Gene coexpression network analysis reveals the role of SRS ...

RESEARCH ARTICLE

Gene coexpression network analysis reveals the role of SRS genesin senescence leaf of maize (Zea mays L.)

BING HE1, PIBIAO SHI2, YUANDA LV3, ZHIPING GAO1 and GUOXIANG CHEN1*

1School of Life Sciences, Nanjing Normal University, Nanjing 210023, Jiangsu, People’s Republic of China2Xinyang Agricultural Experiment Station of Yancheng City, Jiangsu Academy of Agricultural Sciences,Yancheng 224049, Jiangsu, People’s Republic of China3Provincial Key Laboratory of Agrobiology, Institute of Crop Germplasm and Biotechnology, Jiangsu Academyof Agricultural Sciences, Nanjing 210014, Jiangsu, People’s Republic of China*For correspondence. E-mail: [email protected].

Received 5 July 2019; revised 20 September 2019; accepted 4 October 2019

Abstract. Shi-related sequence (SRS) proteins are plant-specific transcription factors that play important roles in developmental pro-cesses, including regulating hormone biosynthesis, response or signal transduction. However, systematical analysis of the SRS gene familyin maize has not yet been conducted. In this study, 11 SRS genes with 13 transcripts were identified and characterized. The characteristics ofthe gene family were analysed in terms of phylogenetic relationships, chromosome distribution and gene structure. RNA-sequencing dataanalysis showed that the expression patterns of SRS genes were quite different from each other in maize, indicating their divergence infunction. Interestingly, the GRMZM2G077752 gene is highly expressed in senescent leaves. Using further coexpression network analysis,we determined that the module containing GRMZM2G077752 were over-represented by genes related to abscisic acid (ABA) stimulus andcarbohydrate metabolic process. This result indicated that GRMZM2G077752 might perceive ABA signal and cause the activation ofcarbohydrate remobilization during leaf ageing. This study provides valuable information for understanding the functions of the SRS genesin maize.

Keywords. gene coexpression network; maize; SRS gene; leaf senescence.

Introduction

The Shi-related sequence (SRS) gene family, also known asthe short internodes (SHI) gene family or SHI/STY/SRSfamily (for short internodes, stylish, and shi-relatedsequence), encode a class of plant-specific transcriptionfactors that have two distinct conserved regions (Fridborget al. 2001). The first region comprises a RING finger-likezinc finger motif and is referred to as the RING domain(Fridborg et al. 2001; Kuusk et al. 2006). The second regionis called the IGGH domain, which is required for homod-

imerization (Eklund et al. 2010a). SHI, the first member ofthe SRS gene family, was identified in the Arabidopsisthaliana dwarf mutant shi and is involved in the response ofgibberellin (GA) (Fridborg et al. 1999, 2001).

There are 11 SRS genes in Arabidopsis: SHI, STY1, STY2,LATERAL ROOT PRIMORDIUM1 (LRP1), SRS3 to SRS8(Smith and Fedoroff 1995; Fridborg et al. 1999, 2001;Kuusk et al. 2002; Sohlberg et al. 2006; Kuusk et al. 2006)and AT1G32730. SRS transcription factors control a diverserange of developmental processes in Arabidopsis, includingroot formation (Smith and Fedoroff 1995), leaf development(Baylis et al. 2013), floral induction and flower development(Fridborg et al. 1999; Sohlberg et al. 2006; Kim et al. 2010;Gomariz-Fernandez et al. 2017), and photomorphogenesis(Yuan et al. 2018). Additionally, many SRS members

Bing He and Pibiao Shi contributed equally to this work.

YL and GC conceived and designed the manuscript. BH and PS performedthe experiments and analysed the data. ZG provided technical andtheoretical support to the manuscript. BH wrote the manuscript.

Electronic supplementary material: The online version of this article (https://doi.org/10.1007/s12041-019-1162-6) contains supplementarymaterial, which is available to authorized users.

Journal of Genetics (2020) 99:3 � Indian Academy of Scienceshttps://doi.org/10.1007/s12041-019-1162-6 (0123456789().,-volV)(0123456789().,-volV)

influence the development of plant organs by regulatinghormone biosynthesis, signal transduction or response. Forexample, the STY1 protein upregulates auxin biosynthesis(Sohlberg et al. 2006; Eklund et al. 2010a), while SHIfunctions as a repressor of the GA response through tran-scriptional control (Fridborg et al. 1999, 2001). Further,when the SRS7 gene is overexpressed, jasmonic acid (JA)signalling is disrupted and anther dehiscence is delayed(Kim et al. 2010). Interestingly, a recent study has suggestedthat SRS5 promotes photomorphogenesis by directly bind-ing with the promoters of photomorphogenesis-promotinggenes (Yuan et al. 2018). The involvement of SRS genes inthe development of plant organs has also been reported inother plant species. For example, SRS transcription factorshave been found to regulate hormones and inflorescencepatterns in barley (Hordeum vulgare L.) (Youssef et al.2017) and root development through auxin signalling inmaize (Zhang et al. 2015), control auxin biosynthesis andsubsequently affect plant growth in the moss Physcomitrellapatens (Eklund et al. 2010b), and shape plant architecture inrice (Oryza sativa L.) (Duan et al. 2019).

Genomewide identification and function analysis of theSRS gene family has been carried out in Arabidopsis byanalysing their mutant phenotypes, expression patterns andphylogeny (Kuusk et al. 2006). By contrast, the researchabout SRS protein is limited in maize. To date, only thefunction of ZmLRP1 has been investigated in detail (Zhanget al. 2015). Maize is one of the important crops in the worldand is valuable for providing forage, food, pharmaceuticals,and other industrial products (Liu et al. 2013). The recentpublication of whole-genome sequences and the massiverelease of transcriptional data provide an opportunity toidentify SRS genes at a genomewide level in maize. In thisstudy, we identified SRS genes in maize and systematicallyanalysed their phylogenetic relationship, chromosome posi-tion and expression level in various tissues. We also studiedthe physical and chemical properties, and conserved domainof SRS proteins. Additionally we carried out gene coexpres-sion network analysis with RNA-sequencing (RNA-seq) dataderived from 12 postflowering leaf tissues (control pollinatedplants and nonpollinated plants) (Sekhon et al. 2012; Stelp-flug et al. 2015) and predicted the putative function of SRSs inmaize. The results will contribute to further investigation ofthe functions of the genes in the maize SRS family.

Materials and methods

Isolation of SRS family genes in maize and theirchromosomal positions

Sequences of 16 Arabidopsis SRS proteins were obtainedfrom the Arabidopsis Information Resource (TAIR; http://www.arabidopsis.org). To acquire all maize SRS genes,BLASTP alignment was carried out in the MaizeGDB(https://www.maizegdb.org/) with the Arabidopsis SRS

protein sequences as seeds. Protein sequences of the putativeSRS family members were downloaded from MaizeGDBwith the e-value \ 1e-10. Each candidate SRS gene wasfurther confirmed using SMART (http://smart.embl-heidelberg.de/) and Pfam (http://pfam.xfam.org). The theo-retical isoelectric point (pI) and molecular weight (MW) ofthe SRS proteins were evaluated using Expasy (http://web.expasy.org/). The physical position, ORF length, full-lengthcDNA and DNA sequence information, were obtained fromthe corresponding GFF format file. The chromosomal posi-tions of SRS genomic data were then visualized usingCIRCOS software (http://circos.ca).

Phylogenetic and gene structural analysis of maize SRSproteins

Six rice SRS proteins, two Physcomitrella patens SRS pro-teins and seven Solanum lycopersicuma SRS proteins wereobtained from the plant transcription factor database (http://planttfdb.cbi.pku.edu.cn/index.php). The amino acidsequences of the SRS proteins of maize, rice, Physcomitrellapatens, Solanum lycopersicuma and Arabidopsis werealigned using ClustalX 2.0 software with the default settings(Thompson et al. 1997). Based on the sequence alignments,an unrooted phylogenetic tree was generated with MEGA Xsoftware (Kumar et al. 2018) using the neighbour-joining(NJ) method and 1000-bootstrap replicates. In addition, aseparate phylogenetic tree was generated for the SRS proteinsequences from maize for further analysis. Besides, thestructures of SRS genes were drawn using GSDS 2.0 (Huet al. 2015) based on the gene annotated file.

Expression profile analysis

To analyse the expression patterns of SRS genes in maize,RNA-seq data of various tissues (NCBI Bioproject ID:PRJNA171684) were downloaded from the NCBI sequenceread archive (SRA) database (https://www.ncbi.nlm.nih.gov/sra/?term=). Expression patterns of SRS genes were dis-played in a heatmap based on transcripts per million (TPM)value using R script (Li and Dewey 2011). The gene coex-pression network was then constructed using the WGCNAprogram (Zhang and Horvath 2005; Langfelder and Horvath2008). For each module, the importance of individual genemembers was quantified by defining module membership asthe correlation between the module eigengene and the geneexpression profile. Lastly, the coexpressed network wasvisualized using the Gephi (Bastian et al. 2009) program.

Enrichment analysis of gene ontology (GO)

Genes were annotated by using the Blastx program againstthe Uniprot protein database (Wu et al. 2006) with a cut-off

3 of 10 Bing He et al.

e-value B 1e-15. GO enrichment analysis was performedusing the singular enrichment analysis (SEA) tool in agriGOv2.0 (Tian et al. 2017). Multiple testing was corrected usingfalse discovery rate (FDR) (Benjamini and Yekutieli 2001),and a cutoff of FDR\ 0.05 was used.

Results

Genomewide identification of SRS family genes in maizeand their chromosomal distribution

A total of 13 putative SRS transcription factors wereidentified in the MaizeGDB database, which were encodedby 11 maize SRS genes (table 1). The amino acid sequencealignment and conserved domain analysis showed that 10of the 13 maize SRS family proteins included both theRING domain and IGGH domain (figure 1).GRMZM2G135783_P01 lacked the IGGH domain, whileAC195955.2_FG006 and AC206191.3_FG001 had onlypartial RING domain. There was significant differences inthe physical and chemical properties of the transcriptionfactors as well. The protein length ranged from 104 aminoacids (aa) to 360 aa, with an average of 285 aa. The MWranged from 10.7 kD to 37.2 kD, with the pI ranging from6.92 to 12.31 (table 1).

The 11 SRS genes were distributed across seven maizechromosomes. Three genes were located on chromosome 6,two on chromosomes 2 and 7, one on chromosomes 1, 4, 5and 8 (table 1; figure 2). No obvious gene cluster was foundon one chromosome. Further, based on an NJ phylogenetictree constructed from the full-length amino acid sequences,the orthologous relationship between the 13 SRS transcrip-tion factors in maize and 16 SRS transcription factors inArabidopsis was investigated (figure 2). No orthologous pairwas found between SRS genes from maize and Arabidopsis,indicating that the two species have significantly divergedduring evolution.

Phylogenetic analysis of SRS gene family

To investigate the phylogenetic relationship among the SRSproteins in various species, a total of 44 SRSs were used toconstruct a NJ phylogenetic tree, including 13 from maize,16 from Arabidopsis, two from Physcomitrella patens, sevenfrom Solanum lycopersicuma and six from rice (figure 3).All of the SRSs were clustered into five subgroups (from I toV). Group I was the largest group containing SRS proteinsfrom all the five species. Group II contained SRS proteins ofmaize and rice, while group III contained maize and Ara-bidopsis SRS proteins. Groups IV and V contained SRSproteins from Arabidopsis and Physcomitrella patens.

We constructed an unrooted phylogenetic tree based onthe maize SRS protein sequences (figure 4a) and analysedthe gene structure based on the available information fromthe maize genome annotation file. The numbers and posi-tions of exons and introns within each SRS gene weredetermined by GSDS 2.0 and displayed in figure 4b. Genestructure varied among SRS genes, with 10 of the 13 tran-scripts having two exons and untranslated regions (UTR).There was a maximum of two introns among SRS genes, andGRMZM2G450459 was the only SRS gene containing threeexons. The gene structure of AC195955.2_FG006 andAC206191.3_FG001 was relatively simple, containing nointron insertion and UTR. However, the gene structure ofSRS genes was highly conserved in the homologous pairsfrom the phylogenetic analysis.

Expression profiles of SRS genes in maize

To understand the putative functions of the maize SRS genes,we analysed the expression profiles of all the maize SRSgenes using publicly available RNA-seq data of the maizeinbred line B73, encompassing 37 tissues. This included 25tissues from different organs and development stages and 12postflowering leaf tissues (Sekhon et al. 2012; Stelpfluget al. 2015). In the maize B73 inbred line, premature

Table 1. SRS gene family in maize.

Gene ID Chromosome location Length (aa) MW (kDa) pI

GRMZM2G042407_T01 Chr1: 207433916–207435929 302 31.52 7.91GRMZM2G042407_T02 Chr1: 207434167–207443393 302 31.52 7.91GRMZM2G080295 Chr2: 196299773–196301206 324 33.08 9.15AC195955.2_FGT006 Chr2: 26998834–26999148 104 10.8 9.26GRMZM2G017606 Chr4: 216534703–216536379 296 30.64 8.08GRMZM2G097683 Chr5: 60161264–60163883 339 33.77 8.77GRMZM2G450459 Chr6: 143217455–143220101 353 36.23 10.16AC206191.3_FGT001 Chr6: 31286644–31286958 104 10.71 9.05GRMZM2G108798 Chr6: 85368896–85371594 331 33.38 8.78GRMZM2G135783_T01 Chr7: 142938485–142940360 348 37.16 12.31GRMZM2G135783_T02 Chr7: 142938485–142940315 321 32.73 8.94GRMZM2G179021 Chr7: 143145425–143146225 221 23.23 6.92GRMZM2G077752 Chr8: 157286610–157288482 360 36.62 7.98

Chr, chromosome; aa, amino acid; MW, molecular weight; pI, isoelectric point.

Coexpression network analysis of SRS genes in maize of 10 3

senescence was induced by preventing pollination (Sekhonet al. 2012). The symptoms of premature senescence firstappeared at 21 days after pollination (DAP), and were welldeveloped by 24 DAP. At 30 DAP, the entire ear leaf bladehad senesced in the plants that were not pollinated, whereasleaves from control plants did not show senescence symp-toms. The 12 postflowering leaf tissues were collected fromcontrol pollinated plants and nonpollinated plants at 0, 6, 12,18, 24 and 30 DAP (Stelpflug et al. 2015).

The gene expression patterns were visualized in the heat-map using R script (figure 4c). We found thatGRMZM2G077752 was highly expressed in roots, stalks(internode), immature leaves, senescent leaves, reproductivetissues (tassels, cobs, silks and anthers), and the embryo.GRMZM2G097683 and GRMZM2G108798 were highlyexpressed in roots, stalks, reproductive tissues and theembryo. However, GRMZM2G179021 had almost noexpression in tested tissues, while transcripts ofAC206191.3_FGT001 and AC195955.2_FGT006 were notdetected in any of these tissues. The rest of the SRS genes wereweakly expressed in roots, stalks, reproductive tissues or theembryo. Notably, GRMZM2G077752 was highly expressedin senescent leaves induced by pollination prevention.

Construction of gene coexpression network

Previous studies have suggested that coexpressed genes areusually members of the same pathway, functionally related orcontrolled by the same transcriptional factor(s) (D’Haeseleeret al. 2000; Ruan et al. 2010; Liao et al. 2011). Therefore,construction of gene coexpression network may help to pre-dict functions of the SRS family genes (Lv et al. 2016). Afterfiltering out all transcripts with low abundance (TPM\ 1),we constructed a weight gene coexpression network (Zhang

and Horvath 2005; Langfelder and Horvath 2008) with RNA-seq data derived from the 12 postflowering leaf tissues (fig-ure 5a). Only one SRS gene, GRMZM2G077752, was inclu-ded in the gene coexpression network. We focussed onanalysing the turquoise module, where GRMZM2G077752was located. The turquoise module was the largest module ofthe network, consisting of 1994 transcripts derived from 1732genes. The expression level of the module remained stable incontrol pollinated plants, whereas it showed sharp activationfrom 24 to 30 DAP in nonpollinated plants (figure 5b).GRMZM2G077752, whose module membership reached0.998 (P = 5.00e-12), belonged to the top 150 hub genes ofthe turquoise module, illustrating that GRMZM2G077752 asone of the central members of the module.

The association between GRMZM2G077752 and othermodule members was calculated with connection weight(Zhang and Horvath 2005), and 1576 significantly correlatedgenes (weight C 0.80) were selected (table 1 in electronicsupplementary material at http://www.ias.ac.in/jgenet/). GOterm analysis (table 2 in electronic supplementary material)showed that these genes were significantly enriched for threebiological processes (FDR\ 0.05), including response toabscisic acid (ABA) stimulus (GO:0009737, FDR = 0.019),carbohydrate metabolic process (GO:0005975, FDR =0.019), and response to stimulus (GO:0050896, FDR =0.029). Specifically, some of the genes in these groups wereinvolved in ABA signalling and remobilization of carbohy-drates. We observed that preventing pollination inducedthree key genes involved in ABA signalling: the proteinphosphatase 2C (PP2C) gene PP2C11, the sucrose nonfer-menting-1-related protein kinase2 (SnRK2) gene SRK2E,and the mitogen-activated protein kinase (MAPK) geneMPK17-2 (Rodriguez et al. 2010; Liu 2012; Zong et al.2016; Wang et al. 2019). The genes involved in carbohy-drate remobilization were related to starch hydrolysis and

Figure 1. Alignment of the RING and IGGH domains in maize SRS proteins. Asterisks indicate the conserved Cys and His zinc ligandresidues in the RING domain. Black box indicated the four IGGH residues in the IGGH domain.


sugar transporters, including three alpha-amylase 1 genes, aswell as bidirectional sugar transporter SWEET genesSWEET13a and SWEET15a (Yang et al. 2001; Eom et al.2015; Thalmann et al. 2016). The five genes were inducedby pollination prevention as well. In addition, two senes-cence-associated genes SAG12 (Liu et al. 2011) were highlyassociated with GRMZM2G077752 and showed strongupregulation during early senescence induced by pollinationprevention. Network connections of GRMZM2G077752 andhighly associated genes were visualized using the Gephiprogram (figure 5c).

Discussion

The SRS gene SHI was first identified in the Arabidopsisdwarf mutant shi in the year 1999 (Fridborg et al. 1999).Since then several SRS genes have been identified across

species, including Arabidopsis (Kuusk et al. 2006), Phys-comitrella patens (Eklund et al. 2010b), Brassica rapa(Hong et al. 2012), maize (Zhang et al. 2015), barley(Youssef et al. 2017) and rice (Duan et al. 2019). However,the genomewide identification of the SRS gene family hasonly been reported in Arabidopsis (Kuusk et al. 2006) thusfar. In this study, a systematical study of the SRS gene familyin maize was performed. We identified 11 SRS genes thatcoded 13 SRS transcription factors (table 1). The 11 SRSgenes were distributed across seven maize chromosomes,and no tandem duplicates were found (figure 2). Products ofmaize SRS genes contained the two conserved domains:RING and IGGH, except AC195955.2_FGT006 andAC206191.3_FGT001, which lacked sequences encodingIGGH domain (figure 1).

The phylogenetic tree of 44 SRS proteins from maize,Arabidopsis, Physcomitrella patens, Solanum lycopersicumaand rice showed that all SRS transcription factors were

Figure 2. Distribution of SRS genes in the genomes of maize and Arabidopsis. The orthologous relationship between maize andArabidopsis SRSs are indicated by red lines.


clustered into five groups (figure 3). Five maize SRS geneswere clustered with Arabidopsis SRS genes in groups I andIII, which indicated that maize SRS genes might participatein regulating organ development. Group II was a monocot-specific group, consisting of maize and rice SRS genes.Groups IVand V were dicot-specific groups, containing onlygenes from Arabidopsis and Solanum lycopersicuma. Thedistribution of SRS genes across maize, Arabidopsis, tomatoand Solanum lycopersicuma may be a result of the differ-ential gene expansion in monocot and dicot plants after theirdivergence from a common ancestor. Further, we constructedan unrooted phylogenetic tree based on the maize SRSprotein sequences and found that the gene structure werehighly similar between SRS genes with strong homology(figure 4, a&b).

SRS genes differed from each other in their expressionpatterns in maize, which suggested that these genes mayparticipate in different biological processes or perform

different biological functions. All the maize SRS genesshowed a tissue-dependent expression (figure 4c), and mostof them were not expressed in leaves. These results areconsistent with previous studies in rice and Arabidopsis.Some Arabidopsis SRSs were highly expressed in flowersand roots but were not expressed in leaves (Kuusk et al.2006; Kim et al. 2010). The high transcript abundance ofrice OsSHI1 was detected in roots and panicles, but not inleaves (Duan et al. 2019).

Noteworthy is that, the GRMZM2G077752 gene washighly expressed in senescent leaves (figure 4c). To inves-tigate the potential function of GRMZM2G077752, weconstructed a gene coexpression network with RNA-se-quencing data derived from ageing leaves induced by pre-venting pollination using WGCNA method. In this method,genes with similar expression profiles were cluster in amodule. It is believed that genes within a module are func-tionally related and that interfacing the module with other

Figure 3. Phylogenetic relationships of SRS transcription factors from maize, Arabidopsis, Physcomitrella patens, Solanumlycopersicuma and rice. The unrooted phylogenetic tree was constructed using MEGA X with the neighbour-joining method. Thebootstrap test was performed with 1000 replicates.


data such as functional annotation and GO allows us topredict the biological function of unknown genes throughknown genes (Zhang and Horvath 2005; Langfelder andHorvath 2008). In our study, GRMZM2G077752 was locatedat the largest module (turquoise module) of the gene coex-pression network (figure 5a). Several lines of evidenceindicated that the module was associated with the senescenceprocess in leaf. In nonpollinated plants, the induced senes-cence first appeared on the leaf at 21 DAP, was welldeveloped by 24 DAP, and spread to the entire leaf blade at30 DAP (Sekhon et al. 2012). The expression level of themodule was upregulated at 24 DAP and further upregulatedat 30 DAP by pollination prevention (figure 5b), which wasconsistent with the development of senescence symptomsinduced by preventing pollination. Additionally, the moduleincluded two orthologues of SAG12, a reliable marker that isinduced during senescence (Lohman et al. 1994) but not dueto stress (Weaver et al. 1998; Miller et al. 1999).

GRMZM2G077752 is one of the central elements of tur-quoise module, indicating that it may play a key role in geneexpression regulation of the module. We then performedgene annotation and GO term analysis to understand thebiological functions involved in the genes that are highly

associated with GRMZM2G077752. Functional enrichmentresults (table 2 in electronic supplementary material) sug-gested associations with response to ABA stimulus (GO:0009737, FDR = 0.019) and carbohydrate metabolic process(GO:0005975, FDR = 0.019). We observed key genesinvolved in ABA signalling and carbohydrate remobiliza-tion, including genes encoding PP2C, SnRK2 and MAPK aswell as SWEETs and alpha-amylase. ABA is known toinduce leaf senescence and premature leaf senescence inplants (Takasaki et al. 2015; Zhao et al. 2016; Mao et al.2017). ABA signalling covers three classes of key proteins(Wang et al. 2019): pyrabactin resistance (PYR)/PYR1-like(PYL)/regulatory component of ABA receptor (RCAR),PP2C and SnRK2. PP2C is a necessary ABA coreceptor.SnRK2 is activated by ABA receptors, and then activates thedownstream substrates, such as ABA responsive elementbinding factors (ABFs), and passes along the ABA signal.Previous studies suggest MAPK is involved in ABA sig-nalling, and might be responsible to convert signals intocellular responses (Liu 2012). SWEETs and alpha-amylaseare considered to be associated with carbohydrate remobi-lization from the leaves (Thalmann et al. 2016; Zhao et al.2017). SWEETs, bidirectional sugar transporters, are vital

Figure 4. Characteristics of maize SRS gene family. (a) The unrooted NJ tree of SRS transcription factors of maize. (b) The exon/introndistribution of maize SRS genes. (c) Expression patterns of SRS genes in maize. The colours represent SRSs expression levels: Log2 (TPM).The phylogenetic relationship is shown on the left.


for pollen nutrition, seed filling and phloem loading (Eomet al. 2015), and alpha -amylase directly degrades starch intoglucose through hydrolytic reactions (Thalmann et al. 2016).Previous studies demonstrated that the expression of bothSWEET genes and alpha-amylase genes is induced by ABA(Chen et al. 2015; Gao et al. 2016; Thalmann et al. 2016). Insummary, the expression level of GRMZM2G077752 and itshighly associated genes increased with the increasing degreeof the ageing leaves, and GRMZM2G077752 coexpressedwith genes encoding ABA signalling proteins and down-stream ABA-responsive genes (figure 5c). These resultsindicated that GRMZM2G077752 might respond to ABAsignal and activate of carbohydrate remobilization duringleaf ageing.

In conclusion, we identified 11 SRS genes with 13 tran-scripts in maize genome. Phylogenetic analysis suggestedthat SRS genes might participate in regulating organ devel-opment in maize. The maize SRS genes exhibited distinctexpression profiles, indicating that they were different inbiological functions. One of the SRS gene,GRMZM2G077752, was highly expressed in senescentleaves, while other maize SRS genes were not expressed in

leaves at all. By gene coexpression network analysis, we findthat GRMZM2G077752 might respond to ABA signal byactivating carbohydrate remobilization in senescent leaves.

Acknowledgements

This work was financially supported by the National Natural Sci-ence Foundation of China [31771813], JAAS Exploratory andDisruptive Innovation Program [ZX (17)2015].

References

Bastian M., Heymann S. and Jacomy M. 2009 Gephi: an opensource software for exploring and manipulating networks.International AAAI Conference on Weblogs and Social Media.

Baylis T., Cierlik I., Sundberg E. and Mattsson J. 2013 Shortinternodes/stylish genes, regulators of auxin biosynthesis, areinvolved in leaf vein development in Arabidopsis thaliana. NewPhytol. 197, 737–750.

Benjamini Y. and Yekutieli D. 2001 The control of the falsediscovery rate in multiple testing under dependency. Ann. Statist.29, 1165–1188.

Figure 5. Network analysis of the function of SRS gene family in maize. (a) Hierarchical cluster tree showing coexpression modulesidentified using WGCNA. Modules corresponded to branches and were labelled by colours as indicated by the colour band underneath thetree. (b) Boxplots shows the distribution of module expression (mean Z-score of all genes within the turquoise module) at different timepoints in control pollinated and nonpollinated plants. (c) Visualization of network connections between GRMZM2G077752 and highlyassociated genes. Highly associated genes are highlighted in orange (for ABA signalling), blue (for carbon remobilization) and purple (forleaf senescence).


Chen L. Q., Lin I. W., Qu X. Q., Sosso D., McFarlane H. E. andLondono A. 2015 A cascade of sequentially expressed sucrosetransporters in the seed coat and endosperm provides nutritionfor the Arabidopsis embryo. Plant Cell 27, 607–619.

D’haeseleer P., Liang S. and Somogyi R. 2000 Genetic networkinference: from co-expression clustering to reverse engineering.Bioinformatics 16, 707–726.

Duan E., Wang Y., Li X., Lin Q., Zhang T., Wang Y. et al. 2019OsSHI1 regulates plant architecture through modulating thetranscriptional activity of IPA1 in rice. Plant Cell 31, 1026–1042.

Eklund D. M., Staldal V., Valsecchi I., Cierlik I., Eriksson C.,Hiratsu K. et al. 2010a The Arabidopsis thaliana STYLISH1protein acts as a transcriptional activator regulating auxinbiosynthesis. Plant Cell 22, 349–363.

Eklund D. M., Thelander M., Landberg K., Staldal V., Nilsson A.,Johansson M. et al. 2010b Homologues of the Arabidopsisthaliana SHI/STY/LRP1 genes control auxin biosynthesis andaffect growth and development in the moss Physcomitrellapatens. Development 137, 1275–1284.

Eom J. S., Chen L. Q., Sosso D., Julius B. T., Lin I. W., Qu X. Q.et al. 2015 SWEETs, transporters for intracellular and intercel-lular sugar translocation. Curr. Opin. Plant Biol. 25, 53–62.

Fridborg I., Kuusk S., Moritz T. and Sundberg E. 1999 TheArabidopsis dwarf mutant shi exhibits reduced gibberellinresponses conferred by overexpression of a new putative zincfinger protein. Plant Cell 11, 1019–1031.

Fridborg I., Kuusk S., Robertson M. and Sundberg E. 2001 TheArabidopsis protein SHI represses gibberellin responses inArabidopsis and barley. Plant Physiol. 127, 937–948.

Gao S., Gao J., Zhu X., Song Y., Li Z., Ren G. et al. 2016 ABF2,ABF3, and ABF4 promote ABA-mediated chlorophyll degrada-tion and leaf senescence by transcriptional activation of chloro-phyll catabolic genes and senescence-associated genes inArabidopsis. Mol. Plant 9, 1272–1285.

Gomariz-Fernandez A., Sanchez-Gerschon V., Fourquin C. andFerrandiz C. 2017 The role of SHI/STY/SRS genes in organgrowth and carpel development is conserved in the distanteudicot species Arabidopsis thaliana and Nicotiana benthami-ana. Front. Plant Sci. 8, 814.

Hong J. K., Kim J. A., Kim J. S., Lee S. I., Koo B. S. and Lee Y. H.2012 Overexpression of Brassica rapa SHI-RELATEDSEQUENCE genes suppresses growth and development inArabidopsis thaliana. Biotechnol. Lett. 34, 1561–1569.

Hu B., Jin J., Guo A. Y., Zhang H., Luo J. and Gao G. 2015 GSDS2.0: an upgraded gene feature visualization server. Bioinformat-ics 31, 1296–1297.

Kim S. G., Lee S., Kim Y. S., Yun D. J., Woo J. C. and Park C. M.2010 Activation tagging of an Arabidopsis SHI-RELATEDSEQUENCE gene produces abnormal anther dehiscence andfloral development. Plant Mol. Biol. 74, 337–351.

Kumar S., Stecher G., Li M., Knyaz C. and Tamura K. 2018MEGA X: Molecular evolutionary genetics analysis acrosscomputing platforms. Mol. Biol. Evol. 35, 1547–1549.

Kuusk S., Sohlberg J. J., Long J. A., Fridborg I. and Sundberg E.2002 STY1 and STY2 promote the formation of apical tissuesduring Arabidopsis gynoecium development. Development 129,4707–4717.

Kuusk S., Sohlberg J. J., Magnus Eklund D. and Sundberg E. 2006Functionally redundant SHI family genes regulate Arabidopsisgynoecium development in a dose-dependent manner. Plant J.47, 99–111.

Langfelder P. and Horvath S. 2008 WGCNA: an R package forweighted correlation network analysis. BMC Bioinformatics 9,559.

Li B. and Dewey C. N. 2011 RSEM: accurate transcript quantifi-cation from RNA-Seq data with or without a reference genome.BMC Bioinformatics 12, 323.

Liao Q., Liu C., Yuan X., Kang S., Miao R., Xiao H. et al. 2011Large-scale prediction of long non-coding RNA functions in acoding–non-coding gene co-expression network. Nucleic AcidsRes. 39, 3864–3878.

Liu X., Li Z., Jiang Z., Zhao Y., Peng J., Jin J. et al. 2011 LSD: aleaf senescence database. Nucleic Acids Res. 39, D1103–D1107.

Liu Y. 2012 Roles of mitogen-activated protein kinase cascades inABA signaling. Plant Cell Rep. 31, 1–12.

Liu S., Wang X., Wang H., Xin H., Yang X., Yan J. et al. 2013Genome-wide analysis of ZmDREB genes and their associationwith natural variation in drought tolerance at seedling stage ofZea mayz L. PLoS Genet. 9, e1003790.

Lohman K. N., Gan S., John M. C. and Amasino R. M. 1994Molecular analysis of natural leaf senescence in Arabidopsisthaliana. Physiol. Plant. 92, 322–328.

Lv Y., Liang Z., Ge M., Qi W., Zhang T., Lin F. et al. 2016Genome-wide identification and functional prediction of nitro-gen-responsive intergenic and intronic long non-coding RNAs inmaize (Zea mays L.). BMC Genomics 17, 350.

Mao C., Lu S., Lv B., Zhang B., Shen J., He J. et al. 2017 A riceNAC transcription factor promotes leaf senescence via ABAbiosynthesis. Plant Physiol. 174, 1747–1763.

Miller J. D., Arteca R. N. and Pell E. J. 1999 Senescence-associatedgene expression during ozone-induced leaf senescence inArabidopsis. Plant Physiol. 120, 1015–1024.

Rodriguez M. C., Petersen M. and Mundy J. 2010 Mitogen-activated protein kinase signaling in plants. Annu. Rev. PlantBiol. 61, 621–649.

Ruan J., Dean A. K. and Zhang W. 2010 A general co-expressionnetwork-based approach to gene expression analysis: comparisonand applications. BMC Syst. Biol. 4, 8.

Sekhon R. S., Childs K. L., Santoro N., Foster C. E., Buell C. R., deLeon N. et al. 2012 Transcriptional and metabolic analysis ofsenescence induced by preventing pollination in maize. PlantPhysiol. 159, 1730–1744.

Smith D. L. and Fedoroff N. V. 1995 LRP1, a gene expressed inlateral and adventitious root primordia of Arabidopsis. Plant Cell7, 735–745.

Sohlberg J. J., Myrenas M., Kuusk S., Lagercrantz U., KowalczykM., Sandberg G. et al. 2006 STY1 regulates auxin homeostasisand affects apical-basal patterning of the Arabidopsis gynoe-cium. Plant J. 47, 112–123.

Stelpflug S. C., Sekhon R. S., Vaillancourt B., Hirsch C. N., BuellC. R., de Leon N. et al. 2015 An expanded maize geneexpression atlas based on RNA sequencing and its use to exploreroot development. Plant Genome 9, 1.

Takasaki H., Maruyama K., Takahashi F., Fujita M., Yoshida T.,Nakashima K. et al. 2015 SNAC-As, stress-responsive NACtranscription factors, mediate ABA-inducible leaf senescence.Plant J. 84, 1114–1123.

Thalmann M., Pazmino D., Seung D., Horrer D., Nigro A., MeierT. et al. 2016 Regulation of leaf starch degradation by abscisicacid is important for osmotic stress tolerance in plants. Plant Cell28, 1860–1878.

Thompson J. D., Gibson T. J., Plewniak F., Jeanmougin F. andHiggins D. G. 1997 The CLUSTAL_X windows interface:flexible strategies for multiple sequence alignment aided byquality analysis tools. Nucleic Acids Res. 25, 4876–4882.

Tian T., Liu Y., Yan H., You Q., Yi X., Du Z. et al. 2017 agriGOv2.0: a GO analysis toolkit for the agricultural community.Nucleic Acids Res. 45, W122–W129.

Wang X., Guo C., Peng J., Li C., Wan F, Zhang S. et al. 2019ABRE-BINDING FACTORS play a role in the feedbackregulation of ABA signaling by mediating rapid ABAinduction of ABA co-receptor genes. New Phytol. 221,341–355.


Weaver L. M., Gan S., Quirino B. and Amasino R. M. 1998 Acomparison of the expression patterns of several senescence-associated genes in response to stress and hormone treatment.Plant Mol. Biol. 37, 455–469.

Wu C. H., Apweiler R., Bairoch A., Natale D. A., Barker W. C.,Boeckmann B. et al. 2006 The Universal Protein Resource(UniProt): an expanding universe of protein information. NucleicAcids Res. 34, 187–191.

Yang J., Zhang J., Wang Z. and Zhu Q. 2001 Activities of starchhydrolytic enzymes and sucrose-phosphate synthase in the stemsof rice subjected to water stress during grain filling. J. Exp. Bot.52, 2169–2179.

Youssef H. M., Eggert K., Koppolu R., Alqudah A. M., Poursare-bani N., Fazeli A. et al. 2017 VRS2 regulates hormone-mediatedinflorescence patterning in barley. Nat. Genet. 49, 157–161.

Yuan T. T., Xu H. H., Zhang Q., Zhang L. Y. and Lu Y. T. 2018 TheCOP1 target SHI-RELATED SEQUENCE5 directly activatesphotomorphogenesis-promoting genes. Plant Cell 30,2368–2382.

Zhang B. and Horvath S. 2005 A general framework for weightedgene co-expression network analysis. Stat. Appl. Genet. Mol.Biol. 4, 17.

Zhang Y., von Behrens I., Zimmermann R., Ludwig Y., Hey S. andHochholdinger F. 2015 LATERAL ROOT PRIMORDIA 1 ofmaize acts as a transcriptional activator in auxin signallingdownstream of the Aux/IAA gene rootless with unde-tectable meristem 1. J. Exp. Bot. 66, 3855–3863.

Zhao Y., Chan Z., Gao J., Xing L., Cao M., Yu C. et al. 2016 ABAreceptor PYL9 promotes drought resistance and leaf senescence.Proc. Natl. Acad. Sci. USA 113, 1949–1954.

Zhao Y., Gao J., Kim J. I., Chen K., Bressan R. A. and Zhu J. K.2017 Control of plant water use by ABA induction of senescenceand dormancy: an overlooked lesson from evolution. Plant CellPhysiol. 58, 1319–1327.

Zong W., Tang N., Yang J., Peng L., Ma S. and Xu Y. 2016Feedback regulation of ABA signaling and biosynthesis by abZIP transcription factor targets drought resistance related genes.Plant Physiol. 171, 2810–2825.

Corresponding editor: H. A. RANGANATH


Gene coexpression network analysis reveals the role of SRS ...

Documents

Transcript of Gene coexpression network analysis reveals the role of SRS ...