Transcriptome characterization and expression profiles of

Gene xxx (2015) xxx–xxx

GENE-40947; No. of pages: 9; 4C:

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Research paper

Transcriptome characterization and expression profiles of the relateddefense genes in postharvest mango fruit againstColletotrichum gloeosporioides

Keqian Hong ⁎, Deqiang Gong, Lubin Zhang, Huigang Hu, Zhiwei Jia, Hui Gu, Kanghua SongKey Laboratory for Postharvest Physiology and Technology of Tropical Horticultural Products of Hainan Province, South Subtropical Crop Research Institute, Chinese Academy of Tropical Agricul-tural Sciences, Zhanjiang, Guangdong, China

Abbreviation: C. gloeosporioides, Colletotrichum gloeospofactor; qRT-PCR, quantitative real-time PCR; RT-PCR, reversamplification of cDNA ends; TF, transcription factor; PRgenes; RCT1, resistance to C. trifolii race 1; Gb, Gigabase pBiotechnology Information; Nr, National Center for Biredundant protein; KEGG, Kyoto Encyclopedia of Genes anRACE, rapid amplification of Cdna.⁎ Corresponding author

E-mail address: [email protected] (K. Hong).

http://dx.doi.org/10.1016/j.gene.2015.10.0410378-1119/© 2015 Published by Elsevier B.V.

Please cite this article as: Hong, K., et al., Trmango fruit against Colletotrichum gloeospor

a b s t r a c t
a r t i c l e i n f o
Article history:Received 7 April 2015Received in revised form 15 September 2015Accepted 13 October 2015Available online xxxx

Keywords:Mango fruitColletotrichum gloeosporioidesTranscriptomePlant defense

Anthracnose, caused by Colletotrichum gloeosporioides, is a major disease of the postharvest mango (Mangiferaindica L.) fruit. However, a lack of transcriptomic and genomic information hinders our understanding of themolecular mechanisms underlying the mango fruit defense response. Here, we studied the host responses ofthe mango fruit against C. gloeosporioides using Illumina paired-end sequencing technology, and expressionprofiles of 35 defense-related genes were further analyzed by qRT-PCR. The results indicated that 5.9 Gigabasepair clean reads were assembled into a total of 131,750 unigenes, of which 89,050 unigenes found to be homol-ogous to genes in the NCBI GenBank database and 61,694 unigenes annonated in the Swiss-Prot database.Orthologous analyses showed that 47,770 unigenes were assigned with one or more Gene Ontology terms and44,145 unigenes were classified into 256 Kyoto Encyclopedia of Genes and Genomes pathways. Moreover,qRT-PCR of 35 defense-related unigenes, including 17 ethylene response factors (ERFs), 6 nucleotide bindingsite-leucine-rich repeats (NBS-LRRs), 6 nonexpressor of pathogenesis-related genes (NPRs) and 6 pathogenesis-related protein (PRs), revealed that most of these defense-related genes were up-regulated after C. gloeosporioidesinfection. Taken together, our study provides a platform to discover newcandidate genes inmango fruit in relationto pathogen resistance.

© 2015 Published by Elsevier B.V.

1. Introduction

Mango (Magnifera indica L.) is a popular fruit in China. Anthracnose,caused by the fungus Colletotrichum gloeosporioides, is a destructivedisease of mango and widely distributed in all mango-growing regionsof theWorld (Ploetz and Prakash, 1997; Chowdhury et al., 2008). Infect-ed blossom or young fruit can result in failure to produce fruit, but themost damaging phase of the disease begins as a latent infection, whenthe fruit is in thepre-climacteric phase of development. Therefore, growthof the pathogen is resumed only after harvest when the fruit starts toripen and postharvest anthracnose develops (Estrada et al., 2000). Thedisease has also caused huge losses in the harvested mango fruits inmany countries, including China (Zeng et al., 2006).

rioides; ERF, ethylene responsivee transcription PCR; RACE, rapid, pathogenesis-related proteinairs; NCBI, National Center for

otechnology Information non-d Genomes; GO, Gene Ontology;

anscriptome characterizationioides, Gene (2015), http://dx

The harvested mango fruit infection with C. gloeosporioides oftenshows a characteristic of sunken, prominent, dark brown to blackdecay spots which can usually coalesce, and eventually penetrate intothe fruit, resulting in extensive fruit rot (Nelson, 2008). Postharvestmanagement strategies to control the disease include storage at lowtemperatures, hotwater dip, vapor heat, heated fungicide dips, biologicalcontrol, pre-harvest salicylic acid, and potassium phosphonate dip(Nelson, 2008; Kefialew and Ayalew, 2008; Dessalegn et al., 2013).These treatment measures could limit or prevent the development andspread of the disease to some extent, thus a better understanding ofthe host responses to pathogen infection is essential to clarify mecha-nisms of plant-microbe interactions and to develop novel and effectivestrategies for anthracnose control.

Based on molecular studies, a large number of genes related todefense response have been identified. For example, RCT1 (resistanceto Colletotrichum trifolii race 1) gene confers resistance to anthracnoseinfection in alfalfa (Yang et al., 2008). Studies have been also conductedto identify some key genes induced by anthracnose pathogen in maize(Vargas et al., 2012). Though the transcriptome analysis was carriedout in the mango fruit during different developmental stages (Wuet al., 2014), there was still lack of enough data available for genesinvolved in the regulation of defense response in the mango fruit. Re-cently, transcriptome profiling of other fruits, such as citrus (Martinelli

and expression profiles of the related defense genes in postharvest.doi.org/10.1016/j.gene.2015.10.041

http://dx.doi.org/10.1016/j.gene.2015.10.041

mailto:[email protected]


http://www.sciencedirect.com/science/journal/03781119

www.elsevier.com/locate/gene


2 K. Hong et al. / Gene xxx (2015) xxx–xxx

et al., 2012) and tomato (Alkan et al., 2015) infected by the different plantpathogens (Candidatus Liberibacter asiaticus and C. gloeosporioides) havebeen characterized. In this study, to elucidate molecular mechanism ofthemango fruit against C. gloeosporioides, we constructed a transcriptomeofmango fruit infectedwith C. gloeosporioides. Nearly 63million reads of atotal of 5.9 billion nucleotides were assembled into 131,750 unigenes. Ofthose unigenes, 89,050 (67.6%) matched the known genes in a BLASTsearch of the NCBI database. Matches included a number of genes relatedto pathogen resistance. These assembled, annonated transcriptomeextended the genomic resources available for researchers studyingon the mango fruit–C. gloeosporioides interaction, and may providea fast approach to identifying candidate genes involved in pathogenresistance. Further, this information will be helpful for improvingcurrent understanding of host–pathogen interaction.

2. Materials and methods

2.1. Fruit and fungal materials

Pre-climacteric mango (Mangifera indica L. cv. Zill) fruits at approxi-mately 80% full stage (about 100 days after anthesis) were obtainedfrom a commercial orchard in South Subtropical Crop Research Institute,and transported to the laboratory within 1 h. Fruits were selected onthe basis of size and color uniformity, and blemished as well as diseasedfruit were discarded. The selected fruits were rinsed twice in tap water.And then, the fruits were allowed to air-dry at 25 °C for 2 h.

C. gloeosporioides was obtained from a decayed mango fruit andcultured on potato dextrose agar (PDA, containing 200 g diced potatoes,20.0 g sucrose, 15.0 g agar and 1.0 L distilled water) for 5 days at 28 °C.The culture plates were flooded with distilled water and the sporesuspensions obtained were adjusted to 105 spores/mL using ahemocytometer.

2.2. Treatment

Mango fruits (200) were divided at random into two groups, eachgroup consisting of 100 fruits. All selected fruits were wounded ateach end using a sterile needle as described by Tang et al. (2010) andeach wound was 1.5 mm in depth. Fruits in Group 1 were inoculatedwith 20 μL (1.0 × 105 spores/mL) of C. gloeosporioides spores in suspen-sion. Fruits in Group 2 (control) were inoculated with 20 μL distilledwater. All fruits (Groups 1 and 2)were subsequently placed into unsealedpolyethylene plastic bags (0.04mm thickness) and stored at 28 °C, with arelative humidity of 90% for 9 days. Each time, mango peel of 30 mmaround the inoculation point from five mangoes was collected for eachsample at regular intervals.

2.3. Plant materials for transcriptome analysis

To obtain an overview of the transcriptome of the postharvestmango fruits infected with C. gloeosporioides, peels of mango fruits inGroup 1 were taken from 30 mm around the inoculation point, whichdid not contain visually discernible fungal hyphae, at various timesafter inoculation: 2 h, 6 h, 1 d, 3 d and 5 d. The RNA from each peelsamplewas isolated, andmixed in equal proportion to generatematerialfor construction of one cDNA library.

2.4. Plant materials for quantitative real-time PCR (qRT-PCR) analysis andmeasurement of disease development

Peels of mango fruits in Groups 1 and 2 were used to investigate theexpression profiles of the related candidate genes in response to C.gloeosporioides.

The lesion diameter (mm) was recorded on each fruit inoculatedwith C. gloeosporioides at regular intervals during 9 days of storage at28 °C.

Please cite this article as: Hong, K., et al., Transcriptome characterizationmango fruit against Colletotrichum gloeosporioides, Gene (2015), http://dx

2.5. RNA extraction, library construction and sequencing

Total RNA was extracted from the mango peel using the hot boratemethod of Wan and Wilkin (1994) and quality checked with Agilent2100 Bioanalyzer (Agilent, USA). Equal amounts of RNA in the mangofruits infected with C. gloeosporioides at various times point: 2 h, 6 h,1 d, 3 d and 5 d were mixed to use for constructing the cDNA library.The process of cDNA library construction was carried out as describedby Zhang et al. (2013a). The cDNA library was sequenced on cBotusing HiSeq 2000 platform (illumina) with 100 bp paired-end readsby CapitalBio Corporation (Beijing, China). Illumina Casava (version1.7) was performed for primary data analysis and base calling.

2.6. De novo assembly and functional annotation

The sequencing and assemblywere performed by CapitalBio Corpora-tion (Beijing, China) using HiSeq 2000 platform (Hayward, CA, USA).Adapter sequences, empty reads and low quality sequences were filteredusing Perl program. After trimming adaptors, 63,242,630 reads withquality Q-value over 20 and sequence length longer than 100 bp wereassembledwith Trinity software (Grabherr et al., 2011). Thedetail processof de novo assembly was described in the paper of Li et al. (2013).

All of the assembled sequences were then searched against severalprotein databases, such as the Nr using the BLASTx program (E-value b

0.00001), the Swiss-Prot protein database (http://www.expasy.ch/sprot), KEGG database (Kanehisa et al., 2006) and the COG database(http://www.ncbi.nih.gov/COG). Functional annotations of the unigenesby GO were carried out according to the method described by Harriset al. (2004), followed by functional classifications using the WEGO soft-ware (Ye et al., 2006) to identify GO terms and view the distribution ofthe unigenes in M. indica in biological process and molecular functionand cellular component.

2.7. Experimental validation

2.7.1. Reverse transcription PCR (RT-PCR) confirmationWe randomly selected 17 putative ethylene response factors (ERFs)

from the RNA-seq to check the validity of the sequences of unigenes. Toobtain full-length cDNAs of these genes, we isolated the remaining 3′-and 5′-cDNA sequences by rapid amplification of cDNA ends (RACE)using cDNA amplification kits (Takara, Shiga, Japan) according to themanufacturer's protocol. Once the sequence of the 3′- and 5′-ends ofeach cDNA had been determined, full-length cDNAs for the ERFs in themango fruit were isolated by RT-PCR.

2.7.2. Sequence confirmationAfter 17 ERFs full-length cDNAs were obtained, their open reading

frames and proteins prediction were performed using NCBI ORF Finder.Alignments of the ERFs with other plants, such as Arabidopsis, tomatoand tobacco, were carried out on DNAMAN software to search theirdomains.

2.8. qRT-PCR analysis

ERF, nucleotide binding site-leucine-rich repeat (NBS-LRR),nonexpressor of pathogenesis-related gene (NPR) and pathogenesis-related protein gene (PR) play an important role in the plant responseto pathogen. 17, 6, 6 and 6 unigenes encoding ERFs, NBS-LRRs, NPRsand PRs, respectively, were randomly selected from the RNA-seq data-base, to explore their expression patterns during infection of themango fruit by C. gloeosporioides using qRT-PCR. Total RNAwas isolatedaccording to the method of Wan and Wilkin (1994) and subsequentlypurified usingDNAse I (Promega,Madison,WI, USA). Reverse transcrip-tion reactionswere performed using PrimerScript™ RT reagent Kit withgDNA Eraser (Takara, Shiga, Japan) according to the manufacturer'sinstructions. The sequences of all primers used are listed in Supplemental


http://www.expasy.ch/sprot

http://www.expasy.ch/sprot

http://www.ncbi.nih.gov/COG


Fig. 1.Unigene size distribution. The left y-axis indicates the number of unigenes to differentunigene length scales, and the right y-axis indicates their percentages.

Table 2Gene Ontology classification of assembled unigenes.

Ontology Class Number of unigene

Biological process Anatomical structure formation 542Biological adhesion 74Biological regulation 4472Cellular component biogenesis 685Cellular component organization 1105Cellular process 22,726Death 56Developmental process 48Establishment of localization 5227Growth 1Immune system process 10Localization 5306Locomotion 1Metabolic process 23,395Multi\-organism process 112Multicellular organismal process 95

3K. Hong et al. / Gene xxx (2015) xxx–xxx

file 1. Reactionswere carried out using SYBR®premix Ex Taq™ II (Takara,Shiga, Japan) on a LightCycler®480 II (Roche, Switzerland) with threereplicates, and Miactin (accession number HQ585999.1) was used as areference gene. All qRT-PCR reactions were normalized using the Ctvalue corresponding to the reference gene. The relative expression levelsof the twenty-three genes were calculated with the formula 2−ΔΔCT

(Livak and Schmittgen, 2001).

3. Results

3.1. Transcriptome generation and assembly

In order to acquire maximum information from RNA-seq data, weisolated RNA frommango fruits infected by C. gloeosporioides at varioustime points: 2 h, 6 h, 1 d, 3 d and 5 d, and mixing these RNAs at equalratio, and constructed one cDNA library. A total of 67,012,956 rawreads with an average length of 90 bp were obtained by Hi-Seq™2000 (Illumina) paired-end sequencing of the mango fruit peel. Thetotal length of the reads was approximately 6.7 Gigabase pairs (Gb).After a stringent filtering process, 63,242,630 higher quality cleanreads (about 5.9 Gb, 87.7% of the raw reads) remained (Table 1). TheQ20 percentage (sequencing error rate b 1%), and GC% were 96.1% and46.1%, respectively. This suggested that the sequencing output andquality were good enough for further analysis.

Based on the clean reads, a total of 131,750 unigenes, ranging from201 bp to 21,184 bp were assembled by the Trinity program(Grabherr et al., 2011). The average length, and N50 of these unigeneswere 1369 bp, and 2316 bp, respectively (Table 1). As shown in Fig. 1,of these, 70,284 unigenes (58.8%) were longer than 500 bp, and 50,725unigenes (42.5%) were longer than 1000 bp. Moreover, 56,137,270(88.8%) clean reads were successfully mapped to the transcriptsequences.

3.2. Sequence annotation

Several complementary approaches were utilized to annotatethe assembled unigenes. All unigenes were searched against the Nr,Swiss-Prot, Nt, KEGG, COG and InterPro databases. A total of 93,325(70.8% of all unigenes) unigenes were annotated (Table 1), where atthe Nr database 89,050 (67.6% of all unigenes) unigeneswere annotated,having the largest match, followed by KEGG (87,059, 66.1%), Nt (84,787,

Table 1Summary of transcriptome data for mango (Mangifera indica L. cv. Zill) fruit as well asdetailed bioinformatics annotations and analyses.

1. Raw sequences and Assembly statisticsRaw reads (paired-end) 67,012,956clean reads (paired-end) 63,242,630GC content percentage 46.1%Total unigenes (average length; N50; min–max length) 131,750 (1369; 2316;

201–21,184)Mean RPKM value of unigenes (min–mix RPKM value) 7.59 (0–81,981)

2. Bioinformatics annotations of mango fruit unigenesProtein coding sequence (CDS) 113,874Gene annotation against Nr (%) 89,050 (67.6)Gene annotation against Swiss-Prot (%) 61,694 (46.8)Gene annotation against Nt (%) 84,787 (64.4)Gene annotation against KEGG (%) 87,059 (66.1)Gene annotation against COG (%) 35,115 (26.7)Gene annotation against InterPro (%) 56,569 (42.9)All unigenes annotated (%) 93,325 (70.8)GO Ontology (%) 47,770 (36.3); 1738 GO

termsBiological process category 30,231; 671 GO termsCellular component category 13,593; 206 GO termsMolecular function category 44,319; 915 GO terms


64.4%), Swiss-Prot (61,694, 46.8%), InterPro (56,569, 42.9%) and COG(35,115, 26.7%) databases. And the rest (38,425, 29.2%) not annotatedto the existing databases was the potential sources of novel genes.

Pigmentation 4258Reproduction 107Reproductive process 105Response to stimulus 1594Viral reproduction 4

Cellular component Cell 13,499Cell part 13,499Envelope 589Extracellular region 60Extracellular region part 5Macromolecular complex 3490Membrane\-enclosed lumen 248Organelle 5350Organelle part 2128Virion 2Virion part 2

Molecular function Antioxidant activity 152Binding 31,455Catalytic activity 23,617Electron carrier activity 624Enzyme regulator activity 628Metallochaperone activity 3Molecular transducer activity 656Nutrient reservoir activity 11Structural molecule activity 1067Transcription regulator activity 1450Translation regulator activity 159Transporter activity 3417

Note: total 131,750 unigenes were categorized into the three main categories: biologicalprocess, cellular component and molecular function.



Fig. 2. COG function classification of the mango fruit transcriptome.


3.3. GO and COG classifications

Based on the protein annotation results from the Nr databasehomology search, 89,050 unigenes (biological process: 30,231 unigenes;molecular function: 44,319 unigenes; cellular component: 13,593unigenes) could be categorized into 44 functional groups (Table 2).Highly represented biological processes were metabolic, cellular, locali-zation, establishment of localization, biological regulation, pigmentationand response to stimulus, where each ontology was represented bymore than 1500 unigenes. From themolecular function ontology analysis,a significant large number of unigenes belonged to binding (31,455unigenes) and catalytic activity (23,617 unigenes) categories. For thecellular component category, cell and cell part represented the majorproportion. As for GO terms, 671 were related to biological processes,206 to cellular components, and 915 to molecular function (Table 1),indicating a large functional diversity of genes in the trancriptomic data.

As shown in Fig. 2, COG classification based on BLAST search againstCOG database resulted in 44,145 unigenes categorized into 24 categorieswith the largest number of unigenes coming under ‘general functionprediction (8126, 27.5%)’, followed by ‘posttranslational modification,protein turnover, chaperones (2439, 8.26%)’, ‘translation, ribosomal struc-ture and biogenesis (1828, 6.19%)’, ‘replication, recombination and repair(1637, 5.55%)’. The functional category ‘nuclear structure (80, 0.27%)’and ‘cell motility (10, 0.03%)’ had the least representation in the wholeRNA-seq data. Among the COG, lots of categories, such asMAPK signalingpathway, carbohydrate transport and metabolism, cell wall (membraneor envelope) biogenesis, transcription factors, plant hormone signal trans-duction, and phenylpropanoid biosynthesis, have been proved beinginvolved in plants against the pathogen infection (Martinelli et al., 2012;Zhuang et al., 2012; Alkan et al., 2015).

Fig. 3. Anthracnose symptoms (A) and lesion diameter (B) of post-harvest mango fruitinoculated with C. gloeosporioides and stored at 28 °C for 9 days. Each value representsthe mean ± SE of three replicates.

3.4. Confirm validity of sequences by RNA-seq

To assess the validity of the sequences from RNA-seq databank, 17ERFs full-length cDNAs in the mango fruit were isolated, and were


predicted to encode proteins of 240, 194, 293, 187, 196, 190, 312, 170,218, 253, 247, 338, 222, 261, 374, 314 and 398 amino acids, respectively,as the unigene order described in Table 1. The deduced amino acidsequences of ERFs comprise a conserved DNA-binding ERF/AP2 domain(ranging from 58 to 59 amino acids) (Supplemental file 2), which ischaracteristic of the plant ERF gene family (Liu et al., 2010). The resultof assignment showed that the 17 ERFs share a highly conserved ERF/AP2 domain with ERF proteins of other plants, including ArabidopsisAtERF1, tomato LeERF1 and tobacco NtERF32, proving that the 17 ERFs



Fig. 4. qRT-PCR analysis of 17 randomly ethylene response factor unigenes in postharvest fruits from C. gloeosporioides inoculated (white bars) and control (gray bars). The expressionlevels of each unigene are expressed as a ratio relative to 0 d of samples, whichwas set at 1. Error bars indicate standard errors of themeans (n=3). Barswith different lower-case lettersindicate significant differences based on a t-test at the P ≤ 0.05 level.


encode the ERF family TF in the mango fruit, and the cDNA sequencesobtained from the RNA-seq are credible.

3.5. Symptoms analysis of fruits infected with C. gloeosporioides

The postharvestmango fruits infectedwith C. gloeosporioides began toshow disease symptoms on day 3 of storage, and symptoms gradually


increased during incubation (Fig. 3), indicating that course of the diseasedeveloped quickly once fruits infected with C. gloeosporioides.

3.6. Expression profilings of ERF, NBS-LRR, NPR and PR genes in fruits infectedwith C. gloeosporioides

Expression levels ofmost selected ERF unigenes (1.5- to 85-fold)wereup-regulated in infected fruits. Down-regulation in the expression of



Fig. 5. qRT-PCR analysis of 6 randomly nucleotide binding site-leucine-rich repeat unigenes in postharvest fruits from C. gloeosporioides inoculated (white bars) and control (gray bars).The expression levels of each unigene are expressed as a ratio relative to 0 d of samples, which was set at 1. Error bars indicate standard errors of the means (n = 3). Bars with differentlower-case letters indicate significant differences based on a t-test at the P ≤ 0.05 level.


a minority of unigenes for comp26782_c1_seq1 and comp30115_c0_seq2 (1.5- to 2.0-fold) occurred under C. gloeosporioides infection.C. gloeosporioides had little effect on the expression levels of unigenesof comp22021_c0_seq1 and comp24119_c0_seq1 (Fig. 4).

Transcripts ofmost selectedNBS-LRR unigenes (1.2- to 20-fold)weremore abundant in infected fruits, but transcript abundance of unigenefor comp31966_c0_seq4 (7- to 16-fold) alone was lower in infectedfruits (Fig. 5).

Among the selected unigenes encoding NPRs, five of them (1.4- to126-fold) were increased whereas only unigenes for comp31345_c1_seq1 (1.3- to 1.5-fold) decreased in infected fruits (Fig. 6).

Expression levels of all the selected unigenes encoding PRs (1.4- to368-fold) were up-regulated in infected fruits (Fig. 7).

4. Discussion

Previous plant biology studies have used transcriptome analysis toinvestigate the molecular mechanisms of fruit response to pathogenattack. However, there have been no reports of transcriptome profilingof mango fruit infected by a pathogen. Recently, Alkan et al. (2015)


have provided a framework for concurrent analysis of fungal-fruitinteraction by transcriptome of tomato fruit infection with C.gloeosporioides. To obtain a maximally informative transcriptome re-source of mango fruit in response to the pathogen, cDNA samples reversetranscription from the different time points of mRNA samples that werepooled from mango fruit infected by C. gloeosporioides were normalizedprior to sequence analysis. In the present study, more than 63 millionhigh quality reads with 96.1% Q20 bases were generated from mangofruit using the Illumina paired-end sequencing technology. This largenumber of reads with paired-end information produced much longerunigene (mean: 1369 bp) (Table 1) than those in previous studies, forexample, with Populus diversifolia (671 bp) (Qiu et al., 2011), persimmon(579 bp) (Luo et al., 2014) and litchi (687 bp) (Zhang et al., 2014).

The sequencingdepth refers to ratio of thenumber of nucleotideswiththe test gene or transcriptome. In this study, the average sequencingdepth was about 21 folds. In addition, the coverage depth of the assem-bled unigenes was also detected against Arabidopsis thaliana orthologs.Most of A. thaliana orthologs could be covered by our unigenes (Supple-mental file 3). Moreover, in our study, 67.6% (89,050 of 131,750) of themango fruit genes had homologs in the Nr protein databases, whereas



Fig. 6. qRT-PCR analysis of 6 randomly nonexpressor of pathogenesis-related unigenes in postharvest fruits from C. gloeosporioides inoculated (white bars) and control (gray bars). Theexpression levels of each unigene are expressed as a ratio relative to 0 d of samples, which was set at 1. Error bars indicate standard errors of the means (n = 3). Bars with differentlower-case letters indicate significant differences based on a t-test at the P ≤ 0.05 level.


in sweet potato (Wang et al., 2010), sesame (Wei et al., 2011) and litchi(Li et al., 2013) using the same approach, only 46%, 54% and 59%unigenes,respectively, has homologs in the Nr database.

To validate of the creditability the RNA-seq data, we randomlyselected 17 ERFs in our transcriptome database, and aligned theirdeduced amino acid sequences with the other plants. The resultsshowed that the ERFs in the mango fruit share a highly conservedERF/AP2 domain with ERF proteins of Arabidopsis, tomato and tobacco(Supplemental file 2).

Taken together, the mango fruit reference transcriptome assembledin this study is comprehensive, accurate and reliable for future geneticresearch of mango fruit.

For gene annotation, the sequence similarity search was performedagainst protein databases, including Nr, Swiss-Prot, GO, COG, andKEGG. Most of our unigenes could match unique known proteins inpublic databases, implying that the transcriptome sequencing yieldeda great number of unique genes in mango fruit. A large number ofunigenes were assigned to a wide range of Gene Ontology categoriesand COG classification, which indicated that our transcriptome datarepresented a broad diversity of transcripts in mango fruit. Similarresults were also reported in other species (Feng et al., 2012; Zhanget al., 2013b; Wei et al., 2013). Based on sequence homology searches


against the KEGG database, 44,145 unigenes could be mapped with256 pathways. Notably, KEGG predictions identified many unigenesassociated with defense responses, including transcription factors,bacterial secretion system, and plant hormone signal transduction,implying that there are a large number of genes involved in defenseresistance in mango fruit. In conclusion, these studies also provide anexcellent resource for gene isolation and gene expression profile analysisin mango fruit.

In response to pathogen attack, plants have evolved complex signalingand defense pathways. Following Colletotrichum infection, many genes,including those encoding defense-associated proteins, enzymes involvedin signaling pathways, and genes encoding transcription factors (TFs)were induced. In our database, there are 373 putative TFs detected.These TFs include MYB (10.7% of all TFs), WRKY (8.0% of all TFs), ERF(16.1% of all TFs). Some of TFs, such as MYB (Chen et al., 2006), NAC(Wang et al., 2009), WRKY (Xu et al., 2006), AP2/EREBP (Martinelliet al., 2012) have been identified as regulators of defense response againstpathogens in other plant species. In this study, members of the ERF familygenes were increased in expression in infected fruits, implying that theyact as positive regulators of resistance to C. gloeosporioides in mangofruit. The result was consistent with previous expression profiling analy-ses of ERF genes from other species, such as tomato (Zhang et al., 2004),




soybean (Zhang et al., 2009), Bupleurum kaoi (Liu et al., 2011) and sweetwormwood (Lu et al., 2013) that also showed positive effects on diseaseresistance. However, several ERFswere down-regulated in infected fruits,suggesting that they may be negative defense response regulators.Similarly, it was reported that others TFs such as WRKY40 acted asnegative regulators of resistance to Pseudomonas syringae in Arabidopsis(Xu et al., 2006). Thus, based on the expression pattern of ERF familygenes reported here, it could be speculated that ERFs could act as activa-tors or repressors in resistance to C. gloeosporioides in mango fruit.

With the exception of TFs being involved in plant defenseresponse, which can be regulated through the other transductionpathways (Kunkel and Brooks, 2002), and these signaling pathwayswere mediated by different signaling molecules, such as jasmonicacid, salicylic acid (SA).NBS-LRR,NPR and PR genes play an importantrole in SA-mediated defense pathways. NBS-LRR proteins mediate resis-tance processing corresponding avirulence (Postel and Kemmerling,2009). In the study, severalNBS-LRRswere up-regulated in infected fruits,implying that they may be receptors triggering fruits against C.gloeosporioides. In addition, almost NPR genes were up-regulated ininfected fruits. C. gloeosporioides up-regulation was observed for PRgenes. Expression change of these genes suggests that NBS-LRR,NPR and PR genes may be involved in the mango fruit against C.gloeosporioides. It has been hypothesized that SA response pathways

Fig. 7. qRT-PCR analysis of 6 randomly pathogenesis-related protein unigenes in postharvest frulevels of each unigene are expressed as a ratio relative to 0 d of samples, whichwas set at 1. Erroindicate significant differences based on a t-test at the P ≤ 0.05 level.


were activated in response to C. gloeosporioides according to Alkanet al. (2015) obtained with tomato fruit.

In conclusion, we employed high-throughput next-generationsequencing technology to obtain the first reference transcriptome dataof postharvest mango fruit against C. gloeosporioides. The results willbe very helpful for the identification of genes involved in plant defenseresponse. In addition, the present work constitutes the largest unigenedataset of mango fruit infected with C. gloeosporioides, which will helpresearchers to explore genes involved in defense response. Furthermore,some of the defense-associated unigenes showed significant differentialexpression, implying that these potential regulators might function inresponse against C. gloeosporioides.

Acknowledgments

We also thank the CapitalBio Corporation at Beijing and GeneDenovo at Guangzhou for its assistance in original data processing andrelated bioinformatics analysis, respectively. This researchwas supportedby the Chinese Special Fund of Basic Scientific Research Projects forState Level and Public Welfare-Scientific Research Institutes (No.1630062012001; No. 1630062014015), Hainan Province ScienceFoundation (No. 20153043; No. 314108) and Research Specific ofPublic Welfare Industry (No. 201503142-13).

its from C. gloeosporioides inoculated (white bars) and control (gray bars). The expressionr bars indicate standard errors of themeans (n=3). Barswith different lower-case letters




Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.gene.2015.10.041.

References

Alkan, N., Friedlander, G., Ment, D., Prusky, D., Fluhr, R., 2015. Simultaneous transcriptomeanalysis of Colletotrichum gloeosporioides and tomato fruit pathosystem reveals novelfungal pathogenicity and fruit defense strategies. New Phytol. 205, 801–815.

Chen, Y.H., Yang, X.Y., He, K., Liu, M.H., Liu, J.Q., et al., 2006. The MYB transcription factorsuperfamily of Arabidopsis: expression analysis and phylogenetic comparison withthe rice MYB family. Plant Mol. Biol. 60, 107–124.

Chowdhury, M.N.A., Rahim, M.A., Khalequzzaman, K.M., Alam, J., Humauan, M.R., 2008.Effect of horticultural practices on incidence of anthracnose on yield and quality ofmango. Int. J. Sustain. Crop Prod. 3, 1–9.

Dessalegn, Y., Ayalew, A.,Woldetsadik, K., 2013. Integrating plant defense inducing chemical,inorganic salt and hot water treatments for the management of postharvest mangoanthracnose. Postharvest Biol. Technol. 85, 83–88.

Estrada, A.B., Dodd, J.C., Jeffries, P., 2000. Effect of humidity and temperature on conidialgermination and appressorium development of two Philippine isolates of the mangoanthracnose pathogen Colletotrichum gloeosporioides. Plant Pathol. 49, 608–618.

Feng, C., Chen, M., Xu, C.J., Bai, L., Yin, X.R., et al., 2012. Transcriptomic analysis of Chinesebayberry (Myrica rubra) fruit development and ripening using RNA-Seq. BMCGenomics13, 19.

Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, D.A., et al., 2011. Full lengthtranscriptome assembly from RNA-Seq data without a reference genome. Nat.Biotechnol. 29, 644–652.

Harris, M.A., Clark, J., Ireland, A., Lomax, J., Ashburner, M., et al., 2004. The Gene Ontology(GO) database and informatics resource. Nucleic Acids Res. 32, D258–D261.

Kanehisa,M., Goto, S., Hattori, M., Aoki-Kinoshita, K.F., Itoh,M., et al., 2006. From genomics tochemical genomics: new developments in KEGG. Nucleic Acids Res. 34, D354–D357.

Kefialew, Y., Ayalew, A., 2008. Postharvest biological control of anthracnose (Colletotrichumgloeosporioides) on mango (Mangifera indica). Postharvest Biol. Technol. 50, 8–11.

Kunkel, B.N., Brooks, D.M., 2002. Cross talk between signaling pathways in pathogendefense. Curr. Opin. Plant Biol. 5, 325–331.

Li, C.Q., Wang, Y., Huang, X.M., Li, J., Wang, H.C., et al., 2013. De novo assembly andcharacterization of fruit transcriptome in Litchi chinensis Sonn and analysis of differ-entially regulated genes in fruit in response to shading. BMC Genomics 14, 552.

Liu, J.X., Li, J.Y., Wang, H.N., Fu, Z.D., Liu, J., et al., 2010. Identification and expression anal-ysis of ERF transcription factor genes in petunia during flower senescence and in re-sponse to hormone treatments. J. Exp. Bot. 62, 825–840.

Liu, W.Y., Chiou, S.J., Ko, C.Y., Lin, T.Y., 2011. Functional characterization of three ethyleneresponse factor genes from Bupleurum kaoi indicates that BkERFs mediate resistanceto Botrytis cinerea. J. Plant Physiol. 168, 375–381.

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-timequantitative PCR and the 2−ΔΔCT method. Methods 25, 402–408.

Lu, X., Jiang, W.M., Zhang, L., Zhang, F., Zhang, F.Y., Shen, Q., Wang, G.F., Tang, K.X., 2013.AaERF1 positively regulates the resistance to Botrytis cinerea in Artemisia annua. PLoSONE 8 (2), e57657.

Luo, C., Qing, L., Zhang, Q.L., Luo, Z.R., 2014. Genome-wide transcriptome analysis of Chinesepollination-constant nonastringent persimmon fruit treated with ethanol. BMCGenomics 15, 112.

Martinelli, F., Uratsu, S.L., Albrecht, U., Reagan, R.L., Phu, M.L., et al., 2012. Transcriptomeprofiling of citrus fruit response to Huanglongbing disease. PLoS ONE 7 (5), e38039.http://dx.doi.org/10.1371/journal.pone.0038039.

Nelson, S.C., 2008. Mango anthracnose (Colletotrichum gloeosporioides). Plant Dis. 48.Ploetz, R.C., Prakash, O., 1997. Foliar, Floral and Soil Borne Diseases. In: Mango: Botany,

Production and Uses. CAB International, Wallingford, UK, pp. 281–326.


Postel, S., Kemmerling, B., 2009. Plant systems for recognition of pathogen-associatedmolecular patterns. Semin. Cell Dev. Biol. 20, 1025–1031.

Qiu, Q., Ma, T., Hu, Q.J., Liu, B.B.,Wu, Y.X., et al., 2011. Genome-scale transcriptome analysis ofthe desert poplar, Populus euphratica. Tree Physiol. 31, 452–461.

Tang, W., Zhu, S., Li, L., Liu, D., Irving, D.E., 2010. Differential expressions of PR1 andchitinase genes in harvested bananas during ripening, and in response to ethephon,benzothiadizole and methyl jasmonate. Postharvest Biol. Technol. 57, 86–91.

Vargas, W.A., Sanz, M., José, M., Rech, G.E., Rivera, L.P., et al., 2012. Plant defensemechanismsare activated during biotrophic and necrotrophic development of Colletotricumgraminicola in maize. Plant Physiol. 158, 1342–1358.

Wan, C.Y., Wilkin, T.A., 1994. A modified hot borate method significantly enhances theyield of high quality RNA from cotton (Gossypium hirsutum L.). Anal. Biochem. 223,7–12.

Wang, X.E., Basnayake, B.M.V.S., Zhang, H.J., Li, G.J., Li, W., et al., 2009. The ArabidopsisATAF1, a NAC transcription factor, is a negative regulator of defense responses againstnecrotrophic fungal and bacterial pathogens. Mol. Plant Microbe Interact. 22,1227–1238.

Wang, Z., Fang, B., Chen, J., Zhang, X., Luo, Z., et al., 2010. De novo assembly and character-ization of root transcriptome using Illumina paired-end sequencing and developmentof cSSR markers in sweetpotato (Ipomoea batatas). BMC Genomics 11, 726.

Wei, W.L., Qi, X.Q., Wang, L.H., Zhang, Y.X., Hua, W., et al., 2011. Characterization of thesesame (Sesamum indicum L.) global transcriptome using Illumina paired-endsequencing and development of EST-SSR markers. BMC Genomics 12, 451.

Wei, X.C., Zhang, X.H., Shen, D., Wang, H.P., Wu, Q.J., et al., 2013. Transcriptome analysis ofBarbarea vulgaris infested with diamondback moth (Plutella xylostella) larvae. PLoSONE 8 (5), e64481. http://dx.doi.org/10.1371/journal.pone.0064481.

Wu, H.X., Ma, X.W., Wang, S.B., Yao, Q.S., Xu, W.T., et al., 2014. Transcriptome and proteomicanalysis of mango (Mangifera indica Linn) fruits. J. Prot. http://dx.doi.org/10.1016/j.jprot.2014.03.030

Xu, X., Chen, C., Fan, B., Chen, Z., 2006. Physical and functional interactions betweenpathogen-induced ArabidopsisWRKY18,WRKY40, andWRKY60 transcription factors.Plant Cell Online. 18, 1310–1326.

Yang, S.M., Gao, M.Q., Xu, C.W., Gao, J.C., Deshpande, S., et al., 2008. Alfalfa benefits fromMedicago truncatula: the RCT1 gene from M. truncatula confers broad-spectrumresistance to anthracnose in alfalfa. Proc. Natl. Acad. Sci. 105, 12164–12169.

Ye, J., Fang, L., Zheng, H., Zhang, Y., Chen, J., et al., 2006. WEGO: a web tool for plotting GOannotations. Nucleic Acids Res. 34, W293–W297.

Zeng, K.F., Cao, J.K., Jiang, W.B., 2006. Enhancing disease resistance in harvested mango(Mangifera indica L. cv. ‘Matisu’) fruit by salicylic acid. J. Sci. Food Agric. 86, 694–698.

Zhang, H.B., Zhang, D.B., Chan, J., Yang, Y.H., Huang, Z.J., et al., 2004. Tomato stress-responsiveactor TSRF1 interacts with ethylene responsive element GCC box and regulatespathogen resistance to Ralstonia solanacearum. Plant Mol. Biol. 55, 825–834.

Zhang, G.Y., Chen, M., Li, L.C., Xu, Z.S., Chen, X.P., et al., 2009. Overexpression of the soybeanGmERF3 gene, an AP2/ERF type transcription factor for increased tolerances to salt,drought, and diseases in transgenic tobacco. J. Exp. Bot. 60, 3781–3796.

Zhang, F., Du, Z., Huang, L., Cruz, C.V., Zhou, Y., et al., 2013a. Comparative transcriptomeprofiling reveals different expression patterns in Xanthomonas oryzae pv. oryzaestrains with putative virulence-relevant genes. PLoS ONE 8 (5), e64267. http://dx.doi.org/10.1371/journal.pone.0064267.

Zhang, M., Yu, H., Yang, Y.Y., Song, C., Hu, X.J., et al., 2013b. Analysis of the transcriptomeof blowfly Chrysomya megacephala (Fabricius) larvae in responses to different edibleoils. PLoS ONE 8 (5), e63168. http://dx.doi.org/10.1371/journal.pone.0063168.

Zhang, H.N., Wei, Y.Z., Shen, J.Y., Lai, B., Huang, X.M., et al., 2014. Transcriptomic analysisof floral initiation in litchi (Litchi chinensis Sonn.) based on de novo RNA sequencing.Plant Cell Rep. http://dx.doi.org/10.1007/s00299-014-1650-3.

Zhuang, X.F., McPhee, K.E., Coram, T.E., Peever, T.L., Chilvers, M.I., 2012. Rapidtranscriptome characterization and parsing of sequences in a non-model host-pathogen interaction; pea-Sclerotinia sclerotiorum. BMC Genomics 13, 668.




http://refhub.elsevier.com/S0378-1119(15)01249-4/rf0005














































http://dx.doi.org/10.1371/journal.pone.0038039














































http://dx.doi.org/10.1007/s00299-014-1650-3





Documents

Transcriptome characterization and expression profiles of