De Novo Transcriptome Sequence Assembly and Identification of AP2/ERF Transcription Factor ...

12 Pages·2014·2.07 MB·English

DeNovoTranscriptome Sequence Assembly and
Identification of AP2/ERF Transcription Factor Related to
Abiotic Stress in Parsley (Petroselinumcrispum)
Meng-Yao Li., Hua-Wei Tan., Feng Wang, Qian Jiang, Zhi-Sheng Xu, Chang Tian, Ai-Sheng Xiong*
StateKeyLaboratoryofCropGeneticsandGermplasmEnhancement,CollegeofHorticulture,NanjingAgriculturalUniversity,Nanjing,China
Abstract
ParsleyisanimportantbiennialApiaceaespeciesthatiswidelycultivatedasherb,spice,andvegetable.Previousstudieson
parsley principally focused on itsphysiologicaland biochemical properties, including phenolic compound and volatileoil
contents. However, little is known about the molecular and genetic properties of parsley. In this study, 23,686,707 high-
quality reads were obtained and assembled into 81,852 transcripts and 50,161 unigenes for the first time. Functional
annotation showed that 30,516 unigenes had sequence similarity to known genes. In addition, 3,244 putative simple
sequencerepeatsweredetectedincurlyparsley.Finally,1,569oftheidentifiedunigenesbelongedto58transcriptionfactor
families.Variousabioticstresseshaveastrongdetrimentaleffectontheyieldandqualityofparsley.AP2/ERFtranscription
factorshaveimportantfunctionsinplantdevelopment,hormonalregulation,andabioticresponse.Atotalof88putative
AP2/ERF factors were identified from the transcriptome sequence of parsley. Seven AP2/ERF transcription factors were
selected in this study to analyze the expression profiles of parsley under different abiotic stresses. Our data provide a
potentially valuableresource that canbe usedforintensiveparsley research.
Citation:LiM-Y,TanH-W,WangF,JiangQ,XuZ-S,etal.(2014)DeNovoTranscriptomeSequenceAssemblyandIdentificationofAP2/ERFTranscriptionFactor
RelatedtoAbioticStressinParsley(Petroselinumcrispum).PLoSONE9(9):e108977.doi:10.1371/journal.pone.0108977
Editor:Zhong-HuaChen,UniversityofWesternSydney,Australia
ReceivedMay13,2014;AcceptedAugust27,2014;PublishedSeptember30,2014
Copyright:(cid:2)2014Lietal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsunrestricted
use,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
DataAvailability:Theauthorsconfirmthatalldataunderlyingthefindingsarefullyavailablewithoutrestriction.Allrelevantdataarewithinthepaperandits
SupportingInformationfiles.
Funding:TheresearchwassupportedbyNationalNaturalScienceFoundationofChina(31272175),NewCenturyExcellentTalentsinUniversity(NCET-11-0670),
JiangsuNaturalScienceFoundation(BK20130027),PriorityAcademicProgramDevelopmentofJiangsuHigherEducationInstitutions,andJiangsuShuangchuang
Project.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*Email:[email protected]
.Theseauthorscontributedequallytothiswork.
Introduction andformcomplexgeneregulatorynetworks[9,14,15].Uptonow,
little is known about the abiotic stress tolerance of parsley.
Parsley(PetroselinumcrispumL.)isabiennialApiaceaespecies Cormack [16] isolated two WRKY transcription factors from
thatisnativetotheMediterraneancoastandwidelycultivatedin
parsley usingthe yeastone-hybrid system. Weisshaar [17] cloned
EuropeandJapan.Parsleyissubdividedintothreeprincipaltypes
three bZIP genes from parsley and found that these genes are
accordingtocultivation:curlyleaftype(subspeciescrispum),plain
involved in the response to environmental changes and disease
leaftype(subspeciesneapolitanum),and‘‘Hamburg’’rootparsley
invasion. However, almost no AP2/ERF members have been
(subspecies tuberosum). The curly leaf and plain leaf types are
identifiedinparsley.AP2/ERFwasoneofthelargesttranscription
cultivatedfortheirfoliage,whereasrootparsleyisgrownasaroot
factorfamiliesinhigherplantsandhasreceivedmuchattentionin
vegetable[1].Parsleyiswidelyutilizedinthecosmetic,medicinal,
recent years. This family can be further classified into four
and food industries because it is an excellent source of phenolic
subfamiles: ERF, DREB, AP2, RAV [18–20]. Numerous reports
compounds, volatile oils,vitamins, andnutrients[2–4].
have demonstrated that the family members can regulate plant
Global challenges, such as climate change, environmental
responses to abiotic stresses [21,22]. JERF3, an ERF member in
degradation, and toxic waste, subject plants to various stresses
tomato,canbeinduced byabscisicacid,ethylene,jasmonicacid,
during growth. Drought, high salinity, and extreme temperature
and low temperature; ectopic overexpression of JERF3 in
are the major limiting factors of higher plant growth and
transgenic tobacco enhances salt tolerance [23]. A DREB-type
production. Numerous genes in higher plants are activated in
gene LsDREB2A, was isolated from lettuce, can increased the
response to these abiotic stresses [5]. Genes can either directly
tolerance ofsalt stress intransgenicplants [24].
respond to stresses or regulate the expression of other genes and
Asofthiswriting,researchonparsleyhasprincipallyfocusedon
signal transcription [6,7]. Transcription factors function in gene
essential oil content [25–27] and flavonoid products [28], but
expression by combining DNA-binding and cis-acting elements
informationonmolecularbiologyandgenefunctionislacking.No
[8,9].Manytranscriptionfactors,suchasAP2/ERF,NAC,bZIP,
genome-sequenced species intheApiaceae familyhas limitedthe
and WRKY, are related to stress resistance in plants [10–13].
research. As far as we know, only three transcriptome sequence
These transcription factors interact to regulate gene expression
PLOSONE | www.plosone.org 1 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
9 4 9
% 30.8 46.2 21.3 1.23 0.12 0.12 100
Total 1002 1500 694 40 4 4 3244 100
.15 135 1 1 0 0 0 137 4.22
8
15 48 0 0 0 0 0 48 1.4
3
14 43 0 0 0 0 0 43 1.3
Figure 1. Size distribution of the assembled contigs, tran-
sdcori:i1p0t.s1,3a71n/djouurnniagl.epnoense..0108977.g001 13 67 0 1 0 0 0 68 2.1
itnhfeorAmpaiaticoenaehfaasmoiblyt,aiwnhedichfrwomerecebleelroyn[g2e9d,3to0]AapnidumcaarnrodtD[3a1u]cuins 12 130 2 0 0 0 0 132 4.07
genus, respectively. The limited resources cannot provide more
hReNlpAt-oSestqudyisthaepfaearssliebyle,wahnicdhiescboenloonmgiecdaltomPoetdreorsnelinseuqmuegnecniunsg. 11 186 79 0 0 0 0 265 8.17
technologyforobtainingtranscriptomicdatainashorttime.This
technology can detect new transcripts that correspond to existing
5
greesnooumrciecssfeoqrugeennceesd;isictovcearny,aelsxoprbesesiounse,dantdoagnennoetraattieons,eaqnudenfcoer 10 393 224 0 0 0 1 618 19.0
discovering simple sequence repeats (SSRs) and single nucleotide
preofleyrmenocrephgiesmnosme(SN[3P2s–)34i]n. RnNonA--mSeoqdehlasorbgeaennismussedwtiothooubttaina 9 0 224 1 0 0 0 225 6.94
transcriptomic data for an increasing number of organisms, such
astobacco[35],grapevine[36],sunflower[37],andsweetpotato
[38].Thismethodisconvenientforintensivestudiesinmolecular 8 0 213 13 1 0 1 228 7.03
biology. In the present study, we performed the first comprehen-
siveanalysisofthetranscriptomeofparsleyusingIlluminapaired-
efonrdinseteqnuseivnecipnagrstelecyhnreosleoagryc,hw.ThihcehAcaPn2/pEroRvFidgeevnaelufaamblielyrewsaosuraclseos 7 0 297 102 2 0 1 402 12.39
analyzed based on the obtained data. Some genes in the AP2/
ERF family were isolated, and their relation to abiotic stress
raensaploynzesethweasmdoeleteccutleadr.mTehceharensiusmltsuonfdtehrilsyisntugdtyhecosutrledssbetoulesreadnctoe me. mbers 6 0 460 168 11 0 1 640 19.73
of parsley. o nu
Methods and Materials anscript Repeat 5 0 0 408 26 4 0 438 13.5
Plant materials tr
y
e
The curly parsley cultivar (P. crispum L. subsp. crispum) was sl
used as plant material (Figure S1). Seeds were sown in a pot ar
p 1
0
ceonnvitraoinnimngentagrsoowil/thvecrhmaimcublietreumndixertuare16(3h:1/)8ihnphaotcoopnetrriooldleda-t urly 77.t0
2rtho5eountCs/swt1oe6rreueCdcaodtlal–ey7c/t0neuidCg,htifmocrmyRcelNed.iAaAteefltxyetrrfar1oc0tzieownne.ienksl,iqlueaidvens,itrsotegmens,, aanndd ypesinc pone.01089
t al.
RRionfNNsTttArAhouestciaistmleiooxplntaRlrsetaNic(oTttAnoeidatanaloRgnfedNRnmA,NliBbAixwereaiedjkirrniyetgs,pvaaemCrcrecihpfpoiilenreaddaria)n.wbtgiyTaoshngteoeeflqoxteutrhrlaeeasnccetttrimqetodyupanehaunonufrsdcaienicsnqigtsuugraaetlnihrt’dyes SSRTable1. Repeattype Mono-nucleotide Di-nucleotide Tri-nucleotide Tetra-nucleotide Penta-nucleotide Hexa-nucleotide Total % doi:10.1371/journ
PLOSONE | www.plosone.org 2 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Table2. Functionalannotationofcurlyparsley unigenesby sequencesimilarity search.
AnnotatedDatabases AnnotatedNumber 300!length,1000 length§1000
Nr 30516 13644 12460
Nt 22871 8948 11225
TrEMBL 30410 13612 12465
Swissprot 23432 9661 10773
GO 26149 11167 11493
COG 9469 3188 5438
KEGG 6569 2519 3139
TotalAnnotated 31658 14256 12513
doi:10.1371/journal.pone.0108977.t002
spectrophotometry (Nanodrop-ND-1000 spectrophotometer, Na- Putative SSR screening
nodrop Technologies Inc., Delaware, USA). mRNA was concen- AlldetectedunigeneswereusedforscreeningputativeSSRsby
trated using oligo(dT) magnetic adsorption and then broken into MIcroSAtellite tool (http://pjrc.ipk-gatersleben.de/misa/) [41].
fragments, which were used as templates to synthesize first- and The putative SSRs contained motifs with one to six nucleotides,
second-strand cDNA. The double-stranded cDNA was further andtheparametersofcontiguousrepeatunitsweresetformono-,
purifiedusingtheQiaQuickPCRextractionkit(Qiagen,Hilden, di-,tri-,tetra-,penta-,andhexa-nucleotidemotifswithaminimum
Germany),resolvedforfinalreparationandpoly(A)addition,and of 10,6,5, 5,5,and 5repeats, respectively.
thenconnectedwithdifferentsequencingadapters.Alibrarywith
a suitable insert length (300bp to 500bp) was sequenced by Functional annotation
Biomarker technologies Co., Ltd. (Beijing, China) using the
A sequence similarity search was performed against seven
Illumina HiSeqTM 2000. The sequence data of parsley were
databases to investigate the putative functions of the unigenes
submitted to NCBI Sequence Read Archive (http://trace.ncbi.
based on sequence or domain alignment. All unigenes were
nlm.nih.gov/Traces/sra/) under the accession number
compared with genes in the NCBI non-redundant protein (Nr),
SRA111430.
NCBI Non-redundant Nucleotide (Nt), Swiss-Prot, TrEMBL,
Gene Ontology (GO, http://www.geneontology.org/), Clusters
De novo assembly of Orthologous Groups (COG), and Kyoto Encyclopedia of
Therawreadswerefirstcleanedbyfilteringadaptorsequences Genes and Genomes (KEGG) databases [42–44]. Homology
andlow-quantityreads(morethan50%ofbaseswithQ-value# searchagainsttheNrdatabasewasperformedtoidentifytop-hit
20). For denovo assembly, the clean reads were mapped back to species by BLASTx with a cut-off E-value of 1e-5. Blast2GO
thecontigsbyTrinity[39]withtheparameterssetatasimilarityof [45] was employed to obtain the functional classification, and
90%. Subsequently, the contigs were assembled to construct WEGO [46] was used to perform the distribution of GO
transcripts with pair-end information and clustered to obtain classification.
unigenes. Open reading frames (ORFs) were identified using the
Getorf program[40].
Figure2.FrequencydistributionofSSRsaccordingtomotiftype.
doi:10.1371/journal.pone.0108977.g002
PLOSONE | www.plosone.org 3 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Figure3.Characteristicsofhomologyofcurlyparsleyunigenes.(A)E-valuedistributionofBLASTxhitsagainstNrdatabaseforunigenes;(B)
SequencesimilaritydistributionofthebestBlastxhitsforallunigenes;(C)ProportionofunigenesmatchedtoeachspeciesbyBLASTx;thetop9
speciesareindicated.
doi:10.1371/journal.pone.0108977.g003
Figure4.GOclassificationofassembledunigenesofP.crispum.Theclassificationsareshownin3principalcategoriesand61functional
groups.
doi:10.1371/journal.pone.0108977.g004
PLOSONE | www.plosone.org 4 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Figure5.COGfunctionclassificationofassembledunigenesofP.crispum.
doi:10.1371/journal.pone.0108977.g005
Transcription abundance analysis Premix Ex-Taq (TaKaRa, Dalian, China). The PCR conditions
The transcription abundance of each unigene in the curly wereasfollows:95uCfor30s;40cyclesof95uCfor5 sand60uC
parsley librarywasmeasured bycalculatingread densityasreads for30s;and65uCfor15s.Theprimersofunigenesandactinare
perkilobaseofthetranscriptpermillionmappedreads(RPKM)to listed in Table S1. The experiments were repeated three bio-
thetranscriptome[47].TheRPKMindicatestheexpressionlevel replicates and tech-replicates, and actin was used as a reference
of each unigene by normalizing the counts of sequenced reads gene.Theexpressionlevelsoftheunigeneswerecalculatedbythe
mappedtoageneagainstthetranscriptlengthandthesequencing 22DDCT method[50].
depth.
Results
Multiplesequencealignmentsandphylogeneticanalyses
Sequencing,de novoassembly,andsequenceanalysisof
of AP2/ERF transcription factors
parsley
HMMERandlocalBLASTwereusedtoscreen thetranscrip-
A cDNA library of curly parsley was constructed for
tionfactorswithE-valuesbelow1e-5.Sequencealignmentsofthe
transcriptome sequencing. Sequence data of 4.78Gb were
AP2/ERF proteins in parsley and Arabidopsis were performed
generated, and 23,686,707 reads were obtained with 95.56%
withClustalW[48]usingdefaultparameters.Aphylogenetictree
Q20 bases as high-quality reads. The high-quality reads were
was constructed with MEGA 5.0 [49] using the neighbor-joining
method withthebootstrap was set to1,000. assembled into 1,224,381 contigs with an N50 length of 126bp
and a mean length of 87bp by Trinity [39]. The contigs were
further assembled into 81,852 transcripts and clustered into
Abiotic stress treatments and quantitative reverse
unigenesusingapaired-endsequencingstrategy.Atotalof50,161
transcription-polymerase chain reaction (qRT–PCR)
unigeneswereobtained.Theseunigeneshadlengthsintherange
Half of the two-month-old curly parsley seedlings were
of 201bp to 15,178bp, an N50 length of 1,344 bp, and a mean
transferred to growth chambers set at 4uC or 38uC, which
length of 802bp. Most of the unigenes (54.41% in curly parsley)
representedlowandhightemperaturestresstreatments.Theother
had lengths in the range of 200bp to 500bp. Up to 9,982
seedlings were irrigated with double-distilled H O (control),
2 (19.90%)unigeneshadlengthsintherangeof500bpto1,000bp,
200mM NaCl (salt treatment), and 20% polyethylene glycol
and12,888(25.59%)unigeneshadlengthsof.1,000 bp.Thesize
6000(droughttreatment).Youngleafsampleswerecollectedat0,
distributionsofthecontigs,transcripts,andunigenesareshownin
1,2,4,8,and24hafterthedifferentstresstreatments.TotalRNA
Figure 1.Getorf[40]wasusedtofindandextracttheORFsofall
was isolated using the total RNA kit (RNAsimply, Tiangen,
the unigenes to obtain the coding and protein sequences. Up to
Beijing,China)andthenreversetranscribedintocDNAusingthe
49,946putativecodingsequenceswereidentified.Thesesequences
PrimeScript RT reagent Kit (TaKaRa, Dalian, China). qRT–
can beusedforgene cloningand functional verification.
PCR was performed using MyiQ Single color Real-Time PCR
Detection System (Bio-rad, Hercules, CA, USA) with SYBR
PLOSONE | www.plosone.org 5 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Figure6.DistributionofeachKEGGpathwaynumberagainsttheKEGGdatabase.EachcolorrepresentsaKEGGpathway.Thetop19
KEGG pathways are indicated. The number of unigenes mapped in each pathway is indicated with brackets. The abbreviations represent the
pathways as follows: ko03010: Ribosome; ko04075: Plant hormone signal transduction; ko03040: Spliceosome; ko04141: Protein processing in
endoplasmic reticulum; ko03013: RNA transport; ko00230: Purine metabolism; ko00190: Oxidative phosphorylation; ko00500: Starch and sucrose
metabolism; ko00010: Glycolysis/Gluconeogenesis; ko04626: Plant-pathogen interaction; ko04120: Ubiquitin mediated proteolysis; ko03018: RNA
degradation; ko03015: mRNA surveillance pathway; ko00240: Pyrimidine metabolism; ko00710: Carbon fixation in photosynthetic organisms;
ko04146:Peroxisome;ko00620:Pyruvatemetabolism;ko04144:Endocytosis;ko00520:Aminosugarandnucleotidesugarmetabolism.
doi:10.1371/journal.pone.0108977.g006
Figure7.Familydistributionofthetranscriptionfactorsoccurringinthecurlyparsleytranscriptome.Thenumberofeachtranscription
factorfamilymemberswasrepresented.Familiescomprisinglessthan15transcriptionfactorsareclassifiedunderothers.
doi:10.1371/journal.pone.0108977.g007
PLOSONE | www.plosone.org 6 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Figure8.PhylogenetictreeofallAP2/ERFtranscriptionfactorsfromparsleyandArabidopsis.AllthecirclesrepresenttheAP2/ERFgenes
inparsley(red)andArabidopsis(green).TheredpentagramsrepresenttheAP2/ERFgeneswhichwereselectedtodetectedresponsetofourabiotic
stresses.
doi:10.1371/journal.pone.0108977.g008
Selection and analysis of putative SSRs numerous A and T repeat elements, showing a strong base
SSRs are repeating DNA sequences of 1bp to 6bp in both preference.
coding and non-coding regions of the genome [51]. SSRs are
commonly used in gene mapping because of their high polymor- Functional classification of curly parsley unigenes
phism,widedistributioninthegenome,andeasyoperation.Inthis Asequencesimilaritysearchwasperformedbasedonsequence-
study, we identified 3,244 putative SSRs in 2,643 curly parsley anddomain-basedalignmentstofunctionallyannotatetheparsley
unigenes, amongwhich473hadmorethan oneSSR. Up to299 transcriptome. All unigenes were searched against seven public
unigenes occurred incompound formation. databases. The sequences that appeared on each database are
AllputativeSSRshaddifferentlengthsbetweendifferentrepeat listed in Table 2. All unigenes were first compared with genes in
types (Table 1). The di-nucleotide SSRs comprised the largest the NCBI non-redundant database based on sequence alignment
fraction(46.24%),followedbymono-nucleotide(30.89%)andtri- using BLASTx with a cut-off E-value of 1e-5. Up to 30,516
nucleotide(21.39%)SSRs.OthertypesofSSRs(tetra-,penta-,and unigenes (60.84% of all assembled unigenes) had sequence
hexa-nucleotide repeats) had a frequency of less than 1.5%. The similarity to known genes. The distributions of E-value and
frequenciesofSSRmotiftypeswerealsoanalyzed(Figure 2).Most sequence similarity were comparable, with 58.41% (E-value
of the mono-nucleotides were of the A/T type, accounting for between 0 and 1e-50) and 19.51% (sequence similarity between
28.42%ofallSSRsandwerealmost15-foldhigherthantheC/G 80% and 100%) showing very strong homology, respectively
type. Di-nucleotide repeat motifs were divided into four classes; (Figures 3Aand3B).Forspeciesdistributionofthebestmatch,P.
the most abundant types were AG/CT and AC/GT, which crispum showed the highest similarity to Vitis vinifera (43.06%),
accounted for32.68%and10.14%of allSSRs,respectively. Tri- followed by Populus trichocarpa (12.57%) and Ricinus comunis
nucleotiderepeatmotifsweredividedinto10categories;themost (11.61%) (Figure 3C). The more detailed information of annota-
abundant types were AAG/CTT and ATC/ATG. The forma- tionswas represented inTable S2.
tions of mono-, di-, and tri-nucleotide repeat types comprised GOwasusedtoclassifytheunigenesintofunctionalcategories
by Blast2GO. A total of 26,149 unigenes were annotated and
PLOSONE | www.plosone.org 7 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
classifiedinto3geneontologycategoriesand61functionalgroups
(Figure 4). In the ‘‘cellular compound’’ category, ‘‘cell part’’
(22.03%) was the most dominant group, followed by ‘‘cell’’
(21.82%) and ‘‘organelle’’ (19.96%). Under the ‘‘molecular
mm function’’ category, ‘‘binding’’ (43.80%) and ‘‘catalytic activity’’
Triticuaestivu 47 57 9 3 1 117 (p3r7o.c5e1s%s’’)cwateergeortyh,e‘‘cmeollsutladropmroincaensst’’g(r1o4u.0p5s.%I)n, ‘‘tmheeta‘‘bboiolilcogpircoa-l
cess’’(13.68%),and‘‘responsetostimulus’’(9.83%)werethemost
dominant groups. According to the COG database, 9,469
unigenes were clustered into 25 functional categories (Figure 5).
‘‘Generalfunctionpredictiononly’’(19.43%)wasthelargestCOG
aea category, followed by ‘‘replication, recombination, and repair’’
zv
Orysati 79 52 26 7 0 164 (10.09%) and ‘‘transcription’’ (9.32%). In addition, all unigenes
were searched against the KEGG pathway database. A total of
n
o 6,569unigenesweremappedto137pathways.Thetop19KEGG
d
yle pathways, which contained over 100 unigenes, are shown in
ot Figure 6. ‘‘Ribosome’’ (PATH:ko03010), ‘‘plant hormone signal
c
Mono Zeamays 107 51 22 3 1 184 t(PraAnTsdHu:cktoio0n3’0’40)w(ePrAeTthHe:kmoo0s4t0d7o5m),inantapnadthway‘s‘.spliceosome’’
Identification of transcription factors in parsley
at. HMMERandlocalBLASTwithE-valuesbelow1e-5wereused
e to screen the transcription factors from curly parsley transcrip-
h
riceandw 6Malusdomestica 127 68 51 6 7 259 tP1fao,lm5ma6nie9lti.eoTTsfrrt(aaFhnniegsscciudrrriieeppntt7tiioiofnniaendfFaducatncTotirogaerbfanlemDesaiSlbti3aee)sbl.oawnTseeghreee(dVcmtleoaross5ssiio8tfinethrdia3gna.h0cslc)cyro[irrp5det2pii]onr.ngeUsfteaopnctttheotdoer
e, transcriptionfactorfamilieswereMYB(172unigenes),PHD(157
z
ai unigenes), bHLH (90 unigenes), and AP2/ERF (88 unigenes).
m
Among these members, MYB and bHLH transcription factors
e,
bean,appl Ricinuscommunis 56 34 19 4 1 114 mA[5Pa32y–/5bE5e]R.Fintvroalnvsecdripintiofnlafvaocntooridsmbaioysybnetihnevsoisl,vewdhienrestarsesPsHreDspoannsde
or Phylogenetic relationship of AP2/ERF transcription
st factors
a
c
s, Transcription factors of the AP2/ERF gene family can be
psi dividedintofoursubgroups(DREB,ERF,RAV,AP2),andSoloist
o
d based on the sequence similarity [18–20]. To confirm the
abi psis subfamily classification and analyze the evolutionary relationship
Ar dona between carrot and Arabidopsis, we used the AP2/ERF amino
arsley, Arabithalia 65 57 18 6 1 147 aFcigidurese8q,uaelnlctehse8to8AgePn2e/rEatReFsawpehreylcolgaesnsiefiteicditnretoe.fivAessushbofawmniliiens
p with the 49 members in ERF subfamily, 22 members in DREB
n
i subfamily,12membersinAP2subfamily,3membersinRAVand
s
or 2 members in Soloist. Compared with other species [18,56–60],
ct the AP2/ERF family in parsley seems to have relatively less
a
f members (Table3). The numbers of each subfamily members
y
mil on um werevariedamongdifferentspecies.ERFisconsistentlythelargest
2/ERFfa Dicotyled Petroselincrispum 49 22 12 3 2 88 saunbdfaSmoliolyistin. these seven plants, followed by DREB, AP2, RAV,
AP 003 Expression analysis of AP2/ERF genes under abiotic
the 8977.t stresses
of 10 The transcript expression levels of all unigenes in the curly
0
y e. parsley library were estimated by calculating read density as
mar pon RPKM[47].TheRPKMsof.50%oftheunigenesrangedfrom
m nal. 1to50,andthoseof.5%oftheunigeneswere.50(TableS4).
SuTable3. Plants Classification ERFsubfamily DREBsubfamily AP2subfamily RAVsubfamily Soloist Total doi:10.1371/jour fpfsaotrrmueeMrssieslasyneunstbygsg(etlrsonutoweduusydp,itiseenssmevwvohpeelraevnevreaedAsteuPrlirene2ecp/,toaEerbhdRtiieogtFdothicttdrthaeesanttmretseccpsttrehseirpdeartetirmuosepnrseeopm,fnoabsnhceestioegr[shr2tso1osb,faf6eol1Aliuon]Prn.it2gyaI/i,nbnEigaoRtnthtiFdoec
PLOSONE | www.plosone.org 8 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
Table4. Selected AP2/ERFgenes putativelyrelatedto stress responsesby GOannotation.
GeneID Subfamily Annotation
Pc16182 ERF responsetocold(GO:0009409);responsetowaterdeprivation(GO:0009414);responsetoabscisicacidstimulus
(GO:0009737)
Pc16943 ERF responsetowaterdeprivation(GO:0009414);responsetoabscisicacidstimulus(GO:0009737);responsetofreezing
(GO:0050826)
Pc24931 ERF responsetowaterdeprivation(GO:0009414);responsetosaltstress(GO:0009651);responsetofreezing(GO:0071497)
Pc32872 AP2 responsetoheat(GO:0009408);responsetowaterdeprivation(GO:0009414);responsetosaltstress(GO:0009651)
Pc33331 AP2 responsetoheat(GO:0009408);responsetowaterdeprivation(GO:0009414);responsetosaltstress(GO:0009651)
Pc37218 RAV no
Pc41893 ERF responsetocold(GO:0009409);responsetowounding(GO:0009611);responsetoabscisicacidstimulus(GO:0009737)
doi:10.1371/journal.pone.0108977.t004
Figure 9. qRT–PCR analysis of AP2/ERF genes in response to different abiotic stresses. Three bio-replicates and tech-replicates were
performed.Thedataarepresentedasthemean6SD.
doi:10.1371/journal.pone.0108977.g009
PLOSONE | www.plosone.org 9 September2014 | Volume 9 | Issue 9 | e108977
TranscriptomeSequenceandGenesRelatedtoAbioticStressinParsley
drought). The genes of Pc16182, Pc16943, Pc24931, and types. This result agrees with the findings in rice and peach but
Pc41893 were belonged to ERF subfamily. The genes of contradicts with the findings in bread wheat and Medicago
Pc32872 and Pc33331 were chose from AP2 subfamily. The truncatula,whereintri-nucleotideSSRswerefoundtobethemost
geneofPc37218wasselectedfromRAVsubfamily.Thosegenes frequent motif type [70–73]. A large number of short repeat
werealsopredictedtoresponsetoabioticstressbyGoannotation sequences are also considered a relatively rapid rate of evolution
(Table 4).TheexpressionlevelsofAP2/ERFgeneswereanalyzed [74,75]. Parsley contains a large number of short repeat motifs.
under different abiotic stresses. Wepredictedthatparsleymaybelocatedonarelativelyhighlevel
As shown in Figure 9, all genes showed sensitivity to cold of biological evolution. The formation of mono-, di-, and tri-
treatment.TheexpressionlevelofPc24931rapidlydecreasedand nucleotide repeat types principally comprised A and T repeat
remained low, whereas those of the other genes increased and elements,indicatingastrongbasepreference.Thispreferencemay
peakedafter8or24h.Pc37218andPc41893wereup-regulated be due to the methylation of C residues, which may result in
by more than 21 and 18 fold, respectively. Pc16182, Pc24931, conversion toT[76].
and Pc41893 were obviously down-regulated under heat stress. Transcription factors have received more attention from
Pc32872, Pc33331, and Pc37218 were initially down-regulated scholarsthatconductedwhole-genomicsequencingandtranscrip-
andthenup-regulated.Bycontrast,Pc16943wasup-regulatedby tomesequencing.Drought,highsalinity,andextremetemperature
9foldin2 handthenwasrapidlydown-regulated.Undersalinity are key factors that contribute to crop failure. Previous studies
stress, Pc16182, Pc16943, Pc24931, and Pc41893 exhibited have shown that AP2/ERF transcription factors are related to
minimal or no change in relative expression, but the other four
plantstressresponse[22,77].TworiceERFgenes,OsERF4aand
genes increased and showed different levels of sensitivity to salt
OsERF10a, confer drought stress tolerance [78]. Some studies
stress. Under drought treatment, Pc24931, Pc32872, Pc33331,
showed that the AP2 and RAV subfamilies respond tostress and
and Pc41893 were up-regulated, whereas the other genes
hormone signals [79,80]. In the present study, 88 AP2/ERF
exhibited no significant change. On the whole, the result was
transcriptionfactorswereidentifiedbaseonparsleytranscriptome
consistent withtheannotations.
sequence. We explored the expression levels of AP2/ERF family
members belonging to different subfamilies under stress treat-
Discussion
ments. All selected genes showed different levels of sensitivity to
stresses,includingPc37218,whichwasnotannotatedtoresponse
Currentdataonthemolecularandgeneticpropertiesofspecies
to stress. Some genes from the same subfamily differed in
in the Apiaceae family are insufficient. Only a few transcriptome
databaseshavebeenestablishedforcelery[29,30]andcarrot[31] expression. Plant stress tolerance is controlled by multiple genes,
in recent years. The lack of reference genomic data has limited andfurtherstudiesarerequiredtoidentifythecomplexregulatory
parsley research. Transcriptome sequencing is a feasible and networks of these AP2/ERFgenes inparsley.
economical technology for creating relatively comprehensive
sequence data in a short time; this technology has become Supporting Information
popular in plant research [35–38]. In the present study, 50,161
Figure S1 Phenotypes of Petroselinum crispum and
unigeneswereassembled.Theobtainedsequencedatacouldserve
Apium graveolens.
asabasisforfurtherstudiesongenecloning,expressionanalysis,
(TIF)
and SSRmarkers.
Morethan60%(31,658of50,161)unigeneswereannotatedby TableS1 qRT–PCR primer sequences andthesubfam-
sequence similarity search in seven public databases. Functional ily of selected AP2/ERFgenes.
annotation can suggest potential gene functions. In our study, (XLS)
qRT-PCR analysis showed that the gene functional annotations
TableS2 Geneannotation byseven publicdatabases.
werereliable.Othersunigenes(approximately40%)weretooshort
(XLS)
to be annotated. The percentages of the unannotated unigenes
were similar to those in rice [62] and tea [63]. Long splicing Table S3 Gene list in each family of transcription
sequences are among the prerequisites for reliable functional factors.
annotation.Theinsufficientinformationonthegenomeandgene (XLS)
functionsintheApiaceaefamilyandthesmallnumberofspecies
TableS4 Transcript expression level of all unigenes in
for referencinghave resulted inlimitedfunctional annotation.
thecurlyparsleylibrary.Gene-ID:geneIDnumber;Length:
Molecular marker techniques, such as restriction fragment
gene length; Depth: the average depth of gene; Coverage:
lengthpolymorphism,random-amplifiedpolymorphicDNA,SNP,
coverage of genes; RPKM (reads per kilobase of exon model per
and SSR, can be used in genetic diversity analysis. SSR markers
millionmappedreads):abundanceofgeneexpression;Totalreads:
are commonly used in genetic linkage map construction and
molecular-assisted breeding because of their good repeatability, the number of reads that hit other genes; Unique reads: the
high reliability, easy operation, and high polymorphism [64,65]. numberofreadsthathitonlyonereferencegene;Multireads:the
To our knowledge, this study is the first to report on the SSR number of reads that hitmultiple locations.
markers in parsley. The SSRs motif types, especially the most (XLS)
abundantrepeats,werecontributedtotheevolutionofgenomesin
variousorganisms[66].Severalresearchstudieshavedocumented Author Contributions
GTisthemostcommontypeinanimalandinvertebrates,whereas
Conceived and designed the experiments: ASX MYL. Performed the
CT and AT are the most common repeats in plants and insects experiments:MYLHWTFWZSXQJCTASX.Analyzedthedata:MYL
[67–69].Inparsley,thedi-nucleotiderepeatcomprisedthelargest HWT ASX. Contributed reagents/materials/analysis tools: ASX. Con-
fraction,whileAG/CTandAC/GTarethemostabundantmotif tributedtothewritingofthemanuscript:MYL.
PLOSONE | www.plosone.org 10 September2014 | Volume 9 | Issue 9 | e108977