Abstract

   Ethylene is crucial in climacteric fruit ripening. The ethylene signal
   pathway regulates several physiological alterations such as softening,
   carotenoid accumulation and sugar level reduction, and production of
   volatile compounds. All these physiological processes are controlled by
   numerous genes and their expression simultaneously changes at the onset
   of ripening. Ethylene insensitive 2 (EIN2) is a key component for
   ethylene signal transduction, and its mutation causes ethylene
   insensitivity. In tomato, silencing SlEIN2 resulted in a non-ripening
   phenotype and low ethylene production. RNA sequencing of
   SlEIN2-silenced and wild type tomato, and differential gene expression
   analyses, indicated that silencing SlEIN2 caused changes in more than
   4,000 genes, including those related to photosynthesis, defense, and
   secondary metabolism. The relative expression level of 28 genes
   covering ripening-associated transcription factors, ethylene
   biosynthesis, ethylene signal pathway, chlorophyll binding proteins,
   lycopene and aroma biosynthesis, and defense pathway, showed that
   SlEIN2 influences ripening inhibitor (RIN) in a feedback loop, thus
   controlling the expression of several other genes. SlEIN2 regulates
   many aspects of fruit ripening, and is a key factor in the ethylene
   signal transduction pathway. Silencing SlEIN2 ultimately results in
   lycopene biosynthesis inhibition, which is the reason why tomato does
   not turn red, and this gene also affects the expression of several
   defense-associated genes. Although SlEIN2-silenced and green wild type
   fruits are similar in appearance, their metabolism is significantly
   different at the molecular level.

Introduction

   The gaseous phytohormone ethylene (C[2]H[4]) is essential for
   developmental and physiological regulation in a variety of higher
   plants [[36]1] and its effect is more obvious during the ripening
   process of climacteric fruits [[37]2]. Physiological and molecular
   genetic analysis have uncovered a pathway for ethylene signal
   transduction, which is evolutionarily conserved from the receptors of
   the endoplasmic reticulum membrane to transcription factors in the
   nucleus [[38]3,[39]4,[40]5,[41]6]. The ethylene receptor (ETR),
   localized on the endoplasmic reticulum, negatively regulates the
   ethylene signal transduction occurring upstream [[42]4,[43]7]. Several
   years later, studies in root hair cells by Dong et al (2010) indicated
   that receptor ETR1 also localizes at the Golgi network and only
   partially at the ER [[44]8]. So far seven ETR genes of tomato (Solanum
   lycopersicum) have been cloned, SlETR1, SlETR2, NR, SlETR4, SlETR5,
   SLETR6, and SlETR7 [[45]9,[46]10,[47]11,[48]12], present different
   expression patterns during tomato development [[49]13]. Arabidopsis
   thaliana constitutive triple response 1 (CTR1), which is a Raf-like
   protein kinase [[50]5], has multiple homologs in tomato. Some CTR1-like
   proteins can interact with ETR proteins and negatively regulate signal
   transduction [[51]14]. In addition, SlCTR1 is sensitive to exo-ethylene
   treatment and its expression increases during tomato development
   [[52]15].

   In A. thaliana, ethylene insensitive 2 (EIN2), also located in the ER,
   is a critical positive regulator in the ethylene signal pathway and its
   C-terminal domain (CEND) can be cleaved and transferred to the nucleus,
   stabilizing another positive regulation protein, ethylene insensitive 3
   (EIN3) [[53]3,[54]16]. ETR1 interacts with EIN2, in which Ser645 is
   phosphorylated by CTR1 to block the cleavage of CEND and its transfer
   to the nucleus [[55]6,[56]17,[57]18]. EIN2 is the only protein whose
   loss-of-function mutation results in complete ethylene insensitivity in
   the ethylene signal transduction pathway between the nucleus and
   cytoplasm [[58]16]. Two F-box proteins, ETP1 and ETP2, destroy EIN2
   through the ubiquitin pathway [[59]19]. In turn, EIN2’s cleaved CEND
   can inhibit the expression of the F-box genes EB1 and EBF2, disrupting
   the accumulation of EIN3 by recognizing their 3′-untranslated regions
   and transferring them to the P-body. This is accompanied by an
   exoribonuclease, EIN5, acting in the cytoplasm at translation level
   [[60]20,[61]21]. Using virus induced gene silence (VIGS), SlEIN2 can be
   silenced in tomato plants, significantly suppressing fruit ripening
   [[62]22]. Silencing only one of the functionally redundant SlEILs
   (coding for EIN3-like proteins) did not produce significantly
   non-ripening phenotypes [[63]23]. After ethylene insensitive-like (EIL)
   proteins bind to the promoter regions of ethylene response elements
   (EREs), the ethylene response factors (ERFs) are able to bind the
   GCC-box, a conserved sequence of ethylene response genes, and activate
   the ethylene-induced pathogenesis-related genes [[64]24,[65]25,[66]26].
   In banana, which is also a climacteric fruit, the GCC-box motif is
   homologous to the cis-acting elements of MaEXP1, suggesting that some
   ERFs might have a role in fruit softening [[67]27].

   Fruit ripening is a complicated process, including the accumulation of
   volatile components, flavonoids formation, pectin degradation, and
   carotenoid biosynthesis [[68]24,[69]28,[70]29,[71]30]. These diverse
   processes are regulated by numerous transcription factors and signal
   transduction pathways [[72]24], among which the ethylene signal pathway
   is typically found. Unlike the functionally redundant ETRs, CTRs, EILs,
   and ERFs, the uniqueness of EIN2 might enable it to participate in many
   ethylene-related metabolic pathways, which probably contributes to the
   complete ethylene insensitivity resulting from a mutation in the
   functional domain of SlEIN2. Previous studies have shown that
   non-ripening tomatoes were obtained by silencing SlEIN2 through VIGS,
   and recently a new paper from Gao et al (2016) showed that in
   yellow-fruited tomato 1(yft1) mutant, lower expressed of EIN2 would
   lead to impaired ethylene biosynthesis and perception, as well as
   abnormal carotenoid production [[73]22,[74]31]. However, no study has
   assessed how SlEIN2 regulates fruit ripening at the gene level.
   Therefore, the present study aims to provide a preliminary analysis of
   SlEIN2 regulation during fruit ripening using RNA sequencing (RNA-seq).
   The results showed that silencing SlEIN2 leads to significant changes
   in the expression of a large number of genes involved in chlorophyll
   binding proteins, ethylene biosynthesis, lycopene production, defense,
   etc., as SlEIN2 and the ethylene signal pathway critically upregulate
   several transcription factors in a feedback loop. The study of SlEIN2
   also increases the knowledge on the molecular mechanisms regulating
   fruit ripening by signal transduction pathways.

Materials and Methods

Plant material and growth conditions

   Ailsa Craig tomato seeds preserved in our laboratory were sown in
   commercially available tomato-cultivation soil and grown in a chamber
   at 25 ± 2°C, with a relative humidity of 75% and under a light:
   darkness cycle of 16:8 h, regulated by fluorescent lamps. Tomato plants
   were watered with a nutrient solution once a week.

Tobacco rattle virus (TRV)-SlEIN2 vector construction

   Vectors used VIGS are based on the TRV pTRV1 and pTRV2 (Liu Y et al.,
   2002). We adopted In-fusion^® (Clontech, Nanjing, China), a new cloning
   technique that does not require T4 DNA ligase and the insertion of a
   silencing fragment, and has high ligation efficiency. To generate
   pTRV2-SlEIN2, pTRV2 plasmids were first linearized through digestion
   with EcoRI and BamHI. As the reverse insertion of the silencing
   fragment can improve silencing efficiency [[75]22], a 348-bp SlEIN2
   fragment was amplified using the forward primer
   5′-TAAGGTTACCGAATTCCCTGAATTGGAGCTGTAC-3′, which included a BamHI
   adaptor (underlined) and the reverse primer 5′-
   GCTCGGTACCGGATCCTGGAAATGTCCCTGTAGG-3′,
   which included an EcoRI adaptor (underlined). The resulting product was
   cloned into pTRV2 using the In-fusion^® kit and following the
   manufacturer’s instructions.

Agrobacterium tumefaciens infiltration

   The vectors pTRV1- and pTRV2-SlEIN2, and the control vectors pTRV1 and
   pTRV2, were transformed into two sets of the A. tumefaciens strain
   GV3101 and cultured at 200 rpm at 28°C in the Luria—Bertani (LB) medium
   containing 10 mM 2-N-morpholino ethanesulfonic acid (MES) and 20 mM
   acetosyringone (AS), with 50 μg/mL of kanamycin, gentamycin, and
   rifampicin antibiotics. After 16 h, both sets of A. tumefaciens cells
   were centrifuged and resuspended in the infiltration buffer (10 mM MES,
   200 μM AS, 10 mM MgCl[2]; pH 5.6), until their OD[600] ranged between 4
   and 7. Bacterial suspensions were set aside for 3–4 h and then combined
   at a 1:1 ratio, before their infiltration in tomato plants’ using a
   1-mL syringe. Ten days after pollination, the carpopodium of tomato
   plants was perforated and 50–100 μL bacterial solution were infiltrated
   at the wound site.

RNA isolation and real-time quantitative PCR

   Five days after breaker (BK), the orange ripe (OR) pTRV and
   pTRV-SlEIN2-inoculated, only the green area sampled, fruits were
   collected, each type including 6 fruits. Mature green (GM) fruits from
   another control group were also collected to evaluate differences
   between gene-silenced and authentic unripe fruits. Control and
   gene-silenced fruits were stored at -80°C before use and their total
   RNA was isolated using the RNeasy^® Mini Kit (Qiagen, Hilden, Germany).
   Unwanted genomic DNA was digested using DNase I (Tiangen Biotech Co.,
   Beijing, China). The concentration and purity of RNA were measured in a
   NAS-99 spectrophotometer (ATCGene Inc., New Jersey, United States). The
   RNA integrity estimated through gel electrophoresis showed a 28S/18S
   brightness ratio of approximately 2:1. Complementary DNA was then
   synthesized from 2 μg RNA using the TransScript^® II One-Step gDNA
   Removal and cDNA Synthesis SuperMix Kit (Transgen Biotech Co., LTD.,
   Beijing, China) with oligo(dTs). Virus vectors were detected by PCR
   using the EasyTaq PCR SuperMix (Transgen Biotech Co., LTD), coat
   protein, and pTRV-RNA2 specific primers. Amplifications were performed
   in a Bio-Rad (Hercules, CA, United States) thermocycler under the
   following conditions: 94°C for 3 min, followed by 30 cycles at 94°C for
   30 s, 55°C for 30 s, and 72°C for 40 s.

   Quantitative PCR (qPCR) was performed in the Bio-Rad CFX96 thermocycler
   using the TransStart Top Green qPCR SuperMix (Transgen Biotech Co.,
   LTD) for 5 min at 95°C, followed by 40 cycles of 15 s at 95°C and 30 s
   at 60°C. Changes in fluorescence intensity were monitored in each
   cycle. Three biological replicates, each including two mixed fruits,
   were included in the PCR and expression levels were determined
   relatively to that of Actin (ACT1), which was used as the internal
   control, and analyzed using the 2^- ΔCt method [[76]32]. All primers
   used are listed in [77]S1 Table.

RNA sequencing and assembly of RNA transcripts

   Total RNA was isolated from green pTRV-SlEIN2 samples and two groups of
   pTRV fruits (biological replicates were the same as in qPCR). Total RNA
   concentration was measured in the NAS-99 spectrophotometer (ATCGene
   Inc.), and an RNA integrity number ≥ 7.0 was confirmed through gel
   electrophoresis. Messenger RNA was then enriched using oligo(dTs)
   coupled with magnetic beads, before being cut into 300-bp fragments
   ([78]Novogene, Tianjing, China). Complementary DNA libraries were
   obtained using random hexamers to synthesize the first strand, and
   adopting DNA polymerase I and dNTPs to generate the second strand.
   Synthesized cDNA was then purified, its ends were repaired, and
   adaptors were ligated. After libraries’ preparation, 150-bp pair-end
   sequencing was performed on an Illumina^® Hiseq PE150 (Illumina, Inc.,
   Beijing, China), generating 6 G raw data for each pair-end sequencing.

   Raw reads were quality checked and trimmed using cutadapt (version
   1.10, [79]https://pypi.python.org/pypi/cutadapt/) and FASTX-Toolkit
   (version 0.0.13.2
   [80]http://hannonlab.cshl.edu/fastx_toolkit/download.html). After
   removing barcode and adaptor sequences, the resulting clean reads were
   checked for quality using the Q < 20 threshold. All clean reads were
   deposited in the National Center for Biotechnology Information (NCBI)
   Sequence Read Archive ([81]http://www.ncbi.nlm.nih.gov/sra/) under the
   accession number [82]SRP076745. Clean reads within each library were
   aligned with the tomato reference genome (version SL2.50,
   [83]ftp://ftp.sgn.cornell.edu/tomato_genome) using TopHat (version
   2.0.8, [84]http://ccb.jhu.edu/software/tophat/index.shtml). Reads with
   less than two mismatches were used to construct transcripts using
   Cufflinks (version 2.0.2,
   [85]http://cole-trapnell-lab.github.io/cufflinks/). Genes in
   pTRV-SlEIN2 and pTRV-GM or pTRV-OR fruits were considered as
   differentially expressed genes (DEGs) if |fold-change| ≥ 2 and Q-value
   < 0.05.

Gene Ontology (GO) enrichment analysis

   GO enrichment analysis was performed using GO-TermFinder (version 0.86,
   [86]http://search.cpan.org/dist/GO-TermFinder/) based on DEGs, gene
   identity in the Sol genomics network database to GO terms, gene
   association, and GO libraries
   ([87]http://geneontology.org/page/downloads). The threshold of the
   corrected P-value was 0.05, and genes were classified into the
   following classes: cellular component, biological process, and
   molecular function.

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis

   Fasta format files containing DEGs cDNA or protein sequences were
   obtained using Perl scripts and KEGG enrichment analysis was then
   performed in KOBAS (version 2.0,
   [88]http://kobas.cbi.pku.edu.cn/download.do), based on native blast
   tools and organism annotation libraries. KEGG pathways with a corrected
   P-value < 0.05 were analyzed.

Results and Discussion

Silencing SlEIN2 produced non-ripening phenotype and differential expression
of several thousand genes

   Ethylene regulates several plant physiological activities including
   development, senescence, flowering, and fruit ripening through signal
   transduction pathways. As SlEIN2 is an important component of signal
   transduction, mutations occurring in this gene will effectively block
   the signal transduction pathway, resulting in plant insensitivity to
   ethylene. To understand the role of SlEIN2 in the development and
   ripening of tomato, we obtained SlEIN2-silenced fruits and analyzed
   their DEGs using RNA-seq.

   According to the sketch presented in [89]Fig 1A, a mixture of A.
   tumefaciens GV3101 cultures containing pTRV1- and pTRV2-SlEIN2 or pTRV1
   and pTRV2 constructs in a 1:1 ratio, were needle-injected into the
   carpopodium of wild-type Ailsa Craig tomato fruits 10 days after
   pollination. The phenotype on the fruit was observed at 5 days after
   breaker stage. The SlEIN2-silenced fruits were still green whereas the
   non-SlEIN2-silenced and control fruits were red ([90]Fig 1B). The
   phenotype presented by SlEIN2-silenced fruits was consistent with our
   previous study [[91]22]. They remained green, i.e., they were not ripe,
   and their appearance was very similar to that of pTRV-injected GM
   fruits.

Fig 1. Acquisition of TRV-SlEIN2 fruits.

   [92]Fig 1
   [93]Open in a new tab

   (A). Main phases of VIGS. (B). Non-ripening phenotype of attached and
   dissected TRV-SlEIN2 tomato fruit compared with the two control groups.
   (C). qPCR analysis of SlEIN2 expression in VIGS and control fruits. The
   error bar indicates the standard deviation, based on three biological
   replicates. Asterisks indicate significant differences, according to
   Student’s t-test (***, P < 0.001). (D). pTRV transmission in fruits. M
   is the 2 kb marker, and C is the control, using the pTRV2 plasmid as
   template. 1, 2, and 3 represent the three biological replicates.

   To confirm SlEIN2 gene silencing at the molecular level, primers
   specific to the SlEIN2 genes outside the region targeted for silencing
   were designed and used in qPCR. Results evidenced a 76% reduction in
   SlEIN2 transcripts in silenced fruits in relation to pTRV-injected OR
   fruits. The expression of SlEIN2 in pTRV-injected GM differed from that
   in pTRV-SlEIN2 fruits, the latter only accounting for 19% of the former
   ([94]Fig 1C). Consideringan average of ~80% reduction of target
   endogenous mRNA is normally achived using TRV-VIGS, the samples were
   feasible for futher studies and analyses [[95]33,[96]34]. As the level
   of ACT1 transcript was similar in tissues infected with pTRV-SlEIN2 and
   control vectors, SlEIN2 seems to play an important role in the ethylene
   signal transduction pathway controlling fruit ripening.

   It was reported that SlEIN2 expression reaches its peak at the GM
   stage, and is reduced after BK, during ripening [[97]35]. The massive
   accumulation of phosphorylated SlEIN2 in the GM stage is probably
   related to the upcoming respiration peak and vast changes in fruit
   substance and color. When fruit development reached the breaker stage,
   there was few EIN2 in VIGS fruits, and thus interrupted fruits ripening
   ([98]Fig 2). The correct size of bands shown in [99]Fig 1D evidenced
   pTRV was well transmitted to the fruits.

Fig 2. Phenotype of SlEIN2-silenced tomato three months after picking, BK+5
phase was initiated.

   [100]Fig 2
   [101]Open in a new tab

   To understand the molecular mechanism of SIEIN2 regulating tomato fruit
   ripening, SlEIN2-silenced and control fruit samples (pTRV-GM and
   pTRV-OR) were analyzed by RNA-seq. All clean reads generated in the
   sequencing experiment were mapped and aligned with the tomato reference
   genome ([102]Table 1) Within each file, 79.56 ± 0.62% of the reads were
   uniquely aligned, suggesting that sequencing results were relatively
   stable. Discarded multiple-mapped reads (0.60 ± 0.07% of total mapped
   reads) and the almost uniform 42% GC-contents of sequences are not
   shown. Selected DEGs (SDEGs) regarding pTRV-GM/pTRV-SlEIN2 and
   pTRV-OR/pTRV-SlEIN2 fruits are listed in [103]S2 Table.

Table 1. Summary of clean read counts and percentage of unique mapped reads.

   Sample Clean reads left/right Left unique mapped Right Unique mapped
   Unique alignment
   TRV-LeEIN2 1 20,795,992
   (100%) 19,032,879
   (91.52%) 17,273,527
   (83.06%) 16,596,614
   (79.81%)
   TRV-LeEIN2 2 21,042,397
   (100%) 19,329,124
   (91.86%) 17,422,732
   (82.80%) 16,772,627
   (79.71%)
   TRV-LeEIN2 3 22,409,355
   (100%) 20,533,360
   (91.63%) 18,780,469
   (83.81%) 18075341
   (80.66%)
   TRV-GM 1 22,011,906
   (100%) 20,122,103
   (91.41%) 18301383
   (83.14%) 17,594,909
   (79.93%)
   TRV-GM 2 22,675,897
   (100%) 20,747,156
   (91.49%) 18,765,171
   (82.75%) 18,053,232
   (79.61%)
   TRV-GM 3 22,025,764
   (100%) 20,100,048
   (91.26%) 18,052,961
   (81.96%) 17,378,831
   (78.90%)
   TRV-OR 1 23,938,433
   (100%) 21,985,222
   (91.84%) 19,547,355
   (81.66%) 18,793,997
   (78.51%)
   TRV-OR 2 27,028,002
   (100%) 24,746,795
   (91.56%) 22,246,381
   (82.31%) 21,423,896
   (79.27%)
   TRV-OR 3 21,571,779
   (100%) 19,700,720
   (91.33%) 17,838,181
   (82.69%) 17,177,796
   (79.63%)
   [104]Open in a new tab

   1, 2, 3, biological replicates.

   A considerable number of genes changed their expression when there was
   a deficiency in SlEIN2 ([105]Fig 3A and 3C). In the SlEIN2-silenced
   fruits, 61.28% of the SDEGs were upregulated and 38.72% were
   downregulated compared to the GM control group ([106]Fig 3B).
   Similarly, in OR fruits, 65.42% of the SDEGs were upregulated and
   34.58% were downregulated in pTRV-SlEIN2 fruits ([107]Fig 3D). Whereas
   901 genes were upregulated in both OU and GU, only 130 genes were
   downregulated in OD and GD ([108]Fig 3E). These results indicate that
   silencing SlEIN2 enhances the expression of more genes.

Fig 3. Analyses of differential expressed genes.

   [109]Fig 3
   [110]Open in a new tab

   (A, C). Volcano diagrams of DEGs. Spots above the threshold line
   (Q-value = 0.05), indicate that differences are significant. Genes
   whose expression was less than a half than that displayed in the
   control group for Q-value < 0.05 are displayed in the green area, while
   those whose expression was more than two-fold that of the control group
   are displayed in the red area. Genes in the grey area were neither
   over- or under-expressed. (B, D). Percentage of up/down regulated
   SDEGs. SDEGs were screened in the green and red areas. (E) Venn diagram
   showing the numbers of non-overlapped and overlapped SDEGs in the four
   conditions tested. OU and OD separately means upregulated and
   downregulated TRV-OR/TRV-SlEIN2 SDEGs. GU and GD represents upregulated
   and downregulated TRV-GR/TRV-SlEIN2 SDEGs respectively.

Classifying SDEGs through GO and Pathway enrichments

   Gene ontology was successfully annotated in molecular function using
   GO-TermFinder ([111]Fig 4A and 4C), while cell component and biological
   process not, since SDEGs were not significantly enriched in the two
   category. A diagram showing molecular functions’ connection is provided
   in [112]S1 Fig. Two groups were highly enriched: catalytic and transfer
   activities. The concentration of catalytic proteins such as
   l-aminocyclopropane-l-carboxylic acid synthase (ACS), which
   biosynthesizes the precursor of ethylene, and the cell wall
   decomposition-related enzyme polygalacturonase, drastically changes
   during fruit ripening [[113]30]. Transfer proteins, such as
   UDP-glucosyltransferases, which modify anthocyanins and flavonoids by
   glycosylation increasing their polarity, water solubility [[114]36],
   and stability, are responsible for fruit ripening, and chlorophyll
   (Chl) a/b binding proteins, form Chl-protein complexes [[115]37] that
   take part in photosynthesis.

Fig 4. Molecular function and pathway enrichment analysis of SDEGs.

   [116]Fig 4
   [117]Open in a new tab

   (A). Molecular function of TRV-GM/TRV-SlEIN2 SDEGs, considering a
   corrected P-value < 0.05. The X axis indicates the gene number, and Y
   represents classification. (B). Top 3 pathway enrichment of
   TRV-GM/TRV-SlEIN2 SDEGs, with a Q-value < 0.05. Rich factor means the
   number of gene from SDEGs/all gene numbers, in a pathway. (C, D).
   Molecular function and top 4 pathway enrichment of TRV-OR/TRV-SlEIN2
   SDEGs. All the graphic descriptions and parameters are identical to
   TRV-GM/TRV-SlEIN2.

   In TRV-GM/TRV-SlEIN2, SDEGs in the class of Ser/Thr kinase were
   enriched while in TRV/TRV-SlEIN2 OR fruits they were not, suggesting
   that most of these genes might not be regulated by SlEIN2. Most SDEGs
   of TRV-OR/TRV-SlEIN2 exclusively related to redox activities were
   enriched. NAD/NADP are representative coenzymes in several metabolic
   activities, better known for tricarboxylic acid cycle and fatty acid
   oxidation. The genes enriched in NAD/NADP include several unannotated
   dehydrogenases and respiratory burst oxidases that are NADPH-oxidase
   homologs and related to plant defense [[118]38], some acyl-CoA
   reductases and fatty acid oxidases that alter the composition of aroma
   volatiles, and the electron carrier ferredoxin ([119]S2 Table). In
   addition, most genes in this class were upregulated in VIGS fruits
   compared to TRV-OR fruits, suggesting that they might prevent fruit
   ripening.

   Results of the KEGG pathway enrichment analyses showed that SlEIN2
   plays a role in regulating the accumulation of chlorophyll binding
   proteins ([120]Fig 4B and 4D). Light-harvest chlorophyll binding
   proteins (LHCPs) and chlorophyll are destabilized during fruit
   senescence through the regulation of staygreen (SGR) [[121]39]. The
   significant enrichment of the SDEGs in the class of
   photosynthesis-antenna proteins indicated that LHCPs expression in
   SlEIN2-silenced fruits differed from both GM and OR fruits. Differences
   in carbon fixation (dark reaction) between pTRV GM and pTRV-SlEIN2
   fruits were found and pTRV OR/pTRV-SlEIN2 SDEGs were enriched on
   photosynthesis (light reaction can be visualized in the link in [122]S3
   Table. Thus, although VIGS fruits were green, the genes they expressed
   in carbon fixation were similar to those of OR fruits, suggesting that
   SlEIN2 influenced light reactions but not dark reactions in
   photosynthesis. In addition, silencing SlEIN2 altered the activities of
   genes associated with anti-pathogen, such as the respiratory burst
   oxidase homolog Solyc03g117980, whose fold change was nearly 4.6 in
   TRV-OR/TRV-SlEIN2 ([123]S2 Table). Although several defense-associated
   genes were differentially expressed between VIGS and GM fruits, these
   differences were not significant according to KEGG pathways.

Ripening-associated transcript factors are influenced by SlEIN2 silencing

   In order to verify the consistency of the RNA-seq results and the gene
   expression pattern in tomato fruit, 28 differentially expressed and
   ripening-associated genes were selected from RNA-seq results, and
   verified by qPCR using RNA-seq materials as template. Gene selection
   was based on SDEGs statistics ([124]S2 Table) and genes identification
   and description are listed in [125]Table 2.

Table 2. SGN ID and discription of genes detected in qPCR.

   Gene SGN ID Discription
   EIN2 Solyc09g007870 Ethylene insensitive 2
   RIN Solyc05g012020 MADS-box transcription factor MADS-RIN
   TDR4 Solyc06g069430 [126]FRUITFULL-like MADS-box 1
   NOR Solyc10g006880 NAC domain protein
   ACS2 Solyc01g095080 [127]1-aminocyclopropane-1-carboxylic acid
   synthase-2
   ACS4 Solyc05g050010 [128]1-aminocyclopropane-1-carboxylic acid
   synthase-4
   ACO1 Solyc07g049530 [129]1-aminocyclopropane-1-carboxylate oxidase 1
   ACO3 Solyc09g089580 1-aminocyclopropane-1-carboxylate oxidase 3
   ETR4 Solyc06g053710 [130]ethylene receptor homolog
   EIL3 Solyc01g096810 Ethylene insensitive 3 class transcription factor
   ERF2 Solyc09g075420 Ethylene responsive transcription factor 2b
   AP2a Solyc03g044300 AP2-like ethylene-responsive transcription factor
   CP26 Solyc06g063370 Chlorophyll a-b binding protein 1A
   CP29 Solyc09g014520 Chlorophyll a-b binding protein 6A
   CAB13 Solyc07g063600 Chlorophyll a-b binding protein 13
   GLK2 Solyc10g008160 Transcription factor (Fragment)
   SGR1 Solyc08g080090 Senescence-inducible chloroplast staygreen protein
   2
   PSY1 Solyc03g031860 Phytoene synthase 1
   PDS Solyc03g123760 Phytoene desaturase
   ZDS Solyc01g097810 Zeta-carotene desaturase
   PMEU1 Solyc03g123630 Pectinesterase
   PG2A Solyc10g080210 Polygalacturonase A
   PL1 Solyc03g111690 Pectate lyase
   TSRF1 Solyc09g089930 Ethylene responsive transcription factor 1a
   PR5 Solyc08g080670 Osmotin-like protein
   PR10 Solyc09g090990 Major allergen Mal d 1
   LoxC Solyc01g006540 Lipoxygenase
   ADH2 Solyc06g059740 Alcohol dehydrogenase 2
   [131]Open in a new tab

   Ripening inhibitor (RIN), a member of MADS (named by four transcription
   factors: MCM1, AG, DEFA and SRF) family, is an indispensible and
   well-known regulator of tomato fruit ripening which positively
   regulates gene expression by directly binding to its promoter resulting
   in other transcription factors expression, ethylene production, cell
   wall decomposition, aroma variation, and RIN expression [[132]40]. The
   self-regulated transcription factors seems have no upstream regulator,
   but the ethylene signal pathway is essential for promoting the
   expression of RIN [[133]41], suggesting that the production of ethylene
   is probably auto-regulated through RIN, and that the initial ripening
   is induced by a developmental factor [[134]42]. When SlEIN2 was
   repressed, the expression of RIN decreased drastically ([135]Fig
   5A–5C), along with its target genes, Non ripening (NOR) and Fruitful 1
   (TDR4) [[136]40]. Rin expression was at least two-fold lower in
   TRV-SlEIN2 than in OR fruits, but not significantly differs from it in
   GM fruits. The NOR is a ripening regulator of the NAC (named by three
   transcription factors: NAM, ATAF and CUC) family, whose mutation causes
   a green phenotype compared to the wild type at red ripe stage [[137]43]
   whereas TDR4 is another transcription factor of the MADS family. It
   interacts with RIN to regulate the accumulation of lycopene and lipid
   metabolism during ripening. After silencing TDR4, ripe fruits remained
   orange whereas wild type were red [[138]44]. Still, pTRV-SlEIN2 was
   green for more than three months, unlike rin- and nor-silenced fruits,
   which are yellowish and yellowish-orange [[139]45]. In summary, SlEIN2
   affects fruit ripening mainly by affecting RIN expression.

Fig 5. Relative expression of 28 genes covering several aspects associated
with ripening.

   [140]Fig 5
   [141]Open in a new tab

   The error bar indicates the standard deviation, based on three
   biological replicates. Asterisks indicate significant differences,
   according to Student’s t-test (*, P<0.05, **, P<0.01, ***, P< 0.001).

SlEIN2 silencing blocked the self-promotion of ethylene biosynthesis and gave
an effect to other components in ethylene signal pathway

   Several genes involved in ethylene biosynthesis are direct RIN targets,
   including ACS2, ACS4, and E8 [[142]41]. The expression of ACS2 and ACS4
   was four times lower in pTRV-SlEIN2 than in OR fruits ([143]Fig 5D and
   5E). In addition, ACO1 and ACO3 gene expressions were also
   significantly reduced when SlEIN2 was silenced ([144]Fig 5F and 5G).
   Studies have confirmed that Rin indirectly regulates the expression of
   ACO1 [[145]19,[146]46], which is a direct target of HB-1; HB-1 is
   directly regulated by RIN [[147]41,[148]47]. The promoter of ACO3 is
   the target of ERF2 [[149]48], an ethylene responsive component at the
   end of the signal pathway. Thus, the abrupt drop in the expressions of
   these genes might have been sufficient to reduce ethylene production in
   gene-silenced fruits.

   ETR is a negative regulation component upstream of SlEIN2. Tomato has
   seven ETRs, some of which are relatively stable while others differ
   from leaf to reproductive tissues [[150]13]. The levels of SlETR4 and
   SlETR5 increase significantly as fruits mature and ripen, but only
   SlETR4 changes in response to ethylene treatment [[151]13,[152]49].
   Ethylene can enhance the mRNA level of SlETR4, while the protein level
   was on opposite. [[153]49]. It appears that mRNA of SlETR4 could
   inhibite its translation. Our study showed that silencing SlEIN2
   reduced SlETR4 expression, as a result of inhibiting ethylene
   biosynthesis ([154]Fig 5H), and this probably would promote the
   accumulation of ETR4 receptors When EIN2 decreased, SlEIL3 was
   downregulated ([155]Fig 5I), and the level of EIL3 in SlEIN2-silenced
   was very closed to it in GM fruits. As ERF2 was also subjected to the
   effect of EIN2 ([156]Fig 5J), it might have led to the low ACO3 level
   in VIGS fruits. APETALA2a (AP2a) is another direct target of RIN and
   positively regulated by this transcription factor [[157]41]. AP2a
   negatively regulates ethylene biosynthesis and signal [[158]50],
   indicating that the accumulation of ethylene can be self-promoted and
   self-limited through ethylene signal pathway. The level of this gene
   decreased more than five-fold ([159]Fig 5K) as RIN reduced.

Silencing SlEIN2 has an effect on chlorophyll binding proteins

   Pathway enrichment unraveled the influence of EIN2 silencing in LHCPs
   (see the link in [160]S3 Table). The important Photosystem II (PSII)
   LHCP named CP29, which is located in the core antenna of PSII, has the
   highest Chl-a/b binding ratio [[161]37]. The expression of CP29
   increased nearly two-fold in pTRV-SlEIN2 tomatoes compared to OR
   fruits, but no significant difference between VIGS and GM fruits
   ([162]Fig 5M). Two minor antenna proteins, CP24 and CP26, affect the
   interactions between PSII subunits [[163]51]. The mRNA level of CP26
   were similar to those of CP29 ([164]Fig 5L). Tomato tolerance to
   continuous light (CL), provided by CAB13, contributes to the increase
   in substantial yield, as CL influences carbohydrate metabolism and
   photosynthesis [[165]52], and SlEIN2 silencing led to a larger increase
   in the expression of CAB13 ([166]Fig 5N) than in normal green fruits.
   GOLDEN-LIKE (GLK) and SGR are not LHCPs but also contribute to the
   photosynthetic capacity, and SGR can directly interact with LHCPII, a
   family of LHCPs belonging to PSII, separating the assembled LHCPII and
   leading to Chl degradation and plant degreening [[167]39]. The SGR1,
   another target of RIN [[168]41], also has physical interaction with
   phytoene synthase 1 (PSY1) and promotes the biosynthesis of carotenoids
   in tomato [[169]53]. The expression of SGR1 was significantly reduced
   in SlEIN2-silenced fruits, indicating its positive regulation by
   ethylene ([170]Fig 5P). Chloroplast development requires GLKs and the
   expression of GLK2 is typically higher than that of GLK1, especially in
   fruit shoulder [[171]54]. The SlEIN2-silenced group presented higher
   levels of GLK2 than GM and OR fruits ([172]Fig 5O). However, RNA-seq
   and relative expression analyses indicated that SlEIN2 and ethylene act
   in Chl degradation and negatively regulate photosynthesis.

Silencing SlEIN2 leads to lycopene reduction and cell wall decomposition

   PSY, phytoene desaturase (PDS), and ζ-carotenedesaturase (ZDS) are
   three of the genes involved in carotenoid metabolism and are closely
   related to the synthesis of lycopene. Although many PSY genes generate
   phytoene in tomato fruit, PSY1 has a direct interaction with SGR1 and
   thus was analyzed here. The results showed that these genes were
   severely inhibited in SlEIN2-silenced fruits, particularly PSY1, whose
   expression was lowered by, at least, 90% ([173]Fig 5Q–5S). Repressing
   PDS is sufficient to cause low lycopene content and prevent tomato from
   reddening [[174]55]. Pectin methylesterase isoenzyme (PMEU1) is a
   ubiquitously expressed pectinesterase contributing to harden the cell
   wall, and the reduction of PMEU1 in fruit enhances softening rate
   [[175]56]. The SlEIN2-silenced fruits presented an increased expression
   of PMEU1, although not as high as that of GM fruits ([176]Fig 5T).
   Polygalacturonase 2A (PG2A) and pectin lyase (PL1) are two enzymes
   related to pectin degradation. Both were expressed in higher amounts
   than Actin in ripening fruits, but SlEIN2-silencing caused their
   extreme reduction ([177]Fig 5U and 5V). Although its relative
   expression was not tested, α-Expansin 1 (EXP1) (Solyc06g051800), which
   is also a booster for cell wall degradation, had a -3.38-fold change in
   TRV-OR/TRV-SlEIN2 according to the RNA-seq data ([178]S2 Table).
   Previous studies have also reported that RIN was associated with the
   promoters of PG2a and EXP1, and might positively regulate their
   expression [[179]41,[180]48]. The results obtained here showed that
   SlEIN2 regulated fruit softening by affecting a series of
   softening-associated genes.

Silencing SlEIN2 altered the mRNA level of several genes involved in defense
and aroma compounds generation

   An in vitro study reported that tomato stress responsive factor 1
   (TSRF1) interacts with the GCC-box and activates the expression of
   pathogenesis-related (PR) genes, to strengthen the resistance of tomato
   to Ralstonia solanacearum, and the study showed that Solyc09g089930 is
   upregulated by ethylene [[181]57]; however, according to the ementary
   file of tomato genome research [[182]58], this upregulation is
   gradually decreased during ripening. Silencing SlEIN2 increased TSRF1
   expression and the expression of the PR genes PR5 and PR10 was also
   enhanced ([183]Fig 5W and 5X). Although not tested, expression of other
   defense-related genes, such as respiratory burst oxidase, the
   defense-related WRKY1 [184]Solyc06g066370, and the RIN4
   [185]Solyc09g059430 were also enhanced, as they were detected in KEGG
   pathways and RNA-seq data. The above results suggested that SlEIN2 is
   involved in transformation of tomato pathogen-defense.

   The genes 13-lipoxygenase (LOXC), hydroperoxide lyase, and alcohol
   dehydrogenase 2 (ADH2) are known to participate in volatiles’
   biosynthesis. Their RNAs increase during ripening [[186]24] and ADH2 is
   a direct RIN target [[187]41]. Silencing SlEIN2 decreased the levels of
   both ADH2 and LOXC ([188]Fig 5a and 5Z), thereby suggesting that this
   gene dowregulates the genes involved in aroma biosynthesis.

Conclusions

   The technique, VIGS, used in the present study is a convenient and
   powerful tool for targeted gene silencing in tomato, producing the
   non-ripening phenotype in a short period. In the present study, we
   assessed SlEIN2 effects on fruit ripening inhibition and this is
   presented in [189]Fig 6. Silencing SlEIN2 leads to the reduction of RIN
   in a feedback regulation process, which is generally found upstream of
   the ethylene signal pathway. In addition, silencing SlEIN2 can decrease
   the degradation of LHCPs and chlorophyll by reducing RIN expression and
   that of its target, SGR1. As a result, few SGR1 proteins will interact
   with PSY1 inhibiting lycopene biosynthesis, which is the reason why
   tomato does not turn red. Downstream regulation of SlEIN2 is conducted
   by ERFs and silencing SlEIN2 altered the expression of several
   defense-associated genes. Although SlEIN2-silenced and GM fruits are
   similar in appearance, there are significant differences in their
   secondary metabolites and in antenna-Chl proteins expression.

Fig 6. Regulation route in SlEIN2-silenced fruits.

   [190]Fig 6
   [191]Open in a new tab

   Red represents upregulated substances and proteins, and blue
   downregulated substances and proteins.

Supporting Information

   S1 Table. Oligonucleotide primers used in the study.

   (DOCX)
   [192]Click here for additional data file.^ (17.1KB, docx)
   S2 Table. Selected differential expressed genes of control vs
   treatment.

   (XLSX)
   [193]Click here for additional data file.^ (721KB, xlsx)
   S3 Table. The results of KEGG pathway enrichment.

   (XLSX)
   [194]Click here for additional data file.^ (17.2KB, xlsx)
   S1 Fig. Gene Association of GO enrichment analysis.

   (PDF)
   [195]Click here for additional data file.^ (8.6MB, pdf)

Acknowledgments