Transcriptome profiling of pepper leaves by RNA-Seq during an incompatible and a compatible pepper-tobamovirus interaction

Upon virus infections, the rapid and comprehensive transcriptional reprogramming in host plant cells is critical to ward off virus attack. To uncover genes and defense pathways that are associated with virus resistance, we carried out the transcriptome-wide Illumina RNA-Seq analysis of pepper leaves harboring the L3 resistance gene at 4, 8, 24 and 48 h post-inoculation (hpi) with two tobamoviruses. Obuda pepper virus (ObPV) inoculation led to hypersensitive reaction (incompatible interaction), while Pepper mild mottle virus (PMMoV) inoculation resulted in a systemic infection without visible symptoms (compatible interaction). ObPV induced robust changes in the pepper transcriptome, whereas PMMoV showed much weaker effects. ObPV markedly suppressed genes related to photosynthesis, carbon fixation and photorespiration. On the other hand, genes associated with energy producing pathways, immune receptors, signaling cascades, transcription factors, pathogenesis-related proteins, enzymes of terpenoid biosynthesis and ethylene metabolism as well as glutathione S-transferases were markedly activated by ObPV. Genes related to photosynthesis and carbon fixation were slightly suppressed also by PMMoV. However, PMMoV did not influence significantly the disease signaling and defense pathways. RNA-Seq results were validated by real-time qPCR for ten pepper genes. Our findings provide a deeper insight into defense mechanisms underlying tobamovirus resistance in pepper.


Results and discussion
Disease symptoms and virus multiplication. ObPV inoculation resulted in the development of typical local necrotic lesions (HR) on the inoculated pepper leaves at 3 days post-inoculation (dpi). These visual symptoms were already shown in an earlier publication 36 . ObPV-inoculated leaves were shed approximately at 7 dpi, but the plants survived the inoculation. In contrast, PMMoV was able to spread into the whole plant, the infection became systemic without any visible disease symptoms according to earlier results. Leaf abscission was not observed at PMMoV inoculated leaves 2,36 .
We isolated total RNA extracts from the ObPV-, PMMoV-and mock-inoculated pepper leaves at 4, 8, 24 and 48 h post-inoculation (hpi), so we obtained 12 RNA libraries. First, we investigated the multiplication of ObPV and PMMoV in the inoculated pepper leaves by measuring the amount of mRNAs encoding viral movement proteins (MP) and coat-proteins (CP) in the total leaf RNA extracts by quantitative, real-time RT-PCR. Generally, the accumulation of PMMoV was markedly stronger than that of ObPV in spite of the absence of any symptoms on PMMoV-inoculated leaves. This result confirmed earlier experiments with other methods 2,38 . The transcript abundances of PMMoV MP and CP genes were 3.8-fold and 97-fold higher than those of ObPV at 48 hpi, respectively (Fig. 1). The expression of ObPV MP and CP genes gradually increased in the inoculated leaves between 4 and 48 hpi. Interestingly, the amounts of mRNAs encoding MP and CP of PMMoV slightly but significantly declined at 8 hpi as compared to 4 hpi, but later they very significantly rose at 24 and 48 hpi (Fig. 1).

Identification of differentially expressed genes (DEGs).
Gene expression patterns of the 12 RNA libraries were investigated by a transcriptome-wide RNA-Seq method and so we obtained expression data of about 31,000 pepper genes in each RNA library. General mapping statistics are available in Supplementary  Table S1. For each RNA library we established a list of differentially expressed genes (DEGs) by filtering all expression data with two threshold values: p ≤ 0.01 and − 1 ≥ log2 fold change (FC) ≥ 1. The cumulative list of DEGs for all 12 libraries as well as the DEG lists separately for each time points are presented in Supplementary  Table S2. Generally, ObPV (incompatible interaction) exerted a much stronger effect on gene expression patterns than PMMoV (compatible interaction) (Fig. 2). A Venn-diagram shows that the proportion of overlapping DEGs between ObPV-and PMMoV-inoculated leaves was only 12.6% ( Fig. 2A). The group of overlapping DEGs showed a high diversity, the most characteristic feature was the down-regulation of a large number of genes related to photosynthesis and carbon fixation. In addition, several overlapping genes were associated with porphyrin and chlorophyll metabolism, hormone signal transduction and ribosomal functions (Supplementary Table S2). In ObPV-inoculated leaves, the number of DEGs generally increased concomitantly with the develop- www.nature.com/scientificreports/ ment of infection, in the case of both up-or down-regulated genes (Fig. 2B). In contrast, the number of DEGs was intriguingly low in PMMoV-inoculated leaves at 8 and 24 hpi but then rapidly increased at 48 hpi, mostly in the case of down-regulated genes (Fig. 2B). An earlier transcriptome-wide analysis of pepper leaves inoculated with the PMMoV-HLD isolate identified 172 up-regulated and 25 suppressed DEGs 33 . However, in this study the RNA samples were taken at 9 days post-PMMoV-inoculation, therefore these results are not comparable to ours. The first overview of DEG lists showed that large, diverse sets of genes are activated or down-regulated by both ObPV and PMMoV (except for 8 and 24 hpi for PMMoV). The most predominantly up-and down-regulated genes identified in ObPV-inoculated pepper leaves at 24 hpi are shown in Table 1. Upon ObPV inoculation, the most strongly induced pepper genes encoded specific transcription factors, pathogenesis-related (PR) proteins, enzymes of phenylpropanoid and terpenoid biosynthesis, enzymes of ethylene biosynthesis, fatty acid desaturases (FADs) and glutathione S-transferases (GSTs) ( Table 1). The massive induction of several pepper FADs by ObPV was already analyzed in detail 39 . Intriguingly, almost all strongly ObPV-induced genes displayed also a marked up-regulation by 5 mM ethephon (ethylene precursor) ( Table 1; Supplementary Table S2). These results clearly showed the prominent importance of ethylene signaling in the up-regulation of defense genes in ObPV-inoculated leaves. The early and strong accumulation of ethylene in ObPV-inoculated pepper leaves at 24 hpi was already reported 2 . Furthermore, numerous strongly ObPV-inducible pepper genes showed marked induction by salicylic acid (Supplementary Table S2), which confirmed its well-known role in antiviral defense. Methyl-jasmonate and ABA, however, seem to play a minor role in the inducibility of defense genes (Supplementary Table S2). In contrast to ObPV, PMMoV induced genes encoding a polyphenol oxidase, PR-proteins, a smotin-like protein, a branched-chain-amino-acid aminotransferase, an expansin and an aquaporin at 24 hpi (Supplementary Table S2 Supplementary Table S7). Reverse transcriptions (RT) of viral RNAs were carried out by using the reverse primers of the MP or CP specific primer pairs instead of an Oligo(dT) primer. Mean values of three independent experiments are shown ± SD. Relative expression values were calculated by dividing each expression value with the expression of ObPV genes at 4 hpi. The symbols *, ** and *** show significant differences between ObPV-and PMMoV-inoculated plants at P < 5%, < 1% and < 0.1%, respectively.  Table S4). The detailed KEGG analysis showed that the most prominent effect of ObPV inoculation was the robust suppression of a large number of genes related to photosynthesis, carbon fixation and photorespiration (Supplementary Table S4). Down-regulation of photosynthesis genes have often been observed in virus-infected plants, but the exact mechanisms of this down-regulation are not known 26,44 . In addition, genes related to porphyrin and chlorophyll metabolism, carotenoid biosynthesis and starch biosynthesis were also considerably suppressed by ObPV. Earlier we already observed the strong inhibition of photochemical energy conversion and a declining chlorophyll a content in ObPV-inoculated pepper leaves 36,41 . On the other hand, ObPV strongly induced several energy-producing pathways like glycolysis, citrate cycle, oxidative phosphorylation and fatty acid degradation (Supplementary Table S4). During virus infection the leaves have an increased energy demand probably to the biosynthesis of virus-induced defense compounds 44 . We observed also the strong up-regulation of several invertase (beta-fructofuranosidase) genes by ObPV (CaCWINV1/XM_016688054; CaAIV-18/XM_016709343; CaBRFUCT /XM_016713500; CaINVA/ XM_016717981) (Supplementary Table S2), which also indicated the elevated energy demand of virus-inoculated leaves 45 . In accordance with the induction of invertase genes, we demonstrated earlier the marked accumulation of glucose and fructose in ObPV-inoculated leaves 41 . Furthermore, ObPV markedly induced the MAPkinase signaling cascades, calcium-dependent signaling, the phosphatidylinositol signaling system, protein catabolism, protein processing in the endoplasmic reticulum, cysteine and methionine metabolism, tyrosine and phenylalanine metabolism, phenylpropanoid biosynthesis, glycerolipid metabolism, and terpenoid biosynthesis (Supplementary Table S4). In contrast to ObPV, PMMoV exerted much weaker effects on pepper metabolic routes. PMMoV also suppressed genes associated with photosynthesis, carbon fixation, photorespiration as well as porphyrin and chlorophyll metabolism but to a much lesser extent than ObPV. PMMoV inoculation induced Ca-dependent signaling, protein processing in the endoplasmic reticulum, and glycerolipid metabolism. However, PMMoV did not influence significantly the genes related to energy production (cellular respiration and  www.nature.com/scientificreports/ beta oxidation), MAP-kinase signaling and the biosynthesis of defense compounds like phenylpropanoids (Supplementary Table S4).
Detailed analysis of six selected DEG groups. For a further, detailed analysis we arbitrarily selected six groups of pepper genes that were ordered into separate lists by manual curation. These datasets comprise genes encoding immune receptors, transcription factors, pathogenesis-related proteins (PR-proteins), as well as enzymes of terpenoid biosynthesis, ethylene metabolism and sulfur metabolism (Supplementary Table S5). We suppose that these robustly virus-inducible genes encode key factors of antiviral defense.
(1) Immune receptors Resistance (R)-proteins generally contain a nucleotide-binding (NB) motif and a leucine-rich repeat (LRR) motif. In pepper, 755 genes were identified that encoded potential NB-LRR-coding protein sequences 46 . The L 3 gene (GenBank LOC107854789) 12 that confers resistance to tobamoviruses was identified among the genes detected in our RNA-Seq experiments, but the expression of L 3 gene was not altered by ObPV or PMMoV (data not shown). Nevertheless, we found two close L 3 homologs, which were significantly up-regulated by ObPV at 48 hpi (CaRPP13L1/XM_016716577, CaRPP13L3/XM_016724966), Supplementary Table S5). The R-proteins encoded by these genes have a typical CC-NBS-LRR structure 12 . In addition, we found two resistance genes encoding TIR-NBS-LRR type R-proteins (TMV resistance protein N-like genes), which were also markedly activated by ObPV at 24 and 48 hpi (CaN_XI/XM_016696219 and CaN/XM_016696926, Supplementary Table S5). Table 1. The most predominantly up-and down-regulated DEGs in pepper leaves at 24 h following inoculation with Obuda pepper virus (ObPV) as detected by RNA-Seq (p ≤ 0.01). Gene expression values were compared to mock-inoculated leaves. Expression data of pepper leaves treated with 5 mM ethephon (ethylene precursor) for 6 h were calculated from publicly available raw data in the Sequence Read Archive (SRP265260) published by Lee et al. (2020) 42 . Ethephon-induced gene expression values were compared to water-treated control. Abbreviations: n.ch.: no significant change; n.d.: no data available. www.nature.com/scientificreports/ PMMoV did not influence the expression of these R-genes. The exact function of these proteins in tobamovirusinoculated pepper leaves is unknown yet. Beside the above R-proteins, which are mainly located in the cytoplasm, numerous cell surface receptor kinases play critical roles in virus resistance as immune receptors 47,48 . We identified numerous ObPV-induced genes that encode leucine-rich repeat receptor protein kinases, serine/threonine-protein receptor kinases, lectin receptor kinases (LecRKs) and other types of receptor kinases (Supplementary Table S5). These receptor kinases are typically transmembrane proteins, which can perceive diverse extracellular signals and transduce them by an intracellular protein kinase domain towards the nucleus by various signal transduction pathways. It is not known whether these plasma membrane-anchored receptors sense viral components through their extracellular or intracellular domain 47 . A particularly large set of LecRKs genes were robustly induced by ObPV. LecRKs possess an extracellular domain sharing sequence similarity with legume lectins. This lectin-like domain can bind various molecules like oligosaccharides or plant hormones 49 . Interestingly, suppressed expression of the LecRK-S.5 gene in pepper led to enhanced susceptibility against TMV and PMMoV 50 .
(2) Transcription factors Generally, six transcription factor families are associated with plant defense mechanisms: AP2/ERF, bHLH, bZIP, MYB, NAC and WRKY 51 . The pepper genome contains 2,153 transcription factors (6.25% of the total genes), which were classified into 80 gene families 27 . In our studies, genes encoding the AP2/ERF, heat-shock transcription factor, NAC, WRKY and ZAT families were robustly up-regulated by ObPV, whereas those of the bHLH and TCP families were suppressed. PMMoV exerted only a negligible effect on the expression of transcription factor genes (Supplementary Table S5).
In the pepper genome 142 ethylene-responsible transcription factors (ERFs) were identified 52 . In A. thaliana, ERFs were classified into 12 groups. Pathogen-inducible A. thaliana ERF genes typically belong to group IX 53 . In our work, several strongly ObPV-inducible pepper ERFs (CaERF1A/XM_016714653, CaPTI5/XM_016706561, CaERF1B/XM_016717675) showed homology to group IX genes of A. thaliana. Interestingly, constitutive overexpression of the tobacco ERF5 gene (homolog to pepper gene CaERF1B/XM_016717675) led to enhanced resistance to TMV, demonstrated by strongly reduced viral RNA level 54 .
Twenty-five heat shock transcription factors (HSTFs) were identified in pepper 55 . ObPV significantly activated the expression of ten HSTFs, particularly those belonging to the B class of HSTFs. Thus, the genes CaHSFB3/ XM_016688344, CaHSFB3/XM_016716424 and CaHSFB1/XM_016705606 were up-regulated 23-, 13-and 7.5fold by ObPV at 24 hpi, respectively (Supplementary Table S5). Although heat shock transcription factors are known to have a pivotal role in the regulation of immune response 56 , very limited information is available about their functions in virus-infected plants 57,58 .
The WRKY family is well characterized in pepper 59 , and their participation in pepper-tobamovirus interactions is also amply demonstrated 5,60 . In our studies, 23 WRKY genes were activated by ObPV at 24 and 48 hpi, and some of them were among the most strongly ObPV-activated transcription factor genes. In contrast, ObPV suppressed the expression of only three WRKYs (in two cases transiently). PMMoV slightly induced five WRKY  1  4  1  3  9  5  1  7  6  7  1  2  1  2  2  2  5  0  2  8  6  3  3  m  s  i  l  o  b  a  t  e  m  e  t  a  r  d  y  h  o  b  r  a  C  0  2  1  7  6  3  2  6  1  7  1  4  5  9  3  1  0  7  1  7  m  s  i  l  o  b  a  t  e  m  d  i  c  a  o  n  i  m  A  9  5  2  1  4  0  1  9  5  8  6  6  7  3  1  2  m  s  i  l  o  b  a  t  e  m  y  g  r  e  n  E  2  2  4  6  9  1  6  4  1  4  6  0  1  2  6  5  8  m  s  i  l  o  b  a  t  e  m  d  i  p  i  L  8  7  2  8  1  2  1  5  1  1  8  7  2  5  2  5  s  e  t  i  l  o  b  a  t  e  m  y  r  a  d  n  o  c  e  s  r  e  h  t  o  f  o  s  i  s  e  h  t  n  y  s  o  i  B  Metabolism of terpenoids and polykeƟdes  6  29  42  6  1  2  4  4  24  46  5  10  0  2  1  7  6  3  2  6  1  7  2  1  2  5  1  m  s  i  l  o  b  a  t  e  m  e  d  i  t  o  e  l  c  u  N  0  2  3  2  4  4  1 Table S5). Furthermore, 13 NAC and 8 ZAT genes were up-regulated by ObPV, particularly robustly in the case of ZATs. NAC and ZAT transcription factors are known to participate in antiviral defense 23,61 , but few information is available about them in pepper. The above results warrant further studies on the role of ObPV-induced transcription factors in the tobamovirus resistance of pepper.
In our studies, the PR-3, PR-9 and PR-10 gene families were particularly highly induced by ObPV (Supplementary Table S5). Notably, seven PR-10 genes were induced by ObPV. The most strongly induced CaPR-10 gene (XM_016710983) was activated 79-and 596-fold by ObPV at 24 and 48 hpi, respectively, as compared to control values. The RNA-Seq analysis of CaCV-infected pepper leaves revealed the high CaCV-inducibility of PR-3, PR-9 and PR-10 genes 34 . In an incompatible pepper-TMV interaction PR-10 transcripts were also robustly induced. The recombinant PR-10 protein exhibited a ribonuclease activity against TMV RNA. In addition, TMV inoculation led to the phosphorylation of PR-10, which increased its ribonucleolytic activity 64 . We assume that PR-10 proteins may participate in the defense of pepper against tobamoviruses due to their ribonuclease activity.
(4) Terpenoid biosynthesis Upon ObPV inoculation, a large number of genes participating in the biosynthesis of various terpenoids (also called isoprenoids) were among the most robustly activated in pepper leaves. The carbon backbone of terpenoids is synthesized by the mevalonate pathway, which produces isopentenyl 5-diphosphate (IPP) from acetyl-CoA, through mevalonic acid. The three steps of the upper mevalonate pathway are catalyzed by acetyl-CoA acetyltransferase (also known as acetoacetyl-CoA thiolase), hydroxymethylglutaryl-CoA synthase (HMGCS) and 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) enzymes, finally resulting mevalonic acid [65][66][67] . In our RNA-Seq studies the expression of an acetyl-CoA acetyltransferase gene and two HMGCS genes were strongly induced by ObPV at 24 and 48 hpi (Supplementary Table S5). Furthermore, we detected the expression of 5 HMGCR genes in pepper leaves, from which four HMGCRs were markedly upregulated by ObPV. The expression of a HMGCR gene (CaHMG-CoA/XM_016702518) was 60-and 326-fold higher in ObPV-inoculated leaves at 24 and 48 hpi, respectively, than in control leaves. The expression of these was not influenced significantly by PMMoV inoculation.
In infected plants, FPP is utilized to the biosynthesis of sesquiterpene (C 15 ) phytoalexins. First, 5-epiaristolochene synthase (EAS) enzymes catalyze the cyclization of FPP to 5-epi-aristolochene [69][70][71] . In our work, five EAS genes were massively up-regulated following ObPV inoculation and some of them were among the most robustly ObPV-inducible genes (up to 2074-fold induction at 48 hpi, CaEAS2/NM_001324691). The transcript abundance of other sesquiterpene cyclases like alpha-farnesene synthase, germacrene synthase and vetispiradiene synthase genes also substantially increased by ObPV (Supplementary Table S5). In the following reaction, 5-epi-aristolochene is converted to capsidiol (C 15 phytoalexin) by 5-epi-aristolochene-1,3-dihydroxylase (EAH) enzymes 70,72 . Interestingly, cytochrome P450 71D (CYP71D) enzymes were shown to possess EAH activity in tobacco 72 . We detected the transcripts of five CYP71D7 genes in the pepper transcriptome and all of them were substantially activated by ObPV, up to a 2296-fold induction in the case of a CYP71D7 gene (CaCYP71D7/ XM_016694256) at 48 hpi. Furthermore, two premnaspirodiene oxygenase genes were also strongly induced by ObPV. Premnaspirodiene oxygenases are also cytochrome P450 enzymes that catalyze the hydroxylations of diverse sesquiterpenes 73 .
We hypothesize that upon ObPV inoculation a blend of several sesquiterpene phytoalexins accumulates in pepper leaves. The strong increase of capsidiol content was already observed a long time ago in pepper and tobacco leaves following elicitor treatments and microbial infections 70 76 . Interestingly, silencing of an ethyleneresponsive transcription factor gene (NaERF2) strongly reduced the expression of EAS and EAH genes as well as the capsidiol content in Nicotiana attenuata 78 . The closest pepper homolog of NaERF2 was identified in our studies as a strongly ObPV-inducible ERF gene (CaERF1A/XM_016714653), Supplementary Table S5). These results show that capsidiol biosynthesis may be transcriptionally regulated by an ERF transcription factor in ObPV-inoculated pepper leaves. Intriguingly, we identified also two beta-amyrin 28-oxidase genes, which were massively induced by ObPV at 24 and 48 hpi. These genes are identical to cytochrome P450 716A genes, which participate in the biosynthesis of various triterpenoids (C 30 compounds) like oleanolic acid, ursolic acid and betulinic acid 79 . In addition, we found a strongly ObPV-inducible beta-amyrin 11-oxidase gene (cytochrome P450 88D-like), which is involved in the biosynthesis of glycyrrhizin 80 . These triterpenoids possess a broad antiviral spectrum 81 .

(5) Ethylene metabolism
A sharp increase of ethylene production has been often observed in virus infected plants 82 . A rapid and massive accumulation of ethylene was also detected in ObPV-inoculated pepper leaves already at 24 hpi, while the ethylene production was not induced by PMMoV 2 . Apart from ethylene, several other defense hormones were also shown to accumulate in ObPV-inoculated leaves including salicylic acid and jasmonic acid 40 . Ethylene is perceived by specific receptors in the endoplasmic reticulum (ER) and triggers downstream responses 22,83 . Ethylene signaling is initiated by the activation of ethylene biosynthesis. The three-step biosynthetic pathway starts from methionine, which is converted to S-adenosyl-L-methionine (SAM) by S-adenosyl-L-methionine synthase (SAMS) enzymes 84,85 . The following steps are catalyzed by 1-aminocyclopropane-1-carboxylate synthase (ACS) and 1-aminocyclopropane-1-carboxylate oxidase (ACO) enzymes 86 . In our present studies, genes encoding a homocysteine S-methyltransferase (CaHMT-2/XM_016710426) and a SAM synthase (CaSAM/ XM_016687895) were markedly induced by ObPV (Supplementary Table S5). In addition, two ACS genes (CaACS1/XM_016710904 and CaACS2/XM_016682648) and eleven ACO genes were also up-regulated by ObPV (some of them robustly), but not by PMMoV. The early up-regulation of several ACS and ACO genes at 24 hpi coincides with the marked accumulation of ethylene 2 , so we presume that transcriptional activation of these genes contributes to an increased ethylene production. Interestingly, the A. thaliana homologs of the two ObPV-inducible ACS genes, which are designated as AtACS2 and AtACS6, were already shown to be specifically pathogen-inducible 84,87 . TMV inoculation also up-regulated the expression of ACO and ACS genes in resistant tobacco leaves 82,86 .
Several genes of the complex ethylene signaling pathway 88 were also significantly induced by ObPV. Thus, the transcript abundance of genes encoding three ethylene receptors, the copper transporter RAN1, the protein kinase CTR1 as well as genes encoding two EIN3 proteins and two EIN3-binding F-box proteins were significantly elevated by ObPV. The master regulator EIN3 protein is the last component of the ethylene signal transduction pathway. EIN3 is located in the nucleus where it can regulate the expression of a large number of genes encoding ERFs 83,89 .

(6) Sulfur metabolism
The homeostasis of cysteine, which is the central metabolite of plant sulfur metabolism, plays an essential role in plant immunity 90 . Cysteine serves as precursor of a wide variety of antimicrobial or antioxidative thiol compounds such as defensins, glucosinolates, glutathione (GSH), GSTs, phytoalexins, S-containing volatiles and thionins 91 . The cysteine-containing tripeptide GSH is a major antioxidant in plants. GSH participates in plant defense mechanism not only as antioxidant, but also as a signaling compound 92 . Nevertheless, very few information is available about the impact of virus infections on cysteine or GSH biosynthetic pathways 93,94 . In our studies, KEGG analysis revealed that ObPV inoculation significantly activated the cysteine and GSH biosynthetic pathways (Supplementary Table S4). In the cysteine biosynthesis route, genes encoding an 5'-adenylylsulfate reductase, two serine-O-acetyltransferases and two O-acetylserine-(thiol)-lyases (OAS-TL) were activated by ObPV. Particularly a cysteine synthase gene (CaCYS/XM_016682985) was induced, the expression of which increased 11-and 16-fold by ObPV at 24 and 48 hpi, respectively (Supplementary Table S5). Furthermore, the expression of two genes participating in GSH biosynthesis was also activated. PMMoV did not influence significantly the cysteine and GSH pathways.
Notably, GSTs were particularly massively induced by ObPV among GSH-related genes. GSTs constitute a large family of soluble proteins that catalyze the covalent binding of GSH to substrates containing a reactive electrophilic centre to form less toxic and more water-soluble conjugates. Although GSTs have a variety of functions, their most likely roles in pathogen-infected plants are the suppression of necrosis by antioxidative reactions, participation in hormone transport and interaction with salicylic acid metabolism 95 . In pepper, eighty-five GST genes were identified 96 . In our studies, 22 GSTs were highly induced while eleven GSTs were suppressed by ObPV, which is represented by a heat-map (Fig. 5). Particularly the genes CaGST/XM_016697096 and CaGST/ XM_016685920 were massively up-regulated by ObPV at 48 hpi, their expressions were 597-and 214-fold higher than in mock samples (Supplementary Table S5). In contrast to GSTs, the genes encoding glutathione peroxidases, glutaredoxins, sulfotransferases, thioredoxins and peroxiredoxins were (with some exceptions) markedly suppressed by ObPV.  Both up-and down-regulated genes were selected to qPCR analyses in order to represent a wide scale of expressional changes during the validation. For qPCR analyses we used the same total RNA preparations that were utilized also for RNA-Seq library preparation. The genes encoding a hydroxymethylglutaryl-CoA synthase (CaHMG-CoA/XM_016702295), a 1-aminocyclopropane-1-carboxylate synthase (CaACS2/XM_016682648), a 3-hydroxy-3-methylglutaryl-coenzyme A reductase (CaHMGR2/XM_016702348), a pathogenesis-related protein 10 (CaPR-10/XM_016710983), a heat shock transcription factor (CaHSFB3/XM_016688343), a phenylalanine ammonia-lyase (CaPAL/NM_001324603), and a WRKY transcription factor (CaWRKY18/XM_016726661) were significantly activated by ObPV, whereas genes encoding a geranylgeranyl diphosphate reductase (CaGGR /XM_016707759), a chlorophyll a-b binding protein 1B (CaLHCB/XM_016700776) and a protochlorophyllide reductase (CaPOR/XM_016689752) were markedly suppressed by ObPV (Fig. 6). In contrast to ObPV, PMMoV exerted only a negligible effect on the expression of these genes (Fig. 6). Next, we compared to gene expression changes obtained for the ten selected genes by RNA-Seq and qPCR. The fold-changes of gene expression shown on log2 scales presented similar trends in qPCR and RNA-Seq (Fig. 7). A high linear regression coefficient (R 2 = 0.884) was obtained between the fold-change values obtained by RNA-Seq and qPCR, which proved that the RNA-Seq data were reliable.

Protein-protein interaction network. Although direct information about the interactions of pep-
per proteins is not available yet, interactions can be predicted based on the Arabidopsis model. By using the BioGRID 98 protein-protein interaction database, we determined the possible interactions of 1307 pepper proteins with 2655 connections. Most likely, the expression levels of these genes and the number of their interacting partners are proportional to their importance. Thus, from the complex graph of all predicted pepper proteinprotein interactions we cut out 5 subgraphs (Groups A to E) based on centrality, measuring the most linked proteins and their primary neighbors ( Fig. 8; Supplementary Table S6). These subgraphs still contain one third of the predicted linkages (507 nodes and 1364 edges). The source code (graphml) of these subgraphs are deposited to GitHub Gist and accessible online via yEd Live (https:// bit. ly/ 3lxFe td). In Group A, we identified five downregulated TCP transcription factors (Fig. 8, Supplementary Table S6). TCP proteins exert a regulatory role in shaping plant morphology but several recent studies have demonstrated that TCPs also function as a cellular hub in plant defense signaling 99 . Downregulation of CIN-TCPs results in delayed maturation of vegetative organs, which increases the survival of the biotrophic phytoplasma 99 . Interestingly, miR319 was shown to regulate TCPs transcription factors 100,101 . Furthermore, Rice ragged stunt virus (RRSV) up-regulated the expression of miR319 gene in rice, which led to the suppression of OsTCP21. Overexpression of OsTCP21 increased the resistance of rice to RRSV 102 . We suppose that the suppression of pepper TCP genes was also induced by increased miRNAs levels following ObPV inoculation.
Examining the Group B and C subgraphs, we found that 50S ribosomal proteins, proteasome subunits, parts of the ubiquitin-proteasome system and many LRR receptor-like serine/threonine-protein kinase (LRR-RLKs) proteins were enriched ( Fig. 8; Supplementary Table S6). Intriguingly, nine of the ten downregulated LRR-RLK proteins interact with upregulated ubiquitin-conjugating enzymes which may be associated with their rapid degradation. In many cases, ubiquitin-ligases (E3) are phosphorylated by LRR-RLKs. Presumably phosphorylation of ubiquitin-ligases regulates their activity or their interaction with target proteins 103 but this control may be less effective at low LRR-RLKs level. E3 ligases play important roles in multiple environmental stresses as regulators of salicylic acid, jasmonic acid, and ethylene signaling pathways 104 . Mutations in the Arabidopsis 26S proteasome regulatory subunit coding genes AtRPN1a and AtRPT2A suppressed edr2-associated powdery mildew resistance phenotype 105 . In our study, we found that pepper homologs of these two genes (CaRPN1a/XM_016689488 and CaRPT2A/ XM_016686630) and the expression of twenty other proteasome subunits were up-regulated following ObPV-inoculation, which suggests their important role in plant innate immunity.
In the D subgraph the three proteins having the most connections are mitogen-activated protein kinases (MAPKs) (Fig. 8; Supplementary Table S6). Virus-induced gene silencing (VIGS) of CaMAPK7 (XM_016705198) significantly enhanced the susceptibility of pepper plants to infection by Ralstonia solanacearum 106 . In our work, the expression of this MAPK and two other MAPKs, CaMPK3 (XM_016720287) and CaMPK4 (XM_016682631) were markedly up-regulated by ObPV (Supplementary Table S2). The Arabidopsis homologs of these pepper MAPKs, AtMPK3 and AtMPK4 are known to play vital roles in immune responses [107][108][109] . Previous studies have already reported the crosstalk between MAPKs and TCPs 110 . We presume that MAPKs and ubiquitination 111 as well as LRR-RLKs and ubiquitination 112 interact during the regulation of defense reactions in ObPV-inoculated pepper plants.

Conclusions. We investigated early (4-48 hpi) reactions of pepper plants to inoculations with two different
tobamoviruses. Transcriptome profiles of virus-infected leaves were examined by Illumina RNA-Seq method. In the incompatible pepper-ObPV interaction, the strong suppression of photosynthesis-related genes and the massive induction of energy-producing pathways (cellular respiration and beta-oxidation of fatty acids) as well as diverse groups of defense-related genes showed that ObPV inoculation caused a metabolic switch from normal metabolism to a defense mode in the infected pepper leaves. Genes encoding immune receptors, specific transcription factors, PR-proteins, fatty acid desaturases, GSTs, enzymes of ethylene metabolism and terpenoid biosynthesis were particularly strongly up-regulated by ObPV. Intriguingly, PMMoV inoculation (compatible interaction) induced only negligible defense reactions in spite of its rapid replication in the infected leaves. The transcriptome-wide RNA-Seq analysis uncovered a large number of novel, interesting candidate genes for fur- www.nature.com/scientificreports/ ther research. In future work, elucidating the function of newly discovered defense genes will be necessary by both molecular and biochemical methods.   In a separate bioinformatical analysis with the above methods we established the lists of pepper genes that were inducible by 5 mM sodium salicylate, 100 μM methyl-jasmonate, 5 mM ethephon (ethylene precursor) and 100 μM abscisic acid (ABA) by using the publicly available raw gene expression data of Lee et al. (2020) 42 from the Sequence Read Archive (SRP265260). We aligned these hormone-inducibility data with the ObPV-and PMMoV-inducibility of pepper genes obtained by our RNA-Seq analyses.   43 , including functional annotations, metabolic pathway classifications and nucleotide/protein sequences. All the annotation data from different sources were combined in a single document (Supplementary Table S2).

Materials and methods
Real-time, quantitative RT-PCR measurements. Quantitative, real-time RT-PCR (qPCR) assays were conducted with the same total RNA samples that were used for RNA-Seq studies. Reverse transcription (RT) of 1.5 µg total RNA was carried out in a total volume of 20 µl with a RevertAid H Minus First Strand cDNA Synthesis kit (Thermo Fisher, Waltham, MA, USA) using an oligo(dT)18 primer. In the case of virus multiplication measurements, RT of viral RNAs were carried out by using the reverse primers of the viral movement protein (MP) or coat protein (CP) gene-specific primer pairs (Supplementary  Table S7). Our preliminary evaluation indicated that expression of CaUBC3 displays little variation across all samples tested, which is in line with earlier results 121 . Since all PCR amplification efficiencies were near 100%, the expression values of target genes were first normalized to CaUBC3 and then the obtained relative expression levels were compared to mock-inoculated values, separately for each sampling time point, according to the method of Livak and Schmittgen (2001) 97 . All RT-qPCR assays were carried out in three independent parallel experiments.
Pepper protein interactome analysis. Our knowledge about interactions between pepper proteins is still rudimentary, therefore the well characterized Arabidopsis thaliana interactome datasets by BioGRID 98 version 3.5.138 were called for help in our analysis. BLASTP similarity searches were performed to identify the best Arabidopsis/pepper orthologs. In the first step, protein sequences of pepper DEGs were searched against the Arabidopsis proteins, then the search was made in the opposite direction, as well. Those pepper proteins, which proved to be similar to Arabidopsis proteins in both directions were considered as genuine orthologs and these pepper proteins were used to replace Arabidopsis proteins in the Arabidopsis interactome. The protein-protein interaction network topology was designed and visualized with the yEd Graph Editor software version 3.21.1 (https:// www. yworks. com/ produ cts/ yed). www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.