Introduction

Caspases are members of the C14 family of cysteine proteinases that specifically target sites containing aspartate residues (Cohen 1997). Caspases have been extensively studied because of their central role in signaling and executing programmed cell death (PCD) in animals (Cryns and Yuan 1998). PCD also occurs during plant development and the response of plants to different stresses (Beers et al. 2000). Although caspase-like proteinase activities have been identified in plants (del Pozo and Lam 1998; Solomon et al. 1999), no true plant caspase has been identified from them. However, plants and fungi have metacaspases, which are a family of caspase-related proteases that have a p20-like subunit containing a histidine and cysteine catalytic diad (Uren et al. 2000). Plant metacaspases are arginine/lysine-specific cysteine proteinases and thus do not have aspartate-specific proteolytic activity like animal caspases (Vercammen et al. 2004; Watanabe and Lam 2005). Thus, they are not able to cleave animal caspase-specific substrates, suggesting that they are not responsible for plant caspase-like enzyme activity. Despite this, plant metacaspases can trigger PCD, such as during embryogenesis in arabidopsis, where PCD was found to be dependent on the enzyme’s proteinase activity (Bozhkov et al. 2005). Also, two arabidopsis metacaspase genes complemented a metacaspase yeast mutant that mediates PCD, indicating that plant metacaspases can activate downstream proteinases with caspase-like activity required for PCD (Watanabe and Lam 2005). Thus, plant metacaspases appear to play a parallel role as animal caspases in PCD.

Relatively little is known about metacaspases in plant–pathogen interactions. Armando et al. (2003) found that a metacaspase gene, AtMCA1, was expressed in wounded and Pseudomonas syringae-infected arabidopsis leaves, and suggested that they participate as caspase-like enzymes during PCD. Watanabe and Lam (2004) examined three type I and six type II metacaspase genes from arabidopsis and showed that two type I metacaspase genes were up-regulated following infiltration by P. syringae pv. maculicola, while expression of two type II metacaspases was not significantly affected. These metacaspase genes were hypothesized to be involved in the activation of PCD. In tomato, one type II metacaspase gene, LeMCA1, rapidly increased during cell death induced by Botrytis cinerea (Hoeberichts et al. 2003). More direct evidence for a role in plant–pathogen interactions comes from an examination of arabidopsis mutants, which showed that there was increased susceptibility in type I metacaspase mutants and decreased susceptibility in type II metacaspase mutants following inoculation by several Botrytis species (van Baarlen et al. 2007).

Nicotiana benthamiana is becoming widely used by molecular biologists because it exhibits more pronounced and longer lasting virus-induced gene silencing (VIGS) than other plants, which permits a high-throughput forward genetics approach for obtaining functional genomics information (Lu et al. 2003). VIGS is a post-transcriptional gene silencing mechanism for sequence-specific targeting and degradation of plant RNA that spreads systemically in the plant, and it has been successfully used in analyzing the function of a number of genes of N. benthamiana (Robertson 2004), including genes involved in plant–pathogen interactions (Baulcombe et al. 2002).

A hemibiotrophic pathogen of N. benthamiana is Colletotrichum destructivum, which has a biotrophic phase of infection followed by a necrotrophic phase, resulting in the formation of water-soaked spots on leaves and petioles that develop into small necrotic lesions by 72 h post-inoculation (HPI) (Shen et al. 2001). In contrast, inoculation of N. benthamiana with P. syringae pv. tomato DC3000 results in an incompatible interaction, which triggers a non-host hypersensitive response (HR) within 24 HPI (Espinosa et al. 2003). Pseudomonas syringae pv. tomato DC3000 is a pathogen of tomato and arabidopsis causing bacterial speck disease (Preston 2000), and several studies have been done on its incompatible interaction with N. benthamiana (Rommens et al. 1995; Scofield et al. 1996; Sessa and Martin 2000; Espinosa et al. 2003).

This work describes a metacaspase cDNA, NbMCA1, amplified from N. benthamiana leaves infected with C. destructivum. The expression of this gene was examined following a compatible interaction of N. benthamiana with C. destructivum and an incompatible interaction with P. syringae pv. tomato. A portion of NbMCA1 was then cloned into a PVX-based VIGS vector (Ruiz et al. 1998) to determine if silencing the gene would affect the response of N. benthamiana to C. destructivum or P. syringae pv. tomato. Silencing may reduce resistance to P. syringae pv. tomato as the HR is associated with resistance and PCD (del Pozo and Lam 1998; Heath 2000). In contrast, it has been suggested that some pathogens induce host PCD to create dead plant tissue as a nutritional source during disease development (Dickman et al. 2001; Richael et al. 2001), and thus silencing a metacaspase may increase resistance by reducing the amount of nutrients to the pathogen.

Materials and methods

Biological materials and pathogen inoculations

Nicotiana benthamiana plants were grown with 16 h light (150 μmol m−2 s−1) at 25°C and 8 h dark at 17°C. Seedlings of N. benthamiana at the sixth true-leaf stage were used for fungal or bacterial inoculation. C. destructivum N150 P3 (Chen et al. 2002) was cultured for 10 days on sodium chloride-yeast extract-sucrose agar medium (SYAS) (Mandanhar et al. 1986) at 22°C under continuous fluorescent light. Conidia were harvested to make a 1 × 105 conidia/ml suspension in sterile distilled water, which was sprayed onto whole plants. These plants were then incubated at room temperature in containers to maintain high humidity. P. syringae pv. tomato DC3000 was grown overnight in King’s medium B (KB) (King et al. 1954) at 22°C and suspended in 10 mM MgCl2 to 108 CFU per ml. The bacterial suspension was infiltrated into each of the two youngest fully developed N. benthamiana leaves per plant using a needle-less syringe (Klement 1963).

The area of the inoculated leaves was measured with a leaf area meter (Model 3100, LI-COR, Lincoln, NE, USA). For quantifying C. destructivum infection, the number of lesions was counted at 60 HPI, and the lesions per cm2 was calculated. For quantifying P. syringae pv. tomato, 1-cm-diameter leaf discs at 24 HPI were ground in 1 ml 10 mM MgCl2 and then dilution-plated onto KB to determine the number of colony forming units. For both pathogens, two inoculated leaves from each of the three replicate plants per treatment were used.

RNA extraction

Leaf samples were immediately frozen in liquid nitrogen and stored at −80°C. Total RNA was prepared following the method of Chen et al. (2000), except that phenol and chloroform were added just prior to the homogenization buffer, which was composed of 200 mM Tris base, 400 mM KCl, 200 mM sucrose, 35 mM MgCl2·6H2O and 25 mM EDTA, pH 9.0. The RNA was resuspended in 25–50 μL DEPC-treated dH2O and stored at −80°C.

Sequence alignments

Putative metacaspase amino acid sequences were obtained from the GenBank NR database (http://www.ncbi.nlm.nih.gov), The Institute for Genomic Research (TIGR) (http://www.tigr.org), Phytome (http://www.phytome.org), The Arabidopsis Information Resource (TAIR) (http://www.arabidopsis.org), and the Beijing Genomics Institute (BGI) (ftp://ftp.genomics.org.cn/pub/ricedb/SynVs9311/Syngenta/Sequence/FgeneSH). The sequence names and origins are listed in Table 1. The sequences were aligned with CLUSTALX (Thompson et al. 1997) using default parameters to assess sequence identity and locate conserved regions for designing the primers described below. A dendrogram was created that also included the predicted protein sequence of NbMCA1, which was obtained in this study. This unrooted dendrogram was generated with the Neighbor-Joining-bootstrap procedure of CLUSTALX using 1,000 bootstrap replications.

Table 1 Cysteine proteinase amino acid sequences of metacaspase families used in sequence analyses
Full size table

Reverse transcriptase PCR (RT-PCR)

Except for the 3′-RACE described below, single-stranded cDNA was synthesized using Moloney Murine Leukemia Virus (MMLV) reverse transcriptase (Invitrogen, Burlington, ON, Canada) and oligo (dT) primer with total RNA following the manufacturer’s instructions. All PCR reactions were performed using a Touchgene Thermo Cycler (Techne, Princeton, NJ, USA). Except where noted, all RT-PCR reactions were done in 15 μL volumes with 2 μL cDNA, 0.25 units Taq polymerase (Promega, Madison, WI, USA), 10× Taq polymerase buffer, 2 mM dNTPs, 2.5 mM Mg2+, and 1.0 mM of each primer. The PCR conditions were 94°C for 3 min followed by 30 cycles of 94°C for 30 s, 62°C for 1 min, and 72°C for 1 min, and a final extension period of 10 min at 72°C.

A degenerate forward primer, CYP11F1, 5′-GGTGAAGAHGATGAYACTGG and a degenerate reverse primer, CYP11R1, 5′-CCTGAACCRGCATAAACCTC, (H = A/C/T, R = A/G, Y = T/C) were designed based on conserved regions of the plant metacaspase nucleotide sequences: LeMCA1 (AAM51555) from Lycopersicon esculentum, HbLAR (AAD13216) from Hevea brasiliensis and At1g79340 (AAL85992) from Arabidopsis thaliana. The primers were used in RT-PCR with cDNA from leaves of N. benthamiana at 96 HPI with C. destructivum. A second forward primer, CYP11F3, 5′-TTCAACAAACTACAGCACGG, was designed based on the sequence of the RT-PCR product that had been amplified with CYP11F1 and CYP11R1, and was paired with the anchor primer supplied in a 5′/3′ RACE kit (Roche, Indianapolis, IN, USA). The RT-PCR fragment obtained with the CYP11F3 primer and 5′/3′ RACE anchor primer was purified following electrophoresis in 1% TAE agarose gels using the GENECLEAN II Kit (Q-BIOgene, Montreal, QUE), and then cloned into pGEMT-easy (MBI Fermentas, Burlington, ON, Canada).

Relative RT-PCR

Relative RT-PCR was done in 15 μL reaction volumes as described previously except that 0.5 mM of each translation elongation factor 1α (EF-1α) primer and 1.2 mM of each metacaspase-specific primer was used. The EF-1α forward primer, TobefIS 5′-CTCCAAGGCTAGGTATGATG, and reverse primer, TobefA 5′-CTTGGTGGTGCATCTCAAC, were designed to amplify a RT-PCR product of 375 bp (Dean et al. 2002). The metacaspase forward primer, CYP11F2, 5′-CACAAGCAAGGCGACGATGA, and reverse primer, CYP11R3, 5′-CTGAATCGCGTTGCTTAGAG, were designed based on the combined sequence of the portion of the metacaspase gene obtained by RT-PCR with primers CYP11F1 and CYP11R1 as well as the portion obtained by RT-PCR with primers CYP11F3 and the 5′/3′ RACE anchor primer. Primers CYP11F2 and CYP11R3 yielded a 630 bp RT-PCR product.

The PCR conditions were 94°C for 3 min followed by 25 cycles of 94°C for 30 s, 62°C for 1 min, and 72°C for 1 min, and a final extension period of 10 min at 72°C. The RT-PCR products were separated in 1% TAE agarose gels, and the images were saved as TIF electronic image files for quantification using NIH Image (Scion Corporation, Frederick, MD, USA). The band intensities were determined for both the metacaspase and EF-1α RT-PCR products, and compared for each gel lane. Relative expression was determined by taking a ratio of the band intensity of metacaspase over the band intensity of the EF-1α. Four plants were assessed at each time point following inoculation with C. destructivum. To confirm the identity of the RT-PCR products, the NbMCA1 RT-PCR product at 96 HPI with C. destructivum were excised after electrophoresis in 1% TAE agarose gels, purified with GENECLEAN II (Q-BIOgene) and then sequenced (Laboratory Services Division, University of Guelph, Guelph, ON, Canada).

Gene silencing

The RT-PCR fragment cloned into pGEMT-easy (MBI Fermentas) was digested with NotI and then subcloned into pGR106 (Jones et al. 1999), which was also digested with NotI. The pGR106 construct was electroporated into A. tumefaciens GV3101, and N. benthamiana leaves were inoculated following Takken et al. (2000). The A. tumefaciens cultures contained either PVX with a fragment of NbMCA1 (PVX-NbMCA1), PVX without any insert as an empty vector control (PVX vector) or water only as a control for A. tumefaciens. A toothpick was used to inoculate A. tumefaciens at six sites per leaf along the main vein of the two youngest fully developed leaves of N. benthamiana at the fourth true-leaf stage. At 20 days after inoculation with A. tumefaciens, the plants had another four fully developed leaves, and one set of plants were spray inoculated with C. destructivum as previously described. The number of lesions per cm2 of leaf for the two youngest fully developed leaves was determined 60 h after spraying. To test the effect of silencing on P. syringae pv. tomato, the two youngest fully developed leaves of another set of plants were infiltrated with a bacterial suspension using a needle-less syringe as previously described, and the number of CFU per cm2 leaf was determined 24 h later. Four plants per treatment were assessed for disease and gene expression.

To test for silencing of NbMCA1 at 20 days after inoculation with A. tumefaciens, the two youngest fully developed leaves per plant were harvested from another set of plants, and relative RT-PCR analysis was done as previously described. This set of plants was only used for relative RT-PCR analysis. To assess the specificity of silencing, the expression of another metacaspase gene of N. benthamiana, NbMCA2, as well as two other cysteine proteinases, NbCYP1 (DQ084022) and NbCYP2 (DQ084023) (Hao et al. 2006), were examined by relative RT-PCR prior to pathogen inoculation as previously described except with using different primers. A 830 bp portion of NbMCA2 was amplified with primers MCA2F1 5′-CCTGCTCTCCATAATCCTCA and MCA2R1 5′-ATACCAGAACGAGGACGATG designed from the sequence of TC7963 (http://www.tigr.org). Based on the primers described by Hao et al. (2006), a 391 bp portion of NbCYP1 was amplified with primers CYP4F3, 5′-GCGGATCTTGCTGGACATTC and CYP4R1 5′-GGGAGTATTGCCACATTCGG, and a 310 bp portion of NbCYP2 was amplified with primers CYP5F1, 5′-CGAAACCGAAACCGACGA and CYP5R1 5′-GTCACGCCAGTCATAATCAG. Quantification and confirmation of the identity of the RT-PCR products was done as previously described.

Results

Amplification and sequencing of NbMCA1

A pair of degenerate primers was designed based on conserved regions of the nucleotide sequence of LeMCA1 (AAM51555) and its two mostly closely related metacaspase genes, HbLAR (AAD13216) and At1g79340 (AAL85992). These conserved regions correspond to the sequences identified by Uren et al. (2000) as being highly conserved among caspases and metacaspases. The primers were used to amplify a single RT-PCR product of 700 bp from cDNA produced from leaves of N. benthamiana at 96 HPI with C. destructivum. Cloning of the RT-PCR product and subsequent sequence analysis of two of the clones revealed the same sequence, which had 87% identity with LeMCA1 from L. esculentum (Table 1). The N. benthamiana sequence was used in 3′ RACE to amplify a 610 bp RT-PCR product, using cDNA from leaves of N. benthamiana at 96 HPI with C. destructivum. This yielded a consensus sequence with a total length of 1,150 bp which was designated NbMCA1 (GenBank accession number DQ084024). Based on its alignment with HbLAR, the 5′-end of the NbMCA1 sequence matched the 3′-end of the highly conserved p20-like subunit of plant metacaspases (Uren et al. 2000). The p20 subunit contains the catalytically active site for caspases (Hoeberichts et al. 2003; Uren et al. 2000).

A search of the N. benthamiana Gene Index at TIGR (http://www.tigr.org/tdb/potato/) using TBLASTX did not reveal any sequences identical to NbMCA1, but did show similarity (expect value < 3E-4) to TC7963, which is composed of nine ESTs and has a tentative annotation as a caspase family protein. This putative metacaspase from the TIGR database was designated NbMCA2. NbMCA1 also had similarity (expect value < 7E-3) to TC8186, which is composed of seven ESTs and also has a tentative annotation as a caspase family protein, and this was designated NbMCA3. Although NbMCA2 and NbMCA3 had low similarity to NbMCA1, a 54 amino acid portion that was most similar with over 35% identity contained the highly conserved p20-like subunit of metacaspases, including the conserved cysteine and histidine catalytic sites of cysteine proteinases (Uren et al. 2000), indicating that NbMCA2 and NbMCA3 are also metacaspases.

The predicted protein sets of the genomes of O. sativa cv. Nipponbare and A. thaliana ecotype “Columbia” were mined for metacaspases using predicted plant metacaspase protein sequences from GenBank as the query sequences in Standalone BLAST version 2.2.6 (Altschul et al. 1997). These sequences plus the predicted N. benthamiana metacaspase proteins and several other putative metacaspases were aligned and a dendrogram was produced (Fig. 1). In this dendrogram, metacaspase sequences from rice and arabidopsis were scattered among various groups, providing evidence that metacaspases arose before the divergence of dicots from monocots. When genomes of other plant species are completed, they are likely to show genes interspersed among the various groups of rice and arabidopsis metacaspases, demonstrating the ancient origin and diversity of this multigene family. The major division in this dendrogram was between types I and II metacaspases which clustered separately from each other with 100% bootstrap support (Fig. 1). NbMCA1 clustered with type II metacaspases, and NbMCA2 and NbMCA3 clustered with type I metacaspases (Fig. 1).

Fig. 1
figure 1

Comparison of metacaspase sequences from Arabidopsis thaliana, Ceratopteris richardii, Hevea brasiliensis, Lycopersicon esculentum, Nicotiana benthamiana, Picea abies, Pinus taeda, Oryza sativa and Solanum tuberosum using Neighbor-Joining analysis to produce an unrooted tree. A description of each sequence is provided in Table 1. Amino acid sequences were aligned using CLUSTALX and a tree generated with the bootstrap NJ-tree procedure of CLUSTALX

Full size image

The division of plant metacaspases into types I and II is based on sequence similarities within their caspase-like domain and their overall domain structure (Uren et al. 2000). Type I metacaspases have a prodomain with a proline-rich repeat motif and a Zn finger motif typical of plant proteins, such as LSD-1, that functions in the HR pathway, whereas type II metacaspases lack those prodomain features, but have a conserved insertion of approximately 180 amino acids directly C-terminal to their p20-like subunit (Uren et al. 2000). An examination of these characteristics confirmed that NbMCA1 is a type II metacaspase, whereas NbMCA2 and NbMCA3 are type I metacaspases.

Within type II metacaspases, NbMCA1 grouped with metacaspases of L. esculentum (AAM51555), H. brasiliensis (AAD13216) and Solanum tuberosum (STUB1528) with 97% bootstrap support, and these were separate from other groups of type II metacaspases of A. thaliana and O. sativa. Phylogenetically, the genera of Lycopersicon, Solanum and Nicotiana are close as they all belong to the Solanaceae, and so it was not unexpected that these type II metacaspases were most similar with high bootstrap support. However, a sister group to the metacaspases from solanaceaeous plants and Hevea contained two conifer sequences (Picea abies, CAD59226 and Pinus taeda, PTAE9659) and a cycad sequence (Cycas rumphii, CRUM1400) with 98% bootstrap support, which indicates that this particular group of type II metascaspases arose before the separation of gymnosperms and angiosperms, possibly over 500 million years ago (Hedges 2002). There was also a more distantly related sequence from a true fern (Ceratopteris richardii, BE642699) among type II metacaspases, demonstrating that the type II metacaspases arose at least prior to the appearance of seed plants.

NbMCA2 and NbMCA3 grouped with A. thaliana (AAP84706 and AAP84707) and O. sativa (OsSFCC034713) type I metacaspases with 100% bootstrap support. The remaining type I metascaspases were represented by predicted proteins of A. thaliana and O. sativa, again reflecting a diverse multigene family.

Expression of NbMCA1 in N. benthamiana following inoculation with C. destructivum or P. syringae pv. tomato DC3000, and the effect of silencing NbMCA1 on disease development

Transcripts of NbMCA1 were detected in healthy N. benthamiana leaves by relative RT-PCR. Following infection with C. destructivum, expression of NbMCA1 showed a slow progressive increase peaking at 72 HPI (Fig. 2). At 72 HPI, expression of NbMCA1 was approximately 40% greater than that at 0 HPI. There was a slight decline at 96 HPI, but that difference was not significant. For the interaction with P. syringae pv. tomato DC3000, relative expression of NbMCA1 showed a transient but significant 48% increase from 0 to 3 HPI (Fig. 3). Infiltration with buffer alone resulted in a 31% increase in NbMCA1 expression at 3 HPI only indicating that changes in expression at 3 HPI were at least partially due to a wound response. At 3 HPI, the leaves appeared slightly wilted, but the leaf regained turgidity by 6 HPI. No necrosis was observed until 20–24 HPI. By 6 HPI, the expression level declined but still remained significantly higher than that at 0 HPI for the remainder of the experiment.

Fig. 2
figure 2

Relative RT-PCR of NbMCA1 expression in N. benthamiana leaves inoculated with C. destructivum. The amount of NbMCA1 mRNA was determined relative to the amount of the translation elongation factor 1α, NbEF-1α . Mean relative expression values are shown with standard error bars calculated from three replications

Full size image
Fig. 3
figure 3

Relative RT-PCR of NbMCA1 gene expression in N. benthamiana inoculated with P. syringae pv. tomato DC3000. The amount of NbMCA1 mRNA was determined relative to the amount of the translation elongation factor 1α, NbEF-1α. Mean relative expression values are shown with standard error bars calculated from three replications

Full size image

PVX-mediated VIGS of NbMCA1 in N. benthamiana resulted in a significant decrease in NbMCA1 expression compared to the water and PVX vector controls (Table 2; Fig. 4). Silencing appeared to be relatively specific as there was no significant difference in the expression of another metacaspase gene, NbMCA2, or two other cysteine proteinase genes, NbCYP1 and NbCYP2, in NbMCA1-silenced plants compared to the water and PVX vector controls (Table 2). NbCYP1 and NbCYP2 are papain-like cysteine proteinases (family CA1) with less than 3% identity to NbMCA1. NbMCA2 is the most similar metacaspase gene to NbMCA1 currently available at the TIGR N. benthamiana Gene Index (http://www.tigr.org/tdb/potato) with 9.2% protein identity over their entire lengths.

Table 2 Relative RT-PCR analysis of silencing of NbMCA1 and related genes in N. benthamiana
Full size table
Fig. 4
figure 4

Relative RT-PCR of virus-induced gene silencing of NbMCA1. RT-PCR gel showing the co-amplification of NbMCA1 and a constitutive control, NbEF-1α (translation elongation factor 1α). Lane M is the 100 bp ladder. The other lanes are WATER, which is the water control where leaves were wound inoculated with water instead of A. tumefaciens; PVX, which is the vector control where leaves were wound inoculated with A. tumefaciens containing potato virus X cloned in a Ti plasmid; and NbMCA1, where leaves were wound inoculated with A. tumefaciens containing potato virus X along with a portion of NbMCA1 cloned in a Ti plasmid

Full size image

The NbMCA1-silenced plants had three to four times more water-soaked lesions induced by C. destructivum compared to the water and PVX vector controls (Table 3). There were no differences in the appearance of the water-soaked lesions among the different treatments. Although the populations of P. syringae pv. tomato DC3000 were higher in NbMCA1-silenced plants at 24 HPI inoculation compared to either the water or PVX controls, the difference was not significant (Table 4). There were also no significant differences among the treatments in the timing or appearance of the hypersensitive response due to infiltration of P. syringae pv. tomato DC3000.

Table 3 Lesion number in water control, PVX vector control or NbMCA1-silenced N. benthamiana plants inoculated with C. destructivum at 60 HPI
Full size table
Table 4 Bacterial populations in water control, PVX vector control or NbMCA1-silenced N. benthamiana plants inoculated with P. syringae pv. tomato DC3000 at 24 HPI
Full size table

Discussion

A new plant metacaspase cDNA, NbMCA1, was amplified from leaves of N. benthamiana during the necrotrophic phase of infection by C. destructivum using a pair of degenerate primers. NbMCA1 is part of a small gene family of metacaspases in N. benthamiana, as there appears to be at least two additional metacaspase genes, which were identified among a collection of 26,918 input ESTs that form the basis of Release 2.0 of the N. benthamiana Gene Index (http://www.tigr.org/tdb/potato/). All of the ESTs comprising the tentative consensus sequences of NbMCA2 and NbMCA3 were obtained from RNA of N. benthamiana from callus, roots in liquid culture or leaves subjected to heat stress, cold stress or challenge with incompatible bacterial pathogens (P. syringae pv. tomato, Xanthomonas campestris pv. campestris, P. syringae pv. phaseolicola and X. campestris pv. vesicatoria).

Plant metacaspases have been divided into types I and II based on the sequence similarity within their caspase-like domain and their overall domain structure (Uren et al. 2000). A dendrogram of various type I and type II metacaspases from plants showed that the sequences of the two types are clearly distinguishable from each other. Nine metacaspase genes have been described from the genome of A. thaliana ecotype “Columbia” (AMC1–9) with three type I genes and six type II genes. There appears to clustering of some of these genes in the genome as the locus tags for the type II metacaspases AMC4 to AMC7 were AT1G79310 to AT1G79340 on chromosome 1. With metacaspase sequences from GenBank as query sequences, nine metacaspases (five type I and four type II) were found in the genome of O. sativa cv. Nipponbare using Standalone BLAST. Like arabidopsis and rice, N. benthamiana probably has a small gene family of both type I and type II metacaspases.

For the interaction with P. syringae pv. tomato DC3000, there was a transient increase in the expression of NbMCA1 at 3 HPI due to the effect of infiltration, but the expression level declined by 6 HPI to a level slightly above that at 0 HPI and then remained relatively constant as the resistance response developed. No peak in NbMCA1 expression was observed immediately preceding or during the PCD-associated HR necrosis at 20–24 HPI. In contrast, transcripts of two arabidopsis type I metacaspases, AtMCP1a and AtMCP1b, were up-regulated at 24 HPI with an incompatible P. syringae strain suggesting that some metacaspase genes are induced during the activation of the HR PCD (Watanabe and Lam 2004).

Following infection with C. destructivum, expression of NbMCA1did not significantly change from 0 to 48 HPI, which is during the biotrophic phase when C. destructivum penetrates and grows as multilobed vesicles in the initially infected leaf epidermal cell (Chen et al. 2002). During this phase, the host cells remain alive with no visible sign of host cell degeneration. At 72 HPI, expression of NbMCA1 significantly increased, and between 60 and 72 HPI, secondary hyphae was observed from the tips of the multilobed vesicles, penetrating adjacent cells and producing host necrosis that is clearly visible in infected cells (Shen et al. 2001). Although the necrotic lesions expanded after 72 HPI, expression of NbMCA1 relative to the constitutive control did not increase from 72 to 96 HPI. By comparison, LeMCA1 mRNA levels were undetectable in healthy tomato leaves but increased at 16 and 32 HPI as primary necrotic lesions formed in a susceptible interaction with the necrotrophic fungus, B. cinerea (Hoeberichts et al. 2003). As disease progression stalled at 48 HPI, expression of LeMCA1 decreased, but this was followed by increases at 72 and 96 HPI as lesions expanded to cover the whole leaf (Hoeberichts et al. 2003). Therefore, expression of LeMCA1 was consistently linked to the appearance of host cell death, unlike NbMCA1 which showed increased expression with the beginning but not the latter stages of the necrotrophic phase of infection by C. destructivum.

Cloning a portion of NbMCA1 into a PVX-based VIGS vector resulted in significantly reduced expression of NbMCA1 in the two youngest fully developed leaves of N. benthamiana compared with that of the controls by 20 days after inoculation with the recombinant virus. However, this silencing had no apparent effect on the HR of such leaves to P. syringae pv. tomato DC3000. The HR is a resistance response to pathogen attack that is a relatively well characterized form of PCD in plants (del Pozo and Lam 1998; Heath 2000). Therefore, if NbMCA1 played an important role in PCD during non-host HR resistance of N. benthamiana to P. syringae pv. tomato DC3000, then reducing NbMCA1 expression by VIGS should reduce the amount of HR necrosis. Although the induction of defense mechanisms can be uncoupled from HR cell death (Lam and del Pozo 2000), the population levels of the incompatible bacteria might also be higher in the silenced plants as the inoculated tissue could be less resistant because of less effective PCD. The lack of a significant difference among NbMCA1-silenced plants compared with the controls in either the timing and degree of HR necrosis or the size of the bacterial populations in inoculated leaves at 24 HPI indicates that NbMCA1 is not an important regulator of PCD during the HR of N. benthamiana to P. syringae pv. tomato DC3000. The discovery that PVX-mediated VIGS of two vacuolar processing proteases of N. benthamiana, that have limited sequence identity to caspases but have caspase activity, resulted in suppression of the HR caused by tobacco mosaic virus demonstrates that there are proteinases other than metacaspases that are important in the PCD of the HR (Hatsugai et al. 2004).

In contrast, virus-induced silencing of NbMCA1 resulted in an increase in the susceptibility of N. benthamiana leaves to C. destructivum. In susceptible interactions, necrotrophic pathogens may induce plant cells to initiate PCD to kill themselves thus creating dead plant tissue that the pathogen can use as a source of nutrition (Dickman et al. 2001; Richael et al. 2001). If NbMCA1 is involved in inducing PCD to create dead host tissue for use during the necrotrophic phase by C. destructivum, then reducing its expression by VIGS should have resulted in the plants becoming less susceptible because there would be less metacaspase available for pathogen-induced host PCD. However, this was not the case as NbMCA1-silenced N. benthamiana exhibited several times more water-soaked lesions of C. destructivum compared to the water and PVX controls.

In other plant–pathogen interactions, metacaspases appear to be involved in host PCD (van Baarlen et al. 2007). Mutants of five arabidopsis type II metacaspases, MCA2 to MCA6, were less susceptible to B. cinerea and B. tulipae, which are pathogenic to arabidopsis, and it was believed that those enzymes are involved in inducing PCD. Thus, the type II metacaspases of arabidopsis appear to have different functions in at least some plant–pathogen interactions compared with NbMCA1. NbMCA1 may act more similarly to the type I metacaspases of arabidopsis, MCA7 and MCA8, since mutants of those metacaspases were more susceptible to infection by Botrytis species. One difference, however, is that the arabidopsis type I metacaspase mutants also showed accelerated senescence, which was not observed in the NbMCA1-silenced plants. It appears that different metacaspases may play different roles depending upon the plant–pathogen interaction.

The increased susceptibility of NbMCA1-silenced N. benthamiana to C. destructivum indicated that reducing the amount of this metacaspase lowered some aspect(s) of host resistance, and thus NbMCA1 may play a role in basal resistance to C. destructivum rather than in pathogen-induced PCD. During infection by C. destructivum, NbMCA1 may have a direct role in host defenses by affecting a virulence factor of the pathogen. Alternatively, NbMCA1 could encode a proteinase that has an indirect role in the host response by functioning to activate and process pro-proteins involved in stress responses (Shimada et al. 1994), eliminate damaged proteins created by oxidative stress (Palma et al. 2002) or degrade proteins to remobilize amino acids to fuel de novo synthesis of enzymes associated with stress adaptation (Forsthoefel et al. 1998). Those roles would be more similar to that of a subgroup of human caspases that are involved in processing cytokines rather than inducing PCD, and cytokine-processing caspases have different cleavage site specificities than the caspases that induce PCD (Reed 2000). Whether it functions directly or indirectly in plant defense or pathogen-induced stress responses, NbMCA1 has a significant role in fungal disease resistance.