Functional characterization of NopT1 and NopT2, two type III effectors of Bradyrhizobium japonicum

Authors

  • Christos T. Fotiadis,

    1. Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Athens, Greece
    Search for more papers by this author
  • Maria Dimou,

    1. Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Athens, Greece
    Search for more papers by this author
  • Dimitrios G. Georgakopoulos,

    1. Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Athens, Greece
    Search for more papers by this author
  • Panagiotis Katinakis,

    1. Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Athens, Greece
    Search for more papers by this author
  • Anastasia P. Tampakaki

    Corresponding author
    • Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Athens, Greece
    Search for more papers by this author

Correspondence: Anastasia P. Tampakaki, Laboratory of General and Agricultural Microbiology, Department of Agricultural Biotechnology, Agricultural University of Athens, Iera Odos 75, Votanikos, 11855 Athens, Greece. Tel.: +302105294346; fax: +302105294344; e-mail: tampakaki@aua.gr

Abstract

NopT1 and NopT2, putative type III effectors from the plant symbiotic bacterium Bradyrhizobium japonicum, are predicted to belong to a family of YopT/AvrPphB effectors, which are cysteine proteases. In the present study, we showed that both NopT1 and NopT2 indeed possess cysteine protease activity. When overexpressed in Escherichia coli, both NopT1 and NopT2 undergo autoproteolytic processing which is largely abolished in the presence of E-64, a papain family-specific inhibitor. Mutations of NopT1 disrupting either the catalytic triad or the putative autoproteolytic site reduce or markedly abolish the protease activity. Autocleavage likely occurs between residues K48 and M49, though another potential cleavage site is also possible. NopT1 also elicitis HR-like cell death when transiently expressed in tobacco plants and its cysteine protease activity is essential for this ability. In contrast, no macroscopic symptoms were observed for NopT2. Furthermore, mutational analysis provided evidence that NopT1 may undergo acylation inside plant cells and that this would be required for its capacity to elicit HR-like cell death in tobacco.

Introduction

Bradyrhizobium japonicum (henceforth B. japonicum or Bja) is a Gram-negative soil bacterium capable of fixing atmospheric nitrogen in symbiosis with specific leguminous plants (e.g. Glycine max). Although the genetic basis of nodulation has been extensively studied, recent findings indicate that the type III secretion system (T3SS) plays a role in symbiosis. Genes encoding T3SSs and putative effector proteins have been identified in several but not all rhizobia by genome sequencing, such as B. japonicum USDA110, Rhizobium sp. NGR234, Mesorhizobium loti MAFF303099, Sinorhizobium fredii HH103, and S. fredii USDA257 (Freiberg et al., 1997; Viprey et al., 1998; Kaneko et al. 2000; 2002; Göttfert et al., 2001; Krishnan et al., 2003; de Lyra Mdo et al., 2009; see http://www.kazusa.jp/rhizobase/). The genes encoding the core components of rhizobial T3SS are called rhc (Rhizobium conserved) and are located in a gene cluster known as tts (type three secretion) (Viprey et al., 1998). Mutational studies on the tts gene clusters provided the first evidence that rhizobia deficient in T3SS are positively or negatively impaired in their ability to form nodules, depending on their hosts (Meinhardt et al., 1993; Bellato et al., 1997; Viprey et al., 1998; Marie et al., 2001; Krause et al., 2002; Deakin & Broughton, 2009).

Most of the rhizobial T3S genes are expressed in response to plant flavonoids. Their promoter regions harbor a regulatory motif known as tts box (Viprey et al., 1998; Krause et al., 2002; Krishnan et al., 2003; Hubber et al., 2004; López-Baena et al., 2008; Zehner et al., 2008; Sánchez et al., 2009), a binding site for the transcriptional activator TtsI whose production is flavonoid dependent (Kobayashi et al., 2004; Marie et al., 2004; Wassem et al., 2008). In B. japonicum, the expression of the T3SS genes is induced by seed extract and genistein that is also an inducer of the nodulation genes (Krause et al., 2002; Süß et al., 2006; Wei et al., 2010). Transcriptional profiling revealed that T3SS genes of B. japonicum are downregulated in bacteroids relative to free living conditions, suggesting that the secretion of proteins via the T3SS may play a role during the nodule initiation (Chang et al., 2007).

Rhizobial proteins secreted by T3SSs are designated Nops (nodulation outer proteins) (Marie et al., 2001). The specific roles of the various effector proteins in nodulation are not yet known, although a few of them have been shown to affect the nodulation in a host-dependent manner (Marie et al., 2003; Skorpil et al., 2005; Dai et al., 2008; Yang et al., 2009).

Despite the importance of type III secretion in pathogenesis, there has been very little work on its function in symbiotic bacteria. Previous studies have shown that B. japonicum contains a functional T3S and > 30 putative T3S effector genes, many of which have homologs in plant and animal pathogenic bacteria (Göttfert et al., 2001; Kaneko et al., 2002; Krause et al., 2002; Süß et al., 2006; Yang et al., 2010). Proteomic analyses have shown that at least 14 effectors are secreted in culture (Süß et al., 2006; Zehner et al., 2008; Hempel et al., 2009). So far, the only secreted protein studied in B. japonicum is NopE1 (Wenzel et al., 2010; Schirrmeister et al., 2011), while a biochemical role of some other T3S secreted proteins can be inferred from the studies of their homologs in other rhizobia.

NopT1 and NopT2, two putative T3S effectors of B. japonicum, share homology to members of the YopT/AvrPphB family. YopT/AvrPphB family members are cysteine autoproteases with a conserved catalytic triad consisting of a cysteine (C), a histidine (H), and an aspartic acid (D) residue and possess an internal putative N-myristoylation site in their N-termini (Shao et al., 2002). Previous studies have shown that myristoylation localizes bacterial T3S effectors to the plasma membrane facilitating access to their substrates/targets (Nimchuk et al., 2000; Shao et al., 2003b; Dowen et al., 2009). Recent studies have shown that NopT from NGR234 functions as a cysteine protease and affects nodulation positively or negatively in a manner dependent on the host (Dai et al., 2008; Kambara et al., 2009).

Here, we report that both NopT1 and NopT2 possess cysteine protease and autoproteolytic activity but only NopT1 is capable of eliciting cell death in Nicotiana species. Mutational analyses of NopT1 provided evidence that the putative acylation sites are essential for both enzymatic and cell death–eliciting activity.

Materials and methods

Bacterial strains, culture conditions, media, and DNA manipulations

Bacterial strains and plasmids used in this study are shown in Table 1. Escherichia coli and Agrobacterium tumefaciens were routinely grown in Luria–Bertani (LB) medium at 37 or 30 °C, respectively. Bradyrhizobium japonicum USDA 110 was grown on yeast extract-mannitol broth (YMB) medium (Daniel & Appleby, 1972) at 28 °C. Antibiotics were usually used at the following concentrations (μg mL−1): ampicillin (Ap), 100; carbenicillin (Cb), 100; kanamycin (Km), 50; rifampicin (Rif), 50; and spectinomycin (Sp), 50. Escherichia coli DH5a was used as a cloning host.

Table 1. Bacteria strains and plasmids
Strains or plasmidsDescriptionSource or reference
  1. a

    ApR, KmR, SpR, CbR, and RifR indicate resistance to ampicillin, kanamycin, spectinomycin, carbenicillin, and rifampicin, respectively.

Bradyrhizobium japonicum USDA 110Wild-type strain, SpRKaneko et al. (2002)
Agrobacterium tumefaciens C58C1 (pGV2260)Strain for plant transient expression, RifR/CbRDeblaere et al. (1985)
Escherichia coli DH5aStrain for cloningInvitrogen
Escherichia coli HB101 (pRK2013)Strain with helper plasmid for triparental mating, KmRDitta et al. (1980)
Escherichia coli BL21 (DE3)Strain for bacterial protein overexpressionNovagene
pT7-7Vector for protein overexpression, φ10 T7 RNA polymerase promoter, ApRTabor & Richardson (1985)
pET26bVector for protein overexpression, lacI, C-terminal His-Tag, KmRNovagene
pBIN-Hyg-TxBinary vector for plant transient expression, KmRGatz et al. (1992)
pT7-7::nopT1Wild-type NopT1 in pT7-7, ApRThis study
pT7-7::nopT2Wild-type NopT2 in pT7-7, ApRThis study
pT7-7::nopT1-C100SNopT1 with cysteine 100 substituted with a serine in pT7-7, AmpRThis study
pT7-7::nopT1-H213ANopT1 with histidine 213 substituted with an alanine in pT7-7, ApRThis study
pT7-7::nopT1-D228ANopT1 with aspartic acid 228 substituted with an alanine in pT7-7, ApRThis study
pT7-7::nopT1-DKMNopT1 with D47, K48, and M49 residues substituted all with alanines (A) in pT7-7, ApRThis study
pT7-7::nopT1-GCCNopT1 with G50, C52, and C53 residues substituted with alanine (A) and serines (S) in pT7-7, ApRThis study
pET26b::nopT1Wild-type NopT1 in pET26b, KmRThis study
pET26b::nopT2Wild-type NopT2 in pET26b, KmRThis study
pBIN::nopT1Wild-type NopT1 in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT2Wild-type NopT2 in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT1-Δ50NopT1 lacking the first 50 N-terminal residues in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmR 
pBIN::nopT1-C100SNopT1 with cysteine 100 substituted with a serine in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT1-H213ANopT1 with histidine 213 substituted with an alanine in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT1-D228ANopT1 with aspartic acid 228 substituted with an alanine in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT1-DKMNopT1 with D47, K48, and M49 residues substituted all with alanines (A) in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study
pBIN::nopT1-GCCNopT1 with G50, C52, and C53 residues substituted with alanine (A) and serines (S) in pBIN-Hyg-Tx under the control of CaMV 35S promoter, KmRThis study

Standard DNA manipulation procedures were used (Sambrook et al., 1989). Genomic DNA was isolated using the GenElute Bacterial Genomic DNA Kit (Sigma). Plasmid isolations and DNA enzyme cleanups were conducted using the Qiagen Spin Miniprep Kit and Clean-up kit, respectively. PCR amplifications were carried out in 50-μL volumes with the GoTaq DNA polymerase (Promega) and were performed in a DNA thermocycler (M. J. Research) using the primers in Table 2. Plasmids transfers were accomplished by triparental mating using the E. coli strain HB101 (pRK2013) (Ditta et al., 1980) or by electroporation (GenePulser; Bio-Rad) following the manufacturer's instructions.

Table 2. Oligonucleotides used in this study
Primers namePrimer sequenceRestriction enzyme
  1. The underlines in the second column denote the restriction enzymes sites indicated in the third column.

  2. a

    The restriction site is split into two halves in each pair of primers.

NopT1-F15′-GTGAATCCATATGTATGATCGAATCGGTGGC-3′NdeI
NopT1-R15′-CATCCGCGAATTCACAGAAGGCGTCAATCAC-3′EcoRI
NopT1-F25′-GAGAGGGGTACCGTCATGTATGATCGAATCGG-3′KpnI
NopT1-R25′-CACCGCGTCTAGAAGGCCCAACTCACTGCATC-3′XbaI
NopT1-R35′-GCCCAACAAGCTTCTGCATCCTTTGCGTCGTG-3′HindIII
NopT2-R35′-CTCGTGAATTCCGATGAGGTTCCGCGCGCGG-3′EcoRI
NopT2-F15′-GGTGAACCCATATGTATAATCGAGTCGATGGCG-3′NdeI
NopT2-R15′-CTGCACGAATTCCAGAGAACTTCCATCGAG-3′EcoRI
NopT2-F25′-GCCGGGGTACCAGCATGTATAATCGAGTCG-3′KpnI
NopT2-R25′-GGGTGAATCTAGACGGGCGATGCGGCTGC-3′XbaI
NopT1-Δ50K-F5′-GTTGCCTGGTACCATGGCCTGCTGCAGCAAG-3′KpnI
NopT1-C100S-F5′-TCCGTCGGGCTGACTGCAGAGTGGTTC-3′PstI
NopT1-C100S-R5′-GATGCCGTCCACATTCGCATCACGCAG-3′
NopT1-H213A-F5′-GCCACAGTGGCCACGTCGGCCTCGAATGG-3′EaeI
NopT1-H213A-R5′-GGCGCCGCCTTCAGCGAAATACAAGCT G-3′
NopT1-D228A-F5′-GGAGAATTTACTGTTCGATCGGATCCTGACC-3′PvuI
NopT1-D228A-R5′-ATAGTTAGGGGCGAAAAGCGTGGTCGTTCC-3′
NopT1-DKM-F5′-CGGCGGGGGCCTGCTGCAGCAAGCCAG-3′SacIIa
NopT1-DKM-R5′-CGGCAGGCAACTCCCCCGATCGTGAGC-3′SacIIa
NopT1-GCC-F5′-AGCAGCAGCAAGCCAGATACCTTGGAT-3′NheIa
NopT1-GCC-R5′-AGCCGCCATCTTGTCAGGCAACTCCCC-3′NheIa

Plasmid constructions and site-directed mutagenesis

Expression constructs for nopT1 (annotated as blr2140) and nopT2 (annotated as blr2058) were made by cloning the full-length nopT1 and nopT2 PCR amplicons derived from B. japonicum genomic DNA template. The primers for nopT1 (NopT1-F1 and NopT1-R1) and nopT2 (NopT2-F1 and NopT2-R1) were designed to contain appropriate restriction sites (NdeI and EcoRI) to facilitate cloning in the corresponding sites of the pT7-7 expression plasmid.

Site-directed mutagenesis was accomplished according to the protocol described by Fisher & Pei (1997) on plasmid pT7-7/nopT1 by amplifying the gene with appropriately designed primers mutating the catalytic triad codons for cysteine (C), histidine (H), and aspartic acid (D). The primers NopT1-C100S-F (forward, mutagenic primer) and NopT1-C100S-R (divergent, nonoverlapping primer) were designed to give a C100S substitution in NopT1 and to create a PstI restriction site. Mutants H213A and D228A were obtained similarly by using the pair of primers NopT1-H213A-F/NopT1-H213A-R and NopT1-D228A-F/NopT1-D228A-R, which simultaneously introduced an EaeI and a PvuI restriction site, respectively. Mutants nopT1-DKM and nopT1-GCC were obtained by PCR amplification as described earlier using the pair of primers NopT1-DKM-F/NopT1-DKM-R and NopT1-GCC-F/NopT1-GCC-R, respectively. The primers were designed to obtain the D47A, K48A, and M49A substitutions in case of the NopT1-DKM mutant and G50A, C52S, and C53S substitutions in case of the NopT1-GCC mutant. All mutations were confirmed by diagnostic restriction digestions taking advantage of SacII and NheI sites designed in the primers and sequencing.

Protein overexpression and purification

C-terminally polyhistidine-tagged wild-type NopT1 and NopT2, as well as mutant derivatives of NopT1, were obtained by cloning the respective coding regions without the stop codons following PCR amplification from the pT7-7 expression constructs with the pair of primers NopT1-F1/NopT1-R3 and NopT2-F1/NopT2-R3, respectively. The amplicons were digested with appropriate restriction enzymes and subcloned into the pET26b vector (Novagen), ligated, and transformed into Ecoli strain BL21 (DE3). For protein expression, E. coli BL21 (DE3) transformants harboring the pET26b constructs were grown in LB medium to an OD600 nm of 0.6 at 37 °C, and protein expression was induced for 4 h at 30 °C by adding 0.5 mM isopropyl β-d-thiogalactopyranoside (IPTG). Bacterial cells were collected by centrifugation, resuspended in lysis buffer (50 mM NaH2PO4, pH 8.0, 300 mM NaCl, 10 mM imidazole) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF), and lysed by the addition of lysozyme followed by sonication. Histidine-tagged wild-type and mutant proteins were expressed in E. coli BL21 (DE3) at 30 °C and purified by Ni2+-NTA affinity chromatography under native conditions according to the standard protocol (Qiagen). Proteins were resolved in 14% SDS-polyacrylamide gel electrophoresis (PAGE) and were visualized by Coomassie blue staining and immunoblotting using alkaline phosphatase (AP)-conjugated anti-His antibody (Qiagen). Protein concentrations were estimated by Coomassie blue staining of SDS-PAGE gels using BSA standards. Prestained molecular size standards (Broad range; New England Bio-Labs) were used to estimate the molecular mass of proteins.

Cysteine protease activity assays

Proteins were purified under nondenaturing conditions as mentioned earlier and lyophilized, and their protease activity was determined using resorufin-labeled casein (Roche) as a substrate. Lyophilized samples were dissolved in different buffers at pH range 5.5–9.5 in final volume of 100 μL and preincubated at 37 °C for 1 h. The enzymatic activity was determined in 50 mM buffers (sodium acetate buffer at pH 5.5; potassium phosphate at pH range 6.5–7.5; Tris at pH range 8.5–9.5) containing 10 mM l-cysteine, 10 mM EDTA, and 0.4% casein in a final volume of 200 μL. The assay mixture was incubated at 37 °C for 24 h, and the reaction was terminated by adding trichloroacetic acid (TCA) to a final concentration of 5%. After incubation at 37 °C for 10 min, the mixture was centrifuged for 5 min and the supernatant was alkalinized by the addition of 0.5 M Tris–HCl, pH 8.8. The concentration of the released resorufin-labeled peptides in the supernatant was measured spectrophotometrically at 574 nm and was used as a measure of cysteine protease activity. For inhibition assays, lyophilized samples were dissolved as mentioned earlier in the optimal assay buffer in the presence/absence of 5 mM E-64 (Sigma) in 200 μL of final volume.

Agrobacterium-mediated transient expression

Wild-type, deletion, and site-directed mutant nopT1 genes were PCR amplified from the corresponding pT7-7 constructs using the primers NopT1-F2 and NopT1-R2 and cloned into the KpnI and XbaI sites of the binary vector pBIN-Hyg-Tx under the control of Cauliflower mosaic virus (CaMV) 35S promoter (Gatz et al., 1992). Similarly, nopT2 wild-type gene was PCR amplified from the pT7-7/nopT2 construct using the primers NopT2-F2 and NopT2-R2 and cloned as KpnI/XbaI fragment in pBIN-Hyg-Tx. To create an N-terminal deletion derivative of NopT1 protein lacking amino acid residues 1–50, a PCR fragment encoding the carboxy-terminal 221 amino acids of NopT1 was amplified from the pT7-7/nopT1 using the primers NopT1-Δ50K-F and NopT1-R2, simultaneously changing the glycine residue at position 50 to methionine. The resulting plasmids were then introduced into A. tumefaciens C58C1 (pGV2260) by triparental mating (Deblaere et al., 1985). Individual transconjugants were grown in 5 mL of LB medium containing the appropriate antibiotics. Following overnight growth at 28 °C, bacteria were centrifuged and resuspended in MMA medium (Murashige–Skoog salts, 10 mM MES pH 5.6, and 200 μM acetosyringone) to a final OD600 nm of 1.0. Cell suspensions were kept at 28 °C for 2 h and were then infiltrated into fully expanded Nicotiana tabacum cv. Xanthi and Nicotiana benthamiana leaves using a needleless syringe.

Results

Bradyrhizobium japonicum contains two putative members of YopT/AvrPphB family of effector proteins

Bradyrhizobium japonicum genome contains two genes, nopT1 and nopT2, encoding proteins with homology to members of the YopT/AvrPphB family. Both genes are located within the symbiotic region but outside of the T3SS gene cluster. Horizontal gene transfer (HGT) analysis of their regions with the Jena prokaryotic genome viewer (http://jpgv.imb-jena.de) showed that nopT1 and nopT2 have a significantly lower GC content, 54.4% and 54.3%, respectively, than the genomic average of 64.1%. This observation together with the fact that both genes are flanked by mobile elements indicates possible acquisition by HGT (Fig. 1a). It is interesting to note that several T3S effector genes of B. japonicum have GC content lower than the genomic average. The nopT1 and nopT2 promoter regions contain potential tts boxes at 72–99 and 71–98 bp upstream from the start codons, respectively (Fig 1b), which suggest transcriptional regulation by TtsI (Marie et al., 2004; Zehner et al., 2008). Previous studies have demonstrated that nopT1 is inducible by the flavonoid genistein and the NopT1 is a type III secreted protein detected in Bradyrhizobium culture supernatants upon induction with genistein (Lang et al., 2008; Zehner et al., 2008; Hempel et al., 2009). NopT1 and NopT2 (271 and 298 residues, respectively) share 48% mutual identity and show 59% and 40% identity, respectively, to NopT of NGR234 or 32% identity to AvrPphB. The presence of a predicted cysteine protease catalytic triad in NopT1 (C100, H213, and D228) and NopT2 (C109, H223, and D238) indicates that these proteins may possess cysteine protease activity. Moreover, in silico analyses showed that both proteins contain putative N-myristoylation and S-palmitoylation sites (Fig. 1c). The glycine residue at position 50 (G50) is a putative internal N-myristoylation site, while the conserved cysteine residues at positions 52 (C52) and 53 (C53) of NopT1 and C52 of NopT2 could be palmitoylated. To our knowledge, there are so far no experimental data available that verify these biochemical features.

Figure 1.

(a) Schematic representation of the loci containing nopT1 and nopT2. Genes are represented by arrows pointing in the direction of translation and are colored according to their GC content. Genes colored red and blue have a GC content value lower than 0.64 (genomic average). Those colored yellow and green have a GC content value > 0.64. The HGT classification of genes is according to Jena prokaryotic genome viewer. The numbers under the lines indicate coordinates. (b) Promoter regions of nopT1 and nopT2 with tts boxes. Numbers indicate distances in bp between the tts box and the start codons of the corresponding genes. Highly conserved nucleotides are shown in capital letters, less conserved nucleotides are shown in lower case letters, and dots indicate any nucleotide. (c) Comparison of NopT1 and NopT2 protein sequences. The catalytic residues are in red, and the putative myristoylation and palmitoylation sites are colored with blue and green, respectively. Numbers indicate the position of the residues in the full-length proteins.

NopT1 and NopT2 possess cysteine protease activity

Previous studies have shown that most members of the YopT family can display cysteine protease activity in vitro when they are overexpressed in E. coli (Puri et al., 1997; Nimchuk et al., 2000; Dowen et al., 2009). To determine whether this was also true for NopT1 and NopT2, we made NopT1-His6 and NopT2-His6 fusions and purified the proteins from E. coli extracts by affinity chromatography using nondenaturing conditions. IPTG induction in E. coli cultures led to the accumulation of two protein bands corresponding to the full-length form (~32 kDa) and a truncated form (~26 kDa) of NopT1 (Fig. 2a). Similarly, NopT2 was produced as a full-length form (~35 kDa) and a truncated form (~30 kDa). These results indicate that both wild-type proteins are processed in E. coli. We have repeatedly observed very low levels of the full-length product in soluble fractions, suggesting that it is also rapidly processed in E. coli cells.

Figure 2.

(a) NopT1 and NopT2 undergo autoproteolytic cleavage in Escherichia coli. (b) Purified NopT1 protein harboring mutations in the catalytic triad residues (C/H/D) are deficient in autoprocessing activity. (c) Autoproteolytic cleavage in E. coli of mutant proteins NopT1-DKM and NopT1-GCC. Samples from purified proteins were subjected to SDS-PAGE (12%), and proteins were visualized by Coomassie blue staining. N-terminally processed NopT1 and NopT2 proteins are indicated with an asterisk to the right of the gel. Molecular weight markers (MW) in kDa are shown to the left of the gel. WT indicates the NopT1 wild-type protein. (d) A schematic view of the various mutant and deletion constructs is shown on the right. Numbers refer to the location of the amino acids in the wild-type protein. Amino acids C100, H213, and D228 of NopT1 present in the catalytic core of cysteine proteases. G50 is the potential myristoylation site and C51 and C52 are putative palmitoylation sites.

To further assess the proteolytic activity of NopT1 and NopT2, we carried out cysteine protease activity assays in vitro using resorufin-labeled casein as a substrate (Twining, 1984). To determine the optimum pH, the activity was monitored by incubation the proteins in constant ionic strength buffers of different pH. Both wild-type proteins displayed maximal activity at pH of 6.5 (Fig 3a). Addition of a well-studied general inhibitor for papain-like cysteine proteases, E-64 (Barrett et al., 1982), abolished the enzymatic activity of each protein (Fig. 3b). These data support the prediction that NopT1 and NopT2 are cysteine proteases belonging to the CA clan.

Figure 3.

Cysteine protease activity of NopT1 and NopT2 proteins. (a) Cysteine protease activity profiles of NopT1 and NopT2 over the pH range of 5.5–9.5. Proteins purified under native conditions were assayed with resorufin-labeled casein. The enzymatic reactions were incubated at 37 °C for 24 h and were then terminated by the addition of TCA. The release of resorufin-labeled peptides from casein, resorufin-labeled was detected spectrophotometrically at 574 nm. White bars represent the activity of the NopT1, while gray bars represent the activity of the NopT2. (b) Proteins were assayed with resorufin-labeled casein at pH of 6.5 in the absence (white bars) or presence (gray bars) of 5 mM E64 inhibitor. (c) Wild-type NopT1 and mutant proteins NopT1-C100S,NopT1-H213A, and NopT1-D228A were assayed for cysteine protease activity as above at pH of 6.5. Each value represents the average of two independent experiments each performed in three replicates. Error bars indicate standard deviations.

NopT1 but not NopT2 elicits HR-like cell death in Nicotiana species

The Agrobacterium-transient expression system has been proven a powerful tool for investigating the potential functions of type III effectors from plant pathogenic bacteria and recently from rhizobial species (Dai et al., 2008). To determine whether NopT1 and NopT2 could elicit hypersensitive response (HR)-like cell death in planta, we carried out agroinfiltration assays. Agrobacterium tumefaciens C58C1 strains carrying the vector pBin-Hyg-Tx, pBin::nopT1, and pBIN::nopT2 were infiltrated into N. tabacum cv. Xanthi and N. benthamiana leaves. NopT1 elicited localized cell death in both Nicotiana species (Fig. 4b). By contrast, leaves infiltrated with A. tumefaciens carrying pBin::nopT2 did not show any visible symptoms (Fig. 4c). No visible symptoms of cell death were observed when Agrobacterium with an empty vector was infiltrated (Fig. 4a). In light of these results, further studies focused on the analysis of NopT1 function.

Figure 4.

Agrobacterium-mediated transient expression of the NopT1 and NopT2 proteins in Nicotiana tabacum cv. Xanthi leaves. (a) Plants were inoculated with Agrobacterium tumefaciens strains C58C1 carrying one of the following plasmids: empty vector pBIN-Hyg-Tx (a), pBIN::NopT1 (b), pBIN::NopT2 (c), pBIN::Δ50-NopT1 (d), pBIN::NopT1-C100S (e), pBIN::NopT1-H213A (f), pBIN::NopT1-D228A (g), pBIN::NopT1-GCC (h), and pBIN::NopT1-DKM (i). The leaves were photographed at 48 h postinoculation (hpi). The experiment was repeated five times with similar results. Similar results were obtained in Nicotiana benthamiana plants.

Mutations in catalytic triad residues of NopT1 abolish the HR-like symptoms

To determine whether the putative catalytic triad (C/H/D) of NopT1 is required for the HR-like cell death in tobacco, we constructed substitutions at positions 100 (C100S), 213 (H213A), and 228 (D228A) with Ala (Fig. 2d). The coding regions of the site-directed mutants were subcloned into a binary Agrobacterium vector and tested for ability to elicit the HR in N. tabacum and N. benthamiana when overexpressed directly within the plant cells via the Agrobacterium-transient expression system. None of the mutants elicited cell death (Fig. 4e–g), whereas the wild-type NopT1 elicited a strong HR (Fig. 4b).

We also examined whether the site-directed mutants retained enzymatic activity. As shown in Fig 2b, all site-directed mutants had lost the NopT1 processing in E. coli, although not completely and their in vitro enzymatic activity was significantly reduced in comparison with wild-type protein (Fig. 3c). These results corroborate further the prediction that that NopT1 is a cysteine protease and requires an intact catalytic triad for both enzymatic and HR-eliciting activity.

Mutation of both acylation sites eliminates the cell death–eliciting activity of NopT1

Previous studies have shown that all YopT/AvrPphB family members identified so far contain an embedded consensus site for eukaryotic fatty acylation which may be exposed following autoproteolytic processing of these effectors (Puri et al., 1997; Nimchuk et al., 2000; Dowen et al., 2009). Similarly, NopT1 possesses putative sites (Fig. 1b) for both N-myristoylation (G50) and S-palmitoylation (C52 and C53) that lack experimental validation. To investigate whether these acylations play a role in cell death elicitation by NopT1, we made deletion and site-directed mutants affecting either one or both sites.

Initially, we made a deletion mutant, Δ50N, in which an ATG codon was introduced just before the A51 codon by replacing the glycine (G) residue at position 50 by a methionine (M) residue. Transient expression via agroinfiltration of this mutant displayed identical necrotic phenotype to that elicited by the full-length protein, in terms of both timing and intensity of the necrotic response (Fig. 4d). Although myristoylation of NopT1 has not been demonstrated biochemically, it is tempting to speculate that an intact myristoylation motif may not be required for HR elicitation by NopT1 at least in plants tested. To examine whether palmitoylation may affect the cell death–eliciting activity of NopT1, we next created a site-directed mutant that destroyed both putative acylation sites by simultaneous substitution of residues G50, C52, and C53 with alanine and serine residues (Fig. 2d). As shown in Fig. 4h, the triple mutant NopT1-GCC was not capable of causing cell death in tobacco following transient expression by Agrobacterium as the wild-type protein did. This result suggests that the putative palmitoylation sites may be more important than myristoylation for plant plasma membrane association and the subsequent cell death in tobacco.

To investigate whether the NopT1 autoprocessing is required to reveal the embedded acylation sites, we created another mutant (NopT1-DKM) by substituting residues D47, K48, and M49 with alanines (Fig. 2d). In this mutant, both acylation sites were intact, while the amino acids immediately preceding the putative NopT1 autocleavage site were modified. This mutant was inactive in eliciting cell death in tobacco (Fig. 4i).

To further test whether the mutant proteins NopT1-GCC and NopT1-DKM are autoprocessed, we expressed them in E. coli and analyzed the purified proteins by SDS-PAGE and Western blotting. The NopT1-DKM was completely resistant to autocleavage (Fig. 2c), suggesting that the residues D47, K48, and M49 are required for autoprocessing of the N-terminal region. In contrast, the protein mutated in residues G50, C52, and C53 (NopT1-GCC) still shows autocleavage (Fig. 2c). It is interesting to note that the wild-type NopT1 was very rapidly processed in E. coli, and we were able to detect the full-length protein only when short times of induction (e.g. 2–4 h) were chosen. In contrast, the full-length protein of the GCC mutant was still detectable in substantial amounts upon induction of protein expression for 12 h in E. coli. Although these results indicate that mutation in the G50, C52, and C53 residues partially affects the autoproteolytic activity of NopT1, significant autocleavage activity is observed for NopT1-GCC protein. Together, the results suggest that autoprocessing of NopT1 is required to unmask its putative acylation sites.

Discussion

In this study, we demonstrated for the first time that NopT1, but not NopT2, of B. japonicum elicits cell death in plants tested. Both proteins possess cysteine protease activity that is essential for the cell death–eliciting activity in the case of NopT1. Many members of the YopT/AvrPphB effector family have been shown to possess cysteine protease activity (Shao et al., 2002; López-Solanilla et al., 2004), although some of them are not autoprocessed or acylated (Dowen et al., 2009). In plant symbiotic bacteria, three genes encoding YopT family members have been found: one in Rhizobium NGR234, named nopT (Dai et al., 2008), and two in B. japonicum (nopT1 and nopT2). Multiple proteases of the YopT family can be found in a single strain, for example, in Pseudomonas syringae pv. tomato strain DC3000 HopC1 and HopN1 effectors (Buell et al., 2003), in P. syringae pv. phaseolicola 1448A annotated as PSPPH_A0122 (HopAW1) and PSPPH_A0129, and in Acidovorax avenae ssp. avenae annotated as Acav_0110 and Acav_4550. Such situation might be indicative of different substrates being targeted in the plant host or the same substrate cleaved at different positions (thus generating different cleavage products) or function in different plant hosts. Other T3S effector genes are also present in duplicate in B. japonicum genome, such as NopE1 and NopE2 (Wenzel et al., 2010). The presence in a single strain of multiple members belonging to a T3S effector family also occurs in phytopathogenic bacteria. For example, multiple members of the YopJ family are present in P. syringae, Xanthomonas campestris, and Ralstonia solanacearum (Ma et al., 2006; Zhou et al., 2009; Szczesny et al., 2010; Lewis et al., 2011). The diversification of effector family members is thought to have been evolved during the co-evolutionary arms race between plants and their attackers via both pathoadaptation and horizontal gene transfer (Ma et al., 2006). Further studies are needed to determine the driving forces in shaping the T3E repertoire of B. japonicum as well as the presence of multiple effector paralogs.

Here, we present evidence that NopT1 and NopT2 are indeed cysteine proteases with autoproteolytic activity which is maximal at pH of 6.5, and it is completely abolished by the class-specific cysteine peptidase inhibitor, E-64. Moreover, single mutations disrupting the catalytic core residues (C100, H213, and D228) of NopT1 diminished both the cysteine protease and the autoprocessing activities. These findings are consistent with previous reports indicating that some T3S effectors classified as clan CA proteases from plant pathogenic and symbiotic bacteria are autoproteolytically cleaved in E. coli (Puri et al., 1997; López-Solanilla et al., 2004; Dai et al., 2008; Dowen et al., 2009; Kambara et al., 2009). It is interesting to note that two previous studies (Dai et al., 2008; Kambara et al., 2009) reached contradictory conclusions regarding the necessity of the catalytic residue H205 in the proteolytic activity of NopT from NGR234. Although these studies reached different conclusions, their primary data are not necessarily mutually exclusive because differences in methods applied might account for the seeming discrepancy.

To gain insight into the role of the residues surrounding the autoproteolytic site in NopT1 autoprocessing, we mutated the amino acids immediately preceding the putative autocleavage sites at P3, P2, and P1 that are occupied by residues D47, K48, and M49, respectively. Triple substitution of P1–P3 sites with alanines completely abolished the autoproteolytic cleavage of NopT1. This finding is consistent with previous studies demonstrating that mutations of P1–P3 residues in AvrPphB, ORF4, NopT, RipT, and PBS1 prevent self-processing (Shao et al., 2003b; Zhu et al., 2004; Dowen et al., 2009). On the other hand, the autocleavage was not inhibited by a triple mutation of the residues G50, C52, and C53 next to the probable autocleavage site. Similarly, a double mutant of AvrPphB (G63A-C64S) is still capable of autoprocessing (Dowen et al., 2009). Taken together, the data suggest that residues at position P1–P3 are absolutely required for maximal autoprocessing while residues G, C, and C do not affect this function. The fact that the triple mutant NopT1-GCC is still capable of autoprocessing implies that the autocleavage site is likely between M49 and G50 or between K48 and M49. In the latter case, N-terminal methionine excision by the methionine aminopeptidase present in E. coli (Frottin et al., 1992) would expose the glycine at position 50 which would then be subjected to myristoylation. The removal of methionine has a high probability to occur because the presence of glycine, just next to methionine, is the most optimal residue at this position (Frottin et al., 1992). Our data are consistent with a recent study (Dai et al., 2008) indicating that the N-terminal sequence of the processed NopT from Rhizobium NGR234 is GCCA.

Transient expression of nopT1 and nopT2 in nonhost Nicotiana plants revealed that NopT1 harbors a cell death–triggering activity, while NopT2 does not (Fig. 4). These results suggest that NopT1 or the products of its action are recognized in Nicotiana species and this system is thus applicable to study the function of NopT1 in a nonhost plant. Recently, the homolog NopT of NGR234 has been shown to cause cell death in tobacco (Dai et al., 2008). In contrast to NopT1, NopT2 did not show any visible phenotype, indicating that either Nicotiana species do not contain the appropriate recognition machinery (e.g. R-like proteins) and that NopT2 is mislocated in plant cells upon in planta transient expression or it does not harbor a cell death–eliciting activity at all. Taken together, these data provide evidence that NopT1 and NopT2 may possess distinct substrate specificities toward plant targets. Furthermore, the fact that mutations in the catalytic triad residues of NopT1 abolished its HR-like cell death–eliciting activity in tobacco indicates that an intact catalytic triad is essential for this ability in tobacco.

Lipid acylation (N-myristoylation and S-palmitoylation) is a common modification of T3S effectors and is responsible for their membrane localization. The presence of both eukaryotic acylation motifs in NopT1 and NopT2 implies that they might be acylated in the host cell cytoplasm and targeted to the plasma membrane. To gain insight into the relevance of the putative acylation sites of NopT1 for its function, we made a deletion mutant, namely Δ50N, in which the glycine residue at position 50 was substituted by a methionine. The observation that the deletion mutant Δ50N is functional in HR cell death assays implies that myristoylation, if it occurs, is not required for NopT1 effector function on tobacco. The aforementioned results are in agreement with previous studies demonstrating that myristoylation-deficient mutants of NopT from NGR234 (Dai et al., 2008) and AvrPphB from P. syringae pv. phaseolicola (Tampakaki et al., 2002) retain cell death activity in tobacco plants. It is likely that the acylation of cysteine residues in the myristoylation-deficient mutants might still direct the proteins to the plasma membrane where they might act. Consistent with this speculation, single and double acylation mutants of AvrPphB are severely reduced in HR elicitation in resistant Arabidopsis plants (Dowen et al., 2009). It is noteworthy that the triple mutant of NopT1-GCC is autoprocessed and is not capable of eliciting cell death in tobacco plants, indicating that disruption of both acylation sites may prevent membrane association and thus proper targeting to its substrate. Considering that first mutation of the glycine residue (G50) of NopT1 does not significantly alter its cell death–eliciting activity and, secondly, that the triple mutant is not functional, our results imply that the putative palmitoylation sites (C52 and C53) of NopT1 are possibly more crucial than the myristoylation one for membrane binding and effector function. Furthermore, these findings provide evidence that autoprocessing is possibly required for unmasking the putative acylation sites, which in turn may facilitate the subsequent membrane association of NopT1. Future experiments are required to clarify whether the proteolytic activity of NopT1 is not only required for processing itself but also required for the proteolysis of another plant substrate, as in the case of AvrPphB (Shao et al., 2003a). Collectively, our data represent the first, although indirect, evidence for possible acylation of NopT1 and suggest that they may play a crucial role in its effector function. Future studies are needed to demonstrate the role of NopT1 and NopT2 in the nodulation process as well as how their mutations in critical residues for their function affect this process.

Acknowledgements

Many thanks to Prof. N. J. Panopoulos for fruitful discussions and critical review of the manuscript.

Ancillary