research article open access transcriptome profiling of … · 2020. 2. 13. · 481,226 (32.7 gb)...

RESEARCH ARTICLE Open Access

Transcriptome profiling of differentiallyexpressed genes in cytoplasmic male-sterileline and its fertility restorer line in pigeonpea (Cajanus cajan L)Swati Saxena1 Sarika Sahu2 Tanvi Kaila1 Deepti Nigam1 Pavan K Chaduvla1 A R Rao2 Sandhya Sanand1N K Singh1 and Kishor Gaikwad1

Abstract

Background Pigeon pea (Cajanus cajan L) is the sixth major legume crop widely cultivated in the Indian sub-continentAfrica and South-east Asia Cytoplasmic male-sterility (CMS) is the incompetence of flowering plants to produce viablepollens during anther development CMS has been extensively utilized for commercial hybrid seeds production in pigeonpea However the molecular basis governing CMS in pigeon pea remains unclear and undetermined In this studytranscriptome analysis for exploring differentially expressed genes (DEGs) between cytoplasmic male-sterile line(AKCMS11) and its fertility restorer line (AKPR303) was performed using Illumina paired-end sequencing

Results A total of 3167 DEGs were identified of which 1432 were up-regulated and 1390 were down-regulated inAKCMS11 in comparison to AKPR303 By querying all the 3167 DEGs against TAIR database 34 pigeon pea homologousgenes were identified few involved in pollen development (EMS1 MS1 ARF17) and encoding MYB and bHLHtranscription factors with lower expression in the sterile buds implying their possible role in pollen sterility Many of theseDEGs implicated in carbon metabolism tricarboxylic acid cycle (TCA) oxidative phosphorylation and elimination ofreactive oxygen species (ROS) showed reduced expression in the AKCMS11 (sterile) buds

Conclusion The comparative transcriptome findings suggest the potential role of these DEGs in pollen development orabortion pointing towards their involvement in cytoplasmic male-sterility in pigeon pea The candidate DEGs identified inthis investigation will be highly significant for further research as they could lend a comprehensive basis in unravellingthe molecular mechanism governing CMS in pigeon pea

Keywords Cytoplasmic male-sterility Differentially expressed genes Illumina sequencing Pigeon pea PollenTranscriptome analysis

BackgroundCytoplasmic male-sterility (CMS) is a maternally inheritedtrait in plants where they are unable to produce functionalpollens [1] It occurs due to unusual mitochondrial openreading frames (orfs) which are chimeric in structure andexpress proteins that restrict mitochondrial function andpollen development [2] Cytoplasmic male-sterility pheno-type is known to be restored by nuclear genes termed asthe restorer of fertility (Rf) and hence result in the

production of functional pollens in a plant with the aber-rant mitochondrial genome [3] CMSRf systems thereforegive a clear view of the interaction and conflicts betweennuclear and mitochondrial genome [4] Cytoplasmic male-sterile systems along with the availability of restorer of fer-tility have emerged as a reliable tool for enhancing the yieldof many crop plants by utilizing hybrid vigor or heterosis[5] Hybrid seed production technology primarily involvesthree breeding systems (a) CMS line carrying male-sterilecytoplasm and devoid of functional nuclear Rf genes (b)Maintainer line containing fertile cytoplasm along withsimilar Rf (c) Restorer line containing fertile cytoplasm

copy The Author(s) 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 40International License (httpcreativecommonsorglicensesby40) which permits unrestricted use distribution andreproduction in any medium provided you give appropriate credit to the original author(s) and the source provide a link tothe Creative Commons license and indicate if changes were made The Creative Commons Public Domain Dedication waiver(httpcreativecommonsorgpublicdomainzero10) applies to the data made available in this article unless otherwise stated

Correspondence kish2012nrcpborg1ICAR-National Institute for Plant Biotechnology New Delhi 110012 IndiaFull list of author information is available at the end of the article

Saxena et al BMC Plant Biology (2020) 2074 httpsdoiorg101186s12870-020-2284-y

along with functional dominant Rf gene(s) The F1 hybridsproduced possess dominant Rf gene to restores the male-fertility and the blend of nuclear genes from the sterile andfertile lines results in hybrid vigor [2] During the 1950sthe first CMS system used for hybrid corn production wasmaize CMS-Texas (CMS-T) system which substantiallyincreased the hybrid seed production and remarkably en-hanced the maize yield [6] Till date impressive advance-ment has been made in plants to comprehend themolecular basis of CMS and fertility restoration A numberof mitochondrial genes related to male-sterility have beenreported such as urf13 from maize [7] pcf associated withatp9 gene from petunia [8 9] and the radish atp8 of OguraCMS [10] However in pigeon pea the gene associated withcytoplasmic male-sterility has not been much exploredIn flowering plants anther development is a highly con-

trolled mechanism which demands proper differentiation ofthe sporogenous tissue and timely programmed microspo-rogenesis [11] The absence of male gametophyte (stamen)or defective development of anther disruption of the pollenmother cells (PMC) or tapetum cells abnormal develop-ment of pollen and failure of anther dehiscence leads topollen abortion [12 13] In all cases CMS due to the dys-functional mitochondrial genome is mainly responsible forthe disruption of pollen development [14] Previous studieson CMS provide definitive evidence that the organellemitochondria are critical for the tapetum and microsporesdevelopment [14] During distinct stages of pollen develop-ment the requirement of energy is extremely high andeven a slight defect in the mitochondrial function becomesdeleterious and results in pollen sterility [15ndash17] Previousstudies have characterized many genes playing a pivotal rolein anther development in Arabidopsis viz SPOROCYTE-LESS (SPL)NOZZEL (NZZ) [18ndash21] EXCESS MICROSPO-ROCYTES 1 (EMS1) [22 23] TAPETUM DETERMINANT1 (TPD1) [24] ABORTED MICROSPORES (AMS) andMALE STERILITY1 (MS1) [25ndash27] MYB gene family (AtMYB103) [28 29] RUPTURED POLLEN GRAIN1 (RPG1)CYTOCHROME P450 (CYP703A2) CYTOCHROME P450(CYP704B1) acyl-coA synthetase5 (acos5) LAP6 and LAP5[30ndash37] redundantly facilitate anther development and cal-lose synthase5 (cals5) is essential for callose synthesis [38]Pigeon pea (Cajanus cajan (L) Millsp) belongs to the

family Fabaceae and regarded as the sixth major legumecrop worldwide [39] It is popularly known for its high pro-tein content [40] and drought tolerance [41] It is an im-portant pulse crop worldwide and predominantly cultivatedin the tropics and sub-tropics (httpfaostatfaoorg) Tomeet the ever-growing demand for pigeon pea extensiveresearch has been done during the last decade a break-through has been accomplished in developing hybrid tech-nology [42] The pre-imperatives for large-scale hybrid seedproduction are pollen transfer method and advancement ofstable cytoplasmic male-sterile (CMS) system The male-

sterility phenomenon could not be much explored in otherlegumes due to their self-pollinating nature Howeverpigeon pea has a considerable natural out-crossing [43] Tilldate eight (A1 to A8) such CMS systems have been bred inpigeon pea by transferring the cytoplasm of the wild rela-tives in the nuclear background of the cultivated pigeon peavia interspecific-hybridization followed by several rounds ofback-crossing [44] Therefore hybrid seed technology basedon cytoplasmic male-sterility has given a chance to over-come the production constraints by efficiently increasingthe yield potential in pigeon pea [45] However among theeight systems A2 and A4 CMS system obtained from thewild-type Cajanus scarabaeoides and Cajanus cajanifoliusrespectively are regarded as highly potential CMS systems[46ndash48] The gene associated with CMS in A4 cytoplasmpigeon pea has been reported [49] Recently the genetics offertility restoration in pigeon pea with A2 cytoplasm wasstudied in detailed [47] Even though A2 cytoplasm hasplayed a significant role in hybrid seed production yet theunderlying molecular processes governing CMS in A2 cyto-plasm pigeon pea is still not knownPresently the next generation sequencing technology has

revolutionized the field of life sciences and offers extraor-dinary speed and cost-effectiveness to study genomic andtranscriptome data [50 51] RNA-Seq has emerged as aneffective approach for transcriptome profiling which pro-vides accurate measurement of gene expression level [52]Transcriptome sequencing offers ease of identification andevaluation of thousands of genes in a single analysis [53]Transcriptome analysis of the cytoplasmic male-sterilelines has been previously reported in many crops such asradish [54] watermelon [55] soybean [56] tomato [57]Brassica napus [58] chili pepper [59]cotton [60ndash62] sweetorange [63] leading to identification of several candidategenes in these systems In this study comparative tran-scriptome analysis of an A2 CMS system derived cytoplas-mic male-sterile and its fertility restorer line in pigeon peawas performed primarily to detect differentially expressedgenes participating in pollen development which mighthelp in understanding the CMS mechanism in this crop

ResultsPollen fertility analysis and phenotypic characterization ofsterile and fertile budsPollens from male-sterile and fertile buds were observedunder a microscope and were evaluated to determine fertil-ity Upon observation it was seen that the pollens grainsshowing red acetocarmine staining were fertile while sterilepollens remained unstained (Fig 1a) During the pheno-typic investigation of the sterile (AKCMS 11) and fertile(AKPR303) flower buds we observed that the filamentsand anthers of the sterile flowers were smaller in contrastto the fertile flowers It was clearly visible that the anthersof the fertile line were denser than the sterile line (Fig 1b)

Saxena et al BMC Plant Biology (2020) 2074 Page 2 of 24

Transcriptome sequencing and de novo assemblyTo create a reference transcriptome we prepared cDNAlibraries from the buds of male-sterile (CMS) and fertile(restorer) lines in pigeon pea (two replicates each) and sub-jected to Illumina paired-end sequencing Workflow dia-gram of the analysis done is shown in Fig 2 Highthroughput sequencing run generated a total of 109600365 and 72575988 raw paired-end reads of ~ 100 bp inthe sterile and fertile buds respectively Also the correl-ation between the replicate datasets of the sterileAKCMS11 and fertile restorer AKPR303 was analyzed bycalculating Pearsonrsquos correlation coefficient (r lt 09) usingan R studio package Correlation analysis indicated that thereplicate datasets of the sterile and fertile lines were posi-tively correlated with Pearsonrsquos coefficient value as r = 072and r = 074 respectively (Additional file 1 Figure S1) Theper base sequence quality and k-mer content of the rawpaired-end reads was also determined (Additional file 1Figure S2 and Table S1) After adapter trimming and re-moval of poor quality reads 101676579 (495 Gb) and 69481226 (327 Gb) clean reads were attained for AKCMS11(sterile) and AKPR303 (fertile) lines respectively Trinitysoftware was further employed for de novo assembly of thepooled reads (171157805) from both the samples Therewas a total of 198587 transcripts with an average length of64134 bp and the N50 value of 1009 bp (L50 = 6 contigs)

We utilized Bowtie2 to align the clean reads to the assem-bled transcriptome alignment score of 9284 and 9284was obtained for sterile AKCMS11 (replicate 1 and repli-cate 2) respectively Similarly alignment score of 9261and 9205 was obtained for fertile AKPR303 (replicate 1and replicate 2) respectively (Additional file 1 Table S2)An overall alignment rate of 9258 was obtained indica-tive of the robust assembly BUSCO analysis revealed thatour assembly was relatively complete with 98 (n = 297)of BUSCOs detected as complete sequences just 2 (n =6) as fragmented sequences and none were missing in theassembly with eukaryotic lineage Similarly 935 (n = 402)of BUSCOs detected as complete sequences 58 (n = 25)as fragmented sequences and just 07 (n = 3) were miss-ing in the assembly with viridiplantae lineage (Table 1)After the de novo assembly 172061 unigenes were ob-tained by using CD-HIT software These unigenes werefurther BLASTN searched against the Rfam database to re-move the non-coding RNAs (rRNA tRNA snoRNAssnRNA etc) Finally a set of 171095 unigenes were ob-tained with an average length and N50 of 56166 bp and757 bp respectively (Table 2) The shortest and longestunigenes identified were 201 bp and 16008 bp respect-ively We obtained 115792 unigenes (677) with a lengthrange from 200 to 500 bp 17692 unigenes (103) weregreater than 1000 bp and 8700 (51) were greater than

Fig 1 a Acetocarmine staining differentiating the pollens of AKCMS 11 (sterile line) and AKPR303 (fertile line) Fertile pollens are darkly stainedwhereas sterile pollens are almost colorless b Phenotypic characterization of the sterile and fertile floral buds A Phenotype of sterile AKCMS11buds B Phenotype of fertile AKPR303 buds


2000 bp (Fig 3a) The average GC content of the unigeneswas 458 The RNA-Seq raw data generated from this in-vestigation has been deposited in the NCBI Sequence ReadAchieve (SRA) database (SRX3740150 SRX3740151SRX3740153 SRX3740152) for sterile and fertile budsrespectively

Functional annotation and classification of the unigenesAnnotation results showed that 102463 (60) 52639(306) 28812 (167) and 7653 (44) unigenes anno-tated in Nr Virdiplantae GO and KEGG databasesrespectively (Fig 3b) In total 102463 (60) unigeneswere annotated in the nr database while the remaining40 unigenes did not show any sequence similarity to thenr database (Additional file 2 Table S1) Sequence similar-ity distribution revealed that 61 of the aligned unigeneshad sequence similarity more than 90 on the other hand386 unigenes had similarity value between 50 to 90and only 04 unigenes had lower than 50 (Fig 3c) Atotal of 52639 unigenes were uniquely mapped to theVirdiplantae database (Additional file 2 Table S2)A total of 28812 unigenes were assigned GO terms and

were distributed into 53 GO categories including 26 bio-logical processes 14 molecular functions and 13 cellularcomponents In the biological process category ldquocellularprocessrdquo (19472 unigenes) was the most predominantfunctional group followed by ldquometabolic processrdquo (17365unigenes) and ldquobiological regulationrdquo (9627 unigenes)Among the molecular function category ldquobindingrdquo (19531

unigenes) ldquocatalytic activityrdquo (14757 unigenes) and ldquotran-scription regulator activityrdquo were the main functionalgroups In regard to the cellular component ldquocellrdquo (22732unigenes) and ldquocell partrdquo (22732 unigenes) were the largestcategories followed by ldquoorganellerdquo (15348 unigenes) andldquomacromolecular complexrdquo respectively (Additional file 2Table S3 and Additional file 3 Figure S1)The unigenes were mapped to the KEGG pathway data-

base 7653 unigenes were assigned KEGG orthologos num-ber involving 132 pathways The result demonstrated thatthe five largest pathways were ldquometabolic pathwaysrdquo (ko01100) ldquobiosynthesis of secondary metabolitesrdquo (ko01110)ldquoplant hormone signal transduction (ko04075) ldquocarbon

Fig 2 Systemic representation showing workflow of the analysis done

Table 1 BUSCO completeness assessment conducted by usingeukaryotic ortholog database (eukaryota_orthoDB9) andviridiplantae ortholog database (viridiplantae_orthoDB10)

BUSCOs (Number of BUSCO units found)

eukaryotic orthologdatabase

viridiplantae orthologdatabase

Complete BUSCOs 297 402

Complete and Single-Copy BUSCOs

186 315

Complete and duplicatedBUSCOs

111 87

Fragmented BUSCOs 6 25

Missing BUSCOs 0 3

Total BUSCOs searched 303 430


metabolismrdquo (ko01200) and ldquobiosynthesis of amino acidsrdquo(ko01230) The 7653 unigenes were allocated into six func-tional categories (Additional file 2 Table S4 and Additionalfile 3 Figure S2) A total of 5499 unigenes were found inldquometabolismrdquo with 1363 (248) unigenes involved inldquometabolic pathwaysrdquo 881 (16) in ldquobiosynthesis of second-ary metabolitesrdquo 870 (158) in ldquocarbohydrate metabolismrdquo665 (121) in ldquoamino acid metabolismrdquo 458 (83) inldquolipid metabolismrdquo and other sub-categories 1222 unigenesaccounted for the category of ldquogenetic information and pro-cessingrdquo in which 449 (367) unigenes were engaged inldquofolding sorting and degradationrdquo followed by ldquotranslationrdquo446 (365) ldquotranscriptionrdquo 181 (148) and ldquoreplicationand repairrdquo 146 (119)ldquoEnvironmental information pro-cessingrdquo was represented by a total of 462 unigenes with416 (90) involved in ldquosignal transductionrdquo and 46 (10) inldquomembrane transportrdquo Additionally 300 and 155 unigeneswere classified in ldquocellular processesrdquo and ldquoorganismalsystemsrdquo categories respectively

Identification of transcription factorsPlantTFcat online tool [64] was utilized in the presentstudy for identification of transcription factors in pigeon

pea A total of 3803 unigenes (22 of the transcriptome)(Fig 3b) were cataloged into 52 putative transcriptionfactors (TF) families Amid the 52 TF family ldquoC2H2rdquo wasfound to be the most predominant category with 806(212) unigenes followed by ldquoWD40-likerdquo (540 unigenes142) ldquoMYB-HB-likerdquo (344 unigenes 9) ldquobHLHrdquo (200unigenes 53) ldquoAP2-EREBPrdquo (163 unigenes 43)ldquoWRKYrdquo (161 unigenes 42) and ldquobZIPrdquo (129 unigenes34) transcription factors (Fig 4a and Additional file 2Table S5)

Identification of differentially expressed genes betweenlines AKCMS11 and AKPR303We identified 3167 differentially expressed genes (DEGs)between the two lines containing 1432 up-regulated and1390 down-regulated in the AKCMS11 sterile buds as com-pared to AKPR303 fertile buds (Additional file 2 Table S6)Additionally 683 DEGs were uniquely expressed in thesterile genotype and 157 DEGs were distinctively expressedin the fertile restorer (Fig 4b)

GO and KEGG enrichment analysis of the DEGsTo gain insight into the putative functions of the DEGs weexamined the significantly enriched Gene Ontology (GO)terms in DEGS The annotated DEGs were assigned to 55major groups covering three main ontologies biologicalprocess cellular component and molecular function(Fig 5) The identified GO terms were further classifiedinto down-regulated and up-regulated groups In thedown-regulated DEGs a total of 221 GO terms wereassigned including 121 59 and 41 GO terms in ldquobiologicalprocessrdquo ldquomolecular functionsrdquo and ldquocellular componentrdquorespectively (Additional file 2 Table S7) In up-regulatedDEGs 174 GO terms were assigned including 78 ldquobio-logical processrdquo 53 ldquomolecular functionsrdquo and 43ldquocellularcomponentrdquo (Additional file 2 Table S8) Overall amongthe biological process category the significantly over-represented GO terms were ldquocellular processrdquo followed byldquometabolic processrdquo ldquoprimary metabolic processrdquo and ldquocel-lular metabolic processrdquo ldquoCatalytic activityrdquo ldquobindingrdquo andldquotransferase activityrdquo were the most significantly expressedGO terms in the molecular functions category Among thecellular component category ldquocell partrdquo ldquocellrdquo and ldquointra-cellular componentrdquo were the most significant GO terms(Figs 6 and 7) Interestingly many GO terms were specificto down-regulated DEGs and were involved in ldquopollen de-velopmentrdquo (10 DEGs) ldquoreproductive structure develop-mentrdquo (41 DEGs) ldquogametophyte developmentrdquo (16 DEGs)and ldquooxidative phosphorylationrdquo (7 DEGs) suggesting thesemay be related to male-sterility in AKCMS11 Hierarchicaltree graph of over-represented GO terms in the biologicalprocess category of down-regulated and up-regulatedDEGs are provided in Additional file 3 Figure S3 and S4

Table 2 Summary of Illumina sequencing and de novotranscriptome

Read processing

AKCMS11 (tworeplicates)

AKPR303 (tworeplicates)

Total

Raw reads 109600365 72575988 182176353

Clean reads (aftertrimming)

101676579 69481226 171157805

Average read length (bp) 100 100 ˗

Trinity based de novo assembly

No of transcripts (n) 198587

Percentage GC 4114

Contig N10 (bp) 3148





L50 value 6

Median contig length 357

Average length 64134

Total assembled bases 127362415

Unigenes

Total number ofunigenes (n)

171095

Average length (bp) 56166

N50 length (bp) 757



KEGG pathway enrichment analysis was performed todiscern the metabolic pathways in which the DEGs wereinvolved using the KEGG pathway database In total 110enriched pathways were identified of which 77 pathwayswere significantly enriched in both down and up-

regulated DEGs whereas 15 and 19 pathways were spe-cific to down and up-regulated DEGs respectively Atotal of 377 down-regulated and 396 up-regulated DEGswere assigned to 92 and 96 KEGG pathways respectively(Additional file 2 Table S9) The top 25 significantly

Fig 3 a Length distribution of the assembled unigenes b Homology search and annotation statistics of pigeon pea unigenes c) Similaritydistribution of the BLASTX hits for each unigenes against nr database

Fig 4 a Transcription factors identified in the transcriptome data of Cajanus cajan b Number of differentially expressed genes (DEGs) in sterileAKCMS11 and restorer AKPR303


enriched pathways for down and up-regulated DEGs arementioned in Fig 8 The down-regulated DEGs were sig-nificantly over-represented in ldquometabolic pathwaysrdquo (72genes) ldquobiosynthesis of secondary metabolitesrdquo (39 genes)ldquooxidative phosphorylationrdquo (13 genes) ldquocarbon metabol-ismrdquo (11 genes) ldquoribosomerdquo (9 genes) and glyceropho-spholipid metabolism (8 genes) Consistent with the GOanalysis ldquooxidative phosphorylation was enriched withdown-regulated DEGs Several down-regulated DEGs par-ticipated in starch and sucrose metabolism glycolysispentose phosphate pathway reactive oxygen species(ROS) generationscavenging and alpha-linolenic acid me-tabolism Among the up-regulated DEGs the most pre-dominant pathway was ldquometabolic pathwaysrdquo (66 genes)followed by ldquobiosynthesis of secondary metabolitesrdquo (34genes) ldquocarbon metabolismrdquo (14 genes) ldquoprotein process-ing in endoplasmic reticulumrdquo (11 genes) and ldquoplanthormone signal transductionrdquo (11 genes) Also some up-regulated genes were involved in ldquoascorbate and aldaratemetabolismrdquo ldquolinolenic acid metabolismrdquo and ldquophotosyn-thesisrdquo etc (Additional file 2 Table S9)Additionally MapMan tool [65] was also used to visualize

significant male-sterility related DEGs involved in differentmetabolic pathways All the 3167 differentially expressedgenes between sterile AKCMS11 and fertile AKPR303 were

annotated to the TAIR database (httpwwwarabidopsisorg) and finally 951 DEGs were identified to be homologsof 930 Arabidopsis genes (Additional file 2 Table S10) Tofurther explore the potential functions of these DEGs inmale-sterility the Arabidopsis homologs genes were studiedin MapMan to identify different metabolic processes inwhich these are involved (Fig 9) In our network the mostsignificant differentially expressed genes were related to En-ergyATP synthesis ROS metabolism hormones second-ary metabolism and cell cycle The expression of genesinvolved in lsquoEnergy (Glycolysis TCA cycle electron trans-port chain transport p- and v-ATPasersquos) was down-regulated in the sterile (AKCMS11) buds Moreover thegenes related to lsquoRedoxrsquo (Ascorbate amp Glutathione Peroxir-edoxin Dismutase amp Catalase) and cell cycle were alsodown-regulated in the sterile buds Additionally the genesimplicated in lsquoAuxin amp Jasmonic acid synthesisrsquo were up-regulated in sterile buds while the ones related to lsquoSignalingpathwayrsquo were down-regulated Also the expression ofgenes involved in lsquoLipidrsquo (Degradation amp FA synthesis) andlsquoCell wallrsquo (degradation amp modification) were mostly up-regulated whereas genes related to lsquoSecondary metabolismrsquo(Flavonoids) were down-regulated in the sterile buds com-pared to fertile buds Among the transcription factorsNAC WRKY (except one) CCAAT SET C2C2(Zn)

Fig 5 GO enrichment analysis of the DEGs X-axis represents the sub-categories Y-axis represents number of genes in each sub-category Blueand green indicates down-regulated and up-regulated genes in a sub-category


Fig 6 Significantly over-represented GO terms in down-regulated DEGs

Fig 7 Significantly over-represented GO terms in up-regulated DEGs


GATA and AP2EREBP (except one) were down-regulatedin the sterile buds whereas the majority of the MYBMADS and C2C2(Zn) Dof transcription factors were up-regulated in comparison to the fertile buds

Candidate DEGs associated with male-sterilityThe bioinformatics analysis gave us a deeper insight intothe differentially expressed genes (DEGs) between thesterile (AKCMS11) and fertile genotypes (AKPR303) Inthis present investigation several subsets of differentiallyexpressed genes were identified which are potentially re-lated to pollen development (Fig 10)

Putative gene related to pollen developmentIn flowering plants pollen development is a complicatedand sequential process of sexual reproduction Arabidopsisthaliana is considered as the model plant to study putativegenes related to pollen development [58] To understandthe mechanism related to male sterility we queried all the3167 DEGs in the TAIR database (httpwwwarabidopsisorg) Additionally all the 3167 DEGs were further com-pared with the Arabidopsis database (httpwwwarabidop-sisorg) to identify homologs using OrthoDB v10 [66](Additional File 2 Table S11) In our study 34 DEGs werefound to be homologs of Arabidopsis genes associated withthe development of pollen (Table 3) We found 10 DEGsinvolved in regulation of pollen development among which1 DEG encoding EXCESS MICROSPOROCYTES 1 (EMS1)1 DEG encoding PHD-finger transcription factor MALE

STERTILITY1 (MS1) 2 DEGs encoding callose synthase 71 DEG encoding AUXIN-RESPONSIVE FACTOR (ARF17)4 DEGs encoding CYTOCHROME P450-like and 1 DEGsencoding aspartic protease 5 DEGs were up-regulatedand 5 were down-regulated in the sterile AKCMS11 incomparison to fertile AKPR303 3 DEGs encoding polyga-lacturonase (PG) and endo-13- beta-glucosidase-like wereinvolved in pollen cell wall remodeling Total 5 DEGsencoding arabinogalactan glycoproteins (AGPs) and GPI-anchored like protein were involved in pollen tube growthSingle stress induced DEG was present encoding late em-bryogenesis abundant protein (LEA) showing significantdown-regulation in AKCMS11 with respect to AKPR303We identified 3 DEGs involved in the cell divisionprocess Additionally 12 DEGs with related functionswere also identified in this study (Table 3)Transcription factors (TFs) are of major importance as

they regulate gene expression in plants Alterations inthe gene expression are associated with modifications inthe expression of transcription factors [67] In our studythere were 16 DEGs representing transcription fac-tors 9 DEGs were down-regulated and 7 were up-regulated in the AKCMS11 The 9 down-regulatedDEGs included 2 bHLH family 2 C2H2zinc-finger 2basic-leucine zipper transcription factor (bZIP) 1 dofzinc-finger and 2 NAC domain-containing transcrip-tion factors Among the 7 up-regulated DEGs 1 werebHLH 2 MYB transcription factors 2 basic-leucinezipper transcription factors (bZIP) 1 NAC domain-

Fig 8 Top 25 significantly enriched KEGG pathways a For down-regulated DEGs b For up-regulated DEGs


containing transcription factorsand 1 WRKY tran-scription factor (Additional file 2 Table S5)

Metabolic pathwaysCarbon fixation and energy metabolism are the most pre-dominant metabolic pathways which are primarily respon-sible for furnishing the requirement of energy and carbonsources in the plants [68] In the present study 12 DEGswere observed to be participating in carbohydrate metabol-ism 6 DEGs related to glycolysisgluconeogenesis pathwayand 6 DEGs related to pentose and glucuronate intercon-versions respectively (Table 4) With respect to glycolysis 4DEGs were down-regulated and 2 DEGs were up-regulatedin AKCMS11 in comparison to AKPR303 Additionallyamong the 6 DEGs representing pentose and glucuronateinterconversions 3 DEGS were down-regulated and 3DEGs were up-regulated in AKCMS11 (Table 4)In this study 2 DEGs participated in the Tricarboxylic

acid cycle (TCA) 1 DEG encoding malate dehydrogenaseand second encoding succinate-CoA synthetase (Table 4)Also we detected 7 DEGs involved in oxidative

phosphorylation majority of the DEGs showed down-regulation with only 2 DEGs showing up-regulation in theAKCMS11 (Table 4) We examined a total of 15 DEGs in-volved in the elimination of reactive oxygen species(ROS) including 2 DEGs encoding peroxidase-like 1DEG encoding superoxide dismutase and 2 DEG encodingperoxiredoxin-2B-like along with 1 DEG encoding gluta-thione S-transferase and 1 DEG encoding for glutathioneperoxidase Additionally there were 5 DEGs encoding formonothiol glutaredoxin and 1 DEG each encoding for as-corbate oxidase thioredoxin and glutaredoxin 10 of theDEGs showed down-regulation in AKCMS11 in compari-son to AKPR303 (Table 4)

Confirmation of DEGs by qRT-PCRWe randomly selected 10 genes for qRT-PCR analysis withthe aim to validate the expression profiles between AKCMS11 and AKPR303 obtained by RNA-Seq The list of DEGsspecific primers used for qRT-PCR analysis is listed in Add-itional file 2 Table S12 The DEGs selected for qRT-PCRconfirmation were related to callose synthase 7-like

Fig 9 Overview of differentially expressed genes (DEGs) involved in diverse metabolic pathways DEGs were selected for the metabolic pathwaysusing MapMan software The colored red and green boxes indicate down-regulated and up-regulated genes respectively The scale bar representfold change values

Saxena et al BMC Plant Biology (2020) 2074 0 of 24

phosphoenolpyruvate carboxykinase polygalacturonaseaspartic protease late embryogenesis abundant (LEA)uncharacterized mitochondrial protein glutathione S-transferase F10 calmodulin binding receptor-like pro-tein mitotic spindle assembly checkpoint protein(MAD1) and pentatricopeptide repeat-containing pro-tein Although the relative gene expression detected byRNA-Seq analysis was higher than those detected byqRT-PCR the overall expression patterns were similar(Fig 11 and Table 5) Linear regression analysis wasperformed and it revealed positive correlation (R2 =09508) between the qRT-PCR and DGE fold changegene expression ratios indicating that the expressionof all the 10 genes depicted by qRT-PCR analysis wasin agreement with the DGE data analysis (Fig 12)These findings confirm the reliability of the IlluminaRNA-Seq data and its analysis

DiscussionCytoplasmic male-sterility (CMS) is considered as an im-portant tool for hybrid seeds production in many plantspecies To elucidate the underlying mechanism govern-ing CMS in pigeon pea the present study was under-taken Comparative transcriptome sequencing andanalysis of the floral buds from sterile (AKCMS11)and fertile (AKPR303) pigeon pea were performed Intotal 3167 genes showed differential expressionbetween the CMS and the fertile restorer Based onthe functional annotation we identified some candi-date DEGs related to pollen development and

metabolism which were previously reported to be in-volved in male-sterility (Fig 13)

DEGs involved in pollen development potentially relatedto CMSWe identified 34 pollen development related genes whichwere found to have homologs in Arabidopsis (Table 3)Pollen formation is a critical developmental mechanism inplants which is highly dependent on co-ordinated metabolicpathways [69] Male-sterility occurs due to the aberrantfunctioning of the genes responsible for tapetum develop-ment pollen formation callose degradation and pollen walldevelopment [70] In our findings several DEGs with apotential role in pollen development were observed (Table3) Pollen development is an intricate and programmedprocess which takes place within specialized structurescalled anthers from reproductive cells (microsporocytes)These microspores (pollens) are dependent on the sporo-phytic tissues for their development finally leading to therelease of functional pollens for fertilization in floweringplants [71] In Arabidopsis EXCESS MICROSPOROCYTES1 (EMS1) proposes a leucine-rich repeat receptor proteinkinase (LRR-RPK) which is actively involved in the deter-mination and differentiation of the tapetal cells and the mi-crosporocytes during plant reproduction It was observedthat the anthers of the EMS mutant (EMS1) exhibited ex-cessive microsporocytes despite lacking the tapetum pro-ducing non-functional pollens and leading to male-sterility[22] We also observed 1 DEG (Ccajan_10186_c0_g1_i1)encoding leucine-rich repeat receptor protein kinase EMS1

Fig 10 Heat map of expression profiles of some of the focused genes related to cytoplasmic male-sterility Color from red to green indicatedthat the FPKM values were small to large red color indicates low level of gene expression and green color indicates high level of gene expression


down-regulated (453-fold) in the AKCMS11 respectivelyand the gene thus may also be related to pollen develop-ment and fertility in pigeon peaDuring the process of pollen development tapetum is of

great significance as it provides nourishment to the devel-oping microspores leading to pollen wall formation [72] In

Arabidopsis MALE STERILITY1 (MS1) and MALE SERI-LITY2 (MS2) genes regulate the stages of tapetum matur-ation and pollen wall biosynthesis and ending with therelease of viable pollen [27 73] The MS1 gene exhibitshomology to PHD-finger transcription factor critical fortapetum and pollen development [74] A PHD-finger

Table 3 Pigeon pea differentially expressed genes (DEGs) homologous to Arabidopsis male-sterilityreproduction genesaQuery ID bLog2FC

cSubject Id E-value Score Symbol Gene annotation

Down-regulated

Ccajan_41968_c0_g1_i1 minus621 AT3G128301 168E-10 558 ARF72 auxin-responsive factor 72 (ARF72)

Ccajan_10186_c0_g1_i1 minus453 AT5G072801 710E-21 863 EMS1 leucine-rich repeat receptor protein kinase EMS1

Ccajan_37653_c0_g4_i5 minus915 AT5G604901 772E-05 32 FLA11 Fasciclin-like arabinogalactan protein 11

Ccajan_38216_c1_g2_i3 minus5 AT1G548601 132E-35 128 GPI-anchored-like proteins

Ccajan_34303_c0_g2_i2 minus427 AT1G772501 217E-33 120 PHD finger like transcription factor (MS1)

Ccajan_44775_c0_g1_i1 minus545 AT1G064902 242E-04 385 CALS7 callose synthase 7

Ccajan_52234_c0_g1_i1 minus638 AT3G145702 0 1100 CALS8 callose synthase 8

Ccajan_83966_c0_g1_i1 minus885 AT1G027901 275E-98 303 PGA3 Polygalacturonase (PG)

Ccajan_61890_c0_g1_i1 minus487 AT2G444805 512E-38 130 BGLU17 endo-13-beta glucosidase 17

Ccajan_7875_c0_g1_i1 minus220 AT2G275001 0 524 BGLU14 endo-13-beta-glucosidase 14

Ccajan_2910_c0_g1_i1 minus455 AT1G059901 756E-47 157 CML7 calmodulin-like protein 7

Ccajan_47280_c0_g1_i1 minus753 AT4G003302 171E-93 285 CRCK2 calmodulin-binding receptor-like protein 2

Ccajan_41720_c0_g1_i1 minus659 AT2G381101 118E-05 416 GPAT6 glycerol-3-phosphate acyltransferase 6 (GPAT6)


Ccajan_13394_c0_g1_i1 minus853 AT2G271103 148E-20 917 FAR3 fatty acid reductase 3 (FAR3)

Ccajan_697_c0_g1_i1 minus233 AT1G613702 366E-11 57 S-locus lectin protein kinase family protein

Ccajan_52697_c0_g1_i1 minus457 AT1G325601 364E-11 601 LEA6 Late embryogenesis abundant protein 6

Ccajan_51888_c0_g1_i1 minus461 AT2G206351 268E-154 447 MAD1 mitotic spindle assembly checkpoint protein (MAD1)

Ccajan_41460_c0_g1_i1 minus586 AT4G332701 407E-73 228 CDC20ndash1 Cell division cycle protein 201

Ccajan_41720_c0_g1_i1 minus660 AT2G381101 118E-05 416 GPAT6 Glycerol-3-phosphate 2-O-acyltransferase 6

Ccajan_47215_c0_g1_i1 minus714 AT2G431401 130E-20 905 Basic helix-loop-helix (bHLH) DNA-binding protein

Ccajan_68741_c0_g2_i1 minus209 AT1G106101 630E-16 705 BHLH90 Transcription factor bHLH90

Ccajan_45806_c0_g1_i1 minus353 AT3G197401 0 1237 P-loop containing nucleoside triphosphate hydrolases


Up-regulated

Ccajan_38216_c3_g1_i1 725 AT1G548601 132E-35 128 GPI-anchored-like proteins

Ccajan_1528_c0_g3_i1 83 AT1G137101 973E-132 387 CYP78A5 Cytochrome P450-like 78A5

Ccajan_88538_c0_g1_i1 823 AT3G263001 388E-54 181 CYP71B34 Cytochrome P450-like 71B34

Ccajan_16351_c0_g1_i1 881 AT5G056903 213E-06 454 CYP90A1 cytochrome P450 90A1-like


Ccajan_24040_c0_g3_i1 1042 AT1G058401 27E-161 471 aspartic protease

Ccajan_37598_c0_g2_i1 9 AT5G472301 734E-42 148 ERF5 ethylene-responsive transcription factor 5 (ERF5)

Ccajan_37057_c0_g1_i1 899 AT4G244703 866E-75 241 ZIM GATA transcription factor 25-like

Ccajan_35517_c1_g2_i5 255 AT5G498801 0 757 MAD1 mitotic spindle assembly checkpoint protein (MAD1)

Ccajan_36686_c0_g1_i4 1005 AT2G384701 207E-39 142 WRKY33 WRKY transcription factor 33 (WRKY33)aQueryId for homologue search using BlastbLog2FC ge 2 represented up-regulated and Log2FC le minus2 were represented down-regulatedcArabidopsis male-sterilitymale-reproduction related genes in TAIR database (httpwwwarabidopsisorg)


Table 4 KEGG pathways enriched of differentially expressed genes (DEGs) between sterile AKCMS11 and fertile AKPR303

Query ID Subject Id Gene Annotation aLog2FCbRegulation cp-value

GlycolysisGluconeogenesis

Ccajan_36409_c0_g2_i1 AT4G378701 Phosphoenolpyruvate carboxykinase minus233 down 004335


Ccajan_37543_c0_g2_i4 AT3G308412 Phosphoglyceratemutase minus2 down 0003866

Ccajan_34880_c0_g2_i1 AT1G436701 Fructose-16-bisphosphatase 257 up 00418388

Ccajan_39910_c0_g1_i1 AT1G231901 phosphoglucomutase minus733 down 000156436

Ccajan_88886_c0_g1_i1 AT4G005701 NAD-dependent malic enzyme 2 228 up 004124277

Pentose and glucuronate interconversions

Ccajan_36483_c0_g1_i4 AT2G377702 aldo-keto reductase family 4 minus854 down 000872853

Ccajan_32476_c1_g2_i1 AT5G624201 Aldoketo reductase family protein 847 up 00292025

Ccajan_30463_c0_g1_i1 AT3G186603 UDP-glucuronosyltransferase 1 883 up 00111043

Ccajan_18517_c0_g1_i1 AT1G223601 UDP-glycosyltransferase 85A2 minus328 down 000017461


Ccajan_38121_c2_g1_i5 AT3G531501 UDP-glucosyl transferase 73D1 424 up 00002212

Tricarboxylic acid cycle

Ccajan_64232_c0_g1_i1 AT1G797501 Malate dehydrogenase minus646 down 0001118

Ccajan_35485_c0_g1_i1 AT2G204201 Succinate-CoA synthetase 230 up 000176114

Starch and sucrose metabolism

Ccajan_7875_c0_g1_i1 AT2G275001 Glucan endo-13-beta-glucosidase 14 minus220 down 001872407

Ccajan_61890_c0_g1_i1 AT2G444805 beta glucosidase 17 minus487 down 002273399

Alpha-linolenic acid metabolism

Ccajan_25458_c0_g1_i1 AT3G518401 Acyl-coenzyme A oxidase 4 347 up 002874461

Oxidative phosphorylation

Ccajan_34619_c0_g1_i1 AT4G389201 Vacuolar H+-ATPase subunit C1 minus993 down 184E-05

Ccajan_36113_c2_g1_i1 AT4G111501 Vacuolar H+-ATPase subunit E1 minus1069 down 0001949

Ccajan_34619_c0_g3_i1 AT4G389201 Vacuolar H+-ATPase subunit C1 102 up 745E-05

Ccajan_36113_c2_g3_i2 AT4G111501 Vacuolar H+-ATPase subunit E 3365 down 0015683

Ccajan_58593_c0_g1_i1 AT5G214301 NADH dehydrogenase subunit B2 minus628 down 0001759

Ccajan_41295_c0_g1_i1 ATMG013601 Cytochrome c oxidase subunit 1 minus646 down 000170655

Ccajan_13673_c0_g1_i1 AT1G224501 cytochrome c oxidase subunit 6b minus598 down 0000415

Reactive oxygen species (ROS) generationscavenging

Ccajan_35344_c0_g3_i4 AT5G067201 peroxidase 53 minus305 down 0005337

Ccajan_12688_c0_g1_i1 AT5G053401 peroxidase 52 minus862 down 882E-05

Ccajan_20878_c0_g1_i1 AT1G659801 peroxiredoxin-2B-like minus441 down 0045412

Ccajan_68963_c0_g2_i1 AT3G109202 superoxide dismutase minus1165 down 144E-11

Ccajan_34224_c0_g1_i2 AT5G211051 ascorbate oxidase 835 up 003753425

Ccajan_90965_c0_g1_i1 AT4G318701 glutathione peroxidase 7 minus520 down 002719158

Ccajan_36196_c0_g2_i3 AT2G308701 glutathione s-transferase F10 241 up 002883

Ccajan_12449_c0_g1_i1 AT5G186001 Monothiol glutaredoxin-S2 minus223 down 002449176

Ccajan_18941_c0_g2_i1 AT4G156801 Monothiol glutaredoxin-S4 (GRXS4) minus866 down 000724279

Ccajan_20878_c0_g1_i1 AT1G659801 Peroxiredoxin-2B minus441 down 004541246


Ccajan_82144_c0_g1_i1 AT5G399501 Thioredoxin H2 minus643 down 000129587

Ccajan_18941_c0_g3_i1 AT4G156801 Monothiol glutaredoxin-S4 (GRXS4) 904 up 000570316


transcription factor MALE STERTILITY1 (MS1) conferringmale-sterility was identified in barley [75] In this study 1DEG (Ccajan_34303_c0_g2_i2) encoding PHD-finger tran-scription factor MALE STERTILITY1 (MS1) was detectedwith 427-fold down-regulation in the AKCMS11 In theMS1 mutant immediately after the release of microsporesfrom the tetrads sudden preterm degeneration of the tap-etum and pollen takes place leading to inviable pollen for-mation and cause male-sterility [27] Similar results werereported in male-sterile tomato mutants ms3 ms15 ms5and ms1035 due to dysfunctional regulation of the tapetum[57 76 77]In flowering plants pollens are secured within the pollen

grains and physically protected by the complex pollen cellwalls The pollen mother cells (PMC) synthesize a polysac-charide callose (β-13-glucan) and the outer protectivecallose wall segregates the newly developed microspore andrestricts them from merging with the neighbouring tap-etum tissues giving the unique tetrad shape [78] Finallythe callose wall surrounding the microspores degeneratesreleasing them into the anther locule [79] The callose

synthase enzyme is required for the synthesis of a tempor-ary callose wall around the developing microspore twelvecallose synthase (cals) genes are reported in Arabidopsis[80 81] The callose synthase 5 (cals5) mutants showed theabsence of callose deposition around the microspores con-firming the importance of this enzyme with reduced fertility[38] Interestingly we noticed 2 DEGs Ccajan_44775c0_g1_i1 and Ccajan_52234_c0_g1_i1 representing callosesynthase with 545 and 638-fold down-regulation in theAKCMS11 in comparison to the AKPR303 thus featuringas a prominent candidate in pollen development These re-sults are consistent with the earlier reports in watermelonand soybean [55 56] Additionally primexine a microfibril-lar polysaccharide matrix is important for the normaldevelopment of functional pollen and is regulated by anAUXIN-RESPONSIVE FACTOR (ARF17) In ArabidopsisARF17 mutant displayed abnormal callose wall formationwith no primexine and pollen abortion [82] A single DEGCcajan_41968_c0_g1_i1 for ARF72 displayed reducedexpression of 621-fold in the AKCMS11 thus featuring as aprominent candidate in pollen development

Table 4 KEGG pathways enriched of differentially expressed genes (DEGs) between sterile AKCMS11 and fertile AKPR303 (Continued)


Ccajan_24088_c0_g1_i1 AT4G330401 Glutaredoxin-C6 408 up 004803176

Ccajan_37604_c0_g1_i3 AT3G629301 Monothiol glutaredoxin-S6 (GRXS6) 1006 up 001361605aLog2FC ge 2 represented up-regulated and Log2FC le minus2 were represented down-regulatedbRegulation direction of the DEGs AKPR303 restorer was the controlcp-value of le005 was considered statistically significant

Fig 11 qRT-PCR validation of differentially expressed unigenes S denotes the sterile AKCMS11 and F denotes the fertile AKPR303 The primaryvertical axis represents qRT-PCR expression values and secondary vertical axis represents RNA-Seq expression values


Sporopollenin is the key constituent of the outer rigidexine wall of the microspores and genes like CYTOCHROME P450 (CYP703A2) CYTOCHROME P450 (CYP704B1) and acyl-coA synthetase 5 (acos5) are responsiblefor regulating the biosynthesis of sporopollenin Down-regulation of the CYP703A2 and CYP704B1 genes depictedthe reduced level of sporopollenin an absence of outerexine layer with unique stripped surface and impairedpollens in Arabidopsis [34 35 73] Surprisingly 4 genescorresponding to CYTOCHROME P450-like were drastic-ally over-expressed in the AKCMS11 in comparison toAKPR303 For example Ccajan_1528_c0_g3_i1 (830-fold)Ccajan_88538_c0_g1_i1 (823-fold) Ccajan_16351_c0_g1_i1 (881-fold) and Ccajan_55694_c0_g9_i1 (636-fold) Thisfinding is supported by a previous study in cotton wherethey proposed that since these genes are involved in thesynthesis of sporopollenin their up-regulation must have

ended up in immoderate accumulation sporopollenin andultimately male-sterility [83] Over-expression of this genemay be related to pollen abortionThe developing microspores (pollen grains) within the

anthers are enclosed by a layer of cells known as tapetum[84 85] it is of considerable importance as it provides thenecessary nourishment to the developing pollens andsecretes components for the outer exine and regulates thepollen wall formation [86 87] During the subsequent laterstages of pollen development tapetum experiences a regu-lated disintegration by programmed cell death (PCD) andreleases all the cellular components required for pollen wallformation into the anther locule [88ndash90] Any delay in thetimely regulation of tapetum PCD results in pollen lethalityand male-sterility [87] Previously a PCS1 gene encodingan anti-cell-death factor known as an aspartic proteasewhich governs PCD in Arabidopsis was reported [91] The

Table 5 Confirmation of the expression profiles of selected genes by qRT-PCR

Gene Id Description Fold change

RNA-Seq qRT-PCR

Ccajan_44775_c0_g1_i1 callose synthase 7-like minus54 minus44

Ccajan_34307_c0_g1_i1 phosphoenolpyruvate carboxykinase minus22 minus17

Ccajan_83966_c0_g1_i1 Polygalacturonase (PG) minus89 minus66

Ccajan_24040_c0_g3_i1 aspartic protease 104 82

Ccajan_52697_c0_g1_i1 Late embryogenesis abundant protein 6 (LEA6) minus45 minus63

Ccajan_12415_c0_g2_i1 uncharacterized mitochondrial protein minus123 minus81

Ccajan_36196_c0_g2_i3 glutathione s-transferase F10 241 20

Ccajan_47280_c0_g1_i1 calmodulin binding receptor-like protein minus75 minus58

Ccajan_51888_c0_g1_i1 mitotic spindle assembly checkpoint protein (MAD1) minus46 minus39

Ccajan_47116_c0_g1_i1 pentatricopeptide repeat-containing protein minus25 minus12

Fold change refers to expression in sterile AKCMS11 in comparison to restorer AKPR303 Negative value refers down-regulated expression in sterile AKCMS11

Fig 12 Linear regression analysis of the fold change of the gene expression ratios between DEG data and qRT-PCR 10 primers were selected forqRT-PCR analysis to confirm the accuracy and reproducibility of the Illumina expression Scatterplots were generated by the log2 expression ratiosfrom DGE sequencing data (x- axis) and qRT-PCR data (y-axis)


over-expression of the PCS1 gene produced excessiveaspartic protease which further inhibited anther dehiscenceand resulted in male-sterility In this study 1 DEG (Cca-jan_24040_c0_g3_i1) was 1042-fold over-expressed in thesterile AKCMS11 than the fertile AKPR303 The higher ex-pression in the sterile line could have delayed the PCD ofthe tapetum which is a must for anther dehiscence and ul-timately lead to pollen sterility in pigeon pea A similar ob-servation was seen in soybean which therefore stands insupport of the present finding [56]In this study 3 DEGs related to cell wall modification

were identified and all were strongly down-regulated in theAKCMS11 as compared to the AKPR303 These genes en-code cell wall-degrading enzymes such as polygalacturonase(PG) known for hydrolyzing pectin and polygalactouronicacid and endo-13-beta-glucosidase-like cellulose- hydro-lyzing enzyme In an earlier report in Arabidopsis QRT1and QRT2 mutant produced non-functional pollen due tothe inefficiency of polygalacturonase in pectin degradationaround the microspores [92] A single DEG (Ccajan_83966_c0_g1_i1) representing polygalacturonase (PG)showed 885-fold down-regulation with 2 DEGs (Ccajan_61890__c0_g1_i1 and Ccajan_7875_c0_g1_i1) encodingendo-13-beta-glucosidase-like were 487-fold and 220-folddown-regulated in the AKCMS11 respectively The resultswere consistent with the findings in tomato and cotton [1193] These under-expressed genes might be involved inpollen abortionArabinogalactan glycoproteins (AGPs) are present in dif-

ferent cells and tissues of the higher plants and are activelyinvolved in the growth and reproduction [94] In Arabi-dopsis AGP6 AGP11 and Fasciclin-like arabinogalactan

proteins (FLA3) are the key genes involved in the synthesisof AGPs [95 96] and are in primarily expressed in thepollen grains and pollen tube with their participation inthe microspore development Alterations in the genes re-sult in obstruction in pollen tube growth and defectivepollen release from the anthers leading to reduced fertility[95 96] In our results we identified 3 differentiallyexpressed transcripts encoding arabinogalactan proteins(AGPs) such as GPI-anchored like protein and Fasciclin-like arabinogalactan 11 (FLA11) Of the 2 DEGs encodingGPI-anchored like proteins 1 was 50-fold down-regulated(Ccajan_38216_c1_g2_i3) in CMS genotype while theother gene showed up-regulation (725-fold) The DEGCcajan_37653_c0_g4_i5 represents FLA11 with 914-folddown-regulation in CMS line with respect to the sterileline Similar results were observed in watermelon cottonand sesame [55 61 97]Calcium ions (Ca2+) ions are known for their considerable

physiological significance in plants [98] They are activelyinvolved in plant reproduction specifically participating inpollen germination and pollen tube growth It was earlierreported that sufficient Ca2+ ions result in normal pollengermination while any changes in the concentration (higheror lower) restrain pollen germination and tube elongation[99 100] In the present study 2 DEGs Ccajan_47280_c0_g1_i1 and Ccajan_2910_c0_g1_i1 representing calmodulin-binding receptor-like proteins were identified 1 showing455-fold down-regulation and the other 753-fold down-regulation Altered expression of these genes in the sterileAKCMS11 might have inhibited pollen germination andtube growth Similar reports in watermelon cotton andkenaf support the present results [55 93 101]

Fig 13 Some of the candidate differentially expressed genes (DEGs) involved in cytoplasmic male-sterility sterility in pigeon pea The upperarrows indicate up-regulation of the DEGs and the down arrows represent down-regulation of the DEGs Each color box depicts an individual metabolism


During the final stages of pollen maturation the pollendehydrates and the mature pollen is ready for germination[102] Late embryogenesis abundant (LEA) proteins are in-volved in conferring desiccation tolerance to the pollen Inlily LP28 is a pollen-specific LEA-like protein is knownwhich slowly accumulates in the developing pollen andgenerously present in the germinated pollen after hydrationwith their probable role in pollen maturation and pollentube growth [103] In this investigation only 1 DEG wasfound related to LEA-like protein for example Ccajan_52697_c0_g1_i1 showing 457-fold down-regulation in theCMS line Abnormal expression of this gene may be relatedto pollen development The results are in accordance withthe results in watermelon and soybean [55 56]Male gametogenesis is a complex process in flowering

plants with rounds of mitotic and meiotic cell divisionsleading to the development of male gametophyte Anyabnormality in the genes regulating the cell division cancause male-sterility [104] In this study 2 mitotic spindleassembly checkpoint protein (MAD1) genes and 1 celldivision protein gene displayed abnormal expression inthe sterile AKCMS11 in comparison to fertile restorerAKPR303 A similar finding was reported in watermelon[55] suggesting it might be involved in pollen develop-ment The findings strongly indicate the potential role ofthese genes responsible for pollen development leadingto CMS in pigeon pea

DEGs encoding transcription factors potentially related toCMSTranscription factors (TFrsquos) are proteins which bind to cis-regulatory specific sequences in the promoter region of thetarget genes and regulate the gene expression [105] Anyalterations in the expression of transcription factors lead tochange in the gene expression causing substantial trans-formation during plant development [106] In Arabidopsis608 transcription factors (TFs) belonging to 34 familieshave been described to be participating in pollen develop-ment [107] In our study 9 and 7 DEGs encoding tran-scription factors showed down-regulation and up-regulation in the sterile AKCMS11 in comparison to itsfertile restorer respectively MYB proteins constitute alarge and functionally diverse family with a major role inplant-specific processes [108] In Arabidopsis AtMYB32and AtMYB103 are expressed in the tapetum and facilitatethe development of pollen by controlling tapetum develop-ment callose dissolution and exine formation [28 29] Anychanges in their expression can lead to early tapetal degen-eration distorted pollens and male-sterility [28 109] Inthis study 2 DEGs Ccajan_30780_c0_g1_i2 and Ccajan_36059_c0_g1_i1 encoding MYB transcription factors wereidentified both up-regulated (359-fold and 311-fold) inthe AKCMS11 Thus pointing towards the possible role ofthese differentially expressed genes in pollen development

Transcription factors like bHLH are significant as theysynchronize the process of normal tapetum developmentin the anthers bHLH- type transcription factors such asDYT1 and AMS in Arabidopsis [110] and UNDEVEL-OPED TAPETUM (OsUDT1) in rice [111] are known tobe crucial for proper tapetum development and finallyleading pollen development In our data we observed 3DEGs encoding bHLH transcription factors among those2 DEGs showed down-regulation in the CMS line Forinstance Ccajan_47215_c0_g1_i1 was 714-fold down-regulated and Ccajan_68741_c0_g2_i1 was 208-folddown-regulated Similar findings were seen in radish ses-ame and kenaf [54 97 101]In Arabidopsis WRKY transcription factors like WRKY2

WRKY34 and WRKY 27 are known for their candidate rolein pollen development during male gametogenesis [112]Over-expression of WRKY 27 showed limited pollen viabil-ity leading to male-sterility [113] In our study there was 1DEG which was over-represented in the AKCMS11 Cca-jan_36686_c0_g1_i4 (1005-fold up-regulated) this was inaccordance with findings in soybean and sesame [56 97]These DEGS recognized in this study might have a possiblerole causing CMS in this crop

DEGs involved in carbohydrate metabolism potentiallyrelated to CMSIn plants carbohydrates in the form of sugars and starchare the substrate for energy supply and growth Sufficientsugar level is highly important for the anther developmentand subsequently during pollen maturation sugars getconverted into starch which later serves as the energysource for pollen germination [114 115] Dysfunctionalsugar metabolism can significantly influence pollen devel-opment producing impaired pollens and results in male-sterility [116] This condition was observed in pepper andcabbage [117 118] In this study altered expression ofgenes involved in glycolysis gluconeogenesis and pentoseand glucuronate interconversions were detected in theAKCMS11 in comparison to the AKPR303 (Table 4)Total 6 DEGs were involved in glycolysis and gluconeo-genesis for instance Ccajan_36409_c0_g2_i1 Ccajan_34307_c0_g1_i1 Ccajan_37543_c0_g2_i4 Ccajan_34880_c0_g2_i1 Ccajan_39910_c0_g1_i1 and Ccajan_88886_c0_g1_i1 all of them were significantly down-regulated in thesterile AKCMS11 except for 1 DEG which showed up-regulation (Table 4) 6 DEGs were related to pentose andglucuronate interconversions 3 of which were down-regulated and 3 DEGs were up-regulated in AKCMS11These genes involved in carbohydrate metabolism exhib-ited lower expression suggesting the low availability ofsugars during pollen development resulting in non-functional pollens and male-sterility The results were inaccordance with the previous report in soybean chili pep-per and cotton [56 59 62]


DEGs involved in the tricarboxylic acid cycle (TCA) andoxidative phosphorylation potentially related to CMSMitochondria are the main energy yielding-power housesof the cells with a number of important metabolic path-ways involving TCA cycle respiratory electron transferand oxidative phosphorylation [119ndash121] Pollen formationis an energy-utilizing process which is highly dependent onmitochondrial respiration and fermentation for satisfyingtheir energy demands [102 122] We identified 2 DEGsparticipating in the TCA cycle Ccajan_64232_c0_g1_i1 en-coding malate dehydrogenase and Ccajan_35485_c0_g1_i1encoding succinate-CoA synthetase (Table 4) Accordingto the previous reports in cotton Brassica napus and cab-bage lower expression of TCA related genes causes pollenabortion and are responsible for CMS [93 123 124] Theseobservations suggest that alterations in the energy-yieldingmetabolic pathways inhibits pollen development and maybe the cause of CMSThe mitochondria generate energy in the form of adeno-

sine triphosphate (ATP) by utilizing two major pathwaysthe respiratory electron transfer chain and oxidative phos-phorylation [125] In this study 7 genes related to respira-tory electron transfer chain and oxidative phosphorylationwere detected showing differential expression in the sterileAKCMS11 in comparison to the fertile AKPR303 (Table4) Cytochrome c oxidase (cox) is the key enzyme involvedin the last stage of the mitochondrial electron transportchain and considered as a major regulating site duringATP synthesis [126] Previous literature on rice and cottonproposed the importance of cox genes in conferring CMS[127 128] Two DEGs Ccajan_13673_c0_g1_i1 and Cca-jan_41295_c0_g1_i1 encoding cytochrome c oxidase sub-unit 6b and subunit 1 were 598-fold and 646-fold down-regulated in the AKCMS11 (Table 4) This was in accord-ance with the earlier reports in chili pepper cotton Bras-sica napus and beet [59 93 123 129] Hence alteredexpression of cox 6 might have resulted in decreased ATPsynthesis and disrupted the pollen formation in this crop

DEGs involved in the elimination of reactive oxygenspecies (ROS) potentially related to CMSIn plants cells reactive oxygen species (ROS) are the by-products of aerobic respiration constantly being gener-ated in the chloroplast mitochondria and peroxisomesand are maintained by ROS scavenging system [130] TheROS are free radicals including superoxide (O2minus) hydro-gen peroxide (H2O2) and malondialdehyde (MDA) andtheir excessive accumulation in the cells can instigate cellapoptosis [131] Previously excessive accumulation of theROS species and remarkable reduction of the antioxidativedefense systems in the anthers of the cotton CMS linecaused male-sterility [132] In rice the abnormal concen-tration of the ROS in the mitochondria induced acute oxi-dative stress during pollen development and leading to

male-sterility [133] In this study 15 DEGs involved in theelimination of ROS were detected with 5 DEGs showingover-expression and 10 DEGs down-expression in thesterile AKCMS11 line in comparison to the fertileAKPR303 (Table 4) Down-regulation of the genes encod-ing oxygen scavenging enzymes in AKCMS11 triggeredthe excessive accumulation of ROS in the sterile budsleading to pollen abortion [134] In broccoli higher-expression of glutathione S-transferase (GST) gene in-duced excessive ROS and resulted in male-sterility [70]Interestingly we came across 1 DEG encoding glutathioneS-transferase (GST) for instance Ccajan_36196_c0_g2_i3showing significant higher-expression 241-fold inAKCMS11 in comparison to AKPR303 We presume ex-pression alterations of the genes involved in ROS scaven-ging probably induced pollen abortion in pigeon pea

ConclusionsTo our knowledge this is the first reported attempt at tran-scriptome sequencing and comparative analysis of the floralbuds of A2 CMS system (Cajanus scarabaeoides L) derivedcytoplasmic male-sterile (AKCMS11) and its fertility re-storer (AKPR303) in pigeon pea The differential gene ex-pression analysis showed 3167 differentially expressed genes(DEGs) between AKCMS11 and AKPR303 among which1432 were up-regulated and 1390 were down-regulated inthe AKCMS11 when compared to the AKPR303 Based onthe functional annotation of these DEGs in GO KEGGdatabases and metabolic pathway analysis combined withprevious studies we hereby conclude that male-sterility ofAKCMS11 is probably related to abnormalities of some ofthe putative DEGs in functional and metabolic pathwayssuch as involved in pollen development encoding transcrip-tion factors elimination of reactive oxygen species (ROS)carbon metabolism oxidative phosphorylation tricarboxylicacid cycle (TCA) etc Further studies of these crucial geneswill focus on clearly understanding their functions in con-ferring male-sterility Therefore this present report will behighly informative and provide support for future studies inelucidating the molecular basis related to cytoplasmic male-sterility in pigeon pea

MethodsPlant material and RNA isolationCytoplasmic male-sterile AKCMS11 (Cajanus scara-baeoides cytoplasm) and fertility restorer line AKPR303(Cajanus cajan cytoplasm) [47] of pigeon pea were usedin this present investigation as described earlier [135]The seeds for both the lines were procured from PulseResearch Unit Dr Panjabrao Deshmukh Krishi Vidya-peeth Krishi Nagar Akola Maharashtra India Theseeds were sown directly in the soil and the plants weremaintained under natural conditions in the experimentalgreenhouse at ICAR- NIPB New Delhi Flower buds of


different-sized were collected from two independentplants (representing replications) of each AKCMS11 andAKPR303 lines respectively The young buds were har-vested frozen in liquid nitrogen and stored at minus 80 degCTotal RNA was extracted from the flower buds of boththe lines (in replicates) using Spectrum Plant Total RNAKit (Sigma-Aldrich USA) following the manufacturerrsquosprotocol

Pollen fertility testPollen analysis was performed to evaluate pollen fertilityin the male-sterile (AKCMS11) and fertile (AKPR303)buds Anthers from the flower buds of AKCMS11 andAKPR303 were removed and squashed on a glass slidein 1 acetocarmine dye Then the glass slide was exam-ined under a microscope

Library preparation and RNA sequencingTotal RNA was checked on 1 denaturing agarose gel andquantified using Nanodrop spectrophotometer (Thermo Fi-scher Scientific USA) The RNA quality was analyzed withRNA 6000 NanoAssay kit (Agilent Technologies USA)using Agilent 2100 Bioanalyzer (Agilent TechnologiesUSA) Total four RNA libraries were constructed from thebuds of sterile and fertile lines (two replicates each) usingTruseq RNA Sample prep kit (Illumina USA) followingthe manufacturerrsquos protocol The libraries were then se-quenced by Illumina paired-end sequencing technology onIllumina HiSeq 1000 sequencer

RNA-Seq data analysis and de novo transcriptomeassemblySequencing raw data was received in the FASTQ formatThe per base sequence quality of the raw reads from ster-ile (41531572 68068793) and fertile (29618927 42957061) lines respectively were determined by FastQC version0114 (httpwwwbioinformaticsbabrahamacukpro-jects) [136] Rcorrector software was also employed forchecking the k-mer content in the data [137] Raw readswith PHRED quality score lt 30 and shorter than 50 bp inlength were filtered and not considered for further tran-scriptome assembly Trimmomatic software version 036[138] was used for trimming of the raw reads to removethe adaptor sequence followed by filtering of low-qualityreads (quality score le 5) and reads with the unknown lsquoNrsquobase (lsquoNrsquo ratio ge 10) to finally obtain the filtered cleanreads Next following the read orientation (R1 and R2)the total clean reads from the sterile and fertile lines wereconcatenated together by using an in-house shell script Ade novo assembly of the pooled clean reads was achievedby Trinity software version v220 [139] using default pa-rameters Further Bowtie2 aligner was used for the valid-ation of the transcriptome assembly by mapping back thefiltered reads onto the assembled transcriptome Then we

evaluated the completeness of our final transcriptome as-sembly with Benchmarking Universal Single-Copy Ortho-logs tool version 3 (BUSCO) [140 141] BUSCO is idealfor assessing assembly completeness because the expect-ation that these genes are found in a given genome as sin-gle copies is reasonable from an evolutionary point ofview [140ndash143] For the present analysis we used thetranscriptome assessment mode with the eukaryotelineage database (eukaryota_orthoDB9) and viridiplantaelineage database (viridiplantae_orthoDB10) Subsequentlynon-redundant unigenes were obtained by employing theCD-HIT software version 461with identity parameter of95 (httpweizhongli-laborgcd-hit) [144] Followed byremoval of rRNAs and other housekeeping non-codingRNAs (tRNAs snRNAs snoRNAs etc) by BLASTNsearch against Rfam database [145] GC content of thetotal unigenes was calculated via GC-Profile tool (httptubictjueducnGC-Profile) [146]

Functional annotation of the unigenesFunctional annotation of the assembled unigenes wasachieved by BLASTX search against Nr (NCBI non-redundant protein sequences) Virdiplantae GO (Geneontology) and KEGG (Kyoto encyclopedia of genes andgenomes) databases with a stringent cut-off E-value of 1eminus3 [147ndash149] Unigenes annotation against Nr database wasperformed with the aid of Blast2GO tool version 313(httpwwwblast2goorg) [150] up to 10 qualified blasthits for each unigene Among the 10 blast hits the onewith higher sequence similarity was accounted as a signifi-cant match for each unigene Transcription factors wereidentified with the help of PlantTFcat online tool (httpplantgrnnobleorgPlantTFcat) [73]

Differential gene expression analysisFirstly RSEM software [74] was used for calculating theread count of each uniquely mapped transcript on to theassembled transcriptome (for each sample) in terms ofFPKM (fragments per kilo base per million) Then differ-entially expressed genes (DEGs) between the sterile andfertile genotypes (two replications for each) were detectedby DESeq2 [61]) Strict criteria of P-value and FDR (Falsediscovery rate) le 005 and Log2 FC (Fold change ratio ster-ile vs fertile) ge 2 and le minus 2 (for up-regulated and down-regulated genes) were used for determining significant dif-ferentially expressed genes (DEGs)

GO and KEGG enrichment analysis of the DEGsGene ontology (GO) annotation of all the differentiallyexpressed genes (DEGs) was performed using SingularEnrichment Analysis Tool in AgriGO software version20 (httpbioinfocaueducnagriGO) with default pa-rameters [151] Based on GO enrichment results theones with P-value le005 were considered as significant


GO terms Then the functional GO classification of allthe DEGs was performed by WEGO software (wegogen-omicsorgcn) and the results were categorized intothree independent hierarchies biological processes mo-lecular functions and cellular components KEGG (Kyotoencyclopedia of genes and genomes) pathway enrich-ment analysis of all the DEGs was conducted by KEGGMapper (httpwwwgenomejpkeggmapperhtml) withdefault parameters The DEGs were compared to the en-tire reference gene set by hypergeometric tests to iden-tify the pathways enriched for DEGs and a P-value le 005indicated significant pathway enrichment AdditionallyMapMan version 33 0 [65] was used for metabolicpathway analysis of the DEGs by searching against Ara-bidopsis thaliana TAIR database (httpwwwarabidop-sisorg)

Validation of DEGs by quantitative real time-PCR (qRT-PCR)Total RNA was isolated from the buds of sterile(AKCMS11) and restorer (AKPR303) line as mentionedearlier First-strand cDNA was prepared from 2 μg of totalRNA using RevertAid First strand cDNA synthesis kit(Thermo Fischer Scientific USA) according to manufac-turerrsquos instructions The first strand cDNA was then furtherdiluted ten times for quantitative real-time PCR (qRT-PCR)reaction qRT-PCR was performed using KAPA SYBRFAST qPCR Master Mix (2X) (KAPA Biosystems USA) onABI PRISM 7500 Real-Time PCR System (Applied Biosys-tems USA) qRT-PCR was performed in three independenttechnical replicates for each genotype The reaction mixture(20 μl) contained 3 μl of cDNA 10 μl of 2X SYBR greenqPCR master mix 020 μl of primers (forward and reverse)and final volume was adjusted with nuclease-free waterPCR conditions were 94 degC for 3min followed by 40 cyclesof 94 degC for 3 s 60 degC for 15 s and 72 degC for 15 s α-tubulinwas used as the internal reference gene The relative levelof gene expression was calculated using the 2(minusΔΔCt) algo-rithm fertile restorer (AKPR303) was used as a calibrator

Supplementary informationSupplementary information accompanies this paper at httpsdoiorg101186s12870-020-2284-y

Additional file 1 Table S1 Rcorrector ouput for k-mer content in theraw paired-end data Table S2 Bowtie2 alignment statistics in sterile(AKCMS11) and fertile restorer (AKPR303) Figure S1 Pearsonrsquos correl-ation coefficient between two replicates in sterile AKCMS11 (left) and fer-tile restorer AKPR303 (right) Figure S2 Graph representing per basesequence quality a) Sterile AKCMS11 (replicates 1 and 2) b) FertileAKPR303 (replicates 1 and 2)

Additional File 2 Table S1 Sequences with significant BLASTX hitsagainst nr protein database Table S2 Sequences with significantBLASTX hits against Virdiplantae database Table S3 Gene Ontologyclassification of the assembled unigenes Table S4 KEGG classification ofCajanus cajan unigenes Table S5 Transcription factors identified in

transcriptome of Cajanus cajan Table S6 List of differentially expressedgenes between sterile AKCMS11 and fertile AKPR303 in pigeon peaTable S7 List of significantly enriched GO terms in down-regulatedDEGs Table S8 List of significantly enriched GO terms in up-regulatedDEGs Table S9 Summary of KEGG annotations for down and up-regulated DEGs Table S10 TAIR database BLASTX results of pigeon pea951 DEGs in MapMan pathway analysis Table S11 OrthoDB homologssearch of 3167 pigeon pea DEGs against TAIR database Table S12 Listof qRT-PCR primers used in this study

Additional file 3 Figure S1 Gene ontology classification of theassembled unigenes The Y-axis indicates the number of unigenes and X-axis indicates the GO categories Figure S2 Functional classification ofKEGG pathways of the assembled unigenes The KEGG pathways wereclassified into six functional categories A- Metabolism B- Genetic Infor-mation Processing C- Environmental Information Processing D- CellularProcesses E- Organismal Systems F- Human Diseases The Y-axis repre-sents the KEGG metabolic pathways The X-axis represents number of uni-genes annotated in that particular pathway Figure S3 Hierarchical treegraph of over-represented GO terms in the biological process category ofdown-regulated DEGS Boxes in the graph represent GO terms labeled ac-cording to their GO ID term definition and statistical information Signifi-cant terms (adjusted P le 005) are in color (red orange or yellow) whilenon-significant terms are shown as white boxes In the diagram the de-gree of color saturation of a box is positively correlated with the enrich-ment level of the term Solid dashed and dotted lines represent twoone and zero enriched terms at both ends connected by the line re-spectively The rank direction of the graph is set from top to bottom Fig-ure S4 Hierarchical tree graph of over-represented GO terms in thebiological process category of up-regulated DEGS Boxes in the graphrepresent GO terms labeled according to their GO ID term definitionand statistical information Significant terms (adjusted P le 005) are incolor (red orange or yellow) while non-significant terms are shown aswhite boxes In the diagram the degree of color saturation of a box ispositively correlated with the enrichment level of the term Solid dashedand dotted lines represent two one and zero enriched terms at bothends connected by the line respectively The rank direction of the graphis set from top to bottom

AbbreviationsACOS acyl-coA synthetase AGPs Arabinogalactan glycoproteinsAMS ABORTED MICROSPORES AP2 APETALA2 ARF AUXIN RESPONSIVEFACTOR ATP Adenosine triphosphate bHLH Basic helix-loop-helix proteinBLAST Basic Local Alignment Search Tool bp Base pair bZIP Basic leucinezipper C2H2 Cys2His2 Cals Callose synthase cDNA Complementary DNACMS Cytoplasmic male-sterility cox Cytochrome c oxidaseCYP703A2 CYTOCHROME P450 CYP704B1 CYTOCHROME P450DEGs Differentially expressed genes DYT DYSFUNCTIONAL TAPETUMEMS EXCESS MICROSPOROCYTES ERF Ethylene-responsive factor FDR Falsediscovery rate FLA Fasciclin-like arabinogalactan FPKM Fragments per kilobase per million GO Gene Ontology GST Glutathione S-transferaseH2O2 Hydrogen peroxide KEGG Kyoto Encyclopedia of Genes andGenomes LAP Polyketide synthase LEA Late embryogenesis abundant Log2FC Fold change LRR-RPK Leucine-rich repeat receptor protein kinaseMAD Mitotic spindle assembly checkpoint protein MDA MalondialdehydeMS MALE STERILITY NCBI National Center for Biotechnology InformationNr NCBI non-redundant protein NZZ NOZZEL O2 Superoxide orfs Openreading frames PCD Programmed cell death PCR Polymerase chainreaction PG Polygalacturonase PMC Pollen mother cells qRT-PCR Quantitative real-time PCR Rf Restorer of fertility RNA-Seq RNASequencing ROS Reactive oxygen species RPG RUPTURED POLLEN GRAINSPL SPOROCYTELESS TCA Tricarboxylic acid cycle TF Transcription factorsTPD TAPETUM DETERMINANT UDT UNDEVELOPED TAPETUM

AcknowledgementsThe authors are grateful to the Director ICAR- National Institute for PlantBiotechnology (NIPB) New Delhi India for providing facilities The authorsalso thank the other members of our research groups and collaborators fortechnical assistance and discussions


Authorsrsquo contributionsSS1 carried out the experiment prepared the cDNA library for sequencingsequencing run qRT-PCR validation GO KEGG pathway analysis identifica-tion of transcription factors and wrote the manuscript SS2 performed theread generation de novo assembly differential expression analysis hom-ology search and annotation TK was involved in cDNA library preparationand sequencing run DN and PKC assisted in the bioinformatics analysis SS1SS2 TK DN and PKC were involved in result interpretation analysis andfinalization of the manuscript NKS and ARR contributed in data analysis andmanuscript finalization SS3 contributed in manuscript finalization KG con-ceived the study designed the experiments and coordinated the work Allthe authors read and approved the final manuscript

FundingThis work was supported by funds from Department of Science ampTechnology (DST) New Delhi India and ICAR- Network Project onTransgenicsin Crops New Delhi India The funders had no role in the designof the study collection data analysis data interpretation and manuscriptwriting

Availability of data and materialsThe dataset generated andor analyzed during the current study areavailable in the NCBI SRA repository (SRX3740150 SRX3740151 SRX3740153SRX3740152) for the sterile and fertile buds(httpswwwncbinlmnihgovsraSRX3740153[accn] httpswwwncbinlmnihgovsraSRX3740152[accn]httpswwwncbinlmnihgovsraSRX3740151[accn] httpswwwncbinlmnihgovsraSRX3740150[accn]

Ethics approval and consent to participateNot applicable

Consent for publicationNot applicable

Competing interestsThe authors declare that they have no competing interests

Author details1ICAR-National Institute for Plant Biotechnology New Delhi 110012 India2ICAR-Indian Agricultural Statistics Research Institute New Delhi 110012India

Received 13 June 2019 Accepted 7 February 2020

References1 Eckardt NA Cytoplasmic male sterility and fertility restoration Plant Cell

200618515ndash72 Schnable PS Wise RP The molecular basis of male sterility and fertility

restoration Trends Plant Sci 19983175ndash803 Bentolila S Alfonso AA Hanson MR A pentatricopeptide repeat-containing

gene restores fertility to cytoplasmic male-sterile plants ProcNatlAcad Sci20029910887ndash92

4 Chen L Liu YG Male sterility and fertility restoration in crops Annu RevPlant Biol 201465579ndash606

5 Mackenzie SA The Influence of Mitochondrial Genetics on Crop BreedingStrategies In Janick J editor Plant Breeding Reviews Wiley 2010 p 115ndash38

6 Levings CS The Texas cytoplasm of maize cytoplasmic male sterility anddisease susceptibility Science 1990250942ndash7

7 Levings CS Thoughts on cytoplasmic male sterility in cms-T maize PlantCell 199351285ndash90

8 Nivison HT Sutton CA Wilson RK Hanson MR Sequencing processing andlocalization of the petunia CMS-associated mitochondrial protein Plant J19945613ndash23

9 Young EG Hanson MR A fused mitochondrial gene associated withcytoplasmic male sterility is developmentally regulated Cell 19875041ndash9

10 Budar F Pelletier G Male sterility in plants occurrence determinismsignificance and use CR AcadSci III 2001324543ndash50

11 Omidvar V Mohorianu I Dalmay T Zheng Y Fei Z Pucci A et alTranscriptional regulation of male-sterility in 7B-1 male-sterile tomatomutant PLoS One 2017121ndash19

12 Shao LH Zheng XW Yi DX Li C Comparative sequence and expressionanalysis of tapetum specific male sterility related genes in Medicagotruncatula GenetMol Res 2016152

13 Chase CD Cytoplasmic male sterility a window to the world of plantmitochondrial-nuclear interactions Trends Genet 20072381ndash90

14 Hanson MR Bentolila S Interactions of mitochondrial and nuclear genesthat affect male gametophyte development Plant Cell 200416154ndash69

15 Bergman P Edqvist J Farbos I Glimelius K Male-sterile tobacco is playsabnormal mitochondrial atp1 transcript accumulation and reduced floralATPADP ratio Plant Mol Biol 200042531ndash44

16 Sabar M Gagliardi D Balk J Leaver CJ ORFB is a subunit of F1F0-ATPsynthase insight into the basis of cytoplasmic male sterility in sunflowerEMBO Rep 200341ndash6

17 Yui R Iketani S Mikami T Kubo T Antisense inhibition of mitochondrialpyruvate dehydrogenase E1asubunit in anther tapetum causes male sterilityPlant J 20033457ndash66

18 Liu X Huang J Parameswaran S Ito T Seubert B Auer M et al TheSPOROCYTELESSNOZZLE gene is involved in controlling stamen identity inArabidopsis Plant Physiol 20091511401ndash11

19 Ito T Wellmer F Yu H Das P Ito H Alves-Ferreira M et al The homeoticprotein AGAMOUS controls microsporogenesis by regulation ofSPOROCYTELESS Nature 2004430356ndash60

20 Schiefthaler U Balasubramanian S Sieber P Chevalier D Wisman E SchneitzK Molecular analysis of NOZZLE a gene involved in pattern formation andearly sporogenesis during sex organ development in Arabidopsis thalianaProcNatlAcad Sci 19999611664ndash9

21 Yang WC Ye D Xu J Sundaresan V The SPOROCYTELESS gene ofArabidopsis is required for initiation of sporogenesis and encodes a novelnuclear protein Genes Dev 1999132108ndash17

22 Zhao DZ Wang GF Speal B Ma H The EXCESS MICROSPOROCYTES1 geneencodes a putative leucine-rich repeat receptor protein kinase that controlssomatic and reproductive cell fates in the Arabidopsis anther Genes Dev2002162021ndash31

23 Jia G Liu X Owen HA Zhao D Signaling of cell fate determination by the TPD1small protein and EMS1 receptor kinase ProcNatlAcad Sci 20081052220ndash5

24 Yang SL Xie LF Mao HZ Puah CS Yang WC Jiang L et al TAPETUMDETERMINANT1 is required for cell specialization in the Arabidopsis antherPlant Cell 2003152792ndash04

25 Xu J Yang C Yuan Z Zhang D Gondwe MY Ding Z et al The ABORTEDMICROSPORES regulatory network is required for post-meiotic malereproductive development in Arabidopsis thaliana Plant Cell 20102291ndash107

26 Sorensen AM Kroumlber S Unte US Huijser P Dekker K Saedler H TheArabidopsis ABORTED MICROSPORES (AMS) gene encodes a MYC classtranscription factor Plant J 200333413ndash23

27 Yang C Vizcay-Barrena G Conner K Wilson ZA MALE STERILITY1 is required forTapetal development and Pollen Wall biosynthesis Plant Cell 2007193530ndash48

28 Preston J Wheeler J Heazlewood J Li SF Parish RW AtMYB32 is required fornormal pollen development in Arabidopsis thaliana Plant J 200440979ndash95

29 Zhang ZB Zhu J Gao JF Wang C Li H Li H et al Transcription factor AtMYB103 isrequired for anther development by regulating tapetum development callosedissolution and exine formation in Arabidopsis Plant J 200752528ndash38

30 Aarts MGM Keijzer CJ Stiekema WJ Pereira A Molecular characterization ofthe CER7 gene of Arabidopsis lnvolved in Epicuticular wax biosynthesis andpollen fertility Am SocPlant Physiol 199572115ndash27

31 Ariizumi T Hatakeyama K Hinata K Sato S Kato T Tabata S et al A novel male-sterile mutant of Arabidopsis thaliana faceless pollen-1 produces pollen with asmooth surface and an acetolysis-sensitive exine Plant Mol Biol 200353107ndash16

32 Guan YF Huang XY Zhu J Gao JF Zhang HX Yang ZN RUPTURED POLLENGRAIN1 a member of the MtN3saliva gene family is crucial for Exinepattern formation and cell integrity of microspores in Arabidopsis PlantPhysiol 2008147852ndash63

33 Morant M Jorgensen K Schaller H Pinot F Moller BL Werck-Reichhart Det al CYP703 is an ancient cytochrome P450 in land plants catalyzing in-chain hydroxylation of Lauric acid to provide building blocks forSporopollenin synthesis in pollen Plant Cell Online 2007191473ndash87

34 Dobritsa AA Shrestha J Morant M Pinot F Matsuno M Swanson R et alCYP704B1 is a long-chain fatty acid ω-hydroxylase essential for Sporopolleninsynthesis in pollen of Arabidopsis Plant Physiol 2009151574ndash89

35 de Azevedo SC Kim SS Koch S Kienow L Schneider K McKim SM et al Anovel fatty acyl-CoA Synthetase is required for pollen development andSporopollenin biosynthesis in Arabidopsis Plant Cell 200921507ndash25


36 Kim SS Grienenberger E Lallemand B Colpitts CC Kim SY Souza Cde Aet al LAP6POLYKETIDE SYNTHASE A and LAP5POLYKETIDE SYNTHASE BEncode Hydroxyalkyl α-Pyrone Synthases Required for Pollen Developmentand Sporopollenin Biosynthesis in Arabidopsis thaliana Plant Cell 2010224045ndash66

37 Dobritsa AA Lei Z Nishikawa SI Urbanczyk-Wochniak E Huhman DV PreussD et al LAP5 and LAP6 encode anther-specific proteins with similarity toChalcone synthase essential for pollen Exine development in ArabidopsisPlant Physiol 2010153937ndash55

38 Dong X Hong Z Sivaramakrishnan M Mahfouz M Verma DPS Callosesynthase (CalS5) is required for exine formation during microgametogenesisand for pollen viability in Arabidopsis Plant J 200542315ndash28

39 Fu YJ Liu W Zu YG Tong MH Li SM Yan MM et al Enzyme assistedextraction of luteolin and apigenin from pigeon pea [Cajanus cajan (L)Millsp] leaves Food Chem 2008111508ndash12

40 Saxena KB Kumar RV Rao PV Pigeon pea nutrition and its improvement InBasra AS Randhawa IS editors Quality improvement in field crops NewYork Food Products Press 2002 p 227ndash60

41 Odeny DA The potential of pigeonpea (Cajanus cajan (L) Millsp) in AfricaNat Resour Forum 200731297ndash05

42 Nadarajan N Ram SG Petchiammal KI Fertility restoration studies in shortduration redgram (Cajanus cajan (L) mill spp) hybrids involving cgmssystem Madras Agric J 200895320ndash7

43 Saxena KB Sharma D Pigeonpea genetics In Nene YL et al editorsPigeonpea Wellingford CAB International 1990 p 137ndash58

44 Saxena KB A novel source of CMS in pigeon pea derived from Cajanusreticulatus Indian J of Genet Plant Breed 201373259ndash63

45 Saxena KB Sultana R Mallikarjuna N Saxena RK Kumar RV Sawargaonkar SLet al Male-sterility systems in pigeonpea and their role in enhancing yieldPlant Breed 2010129125ndash34

46 Saxena KB Nadarajan N Prospects of pigeonpea hybrids in Indianagriculture Elec J Plant Breed 201011107ndash17

47 Meshram MP Patil AN Genetics of Fertility Restoration in A2 Cytoplasmbased Hybrids of Pigeonpea (Cajanus cajan (L) Millsp) Int J CurrMicrobiolApp Sci 20186565ndash71

48 Choudhary AK Singh IP A study on comparative fertility restoration in A2and A4 cytoplasm and its implication in breeding hybrid pigeonpea[Cajanus cajan (L) Millspaugh] Am J Plant Sci 20156385ndash91

49 Sinha P Saxena KB Saxena RK Singh VK Suryanarayana V SK SKCV MAVSKK Khan AW Varshney R Association of Gene with Cytoplasmic MaleSterility in Pigeonpea Plant Genome 20158 httpsdoiorg103835plantgenome2014110084

50 Schuster SC Next-generation sequencing transforms todayrsquos biology NatMethods 2008516ndash8

51 Ansorge WJ Next-generation DNA sequencing techniques New Biotechnol200925195ndash03

52 Love MI Huber W Anders S Moderated estimation of fold change anddispersion for RNA-seq data with DESeq2 Genome Biol 201415550

53 Zenoni S Ferrarini A Giacomelli E Xumerle L Fasoli M Malerba G et alCharacterization of transcriptional complexity during berry development inVitis vinifera using RNA-Seq Plant Physiol 20101521787ndash95

54 Mei S Liu T Wang Z Comparative Transcriptome profile of the cytoplasmicmale sterile and fertile floral buds of radish (Raphanus sativus L) Int J MolSci 20161742

55 Rhee SJ Seo M Jang YJ Cho S Lee GP Transcriptome profiling ofdifferentially expressed genes in floral buds and flowers of male sterile andfertile lines in watermelon BMC Genomics 201516914

56 Li J Han S Ding X He T Dai J Yang S et al Comparative Transcriptomeanalysis between the cytoplasmic male sterile line NJCMS1A and itsmaintainer NJCMS1B in soybean (Glycine max (L) Merr) PLoS One 2015101ndash18

57 Jeong HJ Kang JH Zhao M Kwon JK Choi HS Bae JH et al Tomato Malesterile 1035 is essential for pollen development and meiosis in anthers J ExpBot 2014656693ndash09

58 An H Yang Z Yi B Wen J Shen J Tu J et al Comparative transcriptprofiling of the fertile and sterile flower buds of pol CMS in B napus BMCGenomics 2014152ndash11

59 Liu C Ma N Wang PY Fu N Shen HL Transcriptome sequencing and Denovo analysis of a cytoplasmic male sterile line and its near-isogenicrestorer line in chili pepper (Capsicum annuum L) PLoS One 20138e65209httpsdoiorg101371journalpone0065209

60 Wu J Zhang M Zhang B Zhang X Guo L Qi T et al Genome-widecomparative transcriptome analysis of CMS-D2 and its maintainer andrestorer lines in upland cotton BMC Genomics 201718454

61 Suzuki H Rodriguez-Uribe L Xu J Zhang J Transcriptome analysis ofcytoplasmic male sterility and restoration in CMS-D8 cotton Plant Cell Rep2013321531ndash42

62 Wei M Song M Fan S Yu S Transcriptomic analysis of differentially expressedgenes during anther development in genetic male sterile and wild type cottonby digital gene-expression profiling BMC Genomics 20131497

63 Zheng BB Wu XM Ge XX Deng XX Grosser JW Guo WW Comparativetranscript profiling of a male sterile Cybrid Pummelo and its fertile type revealedaltered gene expression related to flower development PLoS One 201271ndash13

64 Dai X Sinharoy S Udvardi M Zhao PX PlantTFcat An online planttranscription factor and transcriptional regulator categorization and analysistool BMC Bioinformatics 201314321

65 Thimm O Blasing O Gibon Y Nagel A Meyer S Kruger P et al MAPMAN auser-driven tool to display genomics data sets onto diagrams of metabolicpathways and other biological processes Plant J 200437914ndash39

66 Kriventseva EK et al OrthoDB v10 sampling the diversity of animal plantfungal protist bacterial and viral genomes for evolutionary and functionalannotations of orthologs NAR 2018 httpsdoiorg101093nargky1053

67 Yanagisawa S Transcription factors in plants physiological functions andregulation of expression J Plant Res 1998111363ndash71

68 Wu Z Cheng J Qin C Hu Z Yin C Hu K Differential proteomic analysis ofanthers between cytoplasmic male sterile and maintainer lines in Capsicumannuum L Int J Mol Sci 20131422982ndash96

69 Blackmore S Wortley AH Skvarla JJ Rowley JR Pollen wall development inflowering plants New Phytol 2007174483ndash98

70 Pei X Jing Z Tang Z Zhu Y Comparative transcriptome analysis providesinsight into differentially expressed genes related to cytoplasmic malesterility in broccoli (Brassica oleracea var italica) SciHortic 2017217234ndash42

71 Goacutemez JM Perfectti F Abdelaziz M Lorite J Muntildeoz-Pajares AJ Valverde JEvolution of pollination niches in a generalist plant clade New Phytol 2015205440ndash53

72 Heslop-Harrison J Dickinson HG Time relationships of sporopollenin synthesisassociated with tapetum and microspores in Lilium Planta 196984199ndash14

73 Aarts MGM Hodge R Kalantidis K Florack D Wilson ZA Mulligan BJ et alThe Arabidopsis MALE STERILITY 2 protein shares similarity with reductases inelongationcondensation complexes Plant J 199712615ndash23

74 Wilson ZA Morroll SM Dawson J Swarup R Tighe PJ The Arabidopsis MALESTERILITY1 (MS1) gene is a transcriptional regulator of male gametogenesiswith homology to the PHD-finger family of transcription factors Plant J20012827ndash39

75 Goacutemez JF Wilson ZA A barley PHD finger transcription factor that confers malesterility by affecting tapetal development Plant Biotechnol J 201412765ndash77

76 Rick CM Genetics and development of nine male-sterile tomato mutantsHilgardia 194818599ndash33

77 Gorman SW McCormick S Rick C Male sterility in tomato CRC Crit RevPlant Sci 19971631ndash53

78 Scott RJ Spielman M Dickinson HG Stamen structure and function PlantCell 20041646ndash60

79 Ariizumi T Toriyama K Genetic regulation of Sporopollenin synthesis andpollen Exine development Annu Rev Plant Biol 201162437ndash60

80 Hong Z Delauney AJ Verma DPS A cell plate-specific Callose synthase andits interaction with Phragmoplastin Plant Cell 200113755ndash68

81 Enns LC Kanaoka MM Torii KU Comai L Okada K Cleland RE Two callosesynthases GSL1 and GSL5 play an essential and redundant role in plantand pollen development and in fertility Plant Mol Biol 200558333ndash49

82 Yang J Tian L Sun MX Huang XY Zhu J Guan YF et al AUXIN RESPONSEFACTOR17 is essential for Pollen Wall pattern formation in Arabidopsis PlantPhysiol 2013162720ndash31

83 Wu Y Min L Wu Z Yang L Zhu L Yang X et al Defective pollen wallcontributes to male sterility in the male sterile line 1355A of cotton Sci Rep201551ndash8

84 Scott RJ Pollen exine the sporopollenin enigma and the physics of patternIn Scott RJ Stead AD editors Molecular and cellular aspects of plantreproduction Cambridge Cambridge University Press 1994 p 49ndash81

85 Mascarenhas JP The male gametophyte of flowering plants Plant Cell19891657ndash64 httpsdoiorg101105tpc17657

86 Maheshwari P An introduction to the embryology of angiosperms NewYork McGraw-Hill Book Co Inc 1950


87 Kawanabe T Ariizumi T Kawai-Yamada M Uchimiya H Toriyama KAbolition of the tapetum suicide program ruins microsporogenesis PlantCell Physiol 200647784ndash7

88 Rogers LA Dubos C Surman C Willment J Cullis IF Mansfield SD et alComparison of lignin deposition in three ectopic lignification mutants NewPhytol 2005168123ndash40

89 Parish RW Li SF Death of a tapetum a programme of developmentalaltruism Plant Sci 201017873ndash89

90 Vizcay-Barrena G Wilson ZA Altered tapetal PCD and pollen walldevelopment in the Arabidopsis ms1 mutant J Exp Bot 2006572709ndash17

91 Ge X Dietrich C Matsuno M Li G Berg H Xia Y An Arabidopsis asparticprotease functions as an anti-cell-death component in reproduction andembryogenesis EMBO Rep 20056282ndash8

92 Rhee SY Somerville CR Tetrad pollen formation in quartet mutants ofArabidopsis thaliana is associated with persistence of pectic polysaccharidesof the pollen mother cell wall Plant J 19981579ndash88

93 Kong X Liu D Liao X Zheng J Diao Y Liu Y et al Comparative analysis ofthe cytology and transcriptomes of the cytoplasmic male sterility lineH276A and its maintainer line H276B of cotton (Gossypium barbadense L)Int J Mol Sci 2017182240

94 Pennell RI Roberts K Sexual development in the pea is presaged by alteredexpression of arabinogalactan protein Nature 1990334547ndash9

95 Levitin B Richter D Markovich I Zik M Arabinogalactan proteins 6 and 11 arerequired for stamen and pollen function in Arabidopsis Plant J 200856351ndash63

96 Li J Yu M Geng LL Zhao J The fasciclin-like arabinogalactan protein geneFLA3 is involved in microspore development of Arabidopsis Plant J 201064482ndash97

97 Liu H Tan M Yu H Li L Zhou F Yang M et al Comparative transcriptomeprofiling of the fertile and sterile flower buds of a dominant genic malesterile line in sesame (Sesamum indicum L) BMC Plant Biol 201616250

98 Bush DS Calcium regulation in plant cells and its role in signaling AnnuRev Plant Physiol Plant Mol Biol 19954695ndash122

99 Brewbaker JL Kwack BH The essential role of calcium ion in pollengermination and pollen tube growth Am J Bot 196350859ndash65

100 Bednarska E The effect of exogenous Ca2+ ions on pollen graingermination and pollen tube growth investigations with 45Ca2+ togetherwith verapamil La3+ and ruthenium red Sex Plant Reprod 1988253ndash8

101 Chen P Ran S Li R Huang Z Qian J Yu M et al Transcriptome de novoassembly and differentially expressed genes related to cytoplasmic malesterility in kenaf (Hibiscus cannabinus L) Mol Breed 2014341879ndash91

102 Taylor LP Hepler PK Pollen germination and tube growth Annu Rev PlantPhysiol Plant Mol Biol 199748461ndash91

103 Mogami N Shiota H Tanaka I LP28 a lily pollen-specific LEA-like protein islocated in the callosic cell wall during male gametogenesis Sex PlantReprod 20021557ndash63

104 Chen C Marcus A Li W Hu Y Calzada JPV Grossniklaus U et al TheArabidopsis ATK1 gene is required for spindle morphogenesis in malemeiosis Development 20021292401ndash9

105 Guilfoyle TJ The structure of plant gene promoters In Setlow JK editorGenetic engineering New York Plenum Press 1997 p 15ndash47

106 Grotewold E Chamberlin M Snook M Siame B Butler L Swenson J et alEngineering secondary metabolism in maize cells by ectopic expression oftranscription factors Plant Cell 199810721ndash40

107 Honys D Twell D Transcriptome analysis of haploid male gametophytedevelopment in Arabidopsis Genome Biol 20045R85

108 Dubos C Stracke R Grotewold E Weisshaar B Martin C Lepiniec L MYBtranscription factors in Arabidopsis Trends Plant Sci 201015573ndash81

109 Higginson T Li SF Parish RW AtMYB103 regulates tapetum and trichomedevelopment in Arabidopsis thaliana Plant J 200335177ndash92

110 Zhang W Sun Y Timofejeva L Chen C Grossniklaus UMH Regulation ofArabidopsis tapetum development and function by DYSFUNCTIONALTAPETUM1 (DYT1) encoding a putative bHLH transcription factorDevelopment 20061333085ndash95

111 Jung KH Han MH Lee YS Kim YW Hwang I Kim MJ et al Rice UndevelopedTapetum1 is a major regulator of early Tapetum development Plant Cell2005172705ndash22

112 Guan Y Meng X Khanna R LaMontagne E Liu Y Zhang S Phosphorylationof a WRKY transcription factor by MAPKs is required for pollen developmentand function in Arabidopsis PLoS Genet 201410e1004384

113 Mukhtar MS Liu X Somssich IE Elucidating the role of WRKY27 in malesterility in Arabidopsis Plant Signal Behav 201712e1363945

114 Oliver SN Dennis ES Dolferus R ABA regulates apoplastic sugar transportand is a potential signal for cold-induced pollen sterility in rice Plant CellPhysiol 2007481319ndash30

115 Datta R Chamusco KC Chourey PS Starch biosynthesis during pollenmaturation is associated with altered patterns of gene expression in maizePlant Physio 20021301645ndash56

116 Mamun EA Alfred S Cantrill LC Overall RL Sutton BG Effects of chilling onmale gametophyte development in rice Cell Biol Int 200630583ndash91

117 Li YY Wei YY Zhang RH Lu J Studies on the metabolism of male sterilepepper in the development of microspore ActaAgriculturaeBoreali-OccidentalisSinica 200615134ndash7

118 Wang YF Hu CQ Lin ZG Li FY Jin JK et al ActaHort Sin 198437119 Siedow J Umbach A Plant mitochondrial Electron transfer and molecular

biology Plant Cell 19957821ndash31120 Barrientos A Yeast models of human mitochondrial diseases IUBMB Life

20035583ndash95121 Logan DC The mitochondrial compartment J Exp Bot 2006571225ndash43122 Dickinson DB Germination of lily pollen respiration and tube growth

Science 196515030ndash1123 Du K Liu Q Wu X Jiang J Wu J Fang Y et al Morphological structure and

Transcriptome comparison of the cytoplasmic male sterility line in Brassicanapus (SaNa-1A) derived from somatic hybridization and its maintainer lineSaNa-1B Front Plant Sci 201671313

124 Wang S Wang C Zhang XX Chen X Liu JJ Jia XF et al Transcriptomede novo assembly and analysis of differentially expressed genes relatedto cytoplasmic male sterility in cabbage Plant Physiol Biochem 2016105224ndash32

125 Millar AH Whelan J Soole KL Day DA Organization and regulation ofmitochondrial respiration in plants Annu Rev Plant Biol 20116279ndash104

126 Kadenbach B Huumlttemann M Arnold S Lee I Bender E Mitochondrialenergy metabolism is regulated via nuclear-coded subunits of cytochromec oxidase Free RadicBiol Med 200029211ndash21

127 Liu Y Wang X Wang Y Zhuo D Structural variance analysis of mitochondriacoI and coII genes from normal and cytoplasmic male-sterile varieties ofrice Oryza sativa ActaGenetica Sin 198815348ndash54

128 Huang J Yang P Li B Huang JL Yang P et al Study on the activity ofseveral enzymes of cytoplasmic male-sterile cotton line Jin A CottonScience 200416229ndash32

129 Ducos E Touzet P Boutry M The male sterile G cytoplasm of wild beetdisplays modified mitochondrial respiratory complexes Plant J 200126171ndash80

130 Sharma P Jha AB Dubey RS Pessarakli M Reactive oxygen speciesoxidative damage and Antioxidative defense mechanism in plants understressful conditions J Bot 201220121ndash26

131 Maxwell DP Nickels R McIntos L Evidence of mitochondrial involvement inthe transduction of signals required for the induction of genes associatedwith pathogen attack and senescence Plant J 200229269ndash79

132 Jiang P Zhang X Zhu Y Zhu W Xie H Wang X Metabolism of reactiveoxygen species in cotton cytoplasmic male sterility and its restoration PlantCell Rep 2007261627ndash34

133 Wan C Li S Wen L Kong J Wang K Zhu Y Damage of oxidative stress onmitochondria during microspores development in Honglian CMS line ofrice Plant Cell Rep 200726373ndash82

134 Hanson MR Plant mitochondrial mutations and male-sterility Annu RevGenet 199125461ndash86

135 Junaid A Kumar H Rao AR Patil AN Singh NK Gaikwad K Unravelling theepigenomic interactions between parental inbreds resulting in an alteredhybrid methylome in pigeonpea DNA Res 201825361ndash73

136 Andrews S FastQC a quality control tool for high throughput sequencedata 2010 Available online at httpwwwbioinformaticsbabrahamacukprojectsfastqc

137 Song L Liliana F Rcorrector efficient and accurate error correction forIllumina RNA-seq reads Giga Science 2015448

138 Bolger AM Lohse M Usadel B Trimmomatic a flexible trimmer for Illuminasequence data Bioinformatics 2014302114ndash20

139 Grabherr MG Haas BJ Yassour M Levin JZ Thompson DA Amit I et alTrinity reconstructing a full-length transcriptome without a genome fromRNA-Seq data Nat Biotechnol 201129644ndash52

140 Simatildeo FA Waterhouse RM Ioannidis P Kriventseva EV Zdobnov EMBUSCO assessing genome assembly and annotation completeness withsingle-copy orthologs Bioinformatics 2015313210ndash2


141 Waterhouse RM Seppey M Simatildeo FA Manni M Ioannidis P KlioutchnikovG et al BUSCO applications from quality assessments to gene predictionand phylogenomics Mol Biol Evol 201735543ndash8

142 Waterhouse RM Zdobnov EM Kriventseva EV Correlating traits of generetention sequence divergence duplicability and essentiality in vertebratesarthropods and fungi Genome Biol Evol 2011375ndash86

143 Waterhouse RM Tegenfeldt F Li J Zdobnov EM Kriventseva EV OrthoDB ahierarchical catalog of animal fungal and bacterial orthologs Nucleic AcidsRes 201341D358ndashD65

144 Li W Godzik A Cd-hit a fast program for clustering and comparing largesets of protein or nucleotide sequences Bioinformatics 2006221658ndash9

145 Kalvari I Argasinska J Quinones-Olvera N Nawrocki EP Rivas E Eddy SRet al Rfam 130 shifting to a genome-centric resource for non-coding RNAfamilies Nucleic Acids Res 201846D335ndash42

146 Gao F Zhang CT GC-profile a web-based tool for visualizing and analyzingthe variation of GC content in genomic sequences Nucleic Acids Res 200634686ndash91

147 Devani RS Sinha S Banerjee J Sinha RK Bendahmane A Banerjee AK Denovo transcriptome assembly from flower buds of dioeciousgynomonoecious and chemically masculinized female Coccinia grandisreveals genes associated with sex expression and modification BMC PlantBiol 201717241

148 Fu Y Esselink GD Visser RGF Tuyl JM Arens P Transcriptome analysis ofGerbera hybrida including in silico confirmation of defense genes foundFront Plant Sci 20167247

149 Morandin C Pulliainen U Bos N Schultner E De novo transcriptomeassembly and its annotation for the black ant Formica fusca at the larvalstage Sci Data 20185180282

150 Conesa A Goumltz S Garciacutea-Goacutemez JM Terol J Taloacuten M Robles M Blast2GO auniversal tool for annotation visualization and analysis in functionalgenomics research Bioinformatics 2005213674ndash6

151 Tian T Liu Y Yan H You Q Yi X Du Z et al AgriGO v20 a GO analysis toolkit forthe agricultural community 2017 update Nucleic Acids Res 201745W122ndash9

Publisherrsquos NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations


Abstract

Background

Results

Conclusion

Background

Results

Pollen fertility analysis and phenotypic characterization of sterile and fertile buds

Transcriptome sequencing and de novo assembly

Functional annotation and classification of the unigenes

Identification of transcription factors

Identification of differentially expressed genes between lines AKCMS11 and AKPR303

GO and KEGG enrichment analysis of the DEGs

Candidate DEGs associated with male-sterility

Putative gene related to pollen development

Metabolic pathways

Confirmation of DEGs by qRT-PCR

Discussion

DEGs involved in pollen development potentially related to CMS

DEGs encoding transcription factors potentially related to CMS

DEGs involved in carbohydrate metabolism potentially related to CMS

DEGs involved in the tricarboxylic acid cycle (TCA) and oxidative phosphorylation potentially related to CMS

DEGs involved in the elimination of reactive oxygen species (ROS) potentially related to CMS

Conclusions

Methods

Plant material and RNA isolation

Pollen fertility test

Library preparation and RNA sequencing

RNA-Seq data analysis and de novo transcriptome assembly

Functional annotation of the unigenes

Differential gene expression analysis


Validation of DEGs by quantitative real time-PCR (qRT-PCR)

Supplementary information

Abbreviations

Acknowledgements

Authorsrsquo contributions

Funding

Availability of data and materials

Ethics approval and consent to participate

Consent for publication

Competing interests

Author details

References

Publisherrsquos Note

along with functional dominant Rf gene(s) The F1 hybridsproduced possess dominant Rf gene to restores the male-fertility and the blend of nuclear genes from the sterile andfertile lines results in hybrid vigor [2] During the 1950sthe first CMS system used for hybrid corn production wasmaize CMS-Texas (CMS-T) system which substantiallyincreased the hybrid seed production and remarkably en-hanced the maize yield [6] Till date impressive advance-ment has been made in plants to comprehend themolecular basis of CMS and fertility restoration A numberof mitochondrial genes related to male-sterility have beenreported such as urf13 from maize [7] pcf associated withatp9 gene from petunia [8 9] and the radish atp8 of OguraCMS [10] However in pigeon pea the gene associated withcytoplasmic male-sterility has not been much exploredIn flowering plants anther development is a highly con-

trolled mechanism which demands proper differentiation ofthe sporogenous tissue and timely programmed microspo-rogenesis [11] The absence of male gametophyte (stamen)or defective development of anther disruption of the pollenmother cells (PMC) or tapetum cells abnormal develop-ment of pollen and failure of anther dehiscence leads topollen abortion [12 13] In all cases CMS due to the dys-functional mitochondrial genome is mainly responsible forthe disruption of pollen development [14] Previous studieson CMS provide definitive evidence that the organellemitochondria are critical for the tapetum and microsporesdevelopment [14] During distinct stages of pollen develop-ment the requirement of energy is extremely high andeven a slight defect in the mitochondrial function becomesdeleterious and results in pollen sterility [15ndash17] Previousstudies have characterized many genes playing a pivotal rolein anther development in Arabidopsis viz SPOROCYTE-LESS (SPL)NOZZEL (NZZ) [18ndash21] EXCESS MICROSPO-ROCYTES 1 (EMS1) [22 23] TAPETUM DETERMINANT1 (TPD1) [24] ABORTED MICROSPORES (AMS) andMALE STERILITY1 (MS1) [25ndash27] MYB gene family (AtMYB103) [28 29] RUPTURED POLLEN GRAIN1 (RPG1)CYTOCHROME P450 (CYP703A2) CYTOCHROME P450(CYP704B1) acyl-coA synthetase5 (acos5) LAP6 and LAP5[30ndash37] redundantly facilitate anther development and cal-lose synthase5 (cals5) is essential for callose synthesis [38]Pigeon pea (Cajanus cajan (L) Millsp) belongs to the

family Fabaceae and regarded as the sixth major legumecrop worldwide [39] It is popularly known for its high pro-tein content [40] and drought tolerance [41] It is an im-portant pulse crop worldwide and predominantly cultivatedin the tropics and sub-tropics (httpfaostatfaoorg) Tomeet the ever-growing demand for pigeon pea extensiveresearch has been done during the last decade a break-through has been accomplished in developing hybrid tech-nology [42] The pre-imperatives for large-scale hybrid seedproduction are pollen transfer method and advancement ofstable cytoplasmic male-sterile (CMS) system The male-

sterility phenomenon could not be much explored in otherlegumes due to their self-pollinating nature Howeverpigeon pea has a considerable natural out-crossing [43] Tilldate eight (A1 to A8) such CMS systems have been bred inpigeon pea by transferring the cytoplasm of the wild rela-tives in the nuclear background of the cultivated pigeon peavia interspecific-hybridization followed by several rounds ofback-crossing [44] Therefore hybrid seed technology basedon cytoplasmic male-sterility has given a chance to over-come the production constraints by efficiently increasingthe yield potential in pigeon pea [45] However among theeight systems A2 and A4 CMS system obtained from thewild-type Cajanus scarabaeoides and Cajanus cajanifoliusrespectively are regarded as highly potential CMS systems[46ndash48] The gene associated with CMS in A4 cytoplasmpigeon pea has been reported [49] Recently the genetics offertility restoration in pigeon pea with A2 cytoplasm wasstudied in detailed [47] Even though A2 cytoplasm hasplayed a significant role in hybrid seed production yet theunderlying molecular processes governing CMS in A2 cyto-plasm pigeon pea is still not knownPresently the next generation sequencing technology has

revolutionized the field of life sciences and offers extraor-dinary speed and cost-effectiveness to study genomic andtranscriptome data [50 51] RNA-Seq has emerged as aneffective approach for transcriptome profiling which pro-vides accurate measurement of gene expression level [52]Transcriptome sequencing offers ease of identification andevaluation of thousands of genes in a single analysis [53]Transcriptome analysis of the cytoplasmic male-sterilelines has been previously reported in many crops such asradish [54] watermelon [55] soybean [56] tomato [57]Brassica napus [58] chili pepper [59]cotton [60ndash62] sweetorange [63] leading to identification of several candidategenes in these systems In this study comparative tran-scriptome analysis of an A2 CMS system derived cytoplas-mic male-sterile and its fertility restorer line in pigeon peawas performed primarily to detect differentially expressedgenes participating in pollen development which mighthelp in understanding the CMS mechanism in this crop

ResultsPollen fertility analysis and phenotypic characterization ofsterile and fertile budsPollens from male-sterile and fertile buds were observedunder a microscope and were evaluated to determine fertil-ity Upon observation it was seen that the pollens grainsshowing red acetocarmine staining were fertile while sterilepollens remained unstained (Fig 1a) During the pheno-typic investigation of the sterile (AKCMS 11) and fertile(AKPR303) flower buds we observed that the filamentsand anthers of the sterile flowers were smaller in contrastto the fertile flowers It was clearly visible that the anthersof the fertile line were denser than the sterile line (Fig 1b)

Saxena et al BMC Plant Biology (2020) 2074 of 24

















186 315


111 87


Missing BUSCOs 0 3









Read processing



Total

Raw reads 109600365 72575988 182176353


101676579 69481226 171157805




Percentage GC 4114






L50 value 6




Unigenes


171095


N50 length (bp) 757









































Down-regulated

























Up-regulated









































































RNA-Seq qRT-PCR


























































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References


















186 315


111 87


Missing BUSCOs 0 3









Read processing



Total

Raw reads 109600365 72575988 182176353


101676579 69481226 171157805




Percentage GC 4114






L50 value 6




Unigenes


171095


N50 length (bp) 757









































Down-regulated

























Up-regulated









































































RNA-Seq qRT-PCR


























































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References








Read processing



Total

Raw reads 109600365 72575988 182176353


101676579 69481226 171157805




Percentage GC 4114






L50 value 6




Unigenes


171095


N50 length (bp) 757









































Down-regulated

























Up-regulated









































































RNA-Seq qRT-PCR


























































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References








































Down-regulated

























Up-regulated









































































RNA-Seq qRT-PCR


























































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References































































RNA-Seq qRT-PCR


























































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References














































































































































































































Abstract

Background

Results

Conclusion

Background

Results









Metabolic pathways


Discussion






Conclusions

Methods










Abbreviations

Acknowledgements


Funding




Competing interests

Author details

References


research article open access transcriptome profiling of … · 2020. 2. 13. · 481,226 (32.7 gb)...

Documents