Deciphering transcriptional regulation mechanisms underlining fruit development and ripening in Vitis vinifera

Abstract

BACKGROUND:

Grapes (Vitis vinifera) are an important woody crop cultivated in a broad range of environmental conditions. Grapefruit development is a physiological process whose molecular regulatory networks are still not sufficiently investigated.

OBJECTIVE:

The primary aim of the work was to identify which key genes, molecular mechanisms and networks were involved in fruit ripening and development through a comparison of available transcriptomic data at different stages during grape development and ripening. Secondly, we aimed at identifying among these fruit-related genes, which genes play also a functional role in other developmental and physiological processes in reproductive tissues (bud and flower).

METHODS:

The two objectives were obtained through a meta-analysis on 9 RNA-Sequencing (RNA-Seq) studies. Raw data was downloaded from publicly available resource and re-analyzed to find out the significant differentially expressed genes (DEGs) (p-value <0.05). Gene enrichment and functional analysis were done using MapMan and PageMan. DAVID web server was used to identify gene ontology. NetworkAnalyst was used for finding out the protein-protein interaction.

RESULTS:

721 (178 were up-regulated and 534 were down-regulated) differentially regulated genes in fruit development were in common between all the four fruit studies. The gene set enrichment analysis performed on these commonly regulated genes showed that the following biological processes were significantly affected during the fruit development: induction of major CHO metabolism, starch and lactoylglutathione lyase and repression of cell wall-related pathways, photosynthesis and cytokinin synthesis. Some of the key genes involved in ethylene, brassinosteroid and cytokinin were down-regulated in the late stage of fruit ripening, while two key abscisic acid-related genes were up-regulated. Fruit ripening up-regulated some key genes encoding Homeobox (HB17), AP2EREBP (RAP2), C2C2-CO-like, WRKY (WRKY9), MYB (MYB62) while repressing several key genes encoding bHLH, MYBs, WRKYs and C2H2.

CONCLUSION:

This comprehensive meta-analysis allowed identifying which genes should be the main targets of a grape breeding strategy to improve physiological processes linked to fruit development and ripening. These data will be used for future sustainable biotechnologies strategies based on small-molecule treatments and genome editing.

Keywords

Berry ripening fruit development meta-analysis RNA-Seq transcriptomics.

1 Introduction

Grapes (Vitis vinifera) are an important woody crop cultivated in a broad range of environmental conditions. The grape berry is extremely rich in healthy secondary metabolites ranging from anthocyanins, carotenoids, terpenoids, glucosinolates, lignin, and many other organic compounds. Grapefruit development and ripening is the result of complex molecular, physiological and physical modifications such as colour change, fruit softening, volatile emission, primary and secondary metabolism changes. Grapefruit development can be divided into three berry growth stages in which physiological and biochemical changes occur. In the first stage (formation), there is an exponential increase in berry size due to rapid cell division and growth, pericarp cell division, organic acid accumulation (malic acid, tartaric acid, phenolic compounds) occurs. The stage II (lag phase) can be considered as a transition stage between an unripe fruit and the third stage of development. During this stage, pericarp growth is reduced, and the embryo development is finished. Veraison occurs at the end of this lag stage. In the third step (ripening) involves important morphological and physiological changes, like colour development, turgor reduction and berry enlargement and decreased acidity [1]. Transcriptomic works have been conducted to elucidate gene molecular regulatory networks underlining these processes [2 –5]. These metabolic intense changes are modulated by complex regulatory networks where hormones (Abscisic acid (ABA), ethylene, brassinosteroids) play a pivotal role in this orchestrated molecular reprogramming [6]. In addition, auxin, cytokinin and gibberellin are generally high and reach a maximum concentration just before the veraison (lag stage) and then decreases sharply along with ripening [7, 8]. ABA and ethylene amount enhanced just before veraison, and they are reduced when the fruit is mature [9, 10] implying that these two hormones should be strong candidates as ripening inducers. Brassinosteroids are key small molecules regulating berry ripening through the interaction with ABA and auxin [11, 12]. ABA up-regulated ABF2 (ABRE binding factor 2) that promote fruit softening and increases phenylpropanoid [13]. A computational method has been used for the analysis of transcriptomic changes during grape berry development and ripening identifying a hundred genes that might be responsible for the “switching off” of the important physiological and metabolic changes occurring during fruit development and determining the quality of the fruit and the structure of the resulting wine [14].

Within this complex picture, it has been hypothesized recently that sugars should play a regulating role in grape development and ripening [15]. Their role will not only be on the energy supply for cellular respiration but also as small molecule modulating berry development through the interaction with hormones and circadian clock inducing a signaling cascade that will promote senescence and repress photosynthesis [16]. Among the sugar-related genes important for fruit development, there include hexose kinase (HXK1), SNF1 related protein kinase 1.1 (SnRK1.1), a target of rapamycin (TOR) and glucose phosphate transporter (GPT2) [17, 18].

The crucial step in fructification is flowering. The number of flowers formed affects the number of fruits produced. Grapes usually form several flowers which exceeds than the final number of fruits ripened. As in other species which form many flowers [19, 20], the number of flowers is the parameter which usually determines yield in grapes [21]. Though researchers have paid much effort to elucidate the genetic network of flowering induction, differentiation and abscission and to explain their mechanism at the molecular level in different developmental stages including bud dormancy, flowering and fruit development, it is such a complex phenomenon that thousands of genes are involved in [2 , 22]. Indeed, the proper metabolic pathway for each of the development stages remains unclear. The comparison between bud dormancy, flower responses to hormone treatments and fruit development will give insights into the complex biological processes and provide a list of gene candidates for genetic manipulation for research in plant model and for the management of crops. This is essential to stimulate significant progress in molecular breeding. Also, the treatment of gibberellic acid to the flower acts as a trigger for its abscission [23]. The molecular study of this process will identify routes linking the aptitude of an organ to become competent for cell separation, specificities and communication between different pathways leading to plant organ drop.

Transcriptomic studies have arisen during the past few years with the rapid development of the high-throughput sequencing technology. Although single transcriptomic studies are useful to have a wide picture of the molecular reprogramming in relation to an object of study, transcriptomic works have several drawbacks. It is known how gene expression is affected by multiple environmental factors, involved in different metabolic, physiological, developmental, and organ-specific processes. In order to determine the pattern of expression of each gene in different physiological processes, it is essential to compare transcriptomic data dealing with multiple research objects. The presence of many transcriptomic studies using different techniques like RNA-Seq, microarrays, cDNA libraries etc. for several important crops allows gaining insight into common and molecular markers associated with important agronomic and physiological traits [24 –27]. False positive or negative results are frequent because of a) the limited quantitative power of next-generation sequencing, b) the unspecific expression of the gene in relation to the analyzed tissues, c) the high environmental variability affecting transcript abundance (especially in field studies). A meta-analysis of transcriptomic data allows filtering the most meaningful information linked with the object of study, eliminating data affected by environmental variability, reducing false positive results, and virtually increasing the number of replicates. Meta-analysis provides a precise estimate of treatment effect giving due weight to the size of the different studies included. Also, it is to be noted that the number of transcripts is highly affected with the changes in environmental conditions and thereby the expression of genes is modulated along with the changes in sample collection timing, environmental factors, experimental conditions, tissues selected and their developmental stages. The importance of performing a meta-analysis is getting higher because the transcriptomic analyses are generally performed only in one season and consequently highly affected by the environment. Although the large quantity of obtained data renders a bit difficult this analysis, the integration of different omic data and the prediction of protein-protein interactions between differentially regulated genes allow gaining insight into these complex networks. The aim of our study was to compare transcriptomic data previously published and related to grape berry development and ripening in order to identify those genes that are commonly present in multiple studies, should be the candidate of playing a pivotal role in fruit development. Our purpose was to shed lights into the complex and still under-clarified regulatory networks modulating grape berry development and ripening through a comparative analysis of 9 RNA-Seq studies. Finally, we aimed at identifying which of the key genes involved in berry development are also involved in developmental and physiological processes in other reproductive tissues such as bud and flower.

2 Materials and methods

2.1 Search strategy of published study identification for meta-analysis

The published transcriptomic studies in Vitis vinifera were identified from Scopus and PubMed using the combination of keywords “RNA-Seq”, “flower”, “bud”, “fruit”, “transcriptomics” and “Vitis vinifera” in computer-based searches and were published on or before January 2019. We found out a total of 9 studies that satisfy the above criteria. The identified studies were divided into three groups according to the studies described: (i) Bud dormancy (2 articles), (ii) Flower development treated with gibberellin (3 articles), and (iii) Berry development and ripening (4 articles; where, Wong et al., 2016 [28] is divided into two).

We characterized how berries differ in their ripening mechanism, metabolism and transcriptional program in relation to the berry development from stage 1 (early stage) to stage 3 (late stage). Therefore, for the raw data analysis, the dataset for the berry development and ripening were selected and the DEGs were found out by comparing the late stage vs. early stage. The early stage studies selected for the analysis was in a range of 35 days after anthesis (DAA) to 47 DAA and late stages were from 84 DAA to 105 DAA. The article information, cultivar and environment of the Vitis vinifera, objective of the study, organ analyzed, SRA ID, sample information was provided in Table 1.

Table 1
Articles selected, objective of the study, organ used for sampling, sample description and sample selection stages

Article Cultivars & environment Objective Organ Sample description Sample stages

Control (early) Treated (late)

Sara Zenoni et al., 2010 [34] Corvina (Verona, Italy) Berry development Berry Post fruit set1 (SRR029279) Ripening 1 (SRR029282) Samples were collected at 35 and 105 days after anthesis

Post fruit set2 (SRR029281) Ripening2 (SRR029258)

Wong et al., 2016 (a) [28] Merlot (British Columbia, Canada) Berry development Berry 47DAA_Large_R1 (SRR3288837) 103DAA_Large_R1 (SRR3288821) Samples were collected 47 and 103 days after anthesis

47DAA_Large_R2 (SRR3288864) 103DAA_Large_R2 (SRR3288822)

47DAA_Large_R3 (SRR3288865) 103DAA_Large_R2 (SRR3288824)

Wong et al., 2016 (b) [28] Merlot (British Columbia, Canada) Berry development Berry 47DAA_Small_R1 (SRR3288866) 103DAA_Small_R1 (SRR3288825) Samples were collected 47 and 103 days after anthesis

47DAA_Small_R2 (SRR3288867) 103DAA_Small_R2 (SRR3288828)

47DAA_Small_R3 (SRR3288878) 103DAA_Small_R3 (SRR3288830)

Shangguan et al., 2017 [37] Fujiminori (Nanjing, China) Berry development Berry Treated (SRR3124517) Control (SRR3124515) Samples were collected 42 and 84 days after anthesis

Khalil-Ur-Rehman et al., 2017 [88] Shine Muscat (Nanjing, China) Bud dormancy Bud Paradormancy (SRR3113203) Endodormancy (SRR3113201) Samples were harvested on Feb. 2 (EndoDormancy stage) and Aug 8 (ParaDormancy stage)

Min et al., 2017 [91] Cabernet Sauvignon (Yangling, China) Bud dormancy Bud Control1 +Control2+Control3(SRR4449743) Treated1 +Treated2+Treated3 (SRR4449743) Prompt buds were collected on May 25, 2015 and latent bud were collected on November 25, 2015

Cheng et al., 2015 [89] Kyoho (Shaanxi, China) Gibberellin treatment in flower development Flower Control_24 h(1) (SRR1554085) Treated_24 h(1) (SRR1553996) Samples were collected after 24 hours post GA3 application

Control_24 h(2) (SRR1554086) Treated_24 h(2) (SRR1554080)

Upadhyay et al., 2018 [90] Thompson seedless (India) Gibberellin treatment in flower development Flower Control_24h Treated_24 h Samples were collected after 24 hours post GA3 application

Domingos et al., 2016 [92] Thompson seedless (Portugal) Gibberellin treatment in flower development Flower Gac1T5 (SRR1927171) Gac1T7 (SRR1994885) Samples were collected 120 hours post GA3 application

Gac2T5 (SRR1994762) Gac2T7 (SRR1994889)

Gac3T5 (SRR1994880) Gac3T7 (SRR1994892)

Article	Cultivars & environment	Objective	Organ	Sample description	Sample stages
Sara Zenoni et al., 2010 [34]	Corvina (Verona, Italy)	Berry development	Berry	Post fruit set1 (SRR029279)	Ripening 1 (SRR029282)	Samples were collected at 35 and 105 days after anthesis
				Post fruit set2 (SRR029281)	Ripening2 (SRR029258)
Wong et al., 2016 (a) [28]	Merlot (British Columbia, Canada)	Berry development	Berry	47DAA_Large_R1 (SRR3288837)	103DAA_Large_R1 (SRR3288821)	Samples were collected 47 and 103 days after anthesis
				47DAA_Large_R2 (SRR3288864)	103DAA_Large_R2 (SRR3288822)
				47DAA_Large_R3 (SRR3288865)	103DAA_Large_R2 (SRR3288824)
Wong et al., 2016 (b) [28]	Merlot (British Columbia, Canada)	Berry development	Berry	47DAA_Small_R1 (SRR3288866)	103DAA_Small_R1 (SRR3288825)	Samples were collected 47 and 103 days after anthesis
				47DAA_Small_R2 (SRR3288867)	103DAA_Small_R2 (SRR3288828)
				47DAA_Small_R3 (SRR3288878)	103DAA_Small_R3 (SRR3288830)
Shangguan et al., 2017 [37]	Fujiminori (Nanjing, China)	Berry development	Berry	Treated (SRR3124517)	Control (SRR3124515)	Samples were collected 42 and 84 days after anthesis
Khalil-Ur-Rehman et al., 2017 [88]	Shine Muscat (Nanjing, China)	Bud dormancy	Bud	Paradormancy (SRR3113203)	Endodormancy (SRR3113201)	Samples were harvested on Feb. 2 (EndoDormancy stage) and Aug 8 (ParaDormancy stage)
Min et al., 2017 [91]	Cabernet Sauvignon (Yangling, China)	Bud dormancy	Bud	Control1 +Control2+Control3(SRR4449743)	Treated1 +Treated2+Treated3 (SRR4449743)	Prompt buds were collected on May 25, 2015 and latent bud were collected on November 25, 2015
Cheng et al., 2015 [89]	Kyoho (Shaanxi, China)	Gibberellin treatment in flower development	Flower	Control_24 h(1) (SRR1554085)	Treated_24 h(1) (SRR1553996)	Samples were collected after 24 hours post GA3 application
				Control_24 h(2) (SRR1554086)	Treated_24 h(2) (SRR1554080)
Upadhyay et al., 2018 [90]	Thompson seedless (India)	Gibberellin treatment in flower development	Flower	Control_24h	Treated_24 h	Samples were collected after 24 hours post GA3 application
Domingos et al., 2016 [92]	Thompson seedless (Portugal)	Gibberellin treatment in flower development	Flower	Gac1T5 (SRR1927171)	Gac1T7 (SRR1994885)	Samples were collected 120 hours post GA3 application
				Gac2T5 (SRR1994762)	Gac2T7 (SRR1994889)
				Gac3T5 (SRR1994880)	Gac3T7 (SRR1994892)

2.2 Read alignment, gene differential expression and annotation

For the articles selected, the latest available Vitis vinifera genome and its annotation file was downloaded from Phytozome (https://phytozome.jgi.doe.gov). The raw data files were downloaded from NCBI SRA (Sequence Read Archive) (https://www.ncbi.nlm.nih.gov/sra) and EMBL ArrayExpress (https://www.ebi.ac.uk/arrayexpress/) according to the accession number given in the article and converted to FASTQ format using SRA toolkit version 2.3.5. Raw data underwent pre-processing by trimming low-quality bases followed by adaptor sequence removal to obtain high-quality clean reads using Cutadapt version 1.8.1. The pre-processed high-quality reads were mapped to the corresponding genome with HISAT2 version 2.1.0 [29] using the default parameters. The output of HISAT2 was then used for the identification of differentially expressed genes using Cuffdiff tool in Cufflinks version 2.2.1 pipeline with default parameters. Up- and down-regulated genes with q-value (adjusted p-value) <0.05 were considered for downstream functional analysis as performed in other meta-analysis of RNA-Seq data [30]. The DEGs selected were annotated using corresponding crop genome mapping file downloaded from the Phytozome. Custom made in-house Perl script was used for the selection of genes and mapping. The complete workflow of this study was given in Fig. 1.

Fig.1

Workflow explaining the meta-analysis of Vitis vinifera transcriptomic studies in bud dormancy, flower development with gibberellin treatment and fruit development. Functional data mining tools were indicated.

We divided the article Wong et al., 2016 [28] into two: a) study comparing the ripening and early stage of large berry and b) study comparing the ripening and early stage of the small berry. The common genes present in 4 studies in the fruit development category were selected for the further downstream analysis. Among them, we determined those genes that were also significantly regulated in the other works dealing with bud (2 articles) and flower (3 articles). The genes present uniquely in the 4 berry development articles were removed in order to provide an accurate selection of genes corresponding to fruit development.

2.3 Gene enrichment and functional analysis

We used MapMan [31] with the Arabidopsis thaliana mapping file (http://mapman.gabipd.org/) to map the gene IDs and visualize the metabolic overview, hormone regulation, secondary metabolism, and transcription factors using two generated files: 1) related to the common genes among the fruit development studies, 2) fruit-related genes also modulated in bud and flower RNA-Seq studies.

The PageMan [32] analysis plugin of MapMan was used to visualize differences among metabolic pathways using Wilcoxon tests, no correction, and an over-representation analysis (ORA) cutoff value of 3. We considered all the differentially expressed genes present that are commonly regulated in the fruit studies for the PageMan analysis.

2.4 Gene ontology analysis

The DEGs were subjected to the enrichment of gene ontologies (GOs) using Database for Annotation, Visualization and Integrated Discovery (DAVID) version 6.8 Web server (https://david.ncifcrf.gov/). All the commonly regulated gene IDs of the 4 fruit studies were searched against the The Arabidopsis Information Resource (TAIR) ID identifier in DAVID. The gene ontology information related to the biological process was extracted from the DAVID result. GO enrichment analysis was based on the linear search algorithm followed by a 0.05 false discovery rate (FDR) correction for multiple comparisons with a minimum of five entities mapped to each category.

2.5 Protein-protein interaction network

NetworkAnalyst [33], a web-based tool for network-based visual analytics for gene expression profiling, meta-analysis, protein-protein interaction network analysis and visual exploration, was used for individual data annotation and analysis. The common list of homologous TAIR IDs of gene IDs present in the fruit category was uploaded and mapped against the STRING interactome database with default parameters (confident score cutoff = 900 and with experimental evidence) provided in NetworkAnalyst. To study the key connectives and to simplify the large network, we selected “Minimum Network”.

3 Results

The details of the articles selected for the study such as titles, tissues analyzed and the number of up and down regulated genes and number of corresponding unique Arabidopsis thaliana ortholog were listed in Table 2. The analysis resulted in the identification of a total of 14,958 (5,137 were up-regulated and 9,821 were down-regulated) genes in the fruit category, 3,734 (1,156 were up-regulated and 2,578 were down-regulated) in bud category and 2,507 (1,471 were up-regulated and 1,036 were down-regulated) in flower category. A first comparison was performed between the three categories of studies (bud, flower and fruit) (Fig. 2A). 457 genes were common between bud, flower and fruit implying their general role in plant development processes between different tissues. A second comparison was done comparing the 4 works related to fruit development and ripening (Fig. 2B). Venn diagram highlighted shreds of evidence among the 4 fruit-related studies represented by the 1,436 genes that were common in all the four studies. Out of these 1,436 genes, only 721 showed a common expression pattern. All the downstream analysis was done on these 721 genes. From these selected 721 genes, 60 (11 were commonly regulated) genes were common in bud dormancy and flower development. These 60 genes were given special attention in our study (Table 3).

Table 2
Analyzed articles, number of up- or down-regulated genes and no. of corresponding orthologs (Arabidopsis thaliana). The raw data analysis has been done for all the “Group 3” articles

No. Article Organ No. of corresponding unique Arabidopsis orthologs Sample information

Total (DEGs) Up Down

1 Sara Zenoni et al., 2010 Berry 1436 2632 804 1828

2 Wong et al., 2016(a) Berry 1839 3461 1430 2031

3 Wong et al., 2016(b) Berry 1841 3481 1331 2150

4 Shangguan et al., 2017 Berry 9842 11525 7532 3993

5 Khalil-Ur-Rehman et al., 2017 Bud 3694 8949 2969 5980

6 Min et al., 2017 Bud 40 73 41 32

7 Cheng et al., 2015 Flower 2210 2326 1400 926

8 Upadhyay et al., 2018 Flower 178 308 39 269

9 Domingos et al., 2016 Flower 119 344 281 63

Commonly regulated in 4 of 4 fruit studies 712 178 534

No.	Article	Organ	No. of corresponding unique Arabidopsis orthologs	Sample information
1	Sara Zenoni et al., 2010	Berry	1436	2632	804	1828
2	Wong et al., 2016(a)	Berry	1839	3461	1430	2031
3	Wong et al., 2016(b)	Berry	1841	3481	1331	2150
4	Shangguan et al., 2017	Berry	9842	11525	7532	3993
5	Khalil-Ur-Rehman et al., 2017	Bud	3694	8949	2969	5980
6	Min et al., 2017	Bud	40	73	41	32
7	Cheng et al., 2015	Flower	2210	2326	1400	926
8	Upadhyay et al., 2018	Flower	178	308	39	269
9	Domingos et al., 2016	Flower	119	344	281	63
Commonly regulated in 4 of 4 fruit studies		712	178	534

Fig.2

A) Number of genes specifically/commonly regulated by developmental process in the bud, flower and fruit studies. B) Number of genes commonly regulated between the 4 RNA-seq studies dealing with fruit development and ripening.

Table 3

Significantly regulated genes involved in fruit development and ripening playing a role in developmental and physiological processes in bud and flower. Gene, pathways and fold change (log2FC) were indicated

Gene ID	Description	Bud (Log2FC)	Flower (Log2FC)	Fruit (Log2FC)
AT3G49220	PECTIN METHYLESTERASE 34	–4.74013	1.23878686	–4.224331
AT5G47230	ETHYLENE RESPONSIVE ELEMENT BINDING FACTOR 5	–2.7643	1.82771	–1.6957225
AT5G20230	BLUE-COPPER-BINDING PROTEIN	3.4958794	2.333423734	–2.6864452
AT1G08380	PHOTOSYSTEM I SUBUNIT O	–9.110785	1.063502942	–2.75429
AT2G38640	LURP-one-like protein	2.81075	–1.2515	1.777685
AT1G28220	PURINE PERMEASE 3	2.7138084	–1.2878	–3.1175671
AT2G35940	Embryo sac development arrest 29	2.9857870	–1.47393118	–1.5841857
AT2G41640	Glycosyltransferase family 61 protein	1.3650990	1.641546029	2.6948244
AT5G41761	hypothetical protein	–5.109599	1.014355293	–2.4250206
AT5G19530	ACL5 (spermine synthase)	–4.906066	1.510961919	–3.3190356
AT1G80570	RNI-like superfamily protein	1.6769516	–1.21759143	1.0832132
AT1G25510	aspartyl protease family protein	–4.952889	1.443606651	–3.134468
AT4G13840	CER26, ECERIFERUM 26	2.180738	1.389566812	–1.8590045
AT1G68320	MYB DOMAIN PROTEIN 62, MYB62	–1.2079	–1.785875	1.3045583
AT4G28365	EARLY NODULIN-LIKE PROTEIN 3	–9.96364	–2.3296909	–6.130118
AT4G15270	glucosyltransferase-like protein	2.4378880	–1.29471	–1.7734733
AT2G33530	serine carboxypeptidase-like 46	3.4958800	1.378511623	–0.44529
AT1G06410	TREHALOSE -6-PHOSPHATASE SYNTHASE S7	–2.003965	1.597642462	–0.7652651
AT1G58170	Disease resistance-responsive (dirigent-like protein) family protein	1.5342565	–3.058893	–3.9927811
AT5G03380	Heavy metal transport/detoxification superfamily protein	1.180792	–1.25153876	–1.662149
AT3G13310	Chaperone DnaJ-domain superfamily protein	1.7707572	–1.6896598	1.4361250
AT4G09890	mediator of RNA polymerase II transcription subunit	–4.948000	1.39395	–2.6007312
AT5G22920	CHY ZINC-FINGER AND RING PROTEIN 1	–4.242494	1.150559677	–1.497525
AT2G02850	PLANTACYANIN	–1.437482	1.137503524	–3.734371
AT5G20250	RAFFINOSE SYNTHASE 6	–3.178071	–1.1202942	2.602438
AT5G57150	basic helix-loop-helix (bHLH) DNA-binding superfamily protein	–3.310132	1.350497247	–2.335478
AT1G06460	Alpha-crystallin domain 32.1	4.8598898	–1.5145731	–0.563896
AT4G39050	Kinesin-like protein KIN-7D	–5.284577	1.3045110	–0.5616867
AT3G03540	Non-specific phospholipase C5	5.2143212	1.566331592	–1.716816
AT4G10490	DMR6-LIKE OXYGENASE 2	–4.11359	–1.653526	–1.908436
AT4G24000	Glycosyltransferase	1.568741	1.562339167	–2.345427
AT1G80160	Lactoylglutathione lyase	5.225791	–2.37781	3.0666690
AT4G29780	Nuclease	–2.094474	2.03901	–0.783134
AT5G06060	Tropinone reductase homolog	3.374961	1.97085654	–4.561722
AT2G26730	Probable inactive receptor kinase	–3.307231	1.063502942	–2.441607
AT4G18010	Type I inositol polyphosphate 5-phosphatase 2	1.15359	1.042644337	–1.7055957
AT1G22060	Sporulation-specific protein	–2.92056	1.718087584	–1.5048160
AT4G12730	Fasciclin-like arabinogalactan protein 2	–2.61177	1.599317794	–3.526749
AT4G34160	Cyclin-D3-1	–7.10405	1.070389328	–1.552338
AT5G08020	Replication protein A 70 kDa DNA-binding subunit B	–8.97280	1.835924074	–1.572486
AT5G03570	FERROPORTIN 2	–3.6690	–1.18442	–1.544882
AT5G58860	Cytochrome P450 86A1	–3.54349	2.2821772	–1.041560
AT4G18170	WRKY transcription factor 28	2.630180	1.82374936	–3.04522
AT4G29080	Auxin-responsive protein IAA27	–2.069513	1.0635942	–2.648342
AT3G03990	DWARF 14	–2.1806	–1.05889	–1.391506
AT1G24130	Transducin/WD40	–1.47834	–1.785875	–3.459234
AT1G48130	1-CYSTEINE PEROXIREDOXIN 1	5.26761	1.655443027	1.137075
AT5G42500	Dirigent protein 2	6.71035	–1.732469	–1.929931
AT2G27500	Glucan endo-1,3-beta-glucosidase 14	–1.65813	1.3950628	–4.805825
AT5G40610	Glycerol-3-phosphate dehydrogenase	2.10180	–1.286304	–2.30155
AT1G01630	Polyphosphoinositide binding protein	–2.98905	1.1762773	0.758001
AT1G70370	Polygalacturonase 1 beta-like protein 3	–4.6102	1.4276073	–4.172901
AT3G10480	NAC DOMAIN CONTAINING PROTEIN 50	–1.34229	–1.286304	–0.603205
AT5G24470	Two-component response regulator-like APRR5	3.93499	–1.736965	1.541453
AT5G53460	Glutamate synthase 1 [NADH], chloroplastic	–1.02078	1.09085343	0.4932766
AT3G63110	ISOPENTENYLTRANSFERASE 3	–3.38523	–1.5145731	–3.5323252
AT2G19500	CYTOKININ OXIDASE 2	–7.36303	–2.5386112	–2.3589860
AT4G23030	DETOXIFICATION 49	6.1532	5.289096702	–3.4114028
AT1G74160	LONGIFOLIA 3	–3.307149	–1.64385619	–1.193238
AT2G34930	Disease resistance family protein	3.65139	1.201633	–3.47577

3.1 Gene set enrichment analysis

Gene enrichment analysis using PageMan performed on the commonly regulated in the four fruit-related studies showed an induction of major CHO metabolism, starch and lactoylglutathione lyase. It also confirmed the repression of photosynthesis and cytokinin synthesis. Most of the cell wall categories like cellulose synthesis, cell wall proteins and pectin esterases were repressed as expected (Fig. 3).

Fig.3

Pageman analysis of differentially regulated genes during fruit development and ripening. Pathways were identified using Wilcoxon test and an over-representation analysis (ORA) cutoff value of 3.

3.2 Biological process enrichment analysis

DAVID software was used in order to identify the gene ontologies (biological process, cellular component and molecular function) commonly modulated among RNA-Seq datasets on fruit development and ripening (Table 4).

Table 4
Significantly regulated biological process during fruit development and ripening. Down-regulated pathways in late fruit stage compared to early stage which are common in 4 studies were indicated. P-value indicates significantly different according to Fisher Exact test (P-value ≤0.05). Total TAIR Hit indicates the number of Arabidopsis (genes) hit involved in signaling pathway

GO_ID GO_Term Total TAIR Hit p-value

Down regulated

GO:0071555 Cell wall organization 20 6.85E-05

GO:0008152 Metabolic process 18 3.96E-04

GO:0006468 Protein phosphorylation 36 4.07E-04

GO:0007169 Transmembrane receptor protein tyrosine kinase signaling pathway 11 5.85E-04

GO:0006857 Oligopeptide transport 7 0.003438

GO:0009926 Auxin polar transport 6 0.007506

GO:0010200 Response to chitin 9 0.008998

GO:0009833 Plant-type primary cell wall biogenesis 4 0.009003

GO:0016042 Lipid catabolic process 10 0.009862

GO:0003333 Amino acid transmembrane transport 5 0.016638

GO:0030244 Cellulose biosynthetic process 5 0.019254

GO:0009751 Response to salicylic acid 9 0.020982

GO:0006874 Cellular calcium ion homeostasis 4 0.02976

GO:0006816 Calcium ion transport 4 0.03232

GO:0052696 Flavonoid glucuronidation 7 0.036626

GO:0009813 Flavonoid biosynthetic process 8 0.040589

GO:0071554 Cell wall organization or biogenesis 4 0.040664

GO_ID	GO_Term	Total TAIR Hit	p-value
GO:0071555	Cell wall organization	20	6.85E-05
GO:0008152	Metabolic process	18	3.96E-04
GO:0006468	Protein phosphorylation	36	4.07E-04
GO:0007169	Transmembrane receptor protein tyrosine kinase signaling pathway	11	5.85E-04
GO:0006857	Oligopeptide transport	7	0.003438
GO:0009926	Auxin polar transport	6	0.007506
GO:0010200	Response to chitin	9	0.008998
GO:0009833	Plant-type primary cell wall biogenesis	4	0.009003
GO:0016042	Lipid catabolic process	10	0.009862
GO:0003333	Amino acid transmembrane transport	5	0.016638
GO:0030244	Cellulose biosynthetic process	5	0.019254
GO:0009751	Response to salicylic acid	9	0.020982
GO:0006874	Cellular calcium ion homeostasis	4	0.02976
GO:0006816	Calcium ion transport	4	0.03232
GO:0052696	Flavonoid glucuronidation	7	0.036626
GO:0009813	Flavonoid biosynthetic process	8	0.040589
GO:0071554	Cell wall organization or biogenesis	4	0.040664

The main biological process pathways showed a down-regulation in the analysis were as follows: (a) amino acid transmembrane transport, (b) plant type primary cell wall biogenesis, (c) auxin polar transport, (d) calcium transport and (e) cellulose biosynthetic process. Whereas no GO terms were up-regulated in late fruit ripen stage with p-value <0.05.

3.3 Hormone overview

We identified several key genes involved in all plant hormones that should play an important role in fruit development and ripening (Fig. 4). Taken together, auxin-related genes did not show any common trend of expression during berry development. Four auxin-related genes were significantly regulated in the comparison between early and late stage: 2 were up-regulated and 2 were down-regulated in ripe fruits compared with early fruit stage. Among the down-regulated genes, it is worthy to mention well-known genes involved in tissue development such as PIN5 (auxin-hydrogen symporter) and two auxin-responsive genes (AT3G02250, AT1G75590). On the other hand, TORNADO 1 and IAA-amino acid conjugate hydrolase was induced in late stage. Genes involved in brassinosteroid, ethylene, cytokinin and jasmonate were mostly enhanced in the early stage compared to ripe fruit. Among the cytokinin, it is worthy to mention: cytokinin oxidase 5, isopentenyl transferase, cytokinin oxidase 2. All these were upregulated in the early stage of fruit development and repressed at the ripening stage. Two ABA-related genes were repressed (GRAM domain-containing protein (AT5G23350) and HVA22C). As expected, most of the genes involved in ethylene signalling and response were down-regulated in fruits right after veraison stage such as ACC oxidase 2 (1-aminocyclopropane-1-carboxylate oxidase), ERT2 (ERF nuclear protein 2), ERF5, ethylene responsive protein (AT3G20640). Three key biosynthetic genes of gibberellin were upregulated in the early stages (GASA1, and two gibberellins responsive proteins (AT1G22690, AT2G14900)). In addition, genes involved in jasmonate signaling were down regulated at ripening stage: lipoxygenases and LOX5. Finally, brassinosteroid-related gene (CYTOCHROME P450 51G1) shows repression.

Fig.4

Overview of significantly regulated genes during fruit development and ripening involved in hormone-related pathways. “Red” means up-regulated while “green” means down-regulated in late fruit stage compared to early stage.

3.4 Transcription factors

Relating to transcription factor (TF) categories in the ripened stage of the fruit, there were key family members showing an enhanced expression such as AP2-EREBP (RAP2.1, RAP2.6L), HB (HB17), MYB (MYB60), WRKY (WRKY9) and C2C2-CO like (B-Box domain protein 28 and zinc finger (B-Box type) family protein) (Table 5). On the other hand, other families showed the opposite direction of expression between their gene members.

Table 5
Overview of significantly regulated genes during fruit development and ripening encoding transcription factors in late fruit stage compared to early stage

Transcriptions Factors

Bin Name TAIR ID Description Fold Change (log2FC) Expression

AP2/EREBP AT2G46310 CRF5 (CYTOKININ RESPONSE FACTOR 5) –1.9371415 Down

AT1G12610 DDF1 (DWARF AND DELAYED FLOWERING 1) –1.3039219 Down

AT1G46768 RAP2.1 (related to AP2 1) 1.3835801 Up

AT5G13330 Rap2.6L (related to AP2 6L) 1.1867013 Up

AT3G25730 AP2 domain-containing transcription factor –4.4980116 Down

WRKY AT4G31550 WRKY11 –2.0277388 Down

AT1G80840 WRKY40 –1.9740765 Down

AT1G68150 WRKY9 0.7432513 Up

AT4G18170 WRKY28 –3.0452259 Down

AT2G46400 WRKY46 –2.8212848 Down

AT2G38470 WRKY33 –2.308155 Down

bHLH AT5G43650 bHLH92 –1.2454424 Down

AT5G57150 bHLH35 –2.3354785 Down

AT5G46690 bHLH071 (beta HLH protein 71) –2.175655 Down

AT2G22760 basic helix-loop-helix (bHLH) family protein –2.4564972 Down

AT1G61660 basic helix-loop-helix (bHLH) family protein –1.2173105 Down

C2C2(Zn) CO-like AT4G27310 B-BOX DOMAIN PROTEIN 28 3.9653044 Up

AT1G68520 B-BOX DOMAIN PROTEIN 14 –1.7063589 Down

AT3G21890 zinc finger (B-box type) family protein 3.5346227 Up

MYB AT4G38620 MYB4 –1.9843785 Down

AT4G05100 MYB74 –3.3428907 Down

AT1G68320 MYB62 1.3045583 Up

AT1G08810 MYB60 –1.0562247 Down

AT1G57560 MYB50 –3.167689 Down

C2H2 AT5G03510 zinc finger (C2H2 type) family protein 1.9771876 Up

AT5G57520 ZFP2 (ZINC FINGER PROTEIN 2) –1.5527505 Down

AT2G29660 zinc finger (C2H2 type) family protein –2.881028 Down

AT1G14580 zinc finger (C2H2 type) family protein –1.6634076 Down

HB AT2G01430 Homeobox-Leucine zipper protein 17 0.55657 Up

AT5G65310 Homeobox protein 5 –1.64796 Down

AT2G35940 BEL1-Like Homeobox protein 1 –1.584185 Down

Transcriptions Factors
AP2/EREBP	AT2G46310	CRF5 (CYTOKININ RESPONSE FACTOR 5)	–1.9371415	Down
	AT1G12610	DDF1 (DWARF AND DELAYED FLOWERING 1)	–1.3039219	Down
	AT1G46768	RAP2.1 (related to AP2 1)	1.3835801	Up
	AT5G13330	Rap2.6L (related to AP2 6L)	1.1867013	Up
	AT3G25730	AP2 domain-containing transcription factor	–4.4980116	Down
WRKY	AT4G31550	WRKY11	–2.0277388	Down
	AT1G80840	WRKY40	–1.9740765	Down
	AT1G68150	WRKY9	0.7432513	Up
	AT4G18170	WRKY28	–3.0452259	Down
	AT2G46400	WRKY46	–2.8212848	Down
	AT2G38470	WRKY33	–2.308155	Down
bHLH	AT5G43650	bHLH92	–1.2454424	Down
	AT5G57150	bHLH35	–2.3354785	Down
	AT5G46690	bHLH071 (beta HLH protein 71)	–2.175655	Down
	AT2G22760	basic helix-loop-helix (bHLH) family protein	–2.4564972	Down
	AT1G61660	basic helix-loop-helix (bHLH) family protein	–1.2173105	Down
C2C2(Zn) CO-like	AT4G27310	B-BOX DOMAIN PROTEIN 28	3.9653044	Up
	AT1G68520	B-BOX DOMAIN PROTEIN 14	–1.7063589	Down
	AT3G21890	zinc finger (B-box type) family protein	3.5346227	Up
MYB	AT4G38620	MYB4	–1.9843785	Down
	AT4G05100	MYB74	–3.3428907	Down
	AT1G68320	MYB62	1.3045583	Up
	AT1G08810	MYB60	–1.0562247	Down
	AT1G57560	MYB50	–3.167689	Down
C2H2	AT5G03510	zinc finger (C2H2 type) family protein	1.9771876	Up
	AT5G57520	ZFP2 (ZINC FINGER PROTEIN 2)	–1.5527505	Down
	AT2G29660	zinc finger (C2H2 type) family protein	–2.881028	Down
	AT1G14580	zinc finger (C2H2 type) family protein	–1.6634076	Down
HB	AT2G01430	Homeobox-Leucine zipper protein 17	0.55657	Up
	AT5G65310	Homeobox protein 5	–1.64796	Down
	AT2G35940	BEL1-Like Homeobox protein 1	–1.584185	Down

Among the genes involved during the early stages of fruit ripening, there is an induction of MYBs (MYB4, MYB74 MYB60 AND MYB50), WRKYs (WRKY11, WRKY40, WRKY28, WRKY46 AND WRKY33) and two AP2/EREBP domain-containing TFs (Cytokinin response factor 5 and DDF1 (dwarf and delayed flowering 1)). Interestingly, during fruit growth, several key genes encoding homeobox factors were up-regulated such as HB5, BLH1 (BEL1-LIKE HOMEODOMAIN 1), and BEL1 (BELL 1). Among the categories of TF, we observed three bHLH family proteins were repressed at the ripened stage: bHLH92, bHLH35 and bHLH71.

3.5 Secondary metabolism

As expected, transcript abundance of several genes involved in secondary metabolism was commonly affected during ripening between the 4 RNA-Seq studies related with fruit development and ripening (Fig. 5). Transferase family protein of phenylpropanoid genes was mostly down regulated in the ripened stage of the fruit. There were also a few genes with higher expression belonging to different categories of secondary metabolic routes such as geranylgeranyl reductase of Non-MVA pathway and chalcone isomerase. Two genes encoding anthocyanins and two genes involved in dihydroflavonols were repressed at the ripened stage. Relating to the early stage fruit development genes, it is worth mentioning the up-regulation of genes involved in carotenoids (geranylgeranyl-diphosphate geranylgeranyltransferase), flavonols (dihydrokaempferol 4-reductase), glucosinolates (aconitase C-terminal domain containing protein).

Fig.5

Significantly regulated genes during fruit development and ripening involved in secondary metabolism. “Red” means up-regulated while “green” means down-regulated in late fruit stage compared to early stage.

3.6 Fruit related genes involved in the development of other reproductive tissues

The aim of this small part of the meta-analysis was to identify genes that play not only a role in fruit development but also on important developmental and physiological processes in grapes such as bud dormancy and response to gibberellin treatment. We found 11 genes that were commonly regulated (with the same trend of expression) among all the 9 studies while 60 were commonly modulated by without a common trend of expression (Table 3). Among them, it is worth mentioning the repression of cytokinin oxidase 2 and isopentenyltransferase3 involved in cytokinin pathways. Another repressed gene in all three tissues was DWARF14, α/β hydrolase which hydrolyses strigolactone, a plant branching hormone and interacts with the F-box protein D3/MAX2 that should be involved in strigolactone detection too. Only two genes were commonly up-regulated: peroxiredoxin 1 and glycosyltransferase family 61 proteins.

3.7 Protein-protein interaction network analysis involved in fruit development

The aim of this study was to identify which protein play an important role in berry development and ripening at PPI level for their high number of interactions (hub proteins) or for their capacity to connect important hubs. The protein-protein interaction (PPI) network analysis in Vitis vinifera was predicted using NetworkAnalyst based on Arabidopsis knowledgebase. The interaction network among the common genes (721) involved in fruit development was visualized (Fig. 6). The bioinformatic analysis identified some key highly interactive genes involved in pathways linked with fruit development such as RAN1, Cytochrome P450 51, three members of the family of CBL interacting protein kinases (CIPK20, CIPK7 and CIPK14), two members of ubiquitin-related reactions (UBC1 and UBC11) and a highly interactive heat shock protein (HSP70).

Fig.6

Protein-protein interaction network analysis predicted in Vitis vinifera based on Arabidopsis knowledgebase involved in fruit development and ripening mapped against the STRING interactome database. Highly interactive proteins (confident score cutoff = 900) were indicated. Red means up-regulated while green means down-regulated.

4 Discussion

Meta-analysis is a powerful tool to dissect transcriptomic datasets, typically performed only in one season, with no biological replicates and often in only one specific environment. The purpose of this meta-analysis of RNA-Seq was to provide important insights into the complex molecular mechanisms driving berry development and ripening. Extensive research has been recently conducted on this very important physiological aspect exploiting the power of the exponential progress occurred in the development of sequencing platforms [28 , 34–37]. However, these studies have been conducted separately and suffer the typical drawbacks of transcriptomic analysis such as high data variability due to environmental uncontrolled effects, the scarce repeatability and reliability due to the differences in developmental and physiological conditions of the analyzed plant material. Indeed, our meta-analysis is useful to integrate findings obtained by different studies filtering the key information mostly linked to the important physiological and metabolic changes occurring during berry development. The elucidation of which genes should play a major role in the accumulation of key metabolites changing in berry development is of extreme importance to improve the quality of wine and table grapes and render grapevines more resistant to environmental constraints caused by climate change. We focused on the analysis of genes that were commonly modulated with the same trend of expression (up- or down-regulation) in all four datasets dealing with fruit development. We focused our analysis on three kinds of genes playing a key role in transcriptional regulation: 1) transcription factors, 2) hormone-related (signaling and response), 3) secondary metabolism. Considering that our meta-analysis compared late ripe fruits (85–103 days after flowering) vs early developing fruits (around 35–47 days after flowering) (Fig. 7), we expected to see a higher amount of repressed than induced genes. This was because when the fruit already ripe, the metabolic activity of the fruit should be reduced. Unfortunately, there were not enough transcriptomic studies at veraison stage to be used for meta-analysis. This explains the repression of genes involved in hormone-related pathways such as ethylene, cytokinin and gibberellins.

Fig.7

Overview of the stage selection of articles related with grape berry development and ripening along with the significant regulation of key transcription factors.

4.1 Primary metabolism

Gene set enrichment analysis showed that photosynthetic reactions were repressed during fruit development. On the other hand, we found that starch metabolism was up-regulated implying that starch degradation is highly associated with the ripening process of grapefruit. This evidence agreed with recent studies that have shown an up-regulation of genes involved in starch biosynthesis (ADP-glucose pyrophosphorylase and sucrose phosphate synthase) and starch degradation (α- and β-amylases) during grape ripening. Interestingly it has been observed that (in the fruitlets) this expression is regulated by the circadian rhythms [38]. The important role of starch in grape development and ripening has been confirmed by a previous study dealing with the elucidation of primary metabolism changes in grapefruit [39]. This work showed that starch was a primary source of energy for the early stages of fruit development. In addition, the accumulation of starch could have a structural function to maintain cell turgor pressure [40].When the development begins, the fruit needs more energy for its growth and the starch begin to degrade through the action of alpha-amylase and beta-amylase that enhance glucose and fructose levels [39]. Our gene set analysis showed the up regulation of genes involved in starch degradation during fruit development confirmed these previous findings. The up regulation of sugar and starch metabolism is important for grape veraison and softening process.

Metabolomics analysis highlighted the importance of carbohydrate changes during fruit development in peach. This is due to its translocation from the photosynthesizing leaves to fruit under development [41]. Indeed, the high content of galactinol and raffinose has been observed in peach during pit hardening. Our gene set enrichment analysis performed on differentially regulated grape-related genes common between the analyzed studies agreed with previous findings that highlight the up-regulation of sucrose-related genes such as sucrose transport (SUT), sucrose-phosphate synthase (SPS) [42]. Previous studies have shown that changes in activities of sugar-metabolizing enzymes occurring during watermelon fruit development plays an important dynamic role in ripening [43, 44]. Among sugar-regulated genes, raffinose synthase, sucrose-phosphate synthase (SPSs) have been linked with fruit ripening in watermelon. However, it has been shown that sucrose is cleaved quickly in the grapefruit as demonstrated by the increase in glucose and fructose due to the high activity of invertase [39]. Another expected finding of this meta-analysis work was related to cell wall re-structuring processes occurring during fruit development and ripening. Genes encoding enzymes involved in cellulose synthesis and encoding AGPs, pectin methylesterases were repressed during fruit development and ripening. This partially agreed with previous works performed in peach at ripening and post-harvest stage [45, 46].

4.2 Modulation of transcription factors during berry development

Several studies were showed the importance of a finely tuned regulation of ABA, ethylene, auxins and brassinosteroid at the beginning of ripening, through the variations in their concentrations in relation to the timing of ripening [47 –50]. Interestingly we found that three ABA-responsive genes were up-regulated (HVA22E, 9-cis-epoxycarotenoid dioxygenase and a protein phosphatase 2C) while the other two were repressed (HVA22C and GRAM-domain containing protein). On the other hand, the behaviour of auxin signalling and response was less clear: two genes were induced (II- amino acid conjugate hydrolase and TORNADO1) while PIN5 and an ARF were repressed. ARFs are transcription factors that modulate auxin response [51]. ARF4 has been suggested to mediate the response to auxin changes during the beginning of berry ripening and considered a negative regulator of the ripening-related modifications in the pericarp during veraison [52, 53]. Besides, IAA9 and IAA16 were rapidly up-regulated auxin-responsive genes [51]. We observed induction of homeobox TF such as HB17 and repression of HB5 and BEL1 at the late stage of berry ripening. HD-Zip homeobox proteins are characterized by a conserved homeodomain (HD) followed by leucine zipper motifs [54]. They have a highly conserved homeobox DNA domain of 180 bp, encoding a protein with a helix-loop-helix-turn-helix structure, involved in developmental processes [55]. It has been shown that homeobox factors (KNOX) regulate the biosynthesis of gibberellin, cytokinin [56] and auxins [57]. They repress the expression of GA 20-oxidases involved in the biosynthesis of the last step of gibberellin [58]. Based on these pieces of evidence we can speculate that the up regulation of HB5 and BEL1 have a link with the reduction of gibberellin during the early stage of fruit ripening (pericarp cell division). In both Arabidopsis and tobacco, the KNOX proteins directly repress transcription of genes encoding GA 20-oxidases, the enzymes that encode the last step in GA biosynthesis. Homeobox comprise a large family of proteins that can be divided into different families such as HD-Zip (homeodomain associated with a leucine zipper), PHD finger (plant homeodomain characterized by a finger domain), Bell (with a Bell domain), ZF-HD (zinc finger associated with a homeodomain), KNOX (knotted related homeobox) [59, 60]. HD-Zip genes are modulated by ABA and belong to ethylene signaling pathways. The overexpression of this gene showed to reduce the sensitivity of ethylene [61] and repress the biosynthesis of ethylene through the action on ethylene-related genes such as ACO, ERF2, ERF5. These results supported the idea that some key members of homeobox factor play as repressors of ethylene biosynthesis and response. Taken together our hypothesis, the homeobox genes regulation during fruit development and ripening play a role in the crosstalk between ethylene and gibberellin allowing pericarp expansion and maintaining the fruit at the unripe stage during pericarp division. The expression of some other homeobox member from the beginning of veraison should work as promoters of ethylene-independent developmental processes occurring from veraison to mature stage. BELL1 (BEL1) is playing a role in ovule development in Arabidopsis. MDH1, a homeobox gene with a homeodomain like BEL1 is predominantly present in the expanding fruits and leaves. Transgenic over-expression of this gene has been linked to dwarfing, reduced fertility and changes in carpel and fruit (silique) shape in Arabidopsis suggesting that this gene might play a key role in the control of plant fertility [62].

We found repression of three members of AP2-EREBP and an up-regulation of two other members. Some AP2/ERF TF have shown to be involved in ethylene signaling and were associated with fruit coloring, softening, and flavonoid biosynthesis [63, 64]. An early ripening mutant showed intense modulation of these genes in the fruit suggesting that the regulation of AP2/ERF genes plays a role in fruit ripening. Taken together these findings let us speculate that the induction of RAP2.1, RAP2.6L should play a pivotal role in metabolic processes linked to fruit development and veraison. To corroborate this hypothesis, it has been previously shown that the AP2/ERF genes play functional roles in pathways involved fruit growth and in environmental stresses [65].

RAP2 genes encode AP2-like and EREBP-like proteins that are divided by the number of AP2 domains in each polypeptide as well as by two sequence motifs (YRG and RAYD) elements present in the AP2 domain. TINY is a member of the AP2EREBP family that represses cell proliferation during both vegetative and floral organogenesis in transgenic over-expressed plants [66].

Our meta-analysis identified 5 MYBs commonly regulated between the four transcriptomic datasets related to fruit development. Four were repressed such as MYB4, MYB74, MYB60, MYB50 and one was induced (MYB62). From data obtained from different plant species, it has been demonstrated that the anthocyanin biosynthesis and the color have been modulated by an R2R3 MYBs and/or a basic helix-loop-helix (bHLH) TFs. Plant MYBs have been shown to control pathways involved in secondary metabolism (such as anthocyanins), development, and biotic stress resistance [67]. They contain a conserved DNA-binding domain composed by single or multiple imperfect repeats such as two-repeat (R2R3), linked to anthocyanin pathways. Four MYB genes (two of them are mutated in white grapes) have been identified in the grape genome and play a role in fruit color [68]. The expression of different MYBs drives the expression of different colors. For example, purple plants are generated when PAP1 (MYB factor) is over-expressed in Arabidopsis [69]. Besides, MYB1 represses anthocyanin biosynthesis in strawberry. The bHLH TF family has also been shown to play a functional role in association with MYBs. They are characterized by a domain of 18 basic amino acids followed by two regions of hydrophobic helices divided by a loop. It has been hypothesized that the biosynthesis of anthocyanins is modulated by the interaction of bHLH-MYBs TFs through a complex regulatory network at the transcriptional level [69]. Our data suggest that the higher transcript abundance of MYB62 should be linked with grapefruit expansion and ripening. This has been reported for the first time. At our knowledge, there are no previous findings at this regard.

Taken together, we propose a model of transcriptional regulation of grape berry development and ripening where important players in the molecular “orchestra of grape berry development” is played by the repression of bHLHs (5 genes were repressed) and MYBs (4 genes repressed), the repression of C2H2 types, the induction of MYB62, and some other of AP2EREBPs (Table 5). Some AP2/ERF TFs involved in ethylene signaling were linked with fruit veraison softening and flavonoid biosynthesis [70] and a total of 37 and 47 of 149 AP2/ERFs were shown to be respectively up- and down-regulated during fruit ripening [71] suggesting that this family of TFs should play a role in grapefruit ripening. AUX/IAA and ARF should be upregulated during the early stage and drives the promotion of pericarp cell division responsible for the exponential increase of the berry size. While the berry is growing, the levels of auxin and cytokinins should decrease and the expression of these factors should be reduced until veraison occurs. In these processes, the members of bHLH factors should interact with MYBs driving the transcription expression of anthocyanin genes.

The up-regulation of WRKY TFs might be associated with the induction of starch metabolism since SUSIBA2, a member of WRKY family was shown to bind to the SURE and W-box cis elements in the iso1 gene promoter, a key player of starch accumulation in wheat endosperm [72]. In this sugar-WRKYs crosstalk, ABA should play an important role modulating the expression of important functional genes involved in growth and development. It is known how sugar and ABA promote grape ripening through the action of NCED, an ABA-responsive gene [39]. In agreement with our findigs, other reports showed that WRKYs are playing an important role in fruit ripening. Indeed, 19 WRKYs were induced in an early ripening grape mutant [73]. Previous works have shown that WRKY26, WRKY05, WRKY47, WRKY25, WRKY19 were induced during grape development at specific moments, implying that each of these TFs should have specific function during fruit development [74]. In addition, WRKY23 was shown to be induced during grape berry ripening and this increase was linked to nucleosome reorganization and chromatin condensation that has been associated with senescence processes.

4.3 Hormone crosstalk modulating grape berry development and ripening

Several studies have shown that ABA, MeJA and brassinosteroids [75 –77] promote grape veraison while auxin blocks ABA-driven effects on plant growth and delays the onset of ripening through the inhibition of ABA accumulation. The hormones such as Auxin, cytokinin and gibberellin are typically higher at the early stage of fruit development. All these genes should be responsible for the effects of these hormones on the promotion of cell division. Cytokinin are typically higher at the beginning of fruit set and reduced rapidly close to veraison maintaining at low levels in maturing and mature berries [78]. This pattern is due to their positive role in cell division and negative action on ripening. For these reasons, the application of cytokinin-like compounds in vineyards has the function of improving yield obtaining increases of berry size but reduces the accumulation of cytokinins and increased tannins [79, 80]. Gibberellins (GAs) are also promoting cell division and expansion. Indeed, the use of gibberellin at early stages have enhanced berry size. In grapevine, biological active concentrations of GAs were high during early berry development but were reduced throughout the subsequent berry development [81]. The production of a pool of bioactive GAs was modulated by the increase and localization of GA oxidase transcripts. The increased expression of two gibberellic acid receptors, GIDL1 and GIDL2 in Cabernet Sauvignon grapes was observed [2] implying that the activation of gibberellin signaling is involved in the cell growth at an early stage. Interestingly, the accumulation of tartaric acid is occurring at early berry development when auxin, gibberellin and cytokinin are generally present at sufficiently acting levels. This latter assumption is corroborated by our meta-analysis that found repression of three genes involved in cytokinin biosynthesis and metabolism. A role of gibberellins in the modulation of organic acids has been hypothesized. If this action should affect the accumulation of tartaric acid remain to be elucidated. Martinelli et al. have been previously hypothesized a connection between the TCA cycle, glycolysis and gibberellin in citrus fruits [82, 83] and this has been promoted the use of gibberellin as small molecule regulators in response to Huanglongbing disease in Citrus [84]. However, the key players in the connection between gibberellins, cytokinins and tartaric acid and malic acid accumulation have still been found.

As far as it concerns, we found that a gene involved in brassinosteroid pathway was repressed (cytochrome P450 51G1). Brassinosteroids should promote grape ripening interacting with auxins and ABA. Genes involved in brassinosteroid biosynthesis and response have been characterized in grape [11] and they showed to be induced in correspondence with the increase of brassinosteroid levels at the beginning of fruit ripening leading to increased fruit colouration through the induction of downstream genes involved in anthocyanins [53]. Brassinosteroids are perceived by a specific plasma- leucine-rich-repeat membrane named BRI1 that was shown to be up-regulated after veraison [11]. The bind of brassinosteroid with the receptor induced the dissociation of BRI kinase inhibitor and the association of BRI with its co-receptor BAK1. There is a consequent cascade of phosphorylation/dephosphorylations that lead to the accumulation of unphosphorylated BES1/BZR1 transcription factors with the induction of downstream transcriptional networks [85]. Lisso et al. 2006 [86] found that the use of BRs to grape berries significantly induced ripening, while an inhibitor of BR biosynthesis (brassinazole) provoked a delay of fruit ripening. Our analysis identified 3 CIPKs commonly regulated between the four transcriptomic datasets related to fruit development. CBLs and CIPKs are the key components of perceiving plants and thereby relaying calcium signals to play a crucial role in plant response to abiotic stresses. In grapevine, CBL-CIPK network has been indicated to activate shaker inward K+ channel which is very important for fruit development and strongly up-regulated by drought stress. They can specifically target downstream proteins to transfer the perceived calcium signals in response to various environmental stimuli and development processes [87].

Although grapes are non-climacteric fruit, a small increase of ethylene has been observed close to veraison and followed by its rapid decrease. As we have previously indicated, the repression of four genes involved in ethylene responses was expected because the analysis was done among the late stage of berry ripening vs early stage of ripening where veraison has already occurred and consequently the small induction of ethylene too. Ethylene is a promoter of grape ripening [9] and this has been demonstrated by the up regulation of some ACO isoforms at veraison stage. Particularly ACO3 was induced at early developmental stage of early ripening mutants suggesting that these genes should function at lag stage of berry development in grape ripening mutants [35]. The role played by ethylene in grape ripening has been demonstrated by several findings. The ethylene receptor blocking agent (1-MCP) has shown a significant effect on grape ripening [9]. In addition, some genes involved in ethylene biosynthesis were upregulated during grape development. These genes should be linked with others involved in water transport, cell wall metabolism and transcription factors.

The intense modification occurring during grape berry development should be promoted by the complex crosstalk between ethylene, brassinosteroid and ABA-responsive genes. Interestingly, we observed that among the up-regulated genes, three were related to ABA responses highlighting the role of this hormone in berry development. The molecular mechanism of ABA in the modulation of grapefruit development is still under debate. A rapid increase of ABA has been linked with the beginning of grape ripening and this has been associated with an increase in the expression of PYR, a receptor of ABA. This induction should activate the signaling driven by key proteins such as PP2C and SnRK that will promote the activation of downstream genes involved in ripening. Indeed, it has been hypothesized that grape ripening should be driven by a complex signal transduction activated by ABA. Functional crosstalk between ABA and ethylene at the beginning of berry ripening has already been demonstrated [49]. A slight peak of ethylene has been observed when the expression of ABA-responsive gene NCED1 is enhanced at veraison [9]. Indeed, we hypothesize that both ethylene and ABA are playing a pivotal role in fruit ripening and their interplaying may be needed to begin berry ripening. In climacteric fruits like tomato, ABA seems to be an inducer of ripening by promoting ethylene biosynthesis [76]. Ethylene seems to up-regulate ABA, while auxin seems to repress ABA-induced ripening processes [48]. ABA-driven promotion of grape development at early stages should be due to the induction of NCED gene while BG protein should be induced at later stages releasing more active free ABA that facilitates grape ripening [39]. The positive effect of ABA on grape ripening should work through the action of sucrose that is an important source for the biosynthesis of anthocyanins and aroma metabolism. It has been found that ABA promotes aroma production blocking auxin accumulation speeding grape ripening. Expression analysis has identified key ABA-signaling cascade such as ABF2, NAC17, MYB143 [3]. Auxins have an opposite effect on ripening since these hormones inhibit the expression of genes involved in ABA biosynthesis that usually are expressed just before veraison. In agreement with this evidence, it has been found that low amounts of auxins induce seed development and sugar accumulation leading to an increase in grape dry weight. Evidences are suggesting that auxins should play an important role in grape ripening at a later stage [39]. Finally, it is worth notice that most of the phenylpropanoid genes were repressed implying that the biosynthesis of these important secondary metabolites should have already occurred.

5 Conclusion

We may conclude that the comparison between late vs early fruit stages highlights important findings: 1) the repression of key families of transcription factors such as bHLH, HBs and MYBs, 2) the down-regulation of ethylene, gibberellins and cytokinin and induction of ABA responses, 3) the role of CBL interactive protein kinases (CIPKs) along with CBLs in regulating signal transduction pathways at protein-protein interaction level, 4) secondary metabolism is mostly reduced implying that most of these genes should be activated earlier (close to veraison), 5) repression of several WRKYs that are typically involved in stress responses. A limitation in the study of the grape ripening is the impossibility to have ripening mutants and consequently perform a reverse genetic approach. The use of techniques such as agro-infection and genome editing will contribute more to solve some of the doubts in the factor that are really regulating berry ripening.

Conflict of interest

The authors have no conflict of interest to report.

Funding

The authors report no funding.

Footnotes

Acknowledgments

The authors have no acknowledgments.

References

Serrano

, Espinoza

, Armijo

, Inostroza-Blancheteau

, Poblete

, Meyer-Regueiro

, et al. Omics Approaches for Understanding Grapevine Berry Development: Regulatory Networks Associated with Endogenous Processes and Environmental Responses. Frontiers in Plant Science. 2017;8. doi: 10.3389/fpls.2017.01486

Deluc

, Grimplet

, Wheatley

, Tillett

, Quilici

, Osborne

, et al. Transcriptomic and metabolite analyses of Cabernet Sauvignon grape berry development. BMC Genomics. 2007;8:429. doi: 10.1186/1471-2164-8-429

Pilati

, Perazzolli

, Malossini

, Cestaro

, Demattè

, Fontana

, et al. Genome-wide transcriptional analysis of grapevine berry ripening reveals a set of genes similarly modulated during three seasons and the occurrence of an oxidative burst at vèraison. BMC Genomics. 2007;8:428. doi: 10.1186/1471-2164-8-428

Fasoli

, Santo

, Zenoni

, Tornielli

, Farina

, Zamboni

, et al. The Grapevine Expression Atlas Reveals a Deep Transcriptome Shift Driving the Entire Plant into a Maturation Program. The Plant Cell. 2012;24:3489–505. doi: 10.1105/tpc.112.100230

Lijavetzky

, Carbonell-Bejerano

, Grimplet

, Bravo

, Flores

, Fenoll

, et al. Berry Flesh and Skin Ripening Features in Vitis vinifera as Assessed by Transcriptional Profiling. PLoS One. 2012;7. doi: 10.1371/journal.pone.0039547

Kühn

, Abello

, Godoy

, Delrot

, Arce-Johnson

. Differential Behavior within a Grapevine Cluster: Decreased Ethylene-Related Gene Expression Dependent on Auxin Transport Is Correlated with Low Abscission of First Developed Berries. PLoS One. 2014;9. doi: 10.1371/journal.pone.0111258

Conde

, Silva

, Fontes

, et al. Biochemical changes throughout grape berry development and fruit and wine quality. Food. 2007;1(1):1–12.

Böttcher

, Harvey

, Forde

, Boss

, Davies

. Auxin treatment of pre-veraison grape (Vitis vinifera L.) berries both delays ripening and increases the synchronicity of sugar accumulation. Australian Journal of Grape and Wine Research. 2010;17:1–8. doi: 10.1111/j.1755-0238.2010.00110.x

Chervin

, El-Kereamy

, Roustan

J-P

, Latché

, Lamon

, Bouzayen

. Ethylene seems required for the berry development and ripening in grape, a non-climacteric fruit. Plant Science. 2004;167:1301–5. doi: 10.1016/j.plantsci.2004.06.026

10.

Deluc

, Quilici

, Decendit

, Grimplet

, Wheatley

, Schlauch

, et al. Water deficit alters differentially metabolic pathways affecting important flavor and quality traits in grape berries of Cabernet Sauvignon and Chardonnay. BMC Genomics. 2009;10:212. doi: 10.1186/1471-2164-10-212

11.

Symons

. Grapes on Steroids. Brassinosteroids Are Involved in Grape Berry Ripening. Plant Physiology. 2005;140:150–8. doi: 10.1104/105.070706

12.

Zhang

, Cai

, Wang

. The primary signaling outputs of brassinosteroids are regulated by abscisic acid signaling. Proceedings of the National Academy of Sciences. 2009;106:4543–8. doi: 10.1073/pnas.0900349106

13.

Nicolas

, Lecourieux

, Kappel

, Cluzet

, Cramer

, Delrot

, et al. The Basic Leucine Zipper Transcription Factor ABSCISIC ACID RESPONSE ELEMENT-BINDING FACTOR2 Is an Important Transcriptional Regulator of Abscisic Acid-Dependent Grape Berry Ripening Processes. Plant Physiology. 2013;164:365–83. doi: 10.1104/113.231977

14.

Fasoli

, Richter

, Zenoni

, Bertini

, Vitulo

, Santo

, et al. Timing and Order of the Molecular Events Marking the Onset of Berry Ripening in Grapevine. Plant Physiology. 2018;178:1187–206. doi: 10.1104/18.00559

15.

Ghan

, Petereit

, Tillett

, Schlauch

, Toubiana

, Fait

, et al. The common transcriptional subnetworks of the grape berry skin in the late stages of ripening. BMC Plant Biology. 2017;17. doi: 10.1186/s12870-017-1043-1

16.

Lastdrager

, Hanson

, Smeekens

. Sugar signals and the control of plant growth and development. Journal of Experimental Botany. 2014;65:799–807. doi: 10.1093/jxb/ert474

17.

Pourtau

, Jennings

, Pelzer

, Pallas

, Wingler

. Effect of sugar-induced senescence on gene expression and implications for the regulation of senescence in Arabidopsis. Planta. 2006;224:556–68. doi: 10.1007/s00425-006-0243-y

18.

Dingenen

, Milde

, Vermeersch

, Maleux

, Rycke

, Bruyne

, et al. Chloroplasts Are Central Players in Sugar-Induced Leaf Growth. Plant Physiology. 2016;171:590–605. doi: 10.1104/15.01669

19.

Marchese

, Marra

, Costa

, Quartararo

, Fretto

, Caruso

. An investigation of the self- and inter-incompatibility of the olive cultivars ‘Arbequina’ and ‘Koroneiki’ in the Mediterranean climate of Sicily. Australian Journal of Crop Science. 2016;10(1):88–93.

20.

Barone

, La Mantia

, Marchese

, Marra

. Improvement in yield and fruit size and quality of the main Italian table olive cultivar ‘Nocellara del Belice’. Scientia Agricola. 2014;71(1):52–7.

21.

Guardiola

. Overview of flower bud induction, flowering and fruit set. Proceedings of Citrus Flowering and Fruit Short Course, IFAS, Citrus Research and Education Center, University of Florida. 1997;5–21.

22.

Kennedy

, Matthews

, Waterhouse

. Changes in grape seed polyphenols during fruit ripening. Phytochemistry. 2000;55:77–85. doi: 10.1016/s0031-9422(00)00196-5

23.

Domingos

, Scafidi

, Cardoso

, Leitao

, Lorenzo

, Oliveira

, et al. Flower abscission in Vitis vinifera L. triggered by gibberellic acid and shade discloses differences in the underlying metabolic pathways. Frontiers in Plant Science. 2015;6. doi: 10.3389/fpls.2015.00457

24.

Bai

, Saito

, Sakamoto

, Ito

, Fujii

, Moriguchi

. Transcriptome Analysis of Japanese Pear (Pyrus pyrifolia Nakai) Flower Buds Transitioning Through Endodormancy. Plant and Cell Physiology. 2013;54:1132–51. doi: 10.1093/pcp/pct067

25.

Fennell

, Schlauch

, Gouthu

, Deluc

, Khadka

, Sreekantan

, et al. Short day transcriptomic programming during induction of dormancy in grapevine. Frontiers in Plant Science. 2015;6. doi: 10.3389/fpls.2015.00834

26.

Sreekantan

, Mathiason

, Grimplet

, Schlauch

, Dickerson

, Fennell

. Differential floral development and gene expression in grapevines during long and short photoperiods suggests a role for floral genes in dormancy transitioning. Plant Molecular Biology. 2010;73:191–205. doi: 10.1007/s11103-010-9611-x

27.

Zhu

, Li

, Xin

, Chen

, Shao

, Wang

, et al. RNA-Seq-based transcriptome analysis of dormant flower buds of Chinese cherry (Prunus pseudocerasus). Gene. 2015;555:362–76. doi: 10.1016/j.gene.2014.11.032

28.

Wong

DCJ

, Gutierrez

, Dimopoulos

, Gambetta

, Castellarin

. Combined physiological, transcriptome, and cis-regulatory element analyses indicate that key aspects of ripening, metabolism, and transcriptional program in grapes (Vitis vinifera L.) are differentially modulated accordingly to fruit size. BMC Genomics. 2016;17. doi: 10.1186/s12864-016-2660-z

29.

Kim

, Langmead

, Salzberg

. HISAT: A fast spliced aligner with low memory requirements. Nature Methods. 2015;12:357–60. doi: 10.1038/nmeth.3317

30.

Benny

, Pisciotta

, Caruso

, Martinelli

. Identification of key genes and its chromosome regions linked to drought responses in leaves across different crops through meta-analysis of RNA-Seq data. BMC Plant Biology. . 2019;19. doi: 10.1186/s12870-019-1794-y

31.

Thimm

, Bläsing

, Gibon

, Nagel

, Meyer

, Krüger

, et al. mapman: A user-driven tool to display genomics data sets onto diagrams of metabolic pathways and other biological processes. The Plant Journal. 2004;37:914–39. doi: 10.1111/j.1365-313x.2004.02016.x

32.

Usadel

, Nagel

, Steinhauser

, Gibon

, Bläsing

, Redestig

, et al. PageMan: An interactive ontology tool to generate, display, and annotate overview graphs for profiling experiments. BMC Bioinformatics. 2006;7. doi: 10.1186/1471-2105-7-535

33.

Xia

, Benner

, Hancock

REW

. NetworkAnalyst – integrative approaches for protein–protein interaction network analysis and visual exploration. Nucleic Acids Research. 2014;42. doi: 10.1093/nar/gku443

34.

Zenoni

, Ferrarini

, Giacomelli

, Xumerle

, Fasoli

, Malerba

, et al. Characterization of Transcriptional Complexity during Berry Development in Vitis vinifera Using RNA-Seq. Plant Physiology. 2010;152:1787–95. doi: 10.1104/109.149716

35.

, Yang

. Transcriptome profiling and identification of the functional genes involved in berry development and ripening in Vitis vinifera. Gene. 2019;680:84–96. doi: 10.1016/j.gene.2018.09.033

36.

Ghan

, Petereit

, Tillett

, Schlauch

, Toubiana

, Fait

, et al. The common transcriptional subnetworks of the grape berry skin in the late stages of ripening. BMC Plant Biology. 2017;17. doi: 10.1186/s12870-017-1043-1

37.

Shangguan

, Mu

, Fang

, Zhang

, Jia

, Li

, et al. RNA-Sequencing Reveals Biological Networks during Table Grapevine (‘Fujiminori’) Fruit Development. Plos One. 2017;12. doi: 10.1371/journal.pone.0170571

38.

Zhu

, Zhang

, Wu

, Li

, Zhang

, Fang

. Enzyme activities and gene expression of starch metabolism provide insights into grape berry development. Horticulture Research. 2017;4:17018. doi: 10.1038/hortres.2017.18

39.

Jia

, Xie

, Wang

, Shangguan

, Qian

, Cui

, et al. Abscisic acid, sucrose, and auxin coordinately regulate berry ripening process of the Fujiminori grape. Functional & Integrative Genomics. 2017;17:441–57. doi: 10.1007/s10142-017-0546-z

40.

Robinson

, Hewitt

, Bennett

. Sink Metabolism in Tomato Fruit: I. Developmental Changes in Carbohydrate Metabolizing Enzymes. Plant Physiology. 1988;87:727–30. doi: 10.1104/87.3.727

41.

Lombardo

, Osorio

, Borsani

, Lauxmann

, Bustamante

, Budde

, et al. Metabolic Profiling during Peach Fruit Development and Ripening Reveals the Metabolic Networks That Underpin Each Developmental Stage. Plant Physiology. 2011;157:1696–710. doi: 10.1104/111.186064

42.

Baker

, Leach

, Boyer

, Swyers

, Benitez-Alfonso

, Skopelitis

, et al. Sucrose Transporter ZmSut1 Expression and Localization Uncover New Insights into Sucrose Phloem Loading. Plant Physiology. 2016;172:1876–98. doi: 10.1104/16.00884

43.

Yativ

, Harary

, Wolf

. Sucrose accumulation in watermelon fruits: Genetic variation and biochemical analysis. Journal of Plant Physiology. 2010;167:589–96. doi: 10.1016/j.jplph.2009.11.009

44.

Yang

, Mo

, Yang

, Zhang

, Wang

, Li

, et al. Transcriptome Profiling of Watermelon Root in Response to Short-Term Osmotic Stress. Plos One. 2016;11. doi: 10.1371/journal.pone.0166314

45.

Rizzini

, Bonghi

, Chkaiban

, Martinelli

, Tonutti

. Effects Of Postharvest Partial Dehydration And Prolonged Treatments With Ethylene On Transcript Profiling In Skins Of Wine Grape Berries. Acta Horticulturae. 2010:1099–104. doi: 10.17660/actahortic.2010.877.149

46.

Tosetti

, Martinelli

, Tonutti

, Barupal

. Metabolomics Approach To Studying Minimally Processed Peach (Prunus Persica) Fruit. Acta Horticulturae. 2012:1017–21. doi: 10.17660/actahortic.2012.934.135

47.

Coombe

, Hale

. The Hormone Content of Ripening Grape Berries and the Effects of Growth Substance Treatments. Plant Physiology. 1973;51:629–34. doi: 10.1104/51.4.629

48.

Davies

, Boss

, Robinson

. Treatment of Grape Berries, a Nonclimacteric Fruit with a Synthetic Auxin, Retards Ripening and Alters the Expression of Developmentally Regulated Genes. Plant Physiology. 1997;115:1155–61. doi: 10.1104/115.3.1155

49.

Sun

, Fan

X-Y

, Cao

D-M

, Tang

, He

, Zhu

J-Y

, et al. Integration of Brassinosteroid Signal Transduction with the Transcription Network for Plant Growth Regulation in Arabidopsis. Developmental Cell. 2010;19:765–77. doi: 10.1016/j.devcel.2010.10.010

50.

, Hou

, Song

, Zheng

, Guo

, Xiao

, et al. Synergistic and Antagonistic Action of Phytochrome (Phy) A and PhyB during Seedling De-Etiolation in Arabidopsis thaliana. International Journal of Molecular Sciences. 2015;16:12199–212. doi: 10.3390/ijms160612199

51.

Woodward

, Bartel

. Auxin: Regulation, action, and interaction. Ann Bot (Lond). 2005;95:707–35.

52.

Gouthu

, Deluc

. Timing of ripening initiation in grape berries and its relationship to seed content and pericarp auxin levels. BMC Plant Biology. 2015;15:46. doi: 10.1186/s12870-015-0440-6

53.

Fortes

, Teixeira

, Agudelo-Romero

. Complex Interplay of Hormonal Signals during Grape Berry Ripening. Molecules. 2015;20:9326–43. doi: 10.3390/molecules20059326

54.

Sessa

, Morelli

, Ruberti

. The Athb-1 and – 2 HD-Zip domains homodimerize forming complexes of different DNA binding specificities. The EMBO Journal. 1993;12:3507–17. doi: 10.1002/j.1460-2075.1993.tb06025.x

55.

Lin

, Hong

, Yin

, Li

, Zhang

, Grierson

. A tomato HD-Zip homeobox protein, LeHB-1, plays an important role in floral organogenesis and ripening. The Plant Journal. 2008;55:301–10. doi: 10.1111/j.1365-313x.2008.03505.x

56.

Ori

. Leaf Senescence Is Delayed in Tobacco Plants Expressing the Maize Homeobox Gene knotted1 under the Control of a Senescence-Activated Promoter. The Plant Cell Online. 1999;11:1073–80. doi: 10.1105/tpc.11.6.1073

57.

Plesch

, Störmann

, Torres

, Walden

, Somssich

. Developmental and auxin-induced expression of the Arabidopsis prha homeobox gene. The Plant Journal. 1997;12:635–47. doi: 10.1046/j.1365-313x.1997.d01-15.x

58.

Hay

, Kaur

, Phillips

, Hedden

, Hake

, Tsiantis

. The Gibberellin Pathway Mediates KNOTTED1-Type Homeobox Function in Plants with Different Body Plans. Current Biology. 2002;12:1557–65. doi: 10.1016/s0960-9822(02)01125-9

59.

Ariel

, Manavella

, Dezar

, Chan

. The true story of the HD-Zip family. Trends in Plant Science. 2007;12:419–26. doi: 10.1016/j.tplants.2007.08.003

60.

Henriksson

. Homeodomain Leucine Zipper Class I Genes in Arabidopsis. Expression Patterns and Phylogenetic Relationships. Plant Physiology. 2005;139:509–18. doi: 10.1104/105.063461

61.

Manavella

, Arce

, Dezar

, Bitton

, Renou

J-P

, Crespi

, et al. Cross-talk between ethylene and drought signalling pathways is mediated by the sunflower Hahb-4 transcription factor. The Plant Journal. 2006;48:125–37. doi: 10.1111/j.1365-313x.2006.02865.x

62.

Dong

, Yao

, Atkinson

, Putterill

, Morris

, Gardener

. MDH An apple homeobox gene belonging to the BEL1 family. Plant Molecular Biology. 2000;42:623–633.

63.

Karlova

, Rosin

, Busscher-Lange

, Parapunova

, Do

, Fernie

, et al. Transcriptome and Metabolite Profiling Show That APETALA2a Is a Major Regulator of Tomato Fruit Ripening. The Plant Cell. 2011;23:923–41. doi: 10.1105/tpc.110.081273

64.

Licausi

, Giorgi

, Zenoni

, Osti

, Pezzotti

, Perata

. Genomic and transcriptomic analysis of the AP2/ERF superfamily in Vitis vinifera. BMC Genomics. 2010;11:719. doi: 10.1186/1471-2164-11-719

65.

, Tao

, Wei

, Ming

, Huang

, Zhang

, et al. The mining and evolutionary investigation of AP2/ERF genes in pear (Pyrus). BMC Plant Biology. 2018;18. doi: 10.1186/s12870-018-1265-x

66.

Wilson

, Long

, Swinburne

, Coupland

. A Dissociation Insertion Causes a Semidominant Mutation That Increases Expression of TINY, an Arabidopsis Gene Related to APETALA2. The Plant Cell. 1996;8:659. doi: 10.2307/3870342

67.

Jin

, Martin

. Multifunctionality and diversity within the plant MYB-gene family. Plant Mol Biol. 1999;41:577–585.

68.

Walker

, Lee

, Bogs

, Mcdavid

DAJ

, Thomas

, Robinson

. White grapes arose through the mutation of two similar and adjacent regulatory genes. The Plant Journal. 2007;49:772–85. doi: 10.1111/j.1365-313x.2006.02997.x

69.

Allan

, Hellens

, Laing

. MYB transcription factors that colour our fruit. Trends in Plant Science. 2008;13:99–102. doi: 10.1016/j.tplants.2007.11.012

70.

Karlova

, Rosin

, Busscher-Lange

, Parapunova

, Do

, Fernie

, et al. Transcriptome and Metabolite Profiling Show That APETALA2a Is a Major Regulator of Tomato Fruit Ripening. The Plant Cell. 2011;23:923–41. doi: 10.1105/tpc.110.081273

71.

Licausi

, Giorgi

, Zenoni

, Osti

, Pezzotti

, Perata

. Genomic and transcriptomic analysis of the AP2/ERF superfamily in Vitis vinifera. BMC Genomics. 2010;11:719. doi: 10.1186/1471-2164-11-719

72.

Sun

. A Novel WRKY Transcription Factor, SUSIBA2, Participates in Sugar Signaling in Barley by Binding to the Sugar-Responsive Elements of the iso1 Promoter. The Plant Cell Online. 2003;15:2076–92. doi: 10.1105/tpc.014597

73.

, Yang

. Transcriptome profiling and identification of the functional genes involved in berry development and ripening in Vitis vinifera. Gene. 2019;680:84–96. doi: 10.1016/j.gene.2018.09.033

74.

Amato

, Cavallini

, Zenoni

, Finezzo

, Begheldo

, Ruperti

, et al. A Grapevine TTG2-Like WRKY Transcription Factor Is Involved in Regulating Vacuolar Transport and Flavonoid Biosynthesis. Frontiers in Plant Science. 2017;7. doi: 10.3389/fpls.2016.01979

75.

Zhang

, Yuan

, Leng

. The role of ABA in triggering ethylene biosynthesis and ripening of tomato fruit. Journal of Experimental Botany. 2009;60:1579–88. doi: 10.1093/jxb/erp026

76.

Jia

, Zhang

, Pervaiz

, Zhao

, Liu

, Wang

, et al. Jasmonic acid involves in grape fruit ripening and resistant against Botrytis cinerea. Functional & Integrative Genomics. 2015;16:79–94. doi: 10.1007/s10142-015-0468-6

77.

, Mao

, Shi

, Zhou

, Asami

, Yu

. A role of brassinosteroids in early fruit development in cucumber. Journal of Experimental Botany. 2008;59:2299–308. doi: 10.1093/jxb/ern093

78.

Zhang

, Horgan

, Reynolds

PHS

, Jameson

. Cytokinins and bud morphology in Pinus radiata. Physiologia Plantarum. 2003;117:264–9. doi: 10.1034/j.1399-3054.2003.00026.x

79.

Maoz

, Bahar

, Kaplunov

, Zutchi

, Daus

, Lurie

, et al. Effect of the Cytokinin Forchlorfenuron on Tannin Content of Thompson Seedless Table Grapes. American Journal of Enology and Viticulture. 2014;65:230–7. doi: 10.5344/ajev.2014.13095

80.

Peppi

, Fidelibus

. Effects of Forchlorfenuron and Abscisic Acid on the Quality of ‘Flame Seedless’ Grapes. HortScience. 2008;43:173–6. doi: 10.21273/hortsci.43.1.173

81.

Cheng

, Xu

, Singer

, Li

, Zhang

, Gao

, et al. Effect of GA₃ treatment on seed development and seed-related gene expression in grape. PLoS One. 2013;8:e80044. doi: 10.1371/journal.pone.0080044

82.

Martinelli

, Reagan

, Uratsu

, Phu

, Albrecht

, Zhao

, et al. Gene Regulatory Networks Elucidating Huanglongbing Disease Mechanisms. PLoS One. 2013;8. doi: 10.1371/journal.pone.0074256

83.

Martinelli

, Uratsu

, Albrecht

, Reagan

, Phu

, Britton

, et al. Transcriptome Profiling of Citrus Fruit Response to Huanglongbing Disease. PLoS One. 2012;7. doi: 10.1371/journal.pone.0038039

84.

Martinelli

, Ibanez

, Reagan

, Davino

, Dandekar

. Stress responses in citrus peel: Comparative analysis of host responses to Huanglongbing disease and puffing disorder. Scientia Horticulturae. 2015;192:409–20. doi: 10.1016/j.scienta.2015.06.037

85.

Clouse

. Brassinosteroids. Arabidopsis Book. 2011;9:e0151.

86.

Lisso

, Altmann

, Müssig

. Metabolic changes in fruits of the tomato dx mutant. Phytochemistry. 2006;67:2232–8. doi: 10.1016/j.phytochem.2006.07.008

87.

, Liu

, Dong

, Cheng

Z-M(M

. The CBL and CIPK Gene Family in Grapevine (Vitis vinifera): Genome-Wide Analysis and Expression Profiles in Response to Various Abiotic Stresses. Frontiers in Plant Science. 2017;8. doi: 10.3389/fpls.2017.00978

88.

Khalil-Ur-Rehman

, Sun

, Li

C-X

, Faheem

, Wang

, Tao

J-M

. Comparative RNA-seq based transcriptomic analysis of bud dormancy in grape. BMC Plant Biology. 2017;17. doi: 10.1186/s12870-016-0960-8

89.

Cheng

, Jiao

, Singer

, Gao

, Xu

, Zhou

, et al. Gibberellin-induced changes in the transcriptome of grapevine (Vitis labrusca×V.vinifera) cv. Kyoho flowers. BMC Genomics. 2015;16:128. doi: 10.1186/s12864-015-1324-8

90.

Upadhyay

, Maske

, Jogaiah

, Kadoo

, Gupta

. GA3 application in grapes (Vitis vinifera L.) modulates different sets of genes at cluster emergence, full bloom, and berry stage as revealed by RNA sequence-based transcriptome analysis. Functional & Integrative Genomics. 2018;18:439–55. doi: 10.1007/s10142-018-0605-0

91.

Min

, Zhao

, Li

, Yang

, Liu

, Fang

. Comparative transcriptome analysis provides insight into differentially expressed genes related to bud dormancy in grapevine (Vitis vinifera). Scientia Horticulturae. 2017;225:213–20. doi: 10.1016/j.scienta.2017.06.033

92.

Domingos

, Fino

, Cardoso

, Sánchez

, Ramalho

, Larcher

, et al. Shared and divergent pathways for flower abscission are triggered by gibberellic acid and carbon starvation in seedless Vitis vinifera L. BMC Plant Biology. 2016;16. doi: 10.1186/s12870-016-0722-7

Transcriptions Factors
Bin Name	TAIR ID	Description	Fold Change (log2FC)	Expression
AP2/EREBP	AT2G46310	CRF5 (CYTOKININ RESPONSE FACTOR 5)	–1.9371415	Down
	AT1G12610	DDF1 (DWARF AND DELAYED FLOWERING 1)	–1.3039219	Down
	AT1G46768	RAP2.1 (related to AP2 1)	1.3835801	Up
	AT5G13330	Rap2.6L (related to AP2 6L)	1.1867013	Up
	AT3G25730	AP2 domain-containing transcription factor	–4.4980116	Down
WRKY	AT4G31550	WRKY11	–2.0277388	Down
	AT1G80840	WRKY40	–1.9740765	Down
	AT1G68150	WRKY9	0.7432513	Up
	AT4G18170	WRKY28	–3.0452259	Down
	AT2G46400	WRKY46	–2.8212848	Down
	AT2G38470	WRKY33	–2.308155	Down
bHLH	AT5G43650	bHLH92	–1.2454424	Down
	AT5G57150	bHLH35	–2.3354785	Down
	AT5G46690	bHLH071 (beta HLH protein 71)	–2.175655	Down
	AT2G22760	basic helix-loop-helix (bHLH) family protein	–2.4564972	Down
	AT1G61660	basic helix-loop-helix (bHLH) family protein	–1.2173105	Down
C2C2(Zn) CO-like	AT4G27310	B-BOX DOMAIN PROTEIN 28	3.9653044	Up
	AT1G68520	B-BOX DOMAIN PROTEIN 14	–1.7063589	Down
	AT3G21890	zinc finger (B-box type) family protein	3.5346227	Up
MYB	AT4G38620	MYB4	–1.9843785	Down
	AT4G05100	MYB74	–3.3428907	Down
	AT1G68320	MYB62	1.3045583	Up
	AT1G08810	MYB60	–1.0562247	Down
	AT1G57560	MYB50	–3.167689	Down
C2H2	AT5G03510	zinc finger (C2H2 type) family protein	1.9771876	Up
	AT5G57520	ZFP2 (ZINC FINGER PROTEIN 2)	–1.5527505	Down
	AT2G29660	zinc finger (C2H2 type) family protein	–2.881028	Down
	AT1G14580	zinc finger (C2H2 type) family protein	–1.6634076	Down
HB	AT2G01430	Homeobox-Leucine zipper protein 17	0.55657	Up
	AT5G65310	Homeobox protein 5	–1.64796	Down
	AT2G35940	BEL1-Like Homeobox protein 1	–1.584185	Down