Large-Scale Analyses of Angiosperm Nucleotide-Binding Site-Leucine-Rich Repeat Genes Reveal

15 Pages·2016·2.16 MB·English

Large-Scale Analyses of Angiosperm Nucleotide-Binding
Site-Leucine-Rich Repeat Genes Reveal Three Anciently
1
Diverged Classes with Distinct Evolutionary Patterns
Zhu-Qing Shao2, Jia-Yu Xue2, Ping Wu, Yan-Mei Zhang, Yue Wu, Yue-Yu Hang, Bin Wang*, and
Jian-Qun Chen*
State Key Laboratory of Pharmaceutical Biotechnology, School of Life Sciences, Nanjing University, Nanjing,
210023, China (Z.-Q.S., P.W., Y.-M.Z., Y.W., B.W., J.-Q.C.); and Institute of Botany, Jiangsu Province and
Chinese Academy of Sciences, Nanjing, 210014, China (J.-Y.X., Y.-M.Z., Y.-Y.H.)
Nucleotide-bindingsite-leucine-richrepeat(NBS-LRR)genesmakeupthelargestplantdiseaseresistancegenefamily(Rgenes),
with hundreds of copies occurring in individual angiosperm genomes. However, the expansion history of NBS-LRR genes
during angiosperm evolution is largely unknown. By identifying more than 6,000 NBS-LRR genes in 22 representative
angiosperms and reconstructing their phylogenies, we present a potential framework of NBS-LRR gene evolution in the
angiosperm. Three anciently diverged NBS-LRR classes (TNLs, CNLs, and RNLs) were distinguished with unique exon-
intron structures and DNA motif sequences. A total of seven ancient TNL, 14 CNL, and two RNL lineages were discovered
in the ancestral angiosperm, from which all current NBS-LRR gene repertoires were evolved. A pattern of gradual expansion
duringthefirst100millionyearsofevolutionoftheangiospermcladewasobservedforCNLs.TNLnumbersremainedstable
duringthisperiodbutwereeventuallydeletedinthreedivergentangiospermlineages.Weinferredthatanintenseexpansionof
bothTNLandCNLgenes startedfrom theCretaceous-Paleogene boundary.Becausedramaticenvironmental changesandan
explosion in fungal diversity occurred during this period, the observed expansions of R genes probably reflect convergent
adaptive responses of various angiosperm families. An ancient whole-genome duplication event that occurred in an
angiosperm ancestor resulted in two RNL lineages, which were conservatively evolved and acted as scaffold proteins for
defense signal transduction. Overall, the reconstructed framework of angiosperm NBS-LRR gene evolution in this study may
serveasa fundamentalreferenceforbetterunderstandingangiosperm NBS-LRRgenes.
Plantdiseaseresistance(R)genesareasetofgenesthat et al., 2006; Friedman and Baker, 2007). Since the first
conferresistancetovariouspathogens.Amongallknown comprehensive study on NBS genes in Arabidopsis
types of R genes, nucleotide-binding site-leucine-rich re- (Arabidopsisthaliana;Meyersetal.,2003),NBSgeneshave
peat(NBS-LRR,NBSforshort)genesrepresentthelargest been surveyed across various plant genomes, most of
class,encompassingmorethan80%ofthecharacterized which belong to the rosid lineage of eudicots and the
Rgenes(Meyersetal.,1999;Meyersetal.,2005;McHale Poaceae of monocots (Monosi et al., 2004; Zhou et al.,
2004; Yang et al., 2006; Ameline-Torregrosa et al., 2008;
1ThisworkwassupportedbytheNationalNaturalScienceFoun- Yang et al., 2008; Mun et al., 2009; Porter et al., 2009; Li
dationofChina(30930008,31170210,91231102,31300190,31400201, et al., 2010; Zhang et al., 2011; Lozano et al., 2012; Luo
31470327,31500191,and31570217),theNaturalScienceFoundation etal.,2012;Andolfoetal.,2013;Shaoetal.,2014).
ofJiangsuProvince(BK20130565),theFundamentalResearchFunds NBS genes are historically divided into two classes,
for the Central Universities (2062140215, 20620140546, and
namely, TIR-NBS-LRR (TNL) and the non-TIR-NBS-
20620140558),andtheNationalPostdoctoralScienceFoundationof
LRR (nTNL), which are differentiated by the presence
China(2013M540435and2014T70503).
2Theseauthorscontributedequallytothearticle. ofaToll/IL-1Receptor-like(TIR)domainintheprotein
amino terminus (Meyers et al., 1999; Bai et al., 2002;
*[email protected]@nju.
edu.cn. Zhouetal.,2004).BecausemostnTNLgenesencode a
Theauthorresponsiblefordistributionofmaterialsintegraltothe coiled-coil (CC) domain at the N terminus, the nTNL
findingspresented in this article in accordancewith the policy de- genes often are called CC-NBS-LRR (CNL) genes
scribed in the Instructions for Authors (www.plantphysiol.org) is: (Meyers et al., 2003; Ameline-Torregrosa et al., 2008).
Jian-QunChen([email protected]). However,recentstudieshaverevealedthatapartfrom
J.-Q.C.,B.W.,J.-Y.X.,andZ.-Q.S.conceivedanddesignedthepro- CNLgenes,asmallgroupofnTNLgenesthatpossessa
ject;J.-Y.X.,Z.-Q.S.,andB.W.obtainedandanalyzedthedata;P.W.,
special N-terminal domain, RPW8 (resistance to pow-
Y.-M.Z.,Y.W.,andY.-Y.H.participatedinthedataanalysisanddis-
derymildew8)domain,likelyrepresentadistinctclass
cussion;Z.-Q.S.andJ.-Y.X.draftedtheinitialmanuscript;B.W.com-
ofNBSgenes(RPW8-NBS-LRR[RNL];Xiaoetal.,2001;
plementedthewriting;J.-Q.C.supervisedtheanalysisandwriting;
Cannon et al., 2004; Bonardi et al., 2011; Collier et al.,
allauthorscontributedtodiscussionoftheresults,reviewedtheman-
uscript,andapprovedthefinalarticle. 2011; Shao et al., 2014; Zhang et al., 2016). Investiga-
www.plantphysiol.org/cgi/doi/10.1104/pp.15.01487 tionsofNBSgenesintheFabaceae(rosidIlineage)and
Plant Physiology(cid:1), April 2016, Vol. 170, pp. 2095–2109, www.plantphysiol.org (cid:3)2016AmericanSociety ofPlant Biologists. All RightsReserved. 2095
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Shao et al.
Brassicaceae (rosid II lineage) have shown that RNLs providewhenandhowresistancelociemergedordis-
andCNLsdivergedpriortothedivergenceofthesetwo appearedincertainlineages.Inthisstudy,atotalof22
families(Shaoetal.,2014;Zhang etal.,2016).Interest- plant species (Fig. 1) from key systematic clades in
ingly,althoughbothCNLandRNLgenesarepresentin eudicots, monocots, and the early diverging basal an-
monocots and dicots, TNL genes only are present in giosperm were chosen to depict the ancientevolution-
dicots (Meyers et al., 2003; Tarr and Alexander, 2009; ary history of NBS genes in angiosperms. More than
Andolfoetal.,2013;Shaoetal.,2014).Consideringthe 6,000 NBS genes were identified; family level phylo-
abundance of TNLs in various dicot species, their ab- genetictreeswereconstructedfirstandthenexpanded
senceinmonocotsisperplexing. to higher systematic levels to eventually analyze all of
For several years, our knowledge of NBS gene evo- the angiosperms. Ancient taxonomies of NBS genes
lution has been based on investigations of individual weretracedtotheoriginofangiosperms,andatotalof
angiosperm genomes (Meyers et al., 2003; Ameline- 23 ancestral NBS genes were recovered in the last
Torregrosa et al., 2008; Kohler et al., 2008). The num- common ancestor of angiosperms. This angiosperm
ber of NBS genes varied dramatically among different NBS frame provides a fundamental reference for un-
species, even among different accessions of the same derstanding the evolutionary history of NBS genes as
species (Zhang et al., 2010). However, the relative well as insight into the finer details of particular NBS
contributions of recent and ancient expansions to the geneclasses.
abundance of NBS genes in modern plant genomes
remain elusive. Recent studies report the evolutionary
analysesofNBSgenesacrossmultiplespeciesinaplant RESULTS
family(Cannonetal.,2002;Guoetal.,2011;Luoetal.,
2012; Lin et al., 2013; Shao et al., 2014; Zhang et al., Identification andClassificationofNBS Genesfrom
2016). While studies in Brassicaceae and Poaceae both 22AngiospermSpecies
revealed abundant and recent losses of NBS genes in A total of 6,135 NBS genes were identified in 22 an-
variousspeciesafterthedivergenceofthesetwofami-
giosperm species using previously described proce-
lies, a study involvingFabaceae hassuggested theop-
dures(Shaoetal.,2014;Zhangetal.,2016).Fiveofthe
posite scenario in which all surveyed species recently
22 genomes were surveyed in this study, Amborella
underwent a multitude of NBS expansions (Luo et al.,
trichopoda, Musa acuminata (banana), Phyllostachys het-
2012; Shao et al., 2014; Zhang et al., 2016). Although
erocycla(bamboo),Capsicumannuum(hotpepper),and
distinct evolutionarypatternsof NBS geneshave been
Sesamum indicum (sesame), whereas the others have
identified at the family level, hundreds of NBS gene
beenpreviouslyreported(Meyersetal.,2003;Luoetal.,
lineagesaroseintheancestorsofFabaceae,Brassicaceae,
2012; Malacarne et al.,2012; Andolfo et al., 2013; Shao
and Poaceae, indicating that NBS genes may have
et al., 2014; Zhang et al., 2016). A. trichopoda, a basal
probably undergone ancient expansions prior to fa-
angiosperm,hadatotalof105NBSgenes,including15
milial divergence. Therefore, a phylogenetic analysis
TNLs, one RNL, and 89 CNLs (Fig. 1; Supplemental
at a larger scale is necessary to better understand the
TableS1).TheNBSgenesvariedamongthe22surveyed
expansionhistoryofNBSgenesinangiosperms.
species(Fig.1).Ofallspeciessurveyed,Medicagotrun-
Several functional NBS genes have been identified in
catulapossessedthehighestnumberofNBSgenes(571),
angiospermspecies(Liuetal.,2007);however,theirtime
more than 6-fold of that observed in Thellungiella sal-
of emergence and relationship with other NBS genes
suginea (88). RNL genes were detected in all of the
have not been established. For example, several CNL surveyed species, although at lower numbers (1–18).
geneshavebeenidentifiedinbothsoybean(Glycinemax;
A large proportion of NBS genes were either TNLs or
Rpg1b and Rpg1r) and Arabidopsis (RPM1 and RPS2)
CNLs,althoughTNLswerenotdetectedineightof22
against the bacterial pathogen Pseudomonas syringae by
species, including seven monocots and S. indicum. Be-
guardingthehostproteinRIN4(Bentetal.,1994;Mackey
causethebasalangiospermA.trichopodaharboredTNL
et al., 2002; Mackey et al., 2003; Ashfield et al., 2014).
genes,thesegenesweremostlikelyalsopresentinthe
While the Rpg1b and Rpg1r genes were deduced to
ancestorofangiospermsbutindependentlylostincer-
emerge after the divergence of Fabaceae species, the
tainlineages.InspecieswithbothTNLandCNLgenes,
evolutionarytimepointoftheoriginofRPM1andRPS2
CNLs often outnumbered TNLs, except for the four
genesinArabidopsisislargelyunknown(Ashfieldetal.,
Brassicaceae genomes analyzed (Zhang et al., 2016).
2014; Shao et al., 2014). Because all these R genes have
The asterid genomes showed a lower ratio of TNL to
likely adopted a similar strategy to defend against the
CNL genes, with the C. annuum exhibiting the highest
samepathogen,investigatingtheirevolutionmayeluci-
levelofdisparity(TNL:CNL=1:18).
date the mechanism underlying the establishment and
maintenanceofresistanceinvariousplantspecies.
To obtain a more comprehensive understanding of
Phylogenetic AnalysisofNBSGenesinMajor
NBS gene evolution, it is essential to analyze more
AngiospermClades
species/lineages to determine how the current NBS
genes have evolved and diverged from ancestral NBS The 22 angiosperm species surveyed in this study
lineages in the ancestral angiosperm. This also will occupy important phylogenetic positions (Fig. 1). To
2096 Plant Physiol. Vol. 170, 2016
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Evolution of NBS-LRR Genes in Angiosperms
Figure1. Identification andclassification
ofNBSgenesin22angiospermgenomes.
The phylogenetic relationship was con-
structed according to the APG III system
(Bremeretal.,2009).Thedivergencetimes
at different nodes of angiosperms were
combined from previous studies (Lavin
et al., 2005; Stefanovic et al., 2009;
D’Hont et al., 2012; Zhang et al., 2012;
Pengetal.,2013;Yangetal.,2013;Kim
etal.,2014;Wangetal.,2014;Zengetal.,
2014).ThetotalnumberofNBSgenesand
their classification in each species are
shown. The number of ancestral NBS
genesrecoveredbyphylogeneticanalysis
at each node is shown in a green circle,
andanasteriskafterthenumberindicates
thatitisnotdirectlyunraveledbyphylo-
geneticanalysis.
tracetheevolutionaryhistoryofNBSgenes,weexam- lineage designated as Mo-R1, and 56 CNL lineages
inedmultipleangiospermclades,includingrosids(Ro), identified as Mo-C1 to Mo-C56), suggesting that nTNL
asterids (As), eudicots (Ed), monocots (Mo), mesan- genes from the seven monocot genomes were actually
giospermae(Ma),andallangiosperms(An;Fig.1).Itis derivedfrom57ancestralMo-NBSgenes;110MYA(Fig.
widely recognized that the NBS domain effectively 2;SupplementalFig.S1;D’Hontetal.,2012).Fromtheir
distinguishes NBS genes into TNL and nTNL classes commonancestor,thePoaceaeandM.acuminatainherited
(Bai et al., 2002; Meyers et al., 2003). Although a sister 51and32ancestralMo-NBSgenes,respectively.Among
relationship between CNLs and RNLs has been estab- these,26weresharedbybothPoaceaeandM.acuminata.
lishedinrosids,ithasnotbeenexaminedatangiosperm SeveralancestralgenesinthePoaceaeunderwentexten-
level. Therefore, phylogenetic analyses of TNLs and siveduplications(from51to456)duringthetimeinterval
nTNLs (including both CNLs and RNLs) were sepa- of 48 million years. For example, the NBS gene Mo-C56
rately performed in this study. We first conducted duplicated into 197 genes during this time period
phylogenetic analysis of three major angiosperm line- (SupplementalFig.S1).Asmultiplespeciesradiatedfrom
ages, namely, monocots, asterids, and rosids, which thePoaceaecommonancestor,insomespeciesNBSgene
coverseven,four,and10genomes,respectively. numbers contracted drastically (Z. mays, 139 from 456)
The M. acuminata genome, together with six Poaceae andinothersonlyslightly(S.italicaandP.heterocycla,424
genomes,includingOryzasativa,Brachypodiumdistachyon, and344from456,respectively),whereasexpansionwas
P.heterocycla,Zeamays,Sorghumbicolor,andSetariaitalica, onlyfoundintheO.sativagenome(from456to498).
were used to elucidate the evolution of NBS genes in Within the asterids four species were analyzed, in-
monocots. Because these species did not harbor TNL cluding three Solanaceae species, namely, Solanum
genes, phylogenetic analysis of only nTNL genes was lycopersicum,Solanumtuberosum,andC.annuum,aswell
performed.ThereconstructedmonocotnTNLphylogeny as S. indicum from Pedaliaceae. NBS genes from the
(SupplementalFig.S1)revealedthatnTNLgenesfromthe three Solanaceae genomes formed 19 TNL monophy-
six Poaceae species formed 456 monophyletic lineages, letic lineages (Supplemental Fig. S2) and 147 nTNL
indicating the presence of at least 456 ancestral nTNL monophyletic lineages (Supplemental Fig. S3), sug-
genesinthePoaceaecommonancestor(;62millionyears gestingthatatleast166ancestralNBSgeneswerepre-
ago [MYA]; Peng et al., 2013, Zhang et al., 2012). These sent in their last common ancestor. The Solanaceae
Poaceae nTNL genes formed 57 monophyletic lineages nTNL genes formed 55 monophyletic lineages with
whengroupedwithM.acuminatanTNLgenes(oneRNL nTNL genes from S. indicum (two RNL lineages
Plant Physiol. Vol. 170, 2016 2097
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Shao et al.
Figure 2. Cladograms depicting reconciled phylogenetic relationships of ancestral NBS genes in the divergence nodes of
monocots,rosids,andasterids.ThepresenceofaNBSgenelineageindifferentspeciesandfamiliesareindicatedwithspeciesor
lineage-specificicons.ThemaximumlikelihoodtreesofeachcladogramarepresentedinSupplementalFiguresS1toS5.
designated as As-R1 and As-R2, and 53 CNL lineages Rosidsareamajorlineageofeudicotsthatconsistsof
identifiedasAs-C1toAs-C53),indicatingaminimumof three sublineages, rosid I (fabids), rosid II (malvids),
55ancestralgenesintheasteridcommonancestor(Fig.2; and the basal rosids. Previous work analyzed NBS
SupplementalFig.S2).SolanalesinheritedmoreAs-nTNL genes in Fabaceae and Brassicaceae (Shao et al., 2014;
lineages than S. indicum, which represents Lamiales (47 Zhang et al., 2016), which belong to the rosid I sub-
versus30).NoTNLgenesweredetectedintheS.indicum lineageandtherosidIIsublineage,respectively.Atotal
genome;therefore,theaccuratenumberofancestralTNL of 64 nTNL and 55 TNL ancestral lineages were dis-
genesatthisasteridnodewasnotestimated. covered in the last common ancestor of the four
2098 Plant Physiol. Vol. 170, 2016
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Evolution of NBS-LRR Genes in Angiosperms
Fabaceaespecies,and80nTNLand148TNLancestral TNLs (Supplemental Fig. S10). Figure 3 shows that all
lineageswereidentifiedinthelastcommonancestorof TNL genes clustered into seven ancestral angiosperm
the five surveyed Brassicaceae species. In this study, lineages (designated as An-T1 to An-T7), and all nTNL
one representative sequence from each of the 119 genesclusteredinto16ancestrallineages(twoRNLlin-
Fabaceae ancestral NBS lineages and the 228 Brassica- eagesdesignatedasAn-R1andAn-R2,and14CNLlin-
ceaeancestralNBSlineageswereusedtorepresentthe eagesidentifiedasAn-C1toAn-C14),indicatingatleast
diversity of ancient NBS genes in these two families, 23 NBS genes were present in the last angiosperm an-
and then combined with NBS genes identified in Vitis cestor(;225 MYA). ThebasalangiospermA. trichopoda
vinifera (basal rosid) to construct NBS phylogenies for inheritedonlyeightofthe23An-NBSgenes,whereasthe
thewholerosidlineage.Atotalof61ancestralRo-NBS mesangiospermae inherited 19 (Fig. 3).The angiosperm
genes were identified (four RNLs, 49 CNLs, and eight nTNLtreewascomposedoftwo monophyleticgroups,
TNLs)inthecommonancestoroftherosids(;108MYA; namely, RNL and CNL (Supplemental Fig. S9). There-
Fig.2;SupplementalFigs.S4andS5),amongwhichthe fore,despitetheextensivediscrepancyingenenumbers
Fabaceae of rosid I lineage inherited 35 Ro-NBS genes between the two groups, this study revealed that the
(two RNLs, 27 CNLs, and six TNLs), Brassicaceae of RNLisanotherancientclassofNBSgenesthathasasister
rosid II lineage inherited 19 (two RNLs, 12 CNLs, and relationshipwiththeCNLclassinangiosperms.
five TNLs), and basal rosids inherited 33 (two RNLs, The angiosperm NBS gene phylogenies have
29CNLs,andtwoTNLs).LessthanhalfoftheRo-NBS revealedtheevolutionarystateofNBSgenesatcritical
genes(26/61)weredetectedintwoofthelineages,and angiosperm nodes, which indicated that differential
35/61wereobservedinonedescendinglineage(Fig.2), inheritanceofancestralNBSgenesandlineage-specific
thereby reflecting a pattern of frequent NBS gene loss expansionsaremajorthemesofNBSgeneevolution.
after their split from the common rosid ancestor. The
35 and 19 Ro-NBS genes inherited by Fabaceae and CharacterizationofClass-SpecificSignaturesamongThree
Brassicaceaelineagesexpandedto119(55TNLsand64
AncientNBS GeneClasses
nTNLs)and228(148TNLsand80nTNLs)NBSgenesin
theFabaceaeancestor(;54MYA)andtheBrassicaceae Because the results of phylogenetic analysis suggest
ancestor (;42 MYA), respectively (Shao et al., 2014; that RNLs represent an ancient NBS gene class inde-
Zhangetal.,2016).TheNBSgenescontinuedtoexpand pendentofthe CNL and TNLclasses,characterization
in Fabaceae (Shao et al., 2014), whereas these genes ofdistinctivefeaturesofeachclasswasperformed.
contracted as the Brassicaceae diverged into multiple
species (Zhang etal.,2016).AlthoughtheBrassicaceae
Distinctive Signatures in Conserved NBS Motifs Differentiate
inheritedfewerancestralTNLsthanCNLs(fiveversus
theThreeClasses
12)fromtherosidancestor,thislineagevigorouslyex-
pandeditsTNLsduringthetimeintervalof;108MYA The NBS domain is composed of a number of char-
to;42MYA.A30-foldexpansionwasobservedinthe acteristic motifs that are conserved and shared by all
BrassicaceaeTNLs(fromfiveto148),farbeyondthatof NBS genes, whereas other motifs are more variable
legumeTNLs(;9-fold,fromsixto55;Shaoetal.,2014; (DeYoung and Innes, 2006). All three classes of NBS
Zhangetal.,2016). genessharefivefunctionallyimportantmotifs:P-loop,
Kinase-2,RNBS-B,GLPL,andRNBS-D.Theaminoacid
sequence conservation of these five motifs was esti-
ReconstructionofNBSGenePhylogeniesatthe
mated using sequence logos using MEME software
Whole-AngiospermScale
(Fig.4;BaileyandElkan,1995).Allmotifswerehighly
RepresentativeNBSsequencesfromancestrallineages conservedatmostsitesamongallclasses,exceptforthe
ofasterids,rosids,andmonocotswereusedtoconstruct RNBS-Dmotif,whichwaspoorlyconservedintheTNL
ancestral NBS gene lineages at the divergence nodes of and CNL classes. The class-specific features within
eudicotsandmesangiospermae.AtotalofeightEd-TNL these motifs also were identified. The Kinase-2 motif
and49Ed-nTNLgenes(47Ed-CNLsandtwoEd-RNLs) had a conserved “DDVW” sequence in the RNL and
were recovered at the eudicot node (Supplemental Figs. CNLgenes,butfrequentlyappearedas“DDVD”inthe
S6andS7),whereas37Ma-nTNLgenes(35Ma-CNLsand TNL genes. A“TSR”sequence in the RNBS-B motif of
two Ma-RNLs) were detected at the mesangiospermae RNLs was conserved, but frequently appeared as
node (Supplemental Fig. S8). Independent loss and dif- “TTR” and “TTRD” in the CNL and TNL genes, re-
ferential inheritance of ancestral NBS genes were also spectively.CNLdidnothavespecifichighlyconserved
observedatthesetwonodes.Representativegenesfrom motifs, but could be distinguished by a RNL like
each ancestral Ma-nTNL lineage were compiled with Kinase-2motifandaTNLlikeRNBS-Bmotif.
A. trichopoda nTNL genes to reconstruct angiosperm
nTNL phylogeny (Supplemental Fig. S9). Because a
Intron-ExonStructuresAlso DistinguishNBSClasses
mesangiospermae TNL phylogeny could not be con-
structedduetotheabsenceofTNLgenesinmonocots,a Theabsence/presence,position,andphaseofintrons
TNL phylogeny wasgeneratedattheangiospermlevel often are regarded as important parameters in phy-
using eudicot representative TNLs and A. trichopoda logeny(Qiuetal.,1998;Sánchezetal.,2003;Qianetal.,
Plant Physiol. Vol. 170, 2016 2099
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Shao et al.
Figure3. CladogramsdepictingreconciledphylogeneticrelationshipsofancestralNBSgenesinthecommonancestorofan-
giosperms.ThemaximumlikelihoodtreesofeachcladogramareprovidedinSupplementalFiguresS9andS10.
2015). In a previous study, the intron positions and unraveled the ancient NBS gene lineages at multiple
phases were determined to be distinctive among the angiospermevolutionarynodes(Fig.5).Duetotheloss
classesofNBSgenesinbryophytes(Xueetal.,2012).In ofTNLgenesinS.indicumandthemonocotlineage,the
this study, three classes of NBS genes in angiosperms TNLancestralstatesattheasteridandmesangiospermae
alsodifferedinintron characteristics (Fig.4).Threein- nodescouldnotbedirectlyobtainedfromphylogenetic
tronswithconservedpositionsandphaseswereshared analyses. However, based on the recovered ancestral
bymostTNLgenes,withthefirstintronseparatingthe NBS lineages at the divergence nodes of Solanaceae
TIRdomainandtheNBSdomain(phase2),thesecond (Supplemental Fig. S2), dicots (Supplemental Fig. S6),
intronseparatingtheNBSdomainandthewholeLRR and all angiosperms (Supplemental Fig. S10), both
domain (phase 0), and the third intron separating the asterid and mesangiospermae nodes were deduced to
first and the remaining LRR domains (phase 0). An- haveatleastfiveTNLlineages.
giosperm RNLs share four class-specific introns, with Abovethefamilydivergencenodes(An,Ma,Mo,Ed,
the exception of RNLs in monocots, which did not As, Ro, and Er), CNL genes were abundant but only
possess the second intron (Fig. 4). Because the second slightlymorethanthoseofearlierancestralnodes(Fig.
intron is present in both A. trichopoda and dicot RNL 5). This finding indicates that CNL genes adopted a
genes, intron loss likely occurred in the monocot line- gradualexpansionstrategyintheearlystageofangio-
age. For the CNL class, conserved introns were not sperm evolution (from ;225 MYA as angiosperms di-
detectedacrossallCNLlineages,especiallyinthebasal verged to around 100 MYA as asterids, rosids, and
lineages. In fact, a number of CNL genes were intron- monocots diverged; Fig. 1; D’Hont et al., 2012; Wang
less,andwerenotconcentratedinaspecificlineagebut et al., 2014; Zeng et al., 2014). Intensive expansion of
distributedacrossallancestralangiospermnTNLline- CNL genes seems to have occurred in the familial an-
ages. Although some CNL genes had introns, these cestors of Solanaceae and Poaceae, reaching hundreds
were more likely to have been gained at later evolu- of copies. TNL genes apparently followed a different
tionary stages. This finding is in agreement with the evolutionary path, as these were completely lost in
intronlessstructurereportedinbryophytes (Xue etal., monocotsandinS.indicum.Althoughinheritedbydi-
2012). cots, TNL numbers remained relatively constant for a
long evolutionary period, with similar numbers of an-
cestralTNLgenesidentifiedatvariousancestralnodes
TheThree NBSClassesShowDistinct Evolutionary
(An, Ma, Ed, As, Ro, and Er; Fig. 5). The expansion of
Patterns inAngiosperms
TNLs mainly occurred prior to the divergence of fam-
Our step-by-step phylogenetic analysis not only ilies, such as in the Fabaceae ancestor, 55 TNLs were
supports the presence of three distinct NBS classes expandedfromsixofeightRo-TNLs.Asimilarpattern
in the common ancestor of angiosperms, but also was observed in the Brassicaceae ancestor, in which
2100 Plant Physiol. Vol. 170, 2016
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Evolution of NBS-LRR Genes in Angiosperms
Figure4. Motifandexon-intronstructuredifferencesamongthethreeNBSclasses.Top,Theconservationofaminoacidse-
quencesatfivemotifsintheNBSdomainisshown.Bottom,Proposedancientexon-intronstructureofTNL,CNL,andRNL
classes.
fourRo-TNLlineagesexpandingto148genes.TheTNL chromosome fragments containing ADR1 or NRG1
geneswereprobablymaintainedatalowernumberfor orthologs were conserved in angiosperm genomes. Fig-
more than 100 million years (from ;225 MYA as an- ure 6 (detailed in Supplemental Table S2) shows the
giosperm diverged to around 100 MYA as rosids and syntenybetweenADR1-andNRG1-containingchromo-
asteridsdiverged)andonlybegantoexpandthereafter. somal regions between the dicot genomes of Phaseolus
This may explain why TNL genes were easily lost in vulgaris (rosid lineage) and S. lycopersicum (asterid line-
multipleindependentangiospermlineages. age). The syntenic blocks were also detected in the
RNL genes are differentiated from the other two monocotgenomeofS.bicolor,intheabsenceoftheNRG1
classes by their relatively conserved evolutionary pat- gene. Moreover, most RNL flanking gene pairs con-
tern. Nearly all species from basal angiosperm, mono- servedin theduplicatedblocks within the threespecies
cots,andasteridshaveoneortwoRNLgenes,whereas havehigh Ksvalues(.1; Supplemental Table S2), indi-
a slight expansion of RNL genes was observed in cer- cating thatthesegenesmay haveresultedfrom onean-
tain rosid genomes (Fig. 1). Tracing the evolutionary cient WGD event. Because the most recent WGD event
history of RNL genes revealed two anciently diverged sharedbymonocotsanddicotsoccurredinthecommon
lineages in the common ancestor of angiosperms (Fig. ancestor of all angiosperms (Jiao et al., 2011; Vanneste
6). One RNL lineage contained the Arabidopsis ADR1 et al., 2014), the syntenic data and phylogenetic results
gene,whereastheotherincludedthetobacco(Nicotiana support that RNL genes separated into two lineages
benthamiana) NRG1 orthologs. Both lineages were in- sometimebeforethistimepoint.
herited by dicot species, except for S. indicum, which
losttheNRG1lineage,similartoallmonocotgenomes.
DecipheringtheEvolutionofArabidopsisNBSGenes
Several whole-genome duplications (WGDs) occurred
throughout angiosperm evolution, causing gene du- By performing phylogenetic analyses at different
plications and functional divergence (Jiao et al., 2011; evolutionary nodes, we have established a framework
Vannesteetal.,2014).TheseparationoftheADR1and for the evolution of angiosperm NBS genes, which
NRG1lineagesprior tothedivergence ofangiosperms could help decipher the evolution of NBS genes in
prompted us to investigate the involvement of an an- specific species as well as gain a much deeper view of
cient WGD event. The syntenic relationships between theevolutionoffunctionallycharacterizedNBSgenes.
Plant Physiol. Vol. 170, 2016 2101
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Shao et al.
Figure5. Comparisonoftheevolutionary
patternsofCNLandTNLclassesinangio-
sperms.A,ReconstructionoftheNBSgene
loss and gain events during angiosperm
evolution. The number of ancestral NBS
genesateachnodeisdepictedinasolid
circle. The thickness of each terminal
branch is roughly proportional to the
numberofNBSgenesinthegenomeorthe
ancestralgenesinthefamily.Genelossand
gaineventsatdifferentevolutionarystates
areindicatedbyminusandplussymbols,
respectively.Anasteriskafterthenumber
indicates that the number is not directly
recovered from phylogenetic analysis. B,
Extensive expansion of CNL and TNL
classesattheK-Pboundary.Thenumbers
ofCNL(redcircles)andTNL(bluecircles)
genesatdifferentdivergencenodesabove/
atfamilylevel,includingallangiosperms
(An), mesangiospermae (Ma), monocots
(Mo), eudicots (Ed), rosids (Ro), asterids
(As),Poaceae(P),Solanaceae(S),Fabaceae
(F), andBrassicaceae(B), wereplottedto
revealtheirexpansionhistory.
Intotal,fiveoutof14An-CNLgeneswereinherited detected in orthologs of Brassicaceae species. The first
by the Arabidopsis genome (Fig. 7). Gradual duplica- intron also was found in orthologous genes of the le-
tions occurred in four of these An-CNLs (An-C2, -C5, gume family, grape, and the asterid species, but they
-C6, and C8), with An-C5 and An-C6 expanding to a were absent in monocots and A. trichopoda. Therefore,
few copies. Two An-CNLs (An-C2 and An-C8) pro- theexon-intronstructuredetectedintheseArabidopsis
duced more than 80% of Arabidopsis CNL genes, CNL genes gradually formed from two independent
mainlyoccurringpriortothedivergenceoftheBrassi- intron gain events, one in the common ancestor of
caceaefamily.OneCNLgene(At3g07040)represented eudicots andtheother acquiredinthecommonances-
anancientgene(An-C14)thatwasconservedfromthe torofBrassicaceae(Fig.7).
angiosperm ancestor to Arabidopsis without duplica- Furthermore, the angiosperm evolutionary frame-
tion. Together, the evolutionary paths of the five an- work of NBS genes enabled us to determine when a
cestralCNLgenesresultedinthecurrentCNLprofileof functional NBS gene likely arose during angiosperm
Arabidopsisgenome(Fig.7). evolution.FivefunctionalCNLgenesagainstdifferent
The structural changes were also traced using the pathogens have been identified in Arabidopsis (Liu
evolutionaryframeworkofangiospermNBSgenes.For et al., 2007). According to the recovered evolutionary
example,nearlyallArabidopsisCNLgenesshowedan history of Arabidopsis CNL genes in Figure 7, RPM1
intronless structure regardless of their origin (Meyers wasthemostancientoneamongthefivegenes,possi-
etal.,2003),whichisinaccordancewithourproposed bly originating from the common ancestor of angio-
intronless structure for the ancestral angiosperm CNL sperms, whereas the RPP8/HRT was most recent,
gene. However, a group of Arabidopsis CNL genes possibly originating after Arabidopsis speciation. The
were determined to possess two shared introns. All origin of the other three genes could be traced to the
these genes originated from a lineage that diverged divergence nodes of rosids, Brassicaceae, and Arabi-
from An-C8 (Fig. 7) in the common ancestor of dopsis.Amongthefivegenes,RPS2,RPS5,andRPM1
mesangiospermae, with the sister lineage remaining were R genes against P. syringae, whereas RPP8/HRT
intronless. Based on the phylogenies of CNLs at dif- and RPP13 were both against Peronospora parasitica,
ferenttaxonomiclevels,thetwointronswerefrequently indicating functional R genes continuously arose
2102 Plant Physiol. Vol. 170, 2016
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Evolution of NBS-LRR Genes in Angiosperms
Figure6. EvolutionaryhistoryofRNLgenesinferredfromphylogeneticandsyntenicanalyses.Top,Theinheritanceofthetwo
RNLlineagesindifferentangiospermlineages.Theinferredduplicationismarkedwithagreencircle,whereasinferredlossesare
indicatedbyblackcircles.Bottom,Intra-andintergenomesyntenicblockscontainingRNLgenesdetectedwithinandamong
P.vulgaris,S.lycopersicum,andS.bicolorgenomes.
Plant Physiol. Vol. 170, 2016 2103
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Shao et al.
Figure7. AfewancestralangiospermNBSgenesthatareresponsibleingeneratingthecurrentArabidopsisCNLrepertoirevia
gradualexpansion.TheancientstatesofArabidopsisatdifferentevolutionarynodeswereretrievedfromtheconstructedNBS
phylogenies.Offspringgeneratedfromthesameancestralgeneatdifferentdivergencenodesareindicatedbyboxesofthesame
color.Blacktrianglesindicatethebirthofknownfunctionalgenes.Grayarrowsindicateintrongainevents.Theexon-intron
structureofeachancestrallineageisshown.
against the same pathogen during plant evolution. theevolutionaryhistoryoftheNBSgeneinthisspecies.
Actually,apartfromthesegenes,manyotherfunctional Tofacilitatetheuseofthisframeworkintheanalysisof
genesmayhaveappearedinthehistoryandthenbeen NBSgenesinotherplantsthatwerenotincludedinthis
lostduringplantevolution.However,suchcasesoften study, we generated hidden Markov model (HMM)
couldnotbetracedduetotheirabsenceinmodelplant profiles (Supplemental Dataset S1) for the 23 ancestral
genomes. NBS lineages of angiosperms. Taken together, the an-
The adoption of the angiosperm NBS evolutionary giosperm NBS evolutionary framework established in
framework in Arabidopsis provides new insights into this study provides a fundamental reference for the
2104 Plant Physiol. Vol. 170, 2016
Downloaded from on April 1, 2019 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.