Virus induction of heat shock protein 70 reflects a general response to protein accumulation in the plant cytosol

Abstract

Different cytoplasmically replicating RNA viruses were shown to induce a specific subset of heat-inducible heat shock protein 70 (HSP70) genes in Arabidopsis (Arabidopsis thaliana). To identify the inducing principle, a promoterreporter system was developed for the facile analysis of differentially responding Arabidopsis HSP70 genes, by infiltration into Nicotiana benthamiana leaves. Through transient expression of individual viral cistrons or through deletion analysis of a viral replicon, we were unable to identify a unique inducer of HSP70. However, there was a positive correlation between the translatability of the test construct and the differential induction of HSP70. Since these data implied a lack of specificity in the induction process, we also expressed a random series of cytosolically targeted Arabidopsis genes and showed that these also differentially induced HSP70. Through a comparison of different promoterreporter constructs and through measurements of the steady-state levels of the individual proteins, it appeared that the HSP70 response reflected the ability of the cytosol to sense individual properties of particular proteins when expressed at high levels. This phenomenon is reminiscent of the unfolded protein response observed when the induced accumulation of proteins in the endoplasmic reticulum also induces a specific suite of chaperones.

Figure 1.

Induction of Arabidopsis HSP70 genes by heat stress or virus infection. A, Total RNA preparations from Arabidopsis leaf tissues, untreated (lane 1), systemically infected with TuMV (lane 2) or TCV (lane 3), or subjected to heat stress at 42°C (lane 4), were analyzed on RNA blots using probes specific for each of the five cytosol-targeted HSP70 genes (HSC70-1, HSC70-2, HSC70-3, HSP70, and HSP70B). B, Promoter∷reporter fusion Agrobacterium binary constructs were made for AtHSP70 and AtHSP70B utilizing 2.1 kb and 1.98 kb, respectively, of the upstream genomic DNA fused to GUS and the CaMV 19S terminator (term). C, Agrobacterium carrying the binary construct pHSP70 or pHSP70B was mixed with a second Agrobacterium culture carrying either an empty binary vector (non-stress and 42°C) or a binary vector including complete cDNAs for TuMV, TCV, TRV, or TVCV and coinfiltrated into expanded leaves of N. benthamiana. After 4 d, the leaves were stained for GUS activity. Data shown in A and C are representative of at least five independent experiments.

Figure 2.

Induction of pHSP70 by the expression of individual TCV proteins or by TCV mutants. A, The genome of TCV was dissected into its component cistrons (p88, p28, p8, p9, p38) and each cistron cloned after the CaMV 35S promoter as a binary construct. A construct, p38stop, was the same as p38 except that two stop codons were inserted immediately after the p38 ATG. In parallel, virus mutants Δ88/28, Δ88, Δ8, Δ9, and Δ38 were similarly constructed. B, Agrobacterium culture carrying the binary TCV constructs or the empty vector (non-stress and 42°C) was mixed with a culture carrying pHSP70∷GUS and infiltrated into expanded leaves of N. benthamiana. Quantitative GUS activity was measured fluorimetrically after 3 d. Assays were carried out in triplicate on individual extracts of infiltrated leaves from three plants. Bars show ±se of the mean value.

Figure 3.

Induction of pHSP70 following expression of other plant viral gene products. Constructs for the expression of TMV-MP, PSbMV-6K1CI, PSBMV-P1, and CymRSV-p19 were coinfiltrated into N. benthamiana leaves with pHSP70∷GUS or with the empty vector (non-stress and 42°C) as controls. All of these proteins showed significant induction of pHSP70∷GUS as increased GUS activity. Similarly, leaves were coinfiltrated with pHSP70∷GUS and constructs expressing TRV RNA1, RNA2, or RNAs 1 and 2 (RNA1+2). All assays were carried out on three replicates of infiltrated leaves from each of two plants. Bars represent ±se.

Figure 4.

Induction of pHSP70 by the expression of nonviral proteins, relative to protein accumulation and promoter strength. A, Five full-length Arabidopsis cDNAs (At5g08290, At1g74560, At4g16830, At3g23600, At1g71860) fused to eGFP were tested for their abilities to induce pHSP70∷GUS. The predicted functions of the selected genes are as follows: At5g08290, thioredoxin-like protein; At1g74560, nucleosome assembly protein inhibitor; At4g16830, related to cyclic nucleotide regulation of PAI-1 mRNA stability (Tillmann-Bogush et al., 1999); At3g23600, contains KOG3043 (predicted hydrolase related to dienelactone hydrolase); and At1g71860, similar to protein Tyr phosphatase 20. The expression constructs were coagroinfiltrated with pHSP70∷GUS into leaves of N. benthamiana; empty vector (non-stress and 42°C) controls were analyzed in parallel. Quantitative GUS activity was measured fluorimetrically after 3 d. B, Tissues from the experiment in A were analyzed for the concentration of expressed proteins based on a fluorimetric assay for the fused eGFP. C, Two proteins, At3g23600 fused to eGFP and PSbMV 6K1CI, were expressed using the CaMV 35S or the NOS promoters, and the levels of pHSP70∷GUS induction were compared after 3 d. D, The expression level from the CaMV 35S or NOS promoters was assessed for At3g23600 using a fluorimetric assay (left) or immunoblot assay (right) for eGFP to assess accumulation of the protein product. In the immunoblot assay, lanes 1 and 2 are from plants expressing At3g23600.eGFP from the CaMV 35S or NOS promoters, respectively. The predicted size of the fusion protein is indicated (arrow). The panel beneath confirms equal protein loadings as stained plant protein (large subunit of Rubisco). Bars represent ±se.