iTRAQ-based quantitative proteomics reveals a ferroptosis-like programmed cell death in plants infected by a highly virulent tobacco mosaic virus mutant 24A+UPD

Abstract

Plants trigger a highly orchestrated defence mechanism in response to viral infection. In this study, we aimed at understanding the molecular events that lead to more accelerated cell death in Nicotiana benthamiana plants infected with the fast-replicating TMV 24A + UPD in comparison to TMV. TMV 24A + UPD is an artificial mutant that induces more severe symptoms leading to precocious death in plants. We employed the iTRAQ-based quantitative proteomics approach to identify and map the proteomes of TMV and TMV 24A + UPD infected plants at time points that correlate with initiation of early cell death symptoms. TMV 24A + UPD proteome profile revealed 183 highly abundant proteins versus 71 for TMV infected plants. KEGG analysis revealed differentially abundant proteins in the two proteome profiles under cell death, stress signalling, protein folding, sorting, degradation, transport and catabolism. We identified unique differentially abundant proteins in the TMV 24A + UPD profile, in particular under the ferroptosis and glutathione metabolism pathways. For validation, we varied the amount of intracellular iron by supplementing plants with Fe³⁺, employing iron chelators and by virus induced gene silencing of iron storage protein ferritin gene. We also employed potent ferroptosis inhibitors ferostatin-1, liprostatin-1, and transiently silenced glutathione peroxidase 4 gene. TMV 24A + UPD infected plants showed accelerated cell death symptoms when intracellular iron was increased. Decreasing intracellular iron protected the plants from accelerated cell death. We also observed a decrease in TMV 24A + UPD induced cell death when we applied ferroptosis inhibitors. Glutathione peroxidase 4 gene-silenced plants showed enhanced cell death compared to non-silenced control plants. Our study uncovered a link between intracellular iron and accelerated lipid ROS-induced cell death in TMV 24A + UPD infected plants. We propose that the fast-replicating mutant of TMV induces a distinct and potent form of cell death akin to ferroptosis.

Background

Plant viruses have a significant impact on plant production. In response to viruses, plants mount intricate protection systems that include crosstalk between many cellular pathways and immune response. Viruses depend on host cells for successful establishment (Wang and Li 2012). The interplay between virus and host factors continues throughout the lifecycle of the virus and has an impact on pathogenesis.

Such interaction with viruses perturbs the molecular networks in plant cells and triggers expression of hundreds of genes involved in defense, susceptibility and resistance (Di Carli et al. 2012; Das et al. 2018; Garcia-Ruiz 2018). One such plant defence responses is regulated cell death, whereby damaged or unwanted cells are eliminated through specific biochemical pathways (Reape et al. 2008). Cell death in mammals is well characterized into apoptotic or non-apoptotic pathways (Elmore 2007). Although plant and animal immune systems share similarities of events leading to cell death, details in plants are less known (Mukhtar et al. 2016).

The genomes of most tobamoviruses including the Tobacco mosaic virus (TMV) have a 3′ untranslated region (UTR) that consists of an upstream pseudoknot domain (UPD) that plays an important role in translation (Niu et al. 2015). A mutated TMV was constructed by introducing different lengths of internal poly (A) tracts upstream of the UPD (Guo et al. 2015). One such mutant, TMV 24A + UPD, induced more severe leaf necrosis and stem collapse, leading to plant death within a few days. The difference in rate of virus replication and plant death are likely resulted from interactions between the introduced poly (A) tract, UPD and host factors. Host protein elongation factor 1A has been reported to interact with UPD of 3′ UTR (Zeenko et al. 2002). The pathways involved in the accelerated cell death by TMV 24A + UPD have not been identified.

Thus, to understand TMV 24A + UPD interaction with the host and events leading to the altered pathogenesis, we compared the proteome profiles of TMV 24A + UPD or TMV-infected plants using the Isobaric Tags for Relative and Absolute Quantitation (iTRAQ) method. We compared the proteome profiles of TMV- or TMV 24A + UPD inoculated N. benthamiana plants at two different time points (2 and 3 days post inoculation, dpi) that correlate with the first signs of virus-induced cell death.

Interestingly, the KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis revealed differentially abundant proteins (DAPs) unique to glutathione metabolism and ferroptosis pathways in TMV 24A + UPD infected plants. We identified a cytosolic iron chaperone poly (rC)-binding protein 1 (PCBP1) which delivers iron to ferritin. Ferritin is known to store iron in a non-toxic state (Leidgens et al. 2013). We also identified the antioxidant glutathione peroxidase. Ferroptosis is a regulated form of cell death driven by accumulation of lipid based reactive oxygen species (ROS) (Dixon et al. 2012). This cell death is dependent on free intracellular iron, which is required for the accumulation of toxic lipid ROS (Yang et al. 2014). We also identified glutathione peroxidase. In mammalian systems, glutathione peroxidase 4 (GPX4) is a key regulator of ferroptosis (Maiorino et al. 2018; Seibt et al. 2019). Based on the iTRAQ results, we infer that intracellular iron plays a key role in TMV 24A + UPD-ROS induced cell death. Further, our results demonstrated the link between intracellular iron and accelerated ROS induced cell death in TMV 24A+ UPD infected plants.

Results

Protein identification and quantification

There are various host factors involved in activation of immune response during early stages of virus infection (Lietzen et al. 2011). We employed high throughput proteomics to identify host factors that contribute to the rapid cell death induced by TMV 24A + UPD in comparison to TMV in N. benthamiana plants. We confirmed infection before sample processing, TMV coat protein was detected at 2 and 3 dpi in both TMV and TMV 24A + UPD inoculated plants. For each time point, four biological replicates were included in the samples and grouped into two for iTRAQ tagging (Fig. 1).

Fig. 1

A schematic representation of the experimental design based on two iTRAQ 8-plex experiments. a Systemic TMV or TMV 24A + UPD infection in leaves was analysed by western blot detection of viral coat protein, at 1, 2 and 3 days post-inoculation (dpi). b Total protein samples were collected at 2 and 3 dpi and processed immediately for iTRAQ labelling

Overall, proteins identified covered a wide range of biological processes. TMV 24A + UPD induced a significantly higher number of highly abundant proteins (HAPs) compared to that of TMV (Fig. 2a). Of the HAPs, 17 overlapped between TMV and TMV 24A + UPD. For less abundant proteins (LAPs), 56 overlapped between the two viruses (Fig. 2a, b).

Fig. 2

Statistics of proteome identified from iTRAQ experiment. Venn diagrams showing unique and differentially regulated proteins between TMV and TMV 24A + UPD. a Highly abundant proteins. b Lowly abundant proteins. c Total spectra for reversed-phase HPLC and LC-MS/MS identified proteins and differentially accumulated proteins. DAP, differentially abundant proteins; LAP, lowly abundant proteins, HAP, highly abundant proteins

Approximately 77,354 MS spectra identified matched the known spectra. Overall, a total of 15,996 peptides and 2166 proteins were identified (Fig. 2c). A total of 823 proteins showed differential accumulation, 354 were classified as HAPs and 469 as LAPs (Fig. 2c).

To confirm the iTRAQ quantification results, qRT-PCR was performed for 21 selected genes encoding differentially abundant proteins of interest from TMV and TMV 24A + UPD infected plants. Some genes including hypersensitive-induced response, peroxiredoxin-2e-2, and programmed cell death proteins were selected based on their involvement in plant defense and cell death. Chloroplast associated genes were selected based on their low abundance in both proteome profiles. From the TMV infected plants, 5 of the selected genes were up-regulated, while 5 were down-regulated (Additional file 1: Figure S1a). From the TMV 24A + UPD infected plants, 8 genes were up-regulated while 3 were down-regulated (Additional file 1: Figure S1b). From the qRT-PCR results, 19 of the selected genes were up-regulated or down-regulated as expected, thus validating our iTRAQ results. Taken together, we conclude that N. benthamiana proteome is significantly altered upon TMV and TMV 24A + UPD infection.

Gene ontology (GO) enrichment and KEGG analysis of differentially abundant proteins

Proteomic data of identified and quantified proteins were further analysed using different bioinformatics tools. The total differentially expressed genes in TMV and TMV 24A + UPD infected plants were 13% vs 15% for biological processes, 56% vs 58% for cellular processes respectively. Molecular function was 29% for both profiles. When highly abundant proteins were analyzed separately, TMV 24A + UPD revealed more GO terms including sulphur metabolism, immune system process, golgi membranes, plant vacuole, proteasome and protein phosphatase type 2A (Fig. 3a, b). Some terms were common in both profiles, namely ubiquitin ligase complex and ROS. The term P-body was only found in TMV profile (Fig. 3a).

Fig. 3

Gene ontology (GO) enrichment analysis of differentially expressed proteins (DAPs) by TMV and TMV 24A + UPD infection. Panels showing interactive graphs of DAPs in TMV and TMV 24A + UPD infection. a High-abundant differentially expressed proteins in TMV proteome profile. b Highly abundant proteins in TMV 24A + UPD. The sizes of the circles are proportional to the number of proteins associated with the specific term. The interactive network was summarized and plotted following the published REVIGO protocol (http://revigo.irb.hr)

From the plant defence perspectives, 4 main functional groups were analysed (Fig. 4a, b), as classified by KEGG. There were differences in DAPs between TMV and TMV 24A + UPD associated with cellular senescence, apoptosis, necroptosis, glutathione metabolism and ferroptosis (Additional file 2: Table S1). Autophagy proteins were highly abundant with TMV 24A + UPD, while signalling pathways proteins involved in mTOR, PI3K-Akt and sphingolipid signalling pathways were highly abundant in TMV proteome profile (Fig. 4a, b). MAPK signalling pathway proteins were more abundant in TMV 24A + UPD but less so with TMV. SNARE proteins involved in vesicle fusion were highly abundant with TMV 24A + UPD but not with TMV (Fig. 4b). The difference in GO terms and proteins in selected pathways between the two viruses suggest that the internal poly (A) tract in the UPD influences the pattern of plant-pathogen interaction. We conclude that the accelerated cell death in TMV 24A + UPD infected plants is preceded by increased and distinct perturbation, as compared to TMV infection.

Fig. 4

KEGG pathway analysis for TMV and TMV 24A + UPD proteome profiles. Classification of the identified differentially accumulated proteins and their distribution in the KEGG pathways predicted for TMV (a) and TMV 24A + UPD (b) proteome profiles. Highly abundant proteins are shown as up-regulated while lowly abundant proteins are denoted down-regulated. X-axis represents number of proteins identified

Validation of iron-induced oxidative stress in TMV 24A + UPD infected N. benthamiana plants

Our analysis revealed that in contrast to TMV, the TMV 24A + UPD mutant induced accumulation of glutathione peroxidase, and peroxiredoxin ROS regulators (Additional file 2: Table S1). Our data also revealed an intracellular iron regulator, PCBP1, in TMV 24A + UPD samples but not in TMV. We sought to investigate if there was a relationship between iron and oxidative stress induced by TMV 24A + UPD. We first investigated oxidative stress in virus infected plants using TBARS assay. TMV 24A + UPD showed a significantly higher amount of MDA content at 3 dpi compared to mock and TMV infected plants (Fig. 5a).

Fig. 5

Quantification of lipid peroxides, and the effect of Fe³⁺ supplementation in TMV infected Nicotiana benthamiana. a Lipid peroxidation measurement was based on the amount of TBA reactive MDA in infected leaves. MDA levels at 3 dpi was statistically significant at P < 0.01 in comparison to the mock control. Whole plant (b), Close up images (c) of the mock and virus infected leaves. Comparison of TMV and TMV 24A + UPD symptoms on N. benthamiana plants supplemented with 1 mM ethylenediaminetetraacetic acid ferric sodium salt

When plants supplemented with Fe³⁺ in the form of ethylenediaminetetraacetic acid ferric sodium salt were infected with virus, TMV 24A + UPD plants showed an accelerated cell death compared to the control plants (Fig. 5b). TMV 24A + UPD infected leaves showed more severe chlorotic blotches and collapsed quicker than the control plants (Fig. 5c).

After infiltrating the iron chelator deferoxamine mesylate (DFO) into the virus infected plants, cell death was significantly decelerated in TMV 24A + UPD inoculated plants but not in the mock or TMV inoculated plants (Fig. 6a, b). We also observed cell death inhibition with ciclopirox olamine (CPX)-treated TMV 24A + UPD inoculated plants but not in mock and TMV inoculated plants (Additional file 3: Figure S2). To further investigate the role of iron, we silenced ferritin gene by virus induced gene silencing (VIGS). TMV 24A + UPD infected plants exhibited slightly enhanced cell death symptoms compared to plants with non-silenced intact ferritin gene at 6 dpi (Fig. 6c), the symptoms were enhanced more at 12 dpi (Additional file 4: Figure S3). We did not observe a significant difference of symptoms in ferritin-silenced and control plants when TMV was inoculated (Fig. 6c). These results suggest that intracellular iron accelerates accumulation of ROS in TMV 24A + UPD. We infer that a ferroptosis-like phenomenon forms part of the cell death response induced by TMV 24A + UPD in N. benthamiana plants.

Fig. 6

Effects of intracellular iron on virus infected Nicotiana benthamiana plants. Comparison of cell death symptoms in TMV and TMV 24A + UPD infected leaves treated with iron chelators DFO at 6 dpi. a Inoculated leaves. b Whole plants. Control plants were infiltrated with control buffer (DMSO). Plants were treated with CB/DFO (1 mM). Representative data are shown. Red arrows indicate inoculated leaves. c Symptoms of TMV and TMV 24A + UPD in WT, non-silenced, VIGS-silenced ferritin gene in N. benthamiana plants

Effect of ferroptosis inhibition in infected plants

Ferrostatin-1 (Fer-1) and liprostatin-1 (Lip-1) inhibit ferroptosis via blocking the iron induced ROS accumulation, particularly lipid peroxidation (Skouta et al. 2014). Our results showed that Fer-1 and Lip-1 inhibited the strong cell death response elicited in TMV 24A + UPD inoculated plants (Fig. 7a). We observed that inoculated leaves (shown by red arrows) appeared water soaked and shrivelled in the absence of inhibitors and plants collapsed on the flower pot (Fig. 7a). Mock and TMV inoculated plants and leaves treated with inhibitors did not show such symptom of cell death (Fig. 7a).

Fig. 7

Effect of ferroptosis inhibition in virus infected Nicotiana benthamiana plants. a Comparison of cell death symptoms in TMV and TMV 24A + UPD plants treated with ferroptosis inhibitors Lip-1 and Fer-1 at 6 dpi. Control plants were infiltrated with DMSO. Plants were treated with DMSO or Fer-1 (50 μM) or Lip-1 (20 μM). Experiments were performed two times. Representative data are shown. Red arrows show inoculated leaves. b The symptoms of TMV and TMV 24A + UPD in WT, non-silenced, NbGPX4-silenced N. benthamiana. Transcription of gene(s)-of-interest was downregulated via TRV-induced gene silencing

We then sought to silence glutathione peroxidase 4 gene (NbGPX4), a gene implicated in ferroptosis (Conrad et al. 2018). In plants, glutathione peroxidase has several isoforms. We have identified phospholipid hydroperoxide (Q9FXS3-GPX4) of Nicotiana tabacum from uniprot. We cloned a gene from N. benthamiana which has a 100% nucleotide sequence identity to N. tabacum GPX4 (XM_016605968.1, GQ354813.1). A BLAST search showed that NbGPX4 shared high sequence similarity to HsGPX4 (NM_001367832.1), MmGPX4 (NM_001367995.1) and AtGPX6 (NM_117229.4), all of which have been ascribed to ferroptosis regulation. Thus, we selected NbGPX4 for this study.

TMV 24A + UPD infected plants exhibited enhanced cell death symptoms in NbGPX4-silenced plants (Fig. 7b). We did not observe such enhanced symptoms in TMV inoculated plants (Fig. 7b). We conclude that a distinct form of ferroptosis-like cell death occurs in TMV 24A + UPD infected plants.

Using various approaches, our study revealed some physiological features of ferroptosis in TMV 24A + UPD infected N. benthamiana plants (Fig. 8) as previously described (Dixon et al. 2012). We assessed ROS by quantifying malondialdehyde (MDA) content in infected plants, manipulated intracellular iron by supplementing with iron, employing iron chelators and by virus induced gene silencing (VIGS) of iron storage protein ferritin gene. We also employed potent ferroptosis inhibitors ferostatin-1, liprostatin- 1 and silenced GPX4 by VIGS.

Fig. 8

Ferroptosis-like cell death in TMV 24A + UPD infected Nicotiana benthamiana plants. Using iTRAQ, unique DAP’s in the TMV 24A + UPD profile under the ferroptosis and glutathione metabolism pathways were identified. TMV 24A + UPD infected plants showed accelerated cell death symptoms when intracellular iron was increased. Decreasing intracellular iron by employing iron chelators DFO, CPX and silencing of iron storage protein ferritin gene protected the plants from accelerated cell death. We also observed a decrease in TMV 24A + UPD induced cell death when we applied ferroptosis inhibitors ferostatin-1, liprostatin-1. Glutathione peroxidase 4 gene-silenced plants showed enhanced cell death compared to non-silenced control plants

Discussion

Isobaric tag for relative and absolute quantitation LC-MS/MS (liquid chromatography tandem mass spectrometry) technology has been used to study virus-host interactions. Previous tudies have compared virus-infected plants with healthy plants to understand these interactions (Wang et al. 2015; Alexander and Cilia 2016; Chen et al. 2017). In this study, we compared the proteome profiles of plants infected with TMV and a faster-replicating mutant TMV 24A + UPD. The faster replicating mutant induced cell death earlier than TMV. Our aim was to compare proteins involved in cell death from different defense pathways by TMV and its faster replicating mutant TMV 24A + UPD. This study focused on specific proteins that are the cause of accelerated cell death in TMV 24A + UPD treated plants. We therefore compared the proteome profiles of TMV- or TMV 24A + UPD infected N. benthamiana plants at two different time points (2 and 3 dpi) that correlate with the first signs of virus-induced cell death.

The timing and intensity of response to virus infection is crucial to protection against cell death in plants (De Vos et al. 2005). The type and specificity of the plants responses such as regulated cell death also varied accordingly. TMV and TMV 24A + UPD induced defense proteins differed in abundance. The proteome profile of TMV 24A + UPD infected plants revealed a network of defense responses including proteasome, Golgi membrane, mitogen-activated protein kinase (MAPK) and glutathione metabolism (Neumann et al. 2003; Koornneef and Pieterse 2008; País et al. 2009). ROS regulators were identified in protein profiles of both TMV and TMV 24A + UPD infected plants. There were more ROS associated proteins found with TMV 24A + UPD infected plants, as compared to that of TMV (Fig. 4a, b). Peroxiredoxin and glutathione peroxidase were highly abundant only in TMV 24A + UPD infected plants. TMV 24A + UPD appears to induce higher oxidative stress in comparison to TMV. While ROS regulators protect plants from oxidative stress, excessive amount of ROS can cause rapid cell death. Plants have other mechanisms such as RNA silencing to counter invading viruses. Plants infected with TMV showed high abundance of P-body, but TMV 24A + UPD infected plants did not. P-body plays a crucial role in RNA silencing (Eulalio et al. 2007). RNA silencing as a conserved defense pathway, evolves within plants against invading viruses. RNA interference (RNAi) suppression increases virus replication and cell death in mosquitoes (Cirimotich et al. 2009). It suppresses virus replication by targeting the secondary structure of positive-sense RNA genome (Szittya et al. 2002; Molnar et al. 2005). TMV 24A + UPD is an artificial mutant that has an altered secondary structure compared to that of TMV (Guo et al. 2015). It is plausible that TMV 24A + UPD RNA secondary structure suppresses RNAi efficiency and leads to enhanced virus replication and host cell death. Our finding suggests that TMV 24A + UPD enhanced plant defense response which led to more rapid cell death in infected plants.

Differentially abundant proteins identified were grouped into three major cell death pathways; apoptosis, necroptosis and ferroptosis (Additional file 2: Table S1). In plants, apoptosis is not markedly distinctive as compared to that of animal cells (Vanyushin et al. 2004). In this study, we did not find any significant numbers of apoptosis-related proteins. The differentially abundant proteins between TMV and TMV 24A + UPD-infected plants were comparable. Cytochrome C was linked to apoptosis in the protein profile of TMV-infected plants, while peroxiredoxin was linked to that of TMV 24A + UPD-infected plants (Additional file 2: Table S1). Peroxiredoxin guards against oxidative stress caused by abundance of peroxides (Perkins et al. 2015). Cytochrome C released from mitochondria is associated with initiation of programme cell death in plants (Martínez-Fábregas et al. 2013). It is possible that both TMV and TMV 24A + UPD can induce apoptosis in cells of infected plants. Necroptosis associated proteins were high in abundance in TMV 24A + UPD infected plants, but not with TMV. Necroptosis is associated with elevated levels of ROS (Zhou et al. 2012). It results in cell death that leads to uncontrolled disorganization of cellular contents through cell membrane rupture (Dickman and de Figueiredo 2013). Necroptosis is mediated by a complex of proteins that make up the necrosome. It is possible to switch from apopotosis to necroptosis using factors such as autophagy in mammalian cells (Goodall et al. 2016). A similar switch likely exists in plants. Ferroptosis associated proteins were highly abundant in TMV 24A + UPD infected plants. Ferroptosis is first described in mammalian cells (Dixon et al. 2012). It is morphologically and genetically distinct from other forms of cell death (Hao et al. 2018). In plants, ferroptosis-like cell death has been reported in Arabidopsis root hairs subjected to 55 °C heat shock (Distéfano et al. 2017). It is also involved in R-gene mediated resistance against rice blast fungus Magnaporthe oryzae (Dangol et al. 2018). This indicates that ferroptosis can be a conserved pathway for both plants and animals. These results also indicate that cell death from TMV infection can arise from different defense pathways in N. benthamiana plants.

One of the characteristics of ferroptosis is the increased accumulation of intracellular iron. TMV 24A + UPD induced higher abundance of oxidative stress proteins such as the ROS regulators in infected N. benthamiana plants. Many anti-oxidative enzymes require iron to catalyse electron transport (Tewari et al. 2005). Our data revealed that cell death in TMV 24A + UPD infected plants was inhibited in leaves infiltrated with iron chelators. Iron chelators have been shown to protect Arabidopsis against Pseudomonas syringae pv. tomato DC3000 (Aznar et al. 2014). An increase in intracellular iron by exogenous supplementation led to further accelerated cell death in TMV 24A + UPD infected plants. Ferritin protein regulates intracellular iron to protect plants against oxidative stress (Ravet et al. 2009). Reduced transcription levels of ferritin by VIGS in N. benthamiana plants further accelerated cell death in TMV 24A + UPD infected plants. We infer that intracellular iron is functionally associated with ferroptosis-like cell death in N. benthamiana plants infected with the fast-replicating mutant TMV 24A + UPD.

Another characteristic of ferroptosis is iron induced lipid peroxidation (Conrad et al. 2018). Lipid peroxidation in biological membranes is an indicator of oxidative stress in plants. Iron deficient plants have been shown to have decreased lipid peroxidation (Tewari et al. 2005). We quantified the MDA content using TBARS assay. While MDA has been used to ascertain ferroptosis associated lipid peroxidation in plants (Dangol et al. 2018), it is not exclusively generated by lipid peroxidation. We thus further employed ferroptosis inhibitors to ascertain lipid peroxidation and its role in cell death. Ferroptosis has been characterized using inhibitors ferrostatin-1 and liproxstatin-1 (Friedmann Angeli et al. 2014; Linkermann et al. 2014; Distéfano et al. 2017; Dangol et al. 2018). We observed decreased cell death in Fer-1/Lip-1 treated plants which were infected with TMV 24A + UPD. We note that in this study, we used higher concentration of ferroptosis inhibitors relative to other studies (Distéfano et al. 2017; Dangol et al. 2018). Our study utilizes infiltration of the inhibitors in mature leaves while other studies use inhibitors in growth media. We determined optimal concentration to determine the effect of the inhibitors on cell death and avoided concentration that triggered a necrotic cell death response.

To further ascertain lipid peroxidation, we knocked down GPX4 by VIGS. GPX4 is the major antioxidant associated with ferroptosis. GPX4 knockdown in mice has been shown to increase lipid peroxidation dependent cell death (Imai et al. 2017). We observed accelerated cell death in GPX4 knockdown plants infected with TMV 24A + UPD.

TMV induces cell death in N. benthamiana with symptoms of rapid leaf necrosis and stem collapse, leading to plant death within a few days. We assessed these symptoms to show decrease or acceleration of cell death symptoms when we applied inhibitors, increased iron concentration and silenced GPX4 and ferritin gene. As these symptoms are not exclusively associated with ferroptosis, we do not rule out other forms of cell death in addition to the ferroptotic-like cell death inferred from this study. Taken together, our data suggest that TMV 24A + UPD can induce ferroptosis-like programmed cell death in N. benthamiana plants.

Conclusions

We employed the iTRAQ-based quantitative proteomics approach to determine the proteomes of TMV and TMV 24A + UPD-infected plants to compare proteins involved in cell death from different defense pathways by TMV and its faster replicating mutant TMV 24A + UPD. This study provides evidence that TMV 24A + UPD accelerated cell death involves several pathways in N. benthamiana plants. KEGG analysis revealed differences in differentially abundant proteins of the two proteome profiles under cell death, stress signalling, protein folding, sorting, degradation, transport and catabolism. We identified unique differentially abundant proteins in the TMV 24A + UPD profile, in particular under the ferroptosis and glutathione metabolism pathways. We showed a link between intracellular iron and accelerated lipid ROS induced cell death in TMV 24A+ UPD infected plants. These results suggest that this TMV mutant induces a distinct form of cell death similar to that of ferroptosis.

Methods

Plant material and virus inoculation

A total of 32 four-week-old N. benthamiana plants were used for this study. TMV U1 was used for this study. All plants were grown under 16 h light and 8 h dark at 25 °C. Fully expanded leaves were dusted with carborundum and mechanically inoculated with 2.5 μg of in vitro transcribed RNA in GKP buffer (50 mM glycine; 30 mM K₂HPO₄, pH 9.2; 1% bentonite; 1% celite).

Leaf protein preparation

Total protein was extracted from the inoculated and newly emerged leaves of TMV and TMV 24A + UPD treated plants at 2 and 3 dpi, respectively. For each time point, four biological replicates were included in the samples and grouped into two iTRAQ tagging (Fig. 1). Total protein was extracted and quantified using the Bradford assay (Das et al. 2018). Western blotting was used to confirm virus infection using antisera against TMV coat protein.

Sample preparation

The protein concentration was determined using Quick Start™ Bradford Protein Assay and normalized to 100 μg for all the samples. The protein samples were directly incorporated into a polyacrylamide gel matrix without electrophoresis as described (Lu and Zhu 2005). Briefly, proteins were immobilized in a gel matrix and then reduced with Tris (2-carboxyethyl) phosphine (TCEP) and then alkylated with methyl methanethiosulfonate (MMTS). Trypsin in-gel digestion was carried out at 37 °C for 16 h. The proteins were extracted from the gel using 0.5 M triethylammonium bicarbonate (TEAB), pH 8.5. The samples were processed for iTRAQ by labelling with iTRAQ 8PLEX kit (ABI-SCIEX), following the manufacturer’s protocol.

iTRAQ labelling and tandem mass spectrometry

A total of eight samples were tagged (Fig. 1). Labelling was performed as described (Meng et al. 2014). The samples were pooled together for an 8-plex experiment. Sample desalting was carried out using C18 Sep-Pak (Pierce) cartridge followed by protein lyophilisation. Lyophilized protein samples were prepared by dissolving in 1 mL, 20 mM ammonium formate in water, pH 10. For downstream LC–ESI–MS/MS analysis, iTRAQ labelled peptides were subjected to reverse-phase high performance liquid chromatography (RP-HPLC) equipped with C18 column. Mobile phase A comprises of 20 mM ammonium formate water, while the mobile phase B consists of 80% acetonitrile (ACN). Fractions were collected at 1-min intervals with the flow rate set at 0.5 mL/min. The eluents were pooled into five fractions, lyophilized and reconstituted in 2% ACN in water. Each fraction was subjected to MS analysis on a 5600 Triple TOF system (SCIEX). The precursor ion selection parameters were set at mass range 400–1800 m/z and accumulation time 250 ms/sec.

Data analysis was performed as described by Meng et al. (2014). MS/MS data were searched against protein sequence database Nicotiana using ProteinPilot™ software. The protein ratios were weighted and normalized by the median ratio. The ratios with p-values< 0.05, and/or fold changes > 1.5 (up-regulation/highly abundant) and < 0.67 fold (down-regulation/low abundance) were considered significant. Identification and classification of differentially regulated proteins of TMV and TMV 24A + UPD were analysed separately by comparing the protein profiles at 3 dpi against 2 dpi.

Bio-informatics analysis

Differentially abundant proteins (DAPs) were functionally classified by first assigning GO and KO ID using eggnog-mapper (http://eggnogdb.embl.de/#/app/emapper) against eggnog 4.5.1 database, within the taxonomic group Streptophyta. The Gene Ontology interactive network were summarized and plotted following published REVIGO protocol (http://revigo.irb.hr. KEGG pathway analysis was conducted using KEG mapperhttp://www.genome.jp/kegg/tool/map_pathway.html). Venn diagrams were generated using UGen Venn software (http://bioinformatics.psb.ugent.be/webtools/Venn/).

Validation of differentially abundant proteins using real time PCR analysis

Total RNA was isolated from pooling 4 replicate samples for each treatment. The quality and quantity were confirmed by agarose gel electrophoresis and NanoDrop™ measurement, respectively. RNA was converted into cDNA by reverse transcription using oligo d(T) primer. All Primers used are shown (Additional file 2: Table S2). Quantitative real-time RT-PCR was performed to determine the expression of selected host genes of interest, using the KAPA SYBR® FAST universal qPCR kit and CFX384TM real time PCR detection system (Bio-Rad).

Housekeeping genes 5.8S rRNA gene and actin were used as internal controls in the calculation of relative transcript levels at 2 and 3 dpi with TMV and TMV 24A + UPD. Relative quantification of transcripts was conducted using comparative CT method (2^−ΔΔCT method). The experiment was conducted for all biological replicates with three technical triplicates for each sample.

Molecular cloning and quantitative PCR

Ferritin and GPX4 genes were amplified from N. benthamiana cDNA using primers NtFerrF/R and NbGPX4F/R (Additional file 2: Table S3). Cloning of fragments into vector PTRV2 (VIGS) was conducted using Gibson Assembly® (GA) cloning kit from NEB. Primers were designed to overlap both the vector and cloned PCR fragments. Quantitative real-time (qRT-PCR) to determine GPX4 and ferritin gene expression levels was performed. The primers used were DR_qGPX4 F/R, DR_qFerF/R and qGPX4F/R (Additional file 2: Table S3).

Virus-induced gene silencing

Amplified target gene fragments were inserted into pTRV2. Together, pTRV1 and pTRV2 were introduced into Agrobacterium by electroporation. Agrobacteria containing pTRV1 and pTRV2 or pTRV2 derivative plasmids were grown overnight at 28 °C. Bacteria were resuspended in agroinfiltration buffer and mixed at 1:1 ratio as previously described (Liu et al. 2002). After 4-h incubation at room temperature, the mixed agrobacteria cultures were infiltrated into the leaves of 6-leaf-stage N. benthamiana plants. Silenced phenotypes appeared in the upper leaves approximately 10 days post-infiltration. Gene silencing was confirmed by qRT-PCR in the silenced plants. It showed significant down-regulation of the targeted genes relative to non-silenced plants (Additional file 5: Figure S4).

Analyzing the role of ferroptosis in TMV-infected plants

Deferoxamine (deferoxamine mesylate salt; DFO) and ciclopirox olamine (CPX) act as iron chelators, while ferrostatin-1 (Fer-1) and liproxstatin-1 (Lip-1) are potent inhibitors of ferroptosis. The chelators/inhibitors were used at the following concentrations: DFO (1 mM), CPX (250 μM), Fer-1 (50 μM) and Lip-1 (20 μM). First, the plants were inoculated with viral RNA as previously described and then infiltrated with the reagents above after 12 h.

The solution was infiltrated into fully expanded leaves at six-leaf stage using a syringe without a needle. Control plants were infiltrated with DMSO. On each plant, three leaves were infiltrated. Four plants were used for each experiment. Experiments were performed three times.

Thiobarbituric acid reactive substances (TBARS) assay

Estimation of lipid peroxidation in infected N. benthamiana leaves was assayed using the measurement of malondialdehyde (MDA) content (Heath and Packer 1968). Briefly, 3 mL of trichloroacetic acid (TCA) was added into 500 mg of leaf tissue and ground up with a pestle and mortar. The resulting homogenate was centrifuged at 10,000 ×g for 10 min. The supernatant was transferred to a fresh tube. One mL of the supernatant was added into 4 mL of a mixture of 20% TCA and 0.5% thiobarbituric acid (TBA). The mixture was capped tightly and heated in a water bath at 95 °C for 30 min and then quickly cooled on ice. The cooled mixture was centrifuged at 10,000 ×g for 15 min.

The supernatant was transferred into cuvettes and the absorbance was determined using a spectrophotometer at 532 nm. The concentration of MDA was calculated using the Beer Lamberts formula A = Ƹlc: A = Absorbance, Ƹ = extinction coefficient, l = path length. The extinction coefficient of MDA is known to be 155 mM^− 1 cm^− 1 (Hodges et al. 1999).

Fe³⁺supplementation

We supplemented iron to experimental plants through soil by irrigating N. benthamiana plants with ethylenediaminetetraacetic acid ferric sodium solution at a final concentration of 1 mM. Iron is assimilated by the roots from the soil (Briat et al. 2007; Kim and Guerinot 2007). After two rounds of irrigation with equal amount of water, the plants were inoculated with virus. Mock plants did not include ethylenediaminetetraacetic acid ferric sodium salt in the irrigation water. Experiments were performed three times.