dpi, delayed HR cell death was observed in 5 serine, 2 aspartic, and 1 cysteine protease-silenced plants. Delay of the HR cell death was until 3 dpi. Interestingly, none of the metallo-and threonine protease family members exhibited delayed HR cell death in genesilenced N. benthamiana plants ( Fig. 2B and Fig. S2B).

Change of Disease Symptom Development in Proteasesilenced Plants
To better understand the role of proteases in the immune response to pathogen invasion, protease-silenced N. benthamiana plants were also infected with a virulent bacterial strain. P. syringae pv. tabaci, a fire blight pathogen isolated from tobacco plant, produces a monocyclic ß-lactem named tabtoxin and causes loss of chlorophyll, accumulation of ammonia and finally inactivation of glutamine synthetase in plant [41,42,29]. Appearance of the susceptible host plant infected with P. syringae tabaci starts with water-soaking in the apoplast of the infected tissues, localized tissue necrosis and various tissue discolorations [29]. Proteasesilenced and control plants were inoculated with P. syringae pv. tabaci at an OD 600 = 0.005. Disease symptoms were observed at day 3 post-inoculation (3 dpi) in both groups. Sixteen proteasesilenced plants constituting 7 cysteine, 5 serine, 3 metallo-and 1 threonine proteases showed delayed development of disease symptoms ( Fig. 3 and Fig. S3). Disease symptom was delayed until 4 to 5 dpi. Aspartic proteases were not found in this experimental group.

Characterization of Selected Protease-silenced Plants following Pathogens Infection
To confirm the level of gene silencing in each protease subfamily, quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was conducted using primers that anneal outside the sequence used for silencing. Our data indicate that the level of mRNA transcript for all protease members was decreased at least 70% by gene silencing compared to control plants (Fig. S4).
Three protease subfamilies, which showed modulated pathogen responses when silenced, were selected for further characterization, including M18 (metallo-type, aminopeptidase family), M24 (metallo-type, methionyl aminopeptidase family) and S01 (serine type, chymotrypsin family). M18 family is consisted with a member showing delayed symptom development whereas M24 and S01 families are each constituted with a member resulted in enhanced and delayed HR cell death response compared to the control. Silencing of each representative member showed enhanced and delayed HR cell death and delayed disease symptom onset compared to the control, respectively.
The M24 and S01 protease subfamilies possessed with a member showing enhanced and delayed HR cell death after infection with an avirulent pathogen, P. syringae pv. tomato T1, compared to control plants at 1 dpi and 2 dpi (upper panel of Fig. 4A and Fig. 5A). The control plants represented the HR cell death at 2 dpi, while the enhanced HR cell death of Ncn6721 was carried out at 1 dpi while the HR cell death of Ncn5036 was delayed until 3 dpi. Trypan blue staining was performed to assess cell death in the M24 and S01 protease subfamily member (lower panel of Fig. 4A and Fig. 5A). Our data reveal that cell death following Ncn6721-silencing was 20% higher than the control plants and other silenced member from the M24 protease subfamily at 1 dpi (Fig. 4B), suggesting a critical role for this gene product in plants. In contrast, cell death in Ncn5036-silenced plants was 20% lower than the control and Ncn964-silenced plants at 2 dpi (Fig. 5B). The ion leakage of the pathogen-infiltration area was significantly high in Ncn6721and low in Ncn5036-silenced plant than that of the control plant ( Fig. 4C and 5C).
Gene-specific silencing was confirmed with gene-specific primers in the Ncn6721-silenced plants. Among the closest paralogs, the transcript levels of Ncn2132, Ncn881 and Ncn9826 were not reduced in Ncn6721-silenced plants (Fig. 4D). Additionally, the level of Ncn964 mRNA was not decreased in Ncn5036-silenced plants (Fig. 5D). The transcript levels of other M24 and S01 protease members in each protease-silenced plant are shown in Fig. S5. Interestingly, Ncn6721 is predicted to encode methionyl aminopeptidase 2 (MAP2), a cytosolic enzyme responsible for protein N-terminal methionine excision (NME). This cytoplasmic NME machinery displays the same specificity in the plant and animal kingdom and has been established to be required for normal development [43]. Functional studies of MAP2 in plants during development were performed in Arabidopsis [44]. Our finding indicates that MAP2 may play a role in pathogen-induced HR cell death in plants. Moreover, Ncn5036 is described as an At5g40200-type peptidase and predicted to encode degradation of periplasmic proteins 9 (DegP9), a serine type ATP-dependent protease. In Arabidopsis, DegP9 protease is involved in the degradation of D1 protein from photosystem II following photoinhibition and is up-regulated during senescence [45][46][47]. Several recent studies have confirmed that plant chloroplasts are associated with plant pathogen-induced cell death [48]. This result indicates that DegP9 is associated with HR cell death in plants.
The M18 protease subfamily contains a member that is involved in delayed disease symptom development onset following infection with the virulent bacterial pathogen P. syringae pv. tabaci (Fig. 6). Disease symptom of Ncn10583-silenced plant was carried out at 4 to 5 dpi, whereas the control and Ncn8326-silenced plant showed disease symptom at 3 dpi (upper panel of Fig. 6A). Trypan blue analysis revealed that Ncn10583-silenced plants showed 30% less death than the control and Ncn8326-silenced plants at 3 dpi (lower panel of Fig. 6A and Fig. 6B). The bacterial cell growth of Ncn10583-silenced plants was one and a half-fold lower than in control-silenced plants (Fig. 6C). Gene-specific silencing of Ncn10583 was also confirmed by qRT-PCR. The transcript level of the most closely related protease Ncn8326 was not reduced in Ncn10583-silenced plants (Fig. 6D). The level of Ncn10583 mRNA in Ncn8326-silenced plants is shown in Fig. S5. Ncn10583 is an At5g60160-like peptidase and predicted to encode a Zn-dependent exopeptidase superfamily protein also known as aspartyl aminopeptidase (AAP), which is distributed widely in animals and plants and involved in a variety of physiologically important processes [49]. Most studies of AAP have been limited to mammals and focused on its functions and enzymatic activity [50,51]. Nevertheless, our data imply that AAP plays a role as a susceptibility factor of plants against virulent bacterial pathogens.

Discussion
This study presents a larger view of the role of protease in plants. Some of our results pertaining to the function of these enzymes were consistent with previous reports. Among the 61 proteases identified to be involved in the plant developmental process by loss-of function studies, 39 belong to the serine, cysteine, or threonine protease families. Several members from these families have been demonstrated to play roles in plant growth and development [7,8,10,9].
Based on the EST reference dataset, our analysis of representative proteases provided further insight into the role of these enzymes in plant development. Our data demonstrate that silencing of Ncn1998 (serine type) resulted in inhibition of shoot growth (Fig. 1, Fig. S1A, and Table S2). Ncn1998 is predicted to encode a prolyl aminopeptidase (PAP) with a significant degree of sequence identity to other solanaceous PAP proteins in other species, namely S. lycopersicum PAP (88%, Solyc11g044310.1.1), N. benthamiana PAP (74%, NbS00039951g0005.1), Medicago truncatula PAP (92%, XP003593347.1), and Vitis vinifera PAP (84%, XP002271289.2). Despite the ubiquity of PAP in nature, its role remains poorly understood in plants. To date, studies have only shown that the mRNA level of PAP is increased in the shoots of triticale plants in response to suboptimal conditions such as drought and high saline [52]. Our results are the first to indicate that PAP may function in plant growth and development, including the control of shoot apical meristem formation. More specifically, proline, a product of PAP, has various physiological Immune-Related Proteases and Cell Death in Pepper PLOS ONE | www.plosone.org roles in plants, animals, and micro-organisms [53,54]. Our results suggest that decreased expression of PAP may affect proline metabolism and cause an altered phenotype in Ncn1998-silenced plants. Our experiment was limited to observing the pathology of the silenced plant for only 3 or 4 weeks; therefore, either a longer observation period or examination of knock out mutants may needed to determine the role of PAP in reproductive growth processes such as floral development.
Multiple proteases have roles in the pathogenesis and immune defense of plants. Our study identified 34 proteases that function in avirulent pathogen-induced HR cell death; 7 metallo-and 12 cysteine family members represented these proteases. Moreover, we identified 16 proteases, of which 7 were members of the cysteine family, involved in the development of pathogen-induced disease symptoms. Many members of the aspartic, cysteine, and metalloprotease families have been reported as associated with plant defense responses [2,14,39,40,11]. Our results are in close agreement with previous reports. Interestingly, silencing of only 3 aspartic proteases affected HR cell death in response to infection by an avirulent pathogen, while aspartic proteases did not appear to play a role in disease symptom development in plants.
GAT generates ammonia from glutamine amide nitrogen and transfers it to acceptors [55]. One bacterial pathogen we used to infect GAT-1 like protein (Ncn7004)-silenced plants was P. syringae pv. tabaci, which produces tabtoxin to inhibit the plant enzyme glutamine synthetase (GS) [56]. GS is a key enzyme in the photorespiratory nitrogen cycle. Inhibiting this enzyme in plants results in chlorosis, which arises due to necrotic lesions caused by bacteria on leaves [57]. Moreover, tabtoxin-induced symptoms are associated with the accumulation of ammonia under photorespi-    means6SD (n = 3). The values were normalized to NbActin and were calculated to the control. Similar results were obtained from at least two experiments. One representative experiment is shown. Asterisks indicate significant differences relative to the control as determined by Student's t test (*P,0.05, **P,0.01, ***P,0.001). doi:10.1371/journal.pone.0063533.g004 ration [58]. Challenging Ncn7004-silenced plants with virulent bacterial pathogen resulted in delayed disease onset, possibly due to blocking glutamine amide nitrogen from supplying ammonia. This decreased level of ammonia may merely slow the development of disease since ammonia accumulation was not enough to cause the symptom. To date, studies on GAT have mainly focused on its biochemical and structural properties [55,59]. Thus, our finding on the function of GAT may provide additional insight into the molecular basis of disease symptom development in tobacco plants.
In addition, the Ncn7004-silenced plant showed other phenotypes besides cell death and defense responses. The Ncn7004silenced plant displayed inhibited shoot growth and abnormal leaf shape (Fig. S1B). Ncn7004 consists of a GAT type 1 domain and could function similar to what has been reported previously, such as the effect of lower nitrogen metabolism on normal plant growth [60]. Moreover, GAT links nitrogen metabolism to the biosynthetic pathways of several important compounds [55]. Nitrogen supply is also known to affect plant disease development [60]. In previous studies, GAT was reported clustered with At1g15040, which belongs to the same subfamily as GAT1 (GATase 1_2) in Arabidopsis. In the Arabidopsis genome, there are 30 putative genes encoding GAT1 and a previous study suggested that At1g15040 has specific functions in plant branching control [61]. Our results also reveal that Ncn7004 might have roles in plant developmental processes. Furthermore, HR cell death is a kind of programmed cell death (PCD), which is important throughout plant development stages such as senescence, vacuole formation for the degradation of cellular contents, normal seed development, leaf sculpting, cell death in the root cap, and shaping the sexual and nonsexual organs of the flower [62]. GAT1-like protein may play both processes, roles in plant developmental and pathogeninduced cell death processes, as PCD. Together with previous studies, our results suggest that GAT1-like protein (Ncn7004) is closely associated with plant development and the regulation of cell death in response to pathogen invasion in plant immunity.
A number of plant proteases have been implicated in HR development. Some of them were reported as have roles in regulating HR and some of them were firstly described in this study. Orchestration of the multiple proteases in HR cell death is not quite clear yet but each protease may has distinct and diverse roles. From previous studies, some proteases can act at the level of signal perception, transduction and execution in defense response [63]. There are some evidences that type I metacaspase is a initiator or activator of type II metacaspase for progression of PCD [64]. Moreover, AtMC1 (Arabidopsis metacaspase 1) works as a positive regulator, whereas AtMC2 function as a negative regulator of cell death [65]. Even with these studies, there are still more to be found in the events of the chain leading to cell death. Our data only have given the evidence for one-to-one interaction with a protease and a pathogen. Comprehensive studies on the functions of these proteases may be required to uncover the multiple roles of proteases in the complicated machinery of HR cell death in plants. Therefore, our data might have given evidence for the starting Values are means6SD (n = 3). Relative intensity was calculated to the control. C. Ion leakage from leaf discs of S01 protease subfamilysilenced N. benthamiana plant following non-host pathogen infection. Values are means6SD (n = 3). D. Transcript level of S01 protease subfamily member was examined in Ncn5036-silenced plant. Values are means6SD (n = 3). The values were normalized to NbActin and were calculated to the control. Similar results were obtained from at least two experiments. One representative experiment is shown. Asterisks indicate significant differences relative to the control as determined by Student's t test (*P,0.05, **P,0.01, ***P,0.001). doi:10.1371/journal.pone.0063533.g005 point of the cell death event as the diverse roles of protease in HR cell death.
In this study, we provide a phenotypic sketch of protease function in plant developmental processes, pathogen-induced disease symptom development, and HR cell death. Our studies, together with previously published data, provide additional insights into the unknown role of proteases and clues for further investigation into the molecular functions of the protease super family in plants.  Trypan blue staining is in the same area as the photographs were taken at 3 dpi (lower panel). Scale bars = 0.3 mm. B. Relative intensity of trypan blue stained N. benthamiana plants. Relative intensity was calculated to the control. Values are means6SD (n = 3). C. Bacterial cell growth in control and M19 protease subfamily-silenced N. benthamiana plants infiltrated with P. syringae pv. tabaci (1610 8 CFU ml 21 in MgCl 2 ) was determined at 0, 1 and 2 dpi. Values are means6SD (n = 3). Different letters indicate significant differences at the 95% level by Duncan's multiple range tests. D. Transcript level of M18 protease subfamily member was examined in Ncn10583-silenced plant. Values are means6SD (n = 3). The values were normalized to NbActin and were calculated to the control. Similar results were obtained from at least two experiments. One representative experiment is shown. Asterisks indicate significant differences relative to the control as determined by Student's t test (*P,0.05, **P,0.01, ***P,0.001). doi:10.1371/journal.pone.0063533.g006 members with gene specific primers were confirmed by quantitative RT-PCR. The values were normalized to NbActin and were calculated to the control. Values are means6SD (n = 3). Similar results were obtained from at least two experiments. One representative experiment is shown. Asterisks indicate significant differences relative to the control as determined by Student's t test (*P,0.05, **P,0.01, ***P,0.001). (TIF)