Nodulin 22, a novel small heat-shock protein of the endoplasmic reticulum, is linked to the unfolded protein response in common bean

Abstract

The importance of plant small heat shock proteins (sHsp) in multiple cellular processes has been evidenced by their unusual abundance and diversity; however, little is known about their biological role. Here, we characterized the in vitro chaperone activity and subcellular localization of nodulin 22 of Phaseolus vulgaris (PvNod22; common bean) and explored its cellular function through a virus-induced gene silencing-based reverse genetics approach. We established that PvNod22 facilitated the refolding of a model substrate in vitro, suggesting that it acts as a molecular chaperone in the cell. Through microscopy analyses of PvNod22, we determined its localization in the endoplasmic reticulum (ER). Furthermore, we found that silencing of PvNod22 resulted in necrotic lesions in the aerial organs of P. vulgaris plants cultivated under optimal conditions and that downregulation of PvNod22 activated the ER-unfolded protein response (UPR) and cell death. We also established that PvNod22 expression in wild-type bean plants was modulated by abiotic stress but not by chemicals that trigger the UPR, indicating PvNod22 is not under UPR control. Our results suggest that the ability of PvNod22 to suppress protein aggregation contributes to the maintenance of ER homeostasis, thus preventing the induction of cell death via UPR in response to oxidative stress during plant-microbe interactions.

Fig. 1

Expression profile of PvNod22 in bean plants. A, Relative expression levels of PvNod22 in diverse plant tissues of mature plants cultivated in optimal conditions by quantitative polymerase chain reaction (qPCR) using specific primers. First strand cDNA was synthesized and used as template for qPCR amplification. PvNod22 expression levels were normalized against Ef1-α (elongation factor 1-α). The fold change in expression was obtained by comparing the expression ratio of PvNod22 in trifoliate leaves from mature plants versus its expression in all other tissues. P < 0.001, analysis of variance (ANOVA). B, A similar analysis was carried out using 3-week-old stressed Phaseolus vulgaris plants. Bean plants were sprayed with 1 mM hydrogen peroxide or subjected to 4°C (heat shock) for 30 min. After normalization against Ef1-α, the fold change in expression was obtained by calculating the expression ratio of PvNod22 in trifoliate leaves from control plants versus its expression in stressed plants. P < 0.02, ANOVA. Values in both graphs represent the mean and standard deviation of two biological samples with three technical replicates each. In both graphs, letters represent significantly different means compared with empty vector–treated bean plants, according to Tukey’s multiple comparison test (P < 0.01).

Fig. 2

Purified PvNod22 displays chaperone activity. A, The fusion protein was purified by affinity chromatography using Ni-NTA agarose resin (Invitrogen, Carlsbad, CA, U.S.A.) as affinity matrix and was checked on a 13.5% sodium dodecyl sulfate-polyacrylamide electrophoresis gel, followed by immunoblot analysis using rabbit anti-PvNod22 antiserum. Lanes 1 and 4, crude protein extract from Escherichia coli XL1-Blue transformed with pQE30-PvNod22 plasmid; lanes 2 and 5, PvNod22 inclusion bodies in 8 M urea; and lanes 3 and 6, purified 6xHis-tagged PvNod22. B, PvNod22 facilitates luciferase refolding. Firefly luciferase (Luc, 50 nM) was denatured at 42°C and were allowed to refold in the presence of 1 or 3 μM of PvNod22. Registered activity of native (■) or denatured (○) Luc; 3 μM of 6xHis-tagged GFP (◇) or superoxide dismutase (×) as negative control assays. Luc reactivation was observed in the presence of either 1 (△) or 3 μM (▽) of PvNod22. The Luc activity of the unheated sample at the onset of the experiment was defined as 100% (set to 1 in the graph). Normalized data were plotted as the mean and standard deviation of three independent experiments. RLU = relative light units.

Fig. 3

PvNod22 is a resident chaperone of the endoplasmic reticulum (ER). Epithelial cells from tobacco (Nicotiana tabacum SR1 cv. Petit Havana) leaves were infiltrated with Agrobacterium sp. strain GV3101 cultures cotransformed with SP-GFP-PvNod22, or PvNod22-GFP and ERyk (an ER luminal marker), or AtERD2-YFP (a Golgi marker). After 3 days, the subcellular localization of each fluorescent protein was analyzed by confocal microscopy. A and D, SPGFP-PvNod22 or G and J, PvNod22-GFP fluorescence and B and H, the pattern of ERyk and E and K, AtERD2-YFP distribution are shown. Merged images of C, SP-GFP-PvNod22 with ERyk and F, AtERD2-YFP or I, PvNod22-GFP with ERyk or L, AtERD2-YFP are also shown. PvNod22 preferentially co-localized with the ER marker. Bar length = 20 μm.

Fig. 4

Specificity of the VIGS:PvNod22 silencing construct on the expression of PvNod22 and other bean small heat shock proteins (sHsp). A, Phylogenetic relationship between PvNod22 and other Phaseolus vulgaris sHsp. The figure was created by Muscle alignment and is based on full-length amino acid sequences. The sHsp lbpA from Escherichia coli is included as outgroup in this analysis. Colored circles represent putative subcellular localization; in green, chloroplast; in yellow, cytoplasm; in red, mitochondria; in blue, peroxisome; in black, endoplasmic reticulum; and in gray, endomembrane system. B, Amplification of the clade that includes PvNod22 (shaded area in A). C, Relative expression levels of sHsp encoding genes in trifoliate leaves of PvNod22-silenced plants 3 weeks after infection compared with vector-treated plants of the same age. Plotted data represent the sHsp expression ratio between vector- and VIGS: PvNod22-treated bean plants. PvNod22, P < 0.015; Student’s test. D, Relative expression levels of these sHsp in trifoliate leaves of vector-treated plants after heat shock. In all cases, expression levels were normalized against Ef1-α values. Plotted data represent the sHsp expression ratio between heat-treated and untreated empty vector–infected bean plants. P < 0.015; Student’s test. Data and standard deviation values in both quantitative polymerase chain reaction experiments were obtained from two biological samples with three technical replicates each.

Fig. 5

Silencing of PvNod22 in common bean plants induces the development of necrotic lesions. Experimental sets of A and D, mock-, B and E, vector- and C and F, VIGS:PvNod22-inoculated bean plants were generated as described (Díaz-Camino et al. 2011). Inoculation with the mottle virus empty vector was characterized by mild chlorotic mottling and rugosity of the foliage (A and B), whereas PvNod22-silenced (VIGS: PvNod22) plants resulted in the incipient development of necrotic lesions (C). Three weeks after infection, the yellowish mottling caused by viral infection seemed to be accentuated (E compared with F), the foliage showed severe blistering, and the necrotic damage was extended in the PvNod22-silenced plants.

Fig. 6

PvNod22-silenced bean plants do not accumulate more mottle virus (BPMV) than empty vector–infected plants. A, Immunoblot analysis showing levels of the large (42 kDa) coat protein of BPMV in 15 μg of total leaf protein extracts obtained from 3-week-old mock- (lane 1), vector- (lane 2), and VIGS:PvNod22-inoculated (lane 3) plants. B, Protein-loading control; Coomassie blue staining.

Fig. 7

Silencing of PvNod22 induces unfolded protein response (UPR) in common bean. Relative expression levels of key bean genes involved in UPR (PvbZip28, PvBiP1/2, PvUbq10), genes encoding reactive oxygen species detoxifying enzymes (PvCat, PvSOD, PvChSy), and pathogenesis-related genes (PvPR1, PvPR2, PvPR3) were determined in 3-week-old empty vector and VIGS:PvNod22 plants by quantitative polymerase chain reaction (qPCR). Total RNA was isolated from each biological sample. First strand cDNA was synthesized and subjected to qPCR. Expression levels were normalized against Ef1-values. Ratios of expression in empty vector to VIGS:PvNod22 plants are graphed. Asterisks represent significantly different means according to statistical analysis (P < 0.05). These values represent the mean and standard deviation of two biological samples with three technical replicates each.

Fig. 8

PvNod22 expression is not induced by unfolded protein response (UPR)–inducing chemicals in bean plants. Relative expression levels of PvBiP1/2, a key protein involved in UPR, and PvNod22 were determined by quantitative polymerase chain reaction (qPCR) in 2-day-old bean seedlings treated with 10 μg of tunicamycin (Tm) per milliliter or 10 mM dithiotreitol (DTT). Total RNA was isolated from each biological sample at the indicated time points. First-strand cDNA was synthesized and was subjected to qPCR. Expression levels were normalized against Ef1-α values. PvNod22, P = 0.69, not a significant change in expression. Values represent the mean and standard deviation of duplicate experiments.