Mechanism Of Seedlessness In ’Qianyang Seedless’ Ponkan Mandarin And Identification Of Male Sterility Related Genes

Seedlessnesss is one of the most important aims of citrus breeding. Recent years, lots of new seedless citrus germplasms were obtained with somatic hybridization, conventional sexual cross breeding and embryo rescue methods. However, until now most of the germplasms have not acquired new cultivar certifications as sexual cross breeding is time-consuming and citrus possess peculiarities of reproductive biology such as long juvenile period and nucellar polyembryony. In fact, bud sport mutant is still main approach for obtaining seedless citrus varieties. Nevertheless, limited information is available about the mechanism of bud sport mutant and few researches were carried out about the molecular mechanism of citrus seedlessness, which constrained the utilization of the seedless citrus germplasms. Therefore, exploration the mechanism underlying seedless citrus with modern molecular approaches was important for further understanding the seedless bud sport mutants and accelerating the breeding process of seedless citrus. In this study, we identified a seedless ponkan mandarin mutant’Qianyang seedless’ponkan mandarin (QS), and confirmed that the seedless trait was due to male sterility and abnormal development of pistil by cytology and genetics analysis. To analysis the molecular basis of male sterility in QS, we employed SSH-cDNA libraries and microarry techniques to identify differentially expressed genes between QS and a common seedy ponkan mandarin Egan No.1 (EG). Furthermore, RNA-seq approach was applied for analysis genome-wide transcriptional changes during stamen development between QS and EG.And the main results were as follow:1. Exploration about the causes of seedlessness for QS and genetic divergence between QS and EG.Pollen grains viability was evaluated by pollen dying and pollen grain germination assays in vitro, and the results validated that QS was male sterile. We also evaluated the fertility of pistil by tissue section and hand-pollination, and found that the embryo sac of QS was also abnormal. However, both QS and EG showed normal morphology of floral organs, and could produce plenty of pollen grains, only that the number of pollen grains per anther of QS was 31% less than EG. In addition, the pollen dying viability of QS was 6.0% in striking contrast to the high viability of more than 90% for EG. And no pollen grain of QS germinated while about 10% pollen grains of EG could germinate normally. Furthermore, we found that a large amount of abnormal tetrads produced at tetrad stage in QS, and SEM (Scanning electron microscope) assays showed abnormal structures and inconsistent size of the pollen grains of QS. And hand-pollination for QS with viable pollen grains still produced seedless fruits though the pollen gains could germinate on the stigma normally.Various types of DNA markers including SCAR (2 published pairs of primers), SSR (172 pairs of primers), MSAP (96 pairs of primers) and AFLP (13 pairs of primers) were employed to analyze the polymorphism between QS and EG, but no repeatable polymorphic bands were detected. These results suggested that very few nuclear genes were altered during the developmental stages in QS versus EG, and very similar genetic background between QS and EG.In order to identify and clone differentially expressed genes related to seedlessness of QS, SSH-cDNA libraries were constructed with flower samples. The samples were collected from both QS and EG in parallel including 4 continuous phonologically developmental stages including squaring stage (SF, about 20 DBF), medium bud stage (MF, about 10 DBF), flowers at full bloom stage (BF) and young ovaries of 2-3 days after flowering (OV). All the flowers were bagged to prevent cross-pollination when sampled in the field. Total RNA were extracted respectively, and RNA pools from both QS and EG were established in parallel by mixing equal amount of RNA of the four stages. The RNA pools of QS and EG were used to construct SSH-cDNA libraries, both forward (the seedless cultivar QS as tester and the seedy cultivar EG as driver) and reverse (EG as tester while QS as driver) SSH were conducted. In all 10944 clones were obtained and kept in twenty-nine 384-well plates (8832 clones from forward library,2112 clones from reverse library). To screen positive clones form the libraries, about 8000 clones were randomly picked and used as for subsequent PCR templates. A total of 6048 high-quality PCR products (4,195 from the forward library and 1,853 from the reverse one) were obtained and purified for custom-microarray fabrication. In all,255 non-redundant differentially expressed clones (FDR<0.05 and a fold change≥2) were identified from the four developmental stages by microarray screening, and 75% of the clones were down-regulated. After one-single pass EST sequencing and ESTs functional annotation, 133 unigens were obtained and a number of important candidate genes were identified including Male sterile-like protein which involved in the formation of pollen exine and a serial of transcription factors (TFs) related to reproductive development such as AP2-ERF/EREBP, MYB and bHLH-motif protein. In addition, genes associated with lipids translocation and energy metabolism such as LTPs, proton-exporting ATPase were also identified. GO-KEGG annotation analysis showed that the majority of unigens were involved in metabolic process, biological regulation, nucleobase/nucleoside/nucleotide metabolic process and response to stimulus; and genes related to basic metabolic process such as nutrition and energy metabolism were down-regulated in QS. Meanwhile, it suggested that TFs and phytohormones might play important regulatory roles during stamen development To elucidate the regulatory mechanism concerning stamen development of ponkan mandarin at molecular level, we employed massively parallel RNA sequencing (RNA-Seq) approach to analyze differentially expressed genes (DEGs) of flowerlets (diameter 3 mm) (SF) and mature anthers (AN). Over 46 million clear reads were generated from 4 libraries with illumina single molecule array genotyping method. And 680 DEGs (391 up-regulated,508 down-regulated) were identified in SF of QS comparing with EG; and 4734 DEGs (289 up-regulated,4226 down-regulated) were screened in AN. It clearly showed that much more DEGs were down-regulated which might meant that some crucial genes were suppressed and resulted in male sterility in QS.We identified top 20 most significantly DEGs between QS and EG in SF and AN, respectively. In the 10 up-regulated genes of SF,3 genes encoded laccase which involved in lignin biosynthesis and metabolism that could effect the growth and development of plants. And there was one gene encoded pectin methylesterase-like protein and one encoded microtubule-associated proteins which were related to cell wall biogenesis and organization. In fact, the most significantly up-regulated gene was Nucleosome assembly protein that involved in DNA metabolic process. And among the 10 down-regulated genes of SF, seven genes encoded heat shocking proteins (HSPs). While for the top 20 most significantly DEGs of AN, the up-regulated DEGs such as cytochrome P450, MYB 102, senescence-related gene 1 (SRG1) and PAR1 protein were identified; and the down-regulated DEGs included NAC TF, PHS1 (poor homologous synapsisl), LTPs and SA UR-like auxin-responsive protein family and so on.3. To estimate the potential roles of phytohormones on QS stamen development, the hormones content of different tissues including matured leaves,3 mm flowerlets, flower buds at full bloom stage, and anthers and ovaries at full bloom stage were measured with HPLC-mass methods. The results showed the contents of ABA, ACC and SA were too low to analyze, only IAA and JA-Ile were determined. It was found that the endogenous IAA content maintained at relatively low level (2.1-2.7 ng/g) in all tested tissues, though the IAA content was little lower in QS comparing with EG Unlikely, JA-Ile accumulated significantly to relatively high level (29.8-39.5 ng/g) in anthers, while extreme low in other tissues of both QS and EG, and the JA-Ile content of QS was lower than EG in all the tissues except flower buds at full bloom stage. Furthermore, genes involved in the biosynthesis of JAs-Ile were isolated and cloned, and relative expression pattern were analyzed with qRT-PCR, and the results indicated that the alterations of genes expression were coincident well with the JAs-Ile contents changes.4. The full-length cDNA of MSLP and AP2-ERF TF were cloned by RT-PCR, and relative expression profiles were performed with qRT-PCR. The results showed that MSLP was expressed mainly in the ovaries at full bloom stage, and very low expression in the anthers. Throughout the whole development process, the expression of MSLP was significant higher in QS comparing with EG. Likewise, the expression of AP2-ERF TF was also higher in QS, and it expressed mainly in SF and BF. Furthermore, we analyzed the cDNA sequences of both MSLP and AP2-ERF TF in 6 typical citrus varieties. And it found that there were at least 4 transcripts of MSLP in citrus, and only one transcript of AP2-ERF TF. Copy number analysis with southern blotting showed that there were 5-8 copies of MSLP in citrus genome and 1-2 copies for AP2-ERF TF.In conclusion, we had confirmed that the seedlessness of QS was due to male sterility and abnormal development of pistil. By SSH and RNA-seq approaches, many genes related to stamen/pistil development were identified which could applied for father molecular breeding and understanding the insight of molecualr mechanism of seedless citrus.