UBC Faculty Research and Publications

Structure and transcriptional regulation of the major intrinsic protein gene family in grapevine Wong, Darren C J; Zhang, Li; Merlin, Isabelle; Castellarin, Simone D; Gambetta, Gregory A Apr 11, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12864_2018_Article_4638.pdf [ 2.62MB ]
JSON: 52383-1.0365541.json
JSON-LD: 52383-1.0365541-ld.json
RDF/XML (Pretty): 52383-1.0365541-rdf.xml
RDF/JSON: 52383-1.0365541-rdf.json
Turtle: 52383-1.0365541-turtle.txt
N-Triples: 52383-1.0365541-rdf-ntriples.txt
Original Record: 52383-1.0365541-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessStructure and transcriptional regulationof the major intrinsic protein genefamily in grapevineDarren Chern Jan Wong1, Li Zhang2, Isabelle Merlin2, Simone D. Castellarin1 and Gregory A. Gambetta2*AbstractBackground: The major intrinsic protein (MIP) family is a family of proteins, including aquaporins, which facilitatewater and small molecule transport across plasma membranes. In plants, MIPs function in a huge variety ofprocesses including water transport, growth, stress response, and fruit development. In this study, we characterizethe structure and transcriptional regulation of the MIP family in grapevine, describing the putative genomeduplication events leading to the family structure and characterizing the family’s tissue and developmental specificexpression patterns across numerous preexisting microarray and RNAseq datasets. Gene co-expression network(GCN) analyses were carried out across these datasets and the promoters of each family member were analyzed forcis-regulatory element structure in order to provide insight into their transcriptional regulation.Results: A total of 29 Vitis vinifera MIP family members (excluding putative pseudogenes) were identified of whichall but two were mapped onto Vitis vinifera chromosomes. In this study, segmental duplication events were identifiedfor five plasma membrane intrinsic protein (PIP) and four tonoplast intrinsic protein (TIP) genes, contributing to theexpansion of PIPs and TIPs in grapevine. Grapevine MIP family members have distinct tissue and developmentalexpression patterns and hierarchical clustering revealed two primary groups regardless of the datasets analyzed.Composite microarray and RNA-seq gene co-expression networks (GCNs) highlighted the relationships between MIPgenes and functional categories involved in cell wall modification and transport, as well as with other MIPs revealing astrong co-regulation within the family itself. Some duplicated MIP family members have undergone sub-functionalizationand exhibit distinct expression patterns and GCNs. Cis-regulatory element (CRE) analyses of the MIP promoters and theirassociated GCN members revealed enrichment for numerous CREs including AP2/ERFs and NACs.Conclusions: Combining phylogenetic analyses, gene expression profiling, gene co-expression network analyses, andcis-regulatory element enrichment, this study provides a comprehensive overview of the structure and transcriptionalregulation of the grapevine MIP family. The study highlights the duplication and sub-functionalization of the family, itsstrong coordinated expression with genes involved in growth and transport, and the putative classes of TFs responsiblefor its regulation.Keywords: Aquaporin, Berry ripening, Cis-regulatory element, Promoter structure* Correspondence: gregory.gambetta@agro-bordeaux.fr2Bordeaux Science Agro, Institut des Sciences de la Vigne et du Vin,Ecophysiologie et Génomique Fonctionnelle de la Vigne, UMR 1287, F– 33140Villenave d’Ornon, FranceFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Wong et al. BMC Genomics  (2018) 19:248 https://doi.org/10.1186/s12864-018-4638-5BackgroundAquaporins are channel-forming transmembrane proteinspresent in plasma and intracellular membranes in all eu-karyotes and most prokaryotes [1]. Initially, aquaporins’water transport capabilities were discovered and function-ally characterized in human red blood cells [2–4] and laterin plants (Arabidopsis thaliana) with the functionalcharacterization of a vacuolar water-transporting protein,γ-TIP [5]. After the discovery of plant aquaporins, manystudies have been conducted in order to elucidate theirstructure, function, and regulation across numerous plantspecies [6–8]. Aquaporins were first characterized aswater channels, but they are also recognized to contributeto the transport of other small neutral molecules (e.g.,glycerol, urea, boric acid, silicic acid), gases (e.g. CO2, am-monia), and even ions under certain circumstances [7–10].Aquaporins fall within an ancient superfamily of mem-brane proteins called the major intrinsic proteins (MIPs).The MIP family consists of a large number of homologs,and can be subdivided into four major subfamilies basedon sequence similarity, which may also indicate theirsub-cellular localizations [11, 12]. The plasma mem-brane intrinsic proteins (PIPs), the tonoplast intrinsicproteins (TIPs), and the nodulin26-like intrinsic proteins(NIPs), comprise the major subfamilies [6, 8, 13]. Thesethree groups of aquaporins have been intensively studiedand well-documented. The small basic intrinsic proteins(SIPs) include only a few isoforms localized in the ER(e.g., 3 homologs in Arabidopsis) [9, 14]. In additionto these four well-conserved subfamilies present in allplant species, several additional novel types of aqua-porins have been distinguished but with a lessubiquitous presence among plant species. For example, theuncategorized X intrinsic proteins (XIPs) were recently dis-covered but are absent in some higher plants, includingArabidopsis [15–17]. The GlpF-like intrinsic proteins andthe hybrid intrinsic proteins were discovered in moss andalgae, but are absent in vascular plants [9, 13].Aquaporins facilitate water transport through plantcells and tissues and play critical rolls in numerousphysiological processes. At the cell level, aquaporins actin osmoregulation, reactive oxygen species signaling, andintracellular transport and storage processes [9]. At thetissue and organ level, aquaporins contribute to plantwater uptake in roots [18] and facilitate changes in leafhydraulic conductance [19]. Additionally, aquaporinsmodulate changes in plant water relations in response toabiotic stress, including drought, salt, and temperature[9]. In fleshy fruit, there is evidence that aquaporins maycontribute to ripening processes in tomato [20] andgrape [21, 22].The structure of the MIP gene family, like many plantgene families, has undergone numerous gene duplicationsresulting in groups of closely related isogenes [11, 23].These closely related isogenes can have overlappingpatterns of expression, or can have undergone sub-functionalization taking on specific developmental and/ortissue related expression patterns [24]. This is certainlythe case for MIP family members where many isogenesdisplay tissue and/or developmentally specific expressionpatterns. Tissue specific expression of MIP isogenes hasbeen observed in numerous species including poplar [25],corn [1, 26, 27], rice [10, 28], Arabidopsis [29], and tomato[20] among other species. On an even finer scale specificisogenes have been associated with specific cell typeswithin organs [19, 30], although most previous studieswere not comprehensive across all MIP family membersor across organs/tissues.Grapevine is a plant species of economic and culturalimportance and one of the first to have its genomesequenced [31]. This information allowed for thecharacterization of large gene families such as the MIPfamily, and indeed this genome information was immedi-ately utilized to integrate cDNA and genome informationin characterizing the MIP family members in grapevine[32]. Since then the original Pinot noir genome has beengreatly improved and there has been a wealth of micro-array and RNAseq studies examining a plethora of condi-tions (organ specificity, developmental stages, biotic andabiotic stresses, agronomical practices, etc.). Furthermore,new tools and approaches have been developed foranalyzing the nature of genome duplications [33], as wellas gene expression and cis-regulatory element structure[34]. These improvements allow for a more comprehen-sive analysis of the grapevine MIP gene family.In the current study we utilized new tools and ap-proaches to characterize the structure and transcriptionalregulation of the MIP gene family in grapevine. We reas-sessed the MIP family members with the updated genomeinformation describing the putative genome duplicationevents leading to the current family structure. The expres-sion of family members was then assessed across numer-ous preexisting microarray and RNAseq datasets in orderto determine their tissue and developmental specific ex-pression patterns. Co-expression analyses were carried outacross these datasets to determine relevant co-regulationpatterns within the MIP gene family and within the tran-scriptome as a whole. Finally, the promoters of each familymember were analyzed for cis-regulatory element struc-ture in order to provide insight into the possible transcrip-tional regulation of each member.MethodsDendrogram construction and gene duplicationclassificationThe grapevine MIP gene family sequences were re-trieved from the ORCAE 12× grapevine annotationV2 (http://bioinformatics.psb.ugent.be/orcae/) throughWong et al. BMC Genomics  (2018) 19:248 Page 2 of 14a combination of keyword and BLAST searches (using de-fault parameters). For the truncated sequences the sur-rounding regions were visually inspected for sequencehomology to ensure the predicted open reading frameswere correct. Gene nomenclature was created followingthe guidelines established in Grimplet et al. 2014 [35].Orthology assignment between predicted grapevine MIPswith Arabidopsis proteins was performed using theConditional Reciprocal Best (CRB)-BLAST method usingdefault settings [36].Multiple sequence alignments and dendrogram con-structions were carried out with Phylogeny.fr [37]. Thefamily was split into sub-families for alignments in orderto avoid artifacts caused by aligning large groups [38].Sequences were aligned with MUSCLE (v3.8.31) usingthe highest accuracy default settings. After alignmentgaps and/or poorly aligned regions were removed usingGblocks (v0.91b) using the following parameters: mini-mum length of a block after gap cleaning = 5, no gap po-sitions were allowed in the final alignment, all segmentswith contiguous nonconserved positions bigger than 8were rejected, minimum number of sequences for aflank position = 55%. Dendrograms were reconstructedusing the maximum likelihood method implemented inthe PhyML program (v3.1/3.0 aLRT) using default set-tings. Reliability for internal branch was assessed usingthe bootstrapping method (100 bootstrap replicates).Dendrograms were drawn with TreeDyn (v198.3).Analysis of genome structure and duplication analysiswas performed using MCScanX [33] using previouslyestablished parameters [39]. Information pertaining tothe gene duplication type (i.e. singleton, dispersed,proximal, tandem, and segmental; for definition seehttp://chibba.pgml.uga.edu/mcscan2/), detected collinearpairs, and tandem/proximal gene duplicate groups werefurther analyzed. Briefly, all genes are initially assignedas ‘singletons’ and ranked (in ascending order) followingtheir positions along chromosomes. Next, all-vs-allBLASTP is performed and results evaluated. The geneswith BLASTP hits to other genes are assigned with ‘dis-persed’ duplicates. Any two genes are assigned ‘proximal’duplicates if the difference between gene ranks are < 20while a rank = 1 between two genes are assigned as‘tandem’ duplicates. Anchor genes within collinear blocksare assigned as ‘WGD/segmental’ duplicates. In the eventwhere a gene have multiple BLASTP hits, assignment ofduplication mode will be in the order of priority beginningwith WGD/segmental followed by tandem, proximal, andfinally dispersed duplication.RNA-seq data analysisPublicly available grapevine next generation sequencingdatasets were downloaded from NCBI Sequence ReadArchive (http://www.ncbi.nlm.nih.gov/sra). Raw fastqreads (single- and paired-end) were extracted using SRAtoolkit fastq-dump. Read trimming and quality filteringof reads (single- and paired-end) were performed withTrimmomatic v0.36 [40], with the following parameters;LEADING:20, TRAILING:20 SLIDINGWINDOW:4:20,MINLEN:40, AVGQUAL:20. Alignment of filtered readstowards the 12× grapevine reference genome [31] wasperformed using HISAT2 v2.0.5 [41] with default param-eters. Gene-level count summarization was performedusing featureCounts [42] using the grapevine 12× v1(http://genomes.cribi.unipd.it/) reference annotation andsubsequent transcript abundance, expressed as frag-ments per kilobase of transcript per million mappedreads (FPKM), estimated with edgeR [43].Gene co-expression network analysisTwo mutual rank (MR) [44] gene co-expression net-works (GCN) were constructed, one based on RNA-seqdata analyzed in this study and another based on the29 K NimbleGen whole-genome microarray data. RNA-seq GCN was constructed using log-transformed FPKMvalues of 29, 970 genes × 237 conditions obtained in thisstudy. Experiment accessions and publication references ofanalyzed data can be found in Additional file 1: Table S11.Microarray GCN was constructed from an updated inputmatrix of Wong et al. 2016 [39] containing 29, 000 genes ×358 conditions, an additional of 139 conditions comparedto the previous study. Gene-centric co-expression clusterswere created for each MIP gene from both RNA-seq andmicroarray GCNs by considering their top 100 co-expressed genes (ranked by MR value). Visualization of thevarious MIP networks was carried out in Cytoscape v3.0[45]. Enrichment of MapMan BIN categories within co-expression clusters were evaluated for enrichment usingFisher’s exact test adjusted with false discovery rate (FDR)for multiple hypothesis correction according to Wong etal. 2016 and 2017 [34, 39]. MapMan BIN categories wereconsidered significantly enriched within co-expressionclusters with a FDR < 0.05.Cis-regulatory element analysis in promoter regionThe frequencies and position information of selected cis-regulatory elements (CREs) within 1 kb promoter regionfrom the transcription start site of MIP genes were obtainedfrom Wong et al. 2017 [34] and further analyzed for pos-ition bias Z-score considering MIP gene family as a whole/only [46]. The Z-score for each CRE was determined usingthe equation: Z-score = (L/2 + p)/ √[((L- l + 1)^2–1)/n].This strategy takes into account the length of the promoter,L; length of the CRE, l; total number of CRE hits present inall promoters, n; and mean position from all identifiedCRE hits, p. Consideration of these well-established criteriaas a whole improves the likelihood of identifying bona fideCREs in selected promoter groups [46].Wong et al. BMC Genomics  (2018) 19:248 Page 3 of 14ResultsFamily structureA total of 33 Vitis vinifera MIP family members wereidentified (Fig. 1; Additional file 1: Table S1). Of these 33family members we designated 4 of them (VviPIP1–2b,VviPIP2–9, VviNIP9-1a and b) as putative pseudogenes(shown in red in Figs. 1 and 2) because they were bothtruncated and not expressed in any of the RNAseq data-sets we analyzed. These 4 genes were excluded fromsubsequent analyses in this work. Direct orthologous rela-tionships between Vitis vinifera, poplar, and Arabidopsisare extremely difficult to establish as evidenced by thenumerous collapsed dendrogram branches (Fig. 2). Weperformed additional reciprocal BLAST analyses betweenthe Arabidopsis and Vitis vinifera genes to aid in orthol-ogy identification, but again in many cases the orthologycould not be resolved (Additional file 1: Table S1; columnJ “ambiguous”).We examined the nature of duplication eventscontributing to the size of the grapevine MIP gene fam-ily (Fig. 2; Additional file 1: Table S1). A total of 27 of 29grapevine MIP genes were successfully mapped on all 19grapevine chromosomes. Location of the remaining two,VviTIP2–2 and VviTIP2–3, remains unresolved based onthe current 12× genome assembly. In this study, seg-mental (9 of 29) duplication events were identified forfive PIP (VviPIP1–2a, VviPIP1–3, VviPIP2–3, VviPIP2–4,VviPIP2–5) and four TIP (VviTIP1–1, VviTIP1–2, Vvi-TIP1–3,VviTIP1–4) genes, contributing to the expansionof PIPs and TIPs in grapevine. For example, VviPIP2–4is collinear to both VviPIP2–3 and VviPIP2–5 and Vvi-TIP1–1 is collinear with VviTIP1–2. The PIP duplicatesare located on collinear blocks on chromosomes 2/15and 6/8/13 while TIP duplicates are on chromosome 6/8/13. Meanwhile, tandem duplication was observed forVviPIP1–4 and VviPIP1–2a, where the latter is also asegmental duplicate with VviPIP1–3. Proximal duplica-tion was observed for VviXIP2–1 and VviXIP2–2 wherethe two are separated by a disease resistance protein.The remaining were classified as dispersed (16 of 29) du-plicates whereby the specific mode of duplication is un-clear (i.e. other than segmental, tandem, and proximalduplication) and no MIPs were identified as singletons.Tissue and developmental specific expression andsub-functionalizationTissue and developmental specific expression profiles ofthe MIP family members were assessed by examiningtheir expression profiles across the nimblegen grapevineexpression atlas [47] (Fig. 3a; Additional file 1: Table S2)and a wide range of existing RNA-seq datasets (Fig. 3band c; Additional file 1: Table S3). Grapevine MIP familymembers have distinct tissue and developmental expres-sion patterns. Hierarchical clustering revealed two pri-mary groups (Fig. 3 groups 1 and 2) that were similarregardless of the datasets analyzed. Comparing just theexpression atlas (Fig. 3a) with grape berry RNAseq data-sets (Fig. 3b), the composition of several subgroups arenearly identical (Fig. 3 sub-groups 3–6).Generally speaking MIP family members are ubiqui-tously expressed across tissues, although their expressiondiffers across developmental stages (Fig. 3a). This is truewithin subfamilies as well with particular isogenes beingexpressed in almost all tissues, again at specific develop-mental stages. The inflorescence and flower parts tendto have high levels of MIP expression across the wholefamily. The primary groups described above (Fig. 3agroups 1 and 2) generally differ in that group 1 is morehighly expressed.Expression across berry development was examinedmore closely because of the lack of information on aqua-porins’ role in fruit development as well as the wealth ofdatasets available. Of the two primary groups (Fig. 3bgroups 1 and 2), group 1 has a much more dynamic ex-pression pattern across berry development regardless oftissue or genotype. In most cases these family membersare highly expressed early in berry development anddown-regulated as development progresses. However,several members of sub-groups 3 and 4 are up-regulatedat the onset of ripening and later during maturation ofthe berry (Fig. 3b; e.g. VviPIP2–3, VviPIP2–5, VviTIP1–2,VviTIP1–3). In contrast to group 1, group 2 is less dy-namic across berry development with a few exceptions,Fig. 1 Protein sequence relationships within the Vitis vinifera MIPfamily. The six major MIP sub-families are shown: PIP1s, PIP2s, TIPs,NIPs, SIPs, and XIPs. Red numbers represent bootstrap values (100bootstrap replicates). Putative pseudogenes are shown in red.Detailed accession, homology, and duplication information ispresented in Additional file 1: Table S1Wong et al. BMC Genomics  (2018) 19:248 Page 4 of 14most notably a cluster of family members that exhibitpericarp-specific expression (Fig. 3a sub-group 5).Duplicated MIP family members exhibit sub-functionalization, at least at the level of their transcrip-tional regulation. For example,VviXIP2–1 and VviXIP2–2have distinct expression patterns across a variety of data-sets (Fig. 3). This is also true for other examples such asfor VviPIP1–4 and VviPIP1–2a. However some duplicatedfamily members exhibit less distinct expression patternssuch as VviTIP1–1 and VviTIP1–2.Enriched functional categories in grapevine MIP geneco-expression networksTo infer the most representative biological functions ofthis mid-sized gene family, we queried two condition-independent gene co-expression network (GCNs) usingFig. 2 Protein sequence relationships between the Vitis vinifera, Arabidopsis, and poplar MIP families. Six major MIP sub-families: PIP1s (a), PIP2s(b), TIPs (c), NIPs (d), SIPs (e), and XIPs (f). Red numbers represent bootstrap values and the tree was collapsed for all bootstrap values under50 (100 bootstrap replicates). Linked proteins represent gene duplications for Vitis vinifera (green links) and poplar (blue links as detailed in [25]).Putative pseudogenes are shown in red. Detailed accession, homology, and duplication information is presented in Additional file 1: Table S1Wong et al. BMC Genomics  (2018) 19:248 Page 5 of 14Fig. 3 Expression of the grapevine MIP gene family across the NimbleGen grapevine expression atlas (a) and various other RNAseq datasets inberries (b) and other organs (c; note only “control” states are compared). Colored bars group like tissues or genotypes. Heatmap represents theZ-score according to the scale depicted. Like groupings are numbered for clarity (Groups 1& 2, numbered blue circles, Sub-groups 3–6, numberedwhite circles). RNAseq experiment accessions and publication references of analyzed data can be found in Additional file 1: Table S11. RawZ-score values can be found in Additional file 1: File S2 and S3Wong et al. BMC Genomics  (2018) 19:248 Page 6 of 14individual MIP genes as ‘guides’ separately (Additionalfile 1: Table S4 and Additional file 1: Table S5) and ana-lyzed their top 100 correlators in detail for biologicalpathway enrichment (Additional file 1: Table 6 andAdditional file 1: Table S7).In a composite microarray and RNA-seq GCN highlight-ing MIP genes and their high-level (BIN depth ≤ 1)enriched functional categories (Fig. 4), BIN categories suchas transport (BIN34), cell wall (BIN10), miscellaneous en-zyme reactions (BIN26) were commonly enriched in MIPco-expression networks. Conversely, categories such asstress (BIN20), cell organization (BIN31), protein metabol-ism (BIN29), and development (BIN33) were only enrichedin specific MIP networks. Specific categories within trans-port, especially Major Intrinsic Proteins (BIN34.19), wereenriched in 17 MIP co-expression networks. Other cat-egories within transport such as ABC/multi-drug trans-porter (BIN34.16), phosphate (BIN34.7), nitrate (BIN34.4),metal (BIN34.12) were enriched in one (VviNIP7–1), one(VviXIP2–1), two (VviTIP2–3, VviTIP2–2), and three(VviPIP2–4,VviTIP5–1,VviNIP4–1) MIP GCNs, respectively(Additional file 1: Table S6 and Additional file 1: Table S7).Enrichment of cell wall (BIN10) categories observed inmany MIP co-expression networks was unexpected(Fig. 4, Additional file 1: Table S6 and Additional file 1:Table S7). In particular, genes encoding cell wall proteins(BIN10.5), degrading enzymes (BIN10.6), pectin esterases(BIN10.8), cellulose synthesis (BIN10.2), and modification(BIN10.7) belong to categories that were enriched in three(VviTIP5–2, VviPIP2–7, VviTIP1–1), four (VviNIP7–1,VviNIP8–1, VviTIP5–1, VviTIP1–3), three (VviTIP5–1,VviTIP1–1, VviTIP1–3), two (VviPIP2–7, VviPIP2–5), andone (VviTIP1–3) MIP GCNs, respectively. The twocategories that deserve attention within the miscellaneous(BIN26) category relate to the enrichment of glutathione-S-transferase and peroxidase co-expressed genes in four(VviTIP2–3, VviTIP2–2, VviPIP2–4, VviPIP1–1) and five(VviTIP1–4, VviNIP1–2, VviTIP2–3, VviTIP2–2, VviTIP5–Fig. 4 Enriched functional categories identified from composite microarray and RNA-seq MIP gene co-expression networks. Node color representsthe parent functional BIN category at BIN depth ≤ 1. Circle size represents the frequency of microarray and/or RNA-seq MIP gene co-expressionnetworks enriched with the corresponding functional BIN category. Solid and dashed edges represent enriched functional BIN category inrelevant microarray and RNA-seq MIP gene co-expression networks, respectivelyWong et al. BMC Genomics  (2018) 19:248 Page 7 of 141) MIP GCNs, respectively. Meanwhile, seven MIP GCNs(VviNIP8–1, VviXIP2–1, VviTIP1–4, VviTIP2–1, VviPIP1–3, VviTIP2–2, VviTIP2–3) enriched with abiotic stress re-lated genes (BIN20.2) are also of interest.Divergence of enriched functional categories in geneco-expression networks of grapevine MIP duplicatesAs a significant proportion of grapevine MIP duplicatesshowed sub-functionalization of gene expression in atissue-specific manner (Fig. 3), we compared theenriched functional categories in the GCNs (Additionalfile 1: Table S6 and Additional file 1: Table S7) of MIPduplicate pairs (Additional file 1: Table S1). GCNs ofduplicates such as VviXIP2–1 and VviXIP2–2 have to-tally distinct enriched categories. While major intrinsicproteins (BIN34.19) and abiotic stress (BIN20.2) geneswere enriched in VviXIP2–1 GCNs, the latter two cat-egories were absent in VviXIP2–2 GCN. Instead, genesinvolved in the light reaction of photosynthesis (BIN1.1),isoprenoid metabolism (BIN16.1), auxin metabolism(BIN17.2), and CYP450-coding genes (BIN26.1) wereenriched in VviXIP2–2 GCNs. In another example,GCNs of VviTIP1–3 and VviTIP1–4 duplicate pairs havein common enrichment for major intrinsic proteins(BIN34.19), however, enrichment of cell wall pectinesterases (BIN10.8) and modification (BIN10.7) wasobserved for VviTIP1–3 while abiotic stress (BIN20.2),cell cycle (BIN31.3), and hormone (i.e. JA and ABA)metabolism (BIN17) functional categories were amongthe many categories enriched in the GCN of its dupli-cate VviTIP1–4.Conversely, duplicated family members that exhibit lessdivergent expression profiles such as VviPIP2–4 andVviPIP2–5 showed more commonalities. Both VviPIP2–4and VviPIP2–5 have a common enrichment for cell wall(BIN10) and major intrinsic proteins (BIN34.19), albeitsome differences were apparent such as enrichment forglutathione S transferases (BIN26.9) and cell organization(BIN31.1) in VviPIP2–4 and VviPIP2–5, respectively(Additional file 1: Table S6 and Additional file 1:Table S7). Similarly,VviTIP1–1 and VviTIP1–2 share en-richment for cell wall (BIN10) related genes, but categor-ies related to light reaction (BIN1.1) and cell organization(BIN31.1) were enriched in VviTIP1–1 and VviTIP1–2,respectively.Cis-regulatory element structure of grapevine MIPpromotersGenome-wide analysis in grapevine promoters havehighlighted many CREs possessing strong position biastowards the transcription start site (TSS) which were im-plicated in a variety of grapevine development and stressresponses [34]. To determine which CREs are biologic-ally relevant for the regulation of grapevine MIPs weextracted the distribution patterns of 222 CREs (6- to 8-mer in length) in the promoter region for grapevine MIPs(Additional file 1: Table S8) selected from Wong et al. 2017[34]. The frequency of occurrence, the median position ofoccurrence, and position bias Z-score were evaluated. Onthese subset of MIP genes, 6-mer and 7-mer CREs namelyRYCGAC, YAACKG, TTRCGT, and ACGTGKC wereamongst top 10 most highly ranked CREs based Z-score(Fig. 5; Additional file 1: Table S9). The most highly rankedCRE, the RYCGAC – part/variant of the dehydration-responsive element (DRE)/C-repeat elements/low-temperature-responsive element [48] – were present in 14MIP promoters (∑ hits: 23, M position: 262) followed byYAACKG CRE – part/variant of the type I R2R3-MYB rec-ognition sites [49] – that were present in 19 MIP pro-moters (∑ hits: 35, M position: 315). The TTRCGT CRE – themajor NAC TF recognition sites [50] – was also rankedhighly and was present in 9 MIP promoters (∑ hits:13, M position: 286). Longer CREs such as ACGTGKC [51] –a well-known ABA-responsive element (ABRE) – werepresent in 8 MIP promoter (∑ hits: 12, M position: 226).For most of these CREs, a position bias towards the TSS(M position < 300) was also observed considering MIP genes.Enrichment of known cis-regulatory elements ingrapevine MIP gene co-expression networksIn this study, promoters of genes within MIP GCNswere also tested for enrichment for known CREs inorder to identify putative shared TF families within theMIP GCNs that may be responsible for their transcrip-tional regulation. Nineteen MIP GCNs displayed signifi-cant enrichment (FDR < 0.01) for at least one CRE tested(Fig. 6; Additional file 1: Table S10). Of these, six MIPGCNs (i.e. VviNIP8–1, VviPIP1–1, VviXIP2–1, VviTIP1–4,VviTIP2–2, and VviTIP2–3) were commonly enriched forthe PHR1-binding sequence (P1BS, GNATATNC). Severalof these MIP GCNs were also co-enriched with otherCREs. For example, AP2/ERF (GCCGGC) and R2R3-MYB (GKTKGTTR) related CREs were observed in theVviNIP8–1 GCN along with related genes such asAP2/ERF TFs,VviTOE3 (VIT_01s0026g01690),VviERF1L4(VIT_07s0005g03270), and VviMYB82C (VIT_11s0016g05690). Promoters of the VviXIP2–1 GCN are alsoenriched for the GCCGGC CRE correlating with the pres-ence of two AP2/ERF TFs,VviTOE2 (VIT_14s0108g00050)and VviTOE3 (VIT_01s0026g01690). Other co-enrichedCREs of interest include HB (CAATWATT) and extendedDRE elements (DEAR4, CRCCGACA) in promotersof the VviPIP1–4 GCN, coincident with three HBTFs (VIT_08s0007g01290, VIT_16s0100g00670, VIT_18s0001g08410) and two AP2/ERF TFs VviERF061 (VIT_02s0025g01360) and VviERF022 (VIT_18s0001g05850).Interestingly, VviTIP3–1 and VviPIP1 − 2c were thetwo GCNs enriched with many of the CREs tested. ForWong et al. BMC Genomics  (2018) 19:248 Page 8 of 14example, many ACGT-related (e.g. ACGTABREMOTI-FA2OSEM, GADOWNAT, CACGTGMOTIF, ABREAT-CONSENSUS) and RY-related (i.e. RYREPEATBNNAPA, RYREPEATLEGUMINBOX) CREs, but alsoothers involved in Ca2+/calmodulin signaling (CGCGBOXAT) and lateral organ boundary TF binding(TCCGGA) were enriched in the VviTIP3–1 GCNspecifically. A suite of TFs whose homologs target thelatter CREs were also present reaffirming the bio-logical relevance of this broad enrichment pattern(Fig. 6; Additional file 1: Table S10). This includes two B3TFs, VviABI3 (VIT_07s0005g05400) and VviFUS3(VIT_14s0068g01290) whose homologs/orthologs inArabidopsis target the RY motif [52] and VviABI5/VvbZIP25 (VIT_08s0007g03420) that targets variousACGT-related CREs.DiscussionExpansion and sub-functionalization of the grapevine MIPfamilyThe number of grapevine MIPs identified (33) is similar tothe number identified in earlier versions of the grape gen-ome assembly (29 MIPs, [32]), Arabidopsis (35 MIPs, [11]),and rice (33 MIPs, [10]). The MIP family is highly conservedand although many orthologous grapevine-Arabidopsispairs were identified more than half of the orthologous rela-tionships were impossible to resolve. The annotation andgene names presented here differ at times (8 of 33) fromthose established in Shelden et al. (2009) [32]. This is pri-marily due to a much improved genome assembly whichallowed for the identification of previously unidentifiedfamily members, and improved computational methods foridentifying the most likely Arabidopsis orthologs [36].Fig. 5 Number and location of highlighted cis-regulatory elements in the promoters of grapevine MIP family members. The AP2/ERF (orange,blue, pink, green), bZIP (turquoise), NAC (yellow), and R2R3–MYB (beige) were amongst top 10 (of 222) most highly ranked CREs (6- to 8-mer inlength) based Z-score. Each occurrence of the CRE is noted at its position with the appropriate colored line. Complete promoter CRE data can befound in Additional file 1: Table 8 and Additional file 1: Table S9Wong et al. BMC Genomics  (2018) 19:248 Page 9 of 14The grapevine MIP gene family has undergone a num-ber of duplication events consistent with the highlyduplicated nature of plant genomes and grapevine spe-cifically [31]. The duplication events concerning seg-mental and tandem duplications identified in this studyhave also been reported for Arabidopsis [23] and rice[53, 54]. Nonetheless, novel duplication events involvingVviXIP2–1 and VviXIP2–2 may be grapevine-specific.It is commonplace that duplicated genes often take ondifferent expression patterns, with respect to specificportions of development and/or location [24]. In thecurrent study some duplicated MIP gene family mem-bers have distinct patterns of expression. It is likely thatthese duplicates have a similar protein function yet func-tion in different contexts, for example VviXIP2–2 inleaves and VviXIP2–1 in roots (see Fig. 3c). In grapevine,several other gene families have a similar history ofduplication and sub-functionalization [39, 55, 56].Concerning fruit specifically, the expression of mostMIP family members decreases as berry developmentprogresses consistent with earlier studies [21, 22]. Grapeberries become increasingly hydraulically buffered fromthe parent plant during ripening and this buffering isthought to result in part from decreases in hydraulicconductivity [21, 57]. This general downregulation ofMIP family members during ripening may contribute tothese decreases in berry hydraulic conductivity. Incontrast, some specific isogenes (e.g. VviPIP2–3 andVviPIP2–5; note PIP2–5 was previously referred to asPIP2–1) show significant expression and even up-regulation throughout the later stages of berry development[21, 22]. Their role in fruit ripening remains unknown, butsome have speculated that they may facilitate small iontransport and/or osmoregulation [58]. Fleshy fruits likegrape berries undergo rapid growth and sugar accumula-tion during ripening and the role of aquaporins in mediat-ing grape berry water relations is certainly worthy offurther study [59].The grapevine expression atlas [47] is a powerful data-set for examining tissue and developmental specificexpression patterns however caution is warranted espe-cially when examining highly conserved gene families.Microarray based expression analyses can be biased viacross-hybridization [60], and this is why it is importantto include RNAseq based analyses as well. The results ofthe MIP family members presented here show strongparallels between both approaches suggesting that anypotential cross-hybridization did not lead to erroneousresults in the case of the expression atlas.Grapevine MIP co-regulation networksBased on the ‘guilt-by-association’ principle, genes in-volved in related processes often share parallel expres-sion dynamics across a wide range conditions includingdifferent organ/cell types, developmental stages, stress,and hormonal perturbations [61]. Gene co-expressionnetworks (GCNs) analyses, which are built upon the‘guilt-by-association’ principle, have been particularlyuseful for ascribing the most representative biologicalfunctions to both individual gene(s) [62–65] and largegene families [39, 66] in grapevine. This study highlightsthe strong co-expression relationships within the MIPfamily itself, and between MIP family members andgenes involved various processes such as growth, cell-division, and cell redox homeostasis.One of the strongest GCN relationships revealed inthis study was that between the MIP family and genesinvolved in growth and transport processes, namely cellwall modification and cell expansion. Aquaporins havebeen implicated in the growth of rose flower petals andare part of a GCN associated with petal cell expansion[67]. In grape berries, targeted analyses of a limitednumber of aquaporins and cell wall metabolic geneswere shown to have similar patterns of expression thatcorrelated with growth [68]. The treatment of grapeFig. 6 Enriched cis-regulatory elements within shared MIP co-expressionnetworks. Enriched CREs between 6- and 8-mer are depicted asseparate panels. Circle opacity represents the enrichment score(−log10 FDR values) of the corresponding enriched CRE and circle sizeof represents the total number of genes containing the enriched CREWong et al. BMC Genomics  (2018) 19:248 Page 10 of 14berries with exogenous ethylene stimulated growth andassociated micro-array analyses revealed coordinatedchanges in the expression of both aquaporin and cellwall metabolic genes [69]. Among the cell wall metabolicgenes identified by Schlosser et al. (2008) [68] and Cher-vin et al. (2008) [69] were the pectin esterases (BIN10.8)and cellulose synthesis (BIN10.2) identified in this study.The congruence between these previous studies and themore global a priori approach utilized here provides ro-bust evidence for a functional link between these groupsof genes.Our GCN analyses also revealed a strong link betweenthe MIP family and cell division (cell cycle, BIN31.3, andcell organization, BIN31.1). This is a relationship thathas not been studied in plants apart from a few studies.Over-expression of tobacco NtTIP1;1 in cell cultureenhanced cellular expansion and cell-division [70] andspecific aquaporin isoforms have been associated withrapidly proliferating tissues in roots [71, 72]. Cell prolif-eration and growth involves the regulation of source-sink relationships intersecting with turgor driven growth,and one could speculate an important role for MIPfamily members in both of these processes. Outside ofplants there is a growing body of work linking aquaporinfunction with the regulation of cell proliferation [73].Another interesting GCN highlighted in this study wasbetween the MIP family and cell redox homeostasis. Themost obvious link between MIPs and redox homeostasisis the fact that many MIP isoforms transport hydrogenperoxide [74]. Therefore perhaps it should not come asa surprise that MIP family members would be amongcoordinated redox homeostasis genes. Links betweenaquaporin function and redox homeostasis are involved inthe regulation of root water uptake under stress [75–77]but not necessarily through a transcriptional mechanism[78] and the same is true for pathogen responses [79]. Per-haps one of the most interesting observations is the nexusbetween cell expansion, cell division, and redox homeosta-sis [80], where aquaporins may play a cornerstone role incoordinating water fluxes and redox homeostasis in thecontrol of growth.The diversity of bona fide cis-regulatory elements ingrapevine MIP promotersRegulation of plant MIP genes is still poorly understood.This study represents a first attempt of characterizingthe CRE structure of the grapevine MIP family andidentifying putative TFs responsible for its regulation. Aslimitations exists even for well-established statisticalmeasures used for prioritizing CREs, combining severalmetrics may overcome potential caveats of each ap-proach [34].Recent studies have shown that the DRE and GCC-box(GCCGCC) core sequences are critical for the regulationof MIP genes by members of AP2/ERF subgroups I, IV,and V in several plants [81–84]. This is consistent withthe highly prioritized DRE in MIP promoters among allother CREs (Fig. 5) and the co-regulation with AP2/ERFTFs including several predicted grapevine subgroup I, IV,and V members. Some of these regulatory relationshipsare conserved while many others are novel. Knownrelationships include co-regulation of a closely relatedgrapevine homolog of Arabidopsis RAP2.11 (VIT_02s0025g03170) with VviTIP2–1, and co-regulation ofVviTIP3–1 with grapevine homolog of ArabidopsisDREB2D. These examples of co-regulation are consistentwith known and predicted targets of Arabidopsis RAP2.11[85] and DREB2D [86]. The DRE sites within VviTIP2–1and VviTIP3–1 promoter may be important for its regula-tion in grapevine. Unexpectedly, GCC-box elements andother GCC-related CREs (GCCGGC, GCCGTC) were notfound within most MIP promoters within 1000 bases fromthe TSS (Additional file 1: Table S8 and Additional file 1:Table S9). This observation might indicate potential diver-gence in AP2/ERF transcriptional regulatory networks in-volving MIP genes between plant species and the DREmay be preferred in grapevines, and/or that GCC-box andrelated CREs are located beyond the promoter regionsanalyzed in this study.Several bZIP, NAC, and R2R3-MYB transcriptionfactors have been shown to regulate specific MIP genes[67, 87, 88] consistent with many highly co-regulatedTFs of these families in MIP subnetworks. Differences inthe distribution of CREs present in MIP promoters werealso observed (Fig. 5). PIP and TIP promoters containmostly AP2/ERF and bZIP-related CREs while NIP pro-moters contain mostly NAC and R2R3-MYB-relatedCREs, suggesting some degree of transcriptional regula-tion specificity in grapevine aquaporin regulation.Promoter analysis suggests that hormone metabolicpathways such as ABA and ethylene play an importantrole in the regulation of MIP genes. There is evidencefor ethylene-regulated aquaporin expression in rosepetals [67, 89] and aquaporin genes are among thoseregulated by exogenous ethylene treatment in grapeberries [69]. Several studies demonstrate that ABA regu-lates the expression of numerous MIP family members[90–92]. However, it is important to point out thatshort-term modulation of aquaporin activity via ABA,and possibly other hormones such as ethylene, likely oc-curs at the post-translational level [93, 94]. In grapevine,ABA has been shown to differentially regulate the sameaquaporin isogene (VviTIP1–1, VIT_06s0061g00730) de-pending on the organ [95]. These complex relationshipsinvolved in the hormonal regulation of aquaporin geneexpression require further study.The promoters of genes in six MIP GCNs were alsocommonly enriched for the PHR1-binding sequenceWong et al. BMC Genomics  (2018) 19:248 Page 11 of 14(P1BS, GNATATNC). The cognate sequence, such asGAATATTC, is known to be bound by members of theMYB (GARP, G2) TF and is related to the regulation oftranscriptional repressors [96] and accordingly no ho-mologs of MYB (GARP, G2) TFs were represented in re-spective MIP GCNs. The enrichment of this CRE maysuggest a potential role of large-scale transcriptional re-pression in the regulation of MIPs. Conversely, for manyother CREs enrichment profiles were often accompaniedby the presence of TF families known to target them(Fig. 6, Additional file 1: Table S10) suggesting a role oftranscriptional activation of MIP and co-regulated genes.The diversity of enriched CREs also highlights that inaddition to the those shown to be directly implicated inMIP regulation such as AP2/ERF, bZIP, NAC, and R2R3-MYB TFs, regulation of MIPs may involve more TF fam-ilies than previously described. Several genes that belong toHB, LBD, and B3 TF families may also represent novel can-didate regulators of grapevine MIP and co-regulated genes.ConclusionsThe current work utilized the most high quality and up-to-date genome information in characterizing the grape-vine MIP gene family, its structure, and the putative du-plication events involved in its evolution. When pairedwith the GCN analyses conducted here we identifiedthose MIP family members that have undergone duplica-tion and sub-functionalization through characterizingthe tissue and developmental specific expression pat-terns across the family. GCN analyses revealed severalinteresting relationships between MIP family membersand genes involved in cell expansion, cell division, andtransport processes. Characterizing the cis-regulatoryelements in grapevine MIP promoters along with associ-ated GCN members identified the putative classes ofTFs responsible for the regulation of the family and theirassociated GCNs. Combining phylogenetic analyses,gene expression profiling, GCN analyses, and CRE en-richment, this study provides a comprehensive overviewof the structure and transcriptional regulation of thegrapevine MIP family. These results can help guide fu-ture studies aimed at understanding the role of specifictranscription factors in controlling the diverse expres-sion patterns within the MIP family.Additional fileAdditional file 1: Supplementary Tables. (XLSX 748 kb)AbbreviationsCRE: Cis-Regulatory Element; DRE: Dehydration-Responsive Element;GCN: Gene Co-expression Network; MIP: Major Intrinsic Protein;NIP: Nodulin26-like Intrinsic Proteins; PIP: Plasma membrane Intrinsic Protein;SIP: Small basic Intrinsic Proteins; TIP: Tonoplast Intrinsic Protein;TSS: Transcription Start Site; XIP: X Intrinsic ProteinsAcknowledgementsThe authors would like to thank Grant Cramer for his help and technicaladvice and the grapevine research community for making various RNA-seqand microarray datasets publicly available.FundingThis study has been carried out with financial support from the University ofBordeaux’s Initiative of Excellence (IdEx) program, doctoral school of life andhealth sciences, and Cluster of Excellence COTE (ANR-10-LABX-45, within theWater Stress project), as well as the Canada Research Chairs Program,Genome British Columbia (10R21188), and the Natural Sciences andEngineering Research Council of Canada (10R23082). The funding bodiesthemselves were not involved in the design of the study or the collection,analysis, and interpretation of data, or in writing the manuscript.Availability of data and materialsAll data analyzed in this study are available from NCBI, https://www.ncbi.nlm.nih.gov/. The grapevine MIP gene family sequences were retrieved from theORCAE 12× grapevine annotation V2, http://bioinformatics.psb.ugent.be/orcae/.All specific gene/protein accessions and transcriptomic dataset accessions canbe found in Additional file 1: Table S1 and Additional file 1: Table S11, respectively.Authors’ contributionsDCJW, SC, GG designed the experiments; DCJW, IM, SC, and GG carried outthe experiments and analyses; DCJW, IM, and GG analyzed the results; andDCJW, LZ, SC, and GG wrote the manuscript. All authors participated inediting and approving the final work.Ethics approval and consent to participateNot applicable; this study has not directly involved humans, animalsor plants.Competing interestsThe authors declare that they have no competing interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Wine Research Centre, University of British Columbia, 2205 East Mall,Vancouver, BC V6T 0Z4, Canada. 2Bordeaux Science Agro, Institut des Sciencesde la Vigne et du Vin, Ecophysiologie et Génomique Fonctionnelle de la Vigne,UMR 1287, F– 33140 Villenave d’Ornon, France.Received: 29 January 2018 Accepted: 29 March 2018References1. Chaumont F, Barrieu F, Wojcik E, Chrispeels MJ, Jung R. Aquaporinsconstitute a large and highly divergent protein family in maize.Plant Physiol. 2001;125:1206–15.2. Benga G, Popescu O, Borza V, Pop VI, Muresan A, Mocsy I, et al. Waterpermeability in human erythrocytes: identification of membrane proteinsinvolved in water transport. Eur J Cell Biol. 1986;41:252–62.3. Denker BM, Smith BL, Kuhajda FP, Agre P. Identification, purification, andpartial characterization of a novel Mr 28,000 integral membrane proteinfrom erythrocytes and renal tubules. J Biol Chem. 1988;263:15634–42.4. Preston GM, Carroll TP, Guggino WB, Agre P. Appearance of water channelsin Xenopus oocytes expressing red cell CHIP28 protein. Science.1992;256:385–7.5. Maurel C, Reizer J, Schroeder JI, Chrispeels MJ. The vacuolar membraneprotein gamma-TIP creates water specific channels in Xenopus oocytes.EMBO J. 1993;12:2241–7.6. Chaumont F, Tyerman SD. Aquaporins: highly regulated channelscontrolling plant water relations. Plant Physiol. 2014;164:1600–18.7. Tyerman SD, Niemietz CM, Bramley H. Plant aquaporins: multifunctional waterand solute channels with expanding roles. Plant Cell Environ. 2002;25:173–94.8. Maurel C, Verdoucq L, Luu D-T, Santoni V. Plant aquaporins: membranechannels with multiple integrated functions. Annu Rev Plant Biol.2008;59:595–624.Wong et al. BMC Genomics  (2018) 19:248 Page 12 of 149. Maurel C, Boursiac Y, Luu D-T, Santoni V, Shahzad Z, Verdoucq L.Aquaporins in plants. Physiol Rev. 2015;95:1321–58.10. Sakurai J, Ishikawa F, Yamaguchi T, Uemura M, Maeshima M. Identificationof 33 rice aquaporin genes and analysis of their expression and function.Plant Cell Physiol. 2005;46:1568–77.11. Johanson U, Karlsson M, Johansson I, Gustavsson S, Sjövall S, Fraysse L, et al.The complete set of genes encoding major intrinsic proteins in Arabidopsisprovides a framework for a new nomenclature for major intrinsic proteins inplants. Plant Physiol. 2001;126:1358–69.12. Alexandersson E, Fraysse L, Sjövall-Larsen S, Gustavsson S, Fellert M, Karlsson M,et al. Whole gene family expression and drought stress regulation ofAquaporins. Plant Mol Biol. 2005;59:469–84.13. Li G, Santoni V, Maurel C. Plant aquaporins: roles in plant physiology.Biochim. Biophys. Acta - Gen. Subj. 1840;2014:1574–82.14. Ishikawa F, Suga S, Uemura T, Sato MH, Maeshima M. Novel type aquaporinSIPs are mainly localized to the ER membrane and show cell-specificexpression in Arabidopsis thaliana. FEBS Lett. 2005;579:5814–20.15. Gupta A, Sankararamakrishnan R. Genome-wide analysis of major intrinsicproteins in the tree plant Populus trichocarpa: characterization of XIPsubfamily of aquaporins from evolutionary perspective. BMC Plant Biol.2009;9:134.16. Danielson JÅ, Johanson U. Unexpected complexity of the aquaporingene family in the moss Physcomitrella patens. BMC Plant Biol.2008;8:45.17. Lopez D, Bronner G, Brunel N, Auguin D, Bourgerie S, Brignolas F, et al.Insights into Populus XIP aquaporins: evolutionary expansion, proteinfunctionality, and environmental regulation. J Exp Bot. 2012;63:2217–30.Oxford University Press18. Gambetta GA, Knipfer T, Fricke W, McElrone AJ. Aquaporins and root wateruptake. Cham: Springer; 2017. p. 133–53.19. Heinen RB, Ye Q, Chaumont FX. Role of aquaporins in leaf physiology.J Exp Bot. 2009;60:2971–85.20. Reuscher S, Akiyama M, Mori C, Aoki K, Shibata D, Shiratake K. Genome-WideIdentification and Expression Analysis of Aquaporins in Tomato. Boudko D,editor. PLoS One. Public Libr Sci; 2013;8:e79052.21. Choat B, Gambetta GA, Shackel KA, Matthews MA. Vascular function ingrape berries across development and its relevance to apparent hydraulicisolation. Plant Physiol. 2009;151:1677–87.22. Fouquet R, Léon C, Ollat N, Barrieu F. Identification of grapevineaquaporins and expression analysis in developing berries. Plant CellRep. 2008;27:1541–50.23. Cannon SB, Mitra A, Baumgarten A, Young ND, May G. The roles ofsegmental and tandem gene duplication in the evolution of large genefamilies in Arabidopsis thaliana. BMC Plant Biol. 2004;4:10. BioMed Central24. Adams KL, Wendel JF. Novel patterns of gene expression in polyploidplants. Sci Trends Genet. 1999;298:2157–67.25. Cohen D, Bogeat-Triboulot M-B, Vialet-Chabrand S, Merret R, Courty P-E,Moretti S, et al. Developmental and environmental regulation of aquaporingene expression across Populus species: divergence or redundancy?Blazquez MA, editor. PLoS One. 2013;8:e55506. Public Library of Science26. Gaspar M, Bousser A, Sissoeff I, Roche O, Hoarau J, Mahe A. Cloning andcharacterization of ZmPIP1-5b, an aquaporin transporting water and urea.Plant Sci. 2003;165:21–31.27. Opitz N, Marcon C, Paschold A, Malik WA, Lithio A, Brandt R, et al. Extensivetissue-specific transcriptomic plasticity in maize primary roots upon waterdeficit. J Exp Bot. 2016;67:1095–107.28. Sakurai J, Ahamed A, Murai M, Maeshima M, Uemura M. Tissue andcell-specific localization of rice aquaporins and their water transportactivities. Plant Cell Physiol. 2008;49:30–9.29. Weig A, Deswarte C, Chrispeels MJ. The major intrinsic protein family ofArabidopsis has 23 members that form three distinct groups with functionalaquaporins in each group. Plant Physiol. 1997;114:1347–57.30. Kirch H-H, Vera-Estrella R, Golldack D, Quigley F, Michalowski CB, Barkla BJ,et al. Expression of Water Channel proteins in Mesembryanthemumcrystallinum. Plant Physiol. 2000;123:111–24.31. Jaillon O, Aury J-M, Noel B, Policriti A, Clepet C, Casagrande A, et al. Thegrapevine genome sequence suggests ancestral hexaploidization in majorangiosperm phyla. Nature. 2007;449:463–7.32. Shelden MC, Howitt SM, Kaiser BN, Tyerman SD. Identification andfunctional characterisation of aquaporins in the grapevine, Vitis vinifera.Funct Plant Biol. 2009;36:1065–78.33. Wang Y, Tang H, Debarry JD, Tan X, Li J, Wang X, et al. MCScanX: a toolkitfor detection and evolutionary analysis of gene synteny and collinearity.Nucleic Acids Res. 2012;40:e49.34. Wong DCJ, Lopez Gutierrez R, Gambetta GA, Castellarin SD. Genome-wideanalysis of cis-regulatory element structure and discovery of motif-drivengene co-expression networks in grapevine. DNA Res. 2017;24:311–26.35. Grimplet J, Adam-Blondon A-F, Bert P-F, Bitz O, Cantu D, Davies C, et al. Thegrapevine gene nomenclature system. BMC Genomics. 2014;15:1077.36. Aubry S, Kelly S, Kümpers BMC, Smith-Unna RD, Hibberd JM. Deepevolutionary comparison of gene expression identifies parallel recruitmentof trans-factors in two independent origins of C4 photosynthesis. BombliesK, editor. PLoS Genet. 2014;10:e1004365. Public Library of Science37. Dereeper A, Guignon V, Blanc G, Audic S, Buffet S, Chevenet F, et al.Phylogeny.Fr: robust phylogenetic analysis for the non-specialist. NucleicAcids Res. 2008;36:W465–9.38. Boyce K, Sievers F, Higgins DG. Instability in progressive multiple sequencealignment algorithms. Algorithms Mol Biol. 2015;10:26.39. Wong DCJ, Schlechter R, Vannozzi A, Höll J, Hmmam I, Bogs J, et al. Asystems-oriented analysis of the grapevine R2R3-MYB transcription factorfamily uncovers new insights into the regulation of stilbene accumulation.DNA Res. 2016;23:451–66.40. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illuminasequence data. Bioinformatics. 2014;30:2114–20.41. Kim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with lowmemory requirements. Nat Methods. 2015;12:357–60.42. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purposeprogram for assigning sequence reads to genomic features. Bioinformatics.2014;30:923–30.43. Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package fordifferential expression analysis of digital gene expression data.Bioinformatics. 2010;26:139–40.44. Obayashi T, Kinoshita K. Rank of correlation coefficient as a comparablemeasure for biological significance of gene coexpression. DNA Res.2009;16:249–60.45. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al.Cytoscape: a software environment for integrated models of biomolecularinteraction networks. Genome Res. 2003;13:2498–504.46. Ma S, Shah S, Bohnert HJ, Snyder M, Dinesh-Kumar SP. Incorporating motifanalysis into gene co-expression networks reveals novel modular expressionpattern and new signaling pathways. Copenhaver GP, editor. PLoS Genet.2013;9:e1003840.47. Fasoli M, Dal Santo S, Zenoni S, Tornielli GB, Farina L, Zamboni A, et al. Thegrapevine expression atlas reveals a deep transcriptome shift driving theentire plant into a maturation program. Plant Cell. 2012;24:3489–505.48. Yamaguchi-Shinozaki K, Shinozaki K. Transcriptional regulatory networks incellular responses and tolerance to dehydration and cold stresses. Annu RevPlant Biol. 2006;57:781–803.49. Prouse MB, Campbell MM. The interaction between MYB proteins and theirtarget DNA binding sites. Biochim Biophys Acta - Gene Regul Mech. 1819;2012:67–77.50. Lindemose S, Jensen MK, Van de Velde J, O’Shea C, Heyndrickx KS,Workman CT, et al. A DNA-binding-site landscape and regulatory networkanalysis for NAC transcription factors in Arabidopsis thaliana. Nucleic AcidsRes. 2014;42:7681–93.51. Narusaka Y, Nakashima K, Shinwari ZK, Sakuma Y, Furihata T, Abe H, et al.Interaction between two cis-acting elements, ABRE and DRE, in ABA-dependentexpression of Arabidopsis rd29A gene in response to dehydration andhigh-salinity stresses. Plant J. 2003;34:137–48.52. Reidt W, Wohlfarth T, Ellerström M, Czihal A, Tewes A, Ezcurra I, et al. Generegulation during late embryogenesis: the RY motif of maturation-specificgene promoters is a direct target of the FUS3 gene product. Plant J.2000;21:401–8.53. Guyot R, Keller B. Ancestral genome duplication in rice. Genome.2004;47:610–4.54. Yu J, Wang J, Lin W, Li S, Li H, Zhou J, et al. The genomes of Oryza sativa: ahistory of duplications. Bennetzen J, editor. PLoS Biol Sinauer. 2005;3:e38.55. Falginella L, Castellarin SD, Testolin R, Gambetta GA, Morgante M,Di Gaspero G. Expansion and subfunctionalisation of flavonoid 3′,5′-hydroxylasesin the grapevine lineage. BMC Genomics. 2010;11:562.56. Vannozzi A, Dry IB, Fasoli M, Zenoni S, Lucchin M, Henikoff S, et al. Genome-wide analysis of the grapevine stilbene synthase multigenic family: genomicWong et al. BMC Genomics  (2018) 19:248 Page 13 of 14organization and expression profiles upon biotic and abiotic stresses.BMC Plant Biol. 2012;12:130. BioMed Central57. Knipfer T, Fei J, Gambetta GA, McElrone AJ, Shackel KA, Matthews MA.Water transport properties of the grape pedicel during fruit development:insights into xylem anatomy and function using microtomography. PlantPhysiol. 2015;168:1590–602. American Society of Plant Biologists58. Rogiers SY, Coetzee ZA, Walker RR, Deloire A, Tyerman SD. Potassium in thegrape (Vitis vinifera L.) berry: transport and function. Front. Plant Sci. 2017;8:1629.59. Tyerman SD, Chaves MM, Barrieu F. Water Relations of the Grape Berry andAquaporins. In: Gerós H, Chaves MM, Delrot S, editors. Biochem. grape berry.Potomac: Bentham Science; 2012. p. 3–22.60. Cramer GR, Ghan R, Schlauch KA, Tillett RL, Heymann H, Ferrarini A, et al.Transcriptomic analysis of the late stages of grapevine (Vitis vinifera cv. Cabernetsauvignon) berry ripening reveals significant induction of ethylene signaling andflavor pathways in the skin. BMC Plant Biol. 2014;14:370. BioMed Central61. Gillis J, Pavlidis P. “Guilt by association” is the exception rather than the rulein gene networks. Rzhetsky a, editor. PLoS Comput Biol. 2012;8:e1002444.62. Loyola R, Herrera D, Mas A, Wong DCJ, Höll J, Cavallini E, et al. Thephotomorphogenic factors UV-B RECEPTOR 1, ELONGATED HYPOCOTYL 5,and HY5 HOMOLOGUE are part of the UV-B signalling pathway in grapevineand mediate flavonol accumulation in response to the environment. J ExpBot. 2016;67:5429–45.63. Amato A, Cavallini E, Zenoni S, Finezzo L, Begheldo M, Ruperti B, et al. Agrapevine TTG2-like WRKY transcription factor is involved in regulating vacuolartransport and flavonoid biosynthesis. Front Plant Sci. 2017;7:1979. Frontiers.64. Malacarne G, Coller E, Czemmel S, Vrhovsek U, Engelen K, Goremykin V, etal. The grapevine VvibZIPC22 transcription factor is involved in theregulation of flavonoid biosynthesis. J Exp Bot. 2016;67:3509–22.Oxford University Press.65. Vannozzi A, Wong DCJ, Höll J, Hmmam I, Matus JT, Bogs J, et al.Combinatorial Regulation of Stilbene Synthase Genes by WRKY and MYBTranscription Factors in Grapevine (Vitis vinifera L.). Plant Cell Physiol. 2018.66. Wong DCJ, Ariani P, Castellarin S, Polverari A, Vandelle E. Co-expressionnetwork analysis and cis-regulatory element enrichment determine putativefunctions and regulatory mechanisms of grapevine ATL E3 ubiquitin ligases.Sci Rep. 2018;8:3151.67. Pei H, Ma N, Tian J, Luo J, Chen J, Li J, et al. An NAC transcription factorcontrols ethylene-regulated cell expansion in flower petals. Plant Physiol.2013;163:775–91.68. Schlosser J, Olsson N, Weis M, Reid K, Peng F, Lund S, et al. Cellularexpansion and gene expression in the developing grape (Vitis vinifera L.).Protoplasma. 2008;232:255–65.69. Chervin C, Tira-umphon A, Terrier N, Zouine M, Severac D, Roustan J-P.Stimulation of the grape berry expansion by ethylene and effects on relatedgene transcripts, over the ripening phase. Physiol Plant. 2008;134:534–46.70. Okubo-Kurihara E, Sano T, Higaki T, Kutsuna N, Hasezawa S. Acceleration ofvacuolar regeneration and cell growth by overexpression of an aquaporinNtTIP1;1 in tobacco BY-2 cells. Plant Cell Physiol. 2009;50:151–60. OxfordUniversity Press71. Chaumont F, Barrieu F, Herman EM, Chrispeels MJ. Characterization of a maizetonoplast aquaporin expressed in zones of cell division and elongation. PlantPhysiol. 1998;117:1143–52. American Society of Plant Biologists72. Gambetta GA, Fei J, Rost TL, Knipfer T, Matthews MA, Shackel KA, et al.Water uptake along the length of grapevine fine roots: developmentalanatomy, tissue-specific aquaporin expression, and pathways of watertransport. Plant Physiol. 2013;163:1254–65.73. Galán-Cobo A, Ramírez-Lorca R, Echevarría M. Role of aquaporins in cellproliferation: what else beyond water permeability? Channels. 2016;10:185–201.74. Bienert GP, Chaumont F. Aquaporin-facilitated transmembrane diffusion ofhydrogen peroxide. Biochim Biophys Acta - Gen Subj. 1840;2014:1596–604.75. Benabdellah K, Ruiz-Lozano JM, Aroca R. Hydrogen peroxide effects on roothydraulic properties and plasma membrane aquaporin regulation inPhaseolus vulgaris. Plant Mol Biol. 2009;70:647–61. Springer Netherlands76. Aroca R, Amodeo G, Fernández-Illescas S, Herman EM, Chaumont F,Chrispeels MJ. The role of aquaporins and membrane damage in chillingand hydrogen peroxide induced changes in the hydraulic conductance ofmaize roots. Plant Physiol. 2005;137:341–53.77. Lee SH, Chung GC, Steudle E. Gating of aquaporins by low temperature inroots of chilling-sensitive cucumber and chilling-tolerant figleaf gourd.J Exp Bot. 2005;56:985–95.78. Boursiac Y, Boudet J, Postaire O, Luu D-T, Tournaire-Roux C, Maurel C.Stimulus-induced downregulation of root water transport involves reactiveoxygen species-activated cell signalling and plasma membrane intrinsicprotein internalization. Plant J. 2008;56:207–18.79. Tian S, Wang X, Li P, Wang H, Ji H, Xie J, et al. Plant aquaporin AtPIP1;4 linksApoplastic H 2 O 2 induction to disease immunity pathways. Plant Physiol.2016;171:1635–50.80. Schmidt R, Kunkowska AB, Schippers JHM. Role of reactive oxygen speciesduring cell expansion in leaves. Plant Physiol. 2016;172:2098–106. AmericanSociety of Plant Biologists81. Rae L, Lao NT, Kavanagh TA. Regulation of multiple aquaporin genes inArabidopsis by a pair of recently duplicated DREB transcription factors.Planta. 2011;234:429–44.82. Zhu D, Wu Z, Cao G, Li J, Wei J, Tsuge T, et al. TRANSLUCENT GREEN, an ERFfamily transcription factor, controls water balance in Arabidopsis byactivating the expression of aquaporin genes. Mol Plant. 2014;7:601–15.83. Hichri I, Muhovski Y, Clippe A, Žižková E, Dobrev PI, Motyka V, et al. SlDREB2,a tomato dehydration-responsive element-binding 2 transcription factor,mediates salt stress tolerance in tomato and Arabidopsis. Plant Cell Environ.2016;39:62–79.84. Liao X, Guo X, Wang Q, Wang Y, Zhao D, Yao L, et al. Overexpression ofMsDREB6.2 results in cytokinin-deficient developmental phenotypes andenhances drought tolerance in transgenic apple plants. Plant J.2017;89:510–26.85. Kim MJ, Ruzicka D, Shin R, Schachtman DP. The Arabidopsis AP2/ERFtranscription factor RAP2.11 modulates plant response to low-potassiumconditions. Mol Plant. 2012;5:1042–57.86. González-Morales SI, Chávez-Montes RA, Hayano-Kanashiro C, Alejo-Jacuinde G,Rico-Cambron TY, de Folter S, et al. Regulatory network analysis reveals novelregulators of seed desiccation tolerance in Arabidopsis thaliana.Proc Natl Acad Sci U S A. 2016;113:E5232–41.87. Borg M, Brownfield L, Khatab H, Sidorova A, Lingaya M, Twell D. The R2R3MYB transcription factor DUO1 activates a male germline-specific regulonessential for sperm cell differentiation in Arabidopsis. Plant Cell.2011;23:534–49.88. Nicolas P, Lecourieux D, Kappel C, Cluzet S, Cramer G, Delrot S, et al. Thebasic leucine zipper transcription factor ABSCISIC ACID RESPONSE ELEMENT-BINDING FACTOR2 is an important transcriptional regulator of abscisicacid-dependent grape berry ripening processes. Plant Physiol. 2014;164:365–83.89. Ma N, Xue J, Li Y, Liu X, Dai F, Jia W, et al. Rh-PIP2;1, a rose aquaporingene, is involved in ethylene-regulated petal expansion. Plant Physiol.2008;148:894–907. American Society of Plant Biologists90. Olaetxea M, Mora V, Bacaicoa E, Baigorri R, Garnica M, Fuentes M, et al.ABA-regulation of root hydraulic conductivity and aquaporin gene- expressionis crucial to the plant shoot rise caused by rhizosphere humic acids. PlantPhysiol. 2015.91. Jang JY, Kim DG, Kim YO, Kim JS, Kang H. An expression analysis of a genefamily encoding plasma membrane aquaporins in response to abioticstresses in Arabidopsis thaliana. Plant Mol Biol. 2004;54:713–25.92. Aroca R, Ferrante A, Vernieri P, Chrispeels MJ. Drought, abscisic acid andtranspiration rate effects on the regulation of PIP aquaporin gene expressionand abundance in Phaseolus vulgaris plants. Ann Bot. 2006;98:1301–10.93. Morillon R, Chrispeels MJ. The role of ABA and the transpiration stream inthe regulation of the osmotic water permeability of leaf cells. Proc NatlAcad Sci U S A. National Academy of Sciences; 2001;98:14138–14143.94. Grondin A, Rodrigues O, Verdoucq L, Merlot S, Leonhardt N, Maurel C.Aquaporins contribute to ABA-triggered stomatal closure throughOST1-mediated phosphorylation. Plant Cell Online. 2015;95. Rattanakon S, Ghan R, Gambetta GA, Deluc LG, Schlauch KA, Cramer GR.Abscisic acid transcriptomic signaling varies with grapevine organ.BMC Plant Biol. 2016;16:72.96. Wu G, Lin W-C, Huang T, Poethig RS, Springer PS, Kerstetter RA. KANADI1regulates adaxial-abaxial polarity in Arabidopsis by directly repressing thetranscription of ASYMMETRIC LEAVES2. Proc Natl Acad Sci U S A.2008;105:16392–7. National Academy of SciencesWong et al. BMC Genomics  (2018) 19:248 Page 14 of 14


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items