Tissue-Specific Expression Profiling of Seedling Stage in Early-Maturity Mutant Induced by Carbon Ion Beam in Sweet Sorghum

An early-maturity mutant KFJT-1 has been screened out after carbon ion irradiation in sweet sorghum (Sorghum bicolor (L.) Moench). In this study, tissue specific digital gene expression analysis was performed between the KFJT-1 mutant and the wild type KFJT-CK at seedling stage. The results showed that a total of 717, 2160 and 2,331 tags-mapped genes were differently expressed in roots, stems and leaves of young seedling, respectively. In KFJT-1, 557 (77.7%) genes were up-regulated and 160 (22.3%) genes were down-regulated in young root; 1,232 (57.0%) genes were up-regulated and 928 (43.0%) were down-regulated in young stem; and 1,577 (67.7%) genes were up-regulated and 754 (32.3%) genes were down-regulated in young leaf. Functional annotation revealed that most induced genes functioned as “binding”, “synthase activity”, “transferase” and “transporter activity” which involved in the biological processes of metabolic and response to stimulus. Surprisingly, the up-regulated genes in KFJT-1 were classified into four KEGG pathways: “alpha-Linolenic acid metabolism”, “flavonoid biosynthesis”, “inositol phosphate metabolism” and “fatty acid biosynthesis”, which related to the stress resistance and supported the outstanding agronomic traits of KFJT-1 in the process of plant growth and development. Among the DEGs, a critical photoreceptor from photoperiod pathway PHYA gene was significantly up-regulated in leaf and root of KFJT-1, suggesting the mutation could occur on the genomic upstream of PHYA. This work may provide helpful insights to further understand the mutation mechanism in sweet sorghum.


Introduction
Sweet sorghum (Sorghum bicolor (L.) Moench) is a useful energy crop because of high photosynthetic efficiency, high biomass-and sugar-yielding (Billa et al., 1997).However, being a short-day plant, the grains cannot mature under long day condition.We previously isolated an early-maturity mutant KFJT-1 from wild type plants KFJT-CK by heavy ion beam irradiation.Resistance experiment showed that proline content was increased by 11.05% with drought stress, which showed that the tolerance of KFJT-1 to the stress is advantage to KFJT-CK (Dong & Li, 2012).
The biological effects of heavy-ion radiation encompass a wide range of alterations, including developmental abnormalities (Kranz, 1994), chromosomal aberrations (Kawat et al., 2001;Kikuchi et al., 2009;Wei et al., 2006) and genomic structural variation (Mei et al., 2011;Xu et al., 2006).Many studies has shown that the carbon ions beam induce more effective structural alterations in DNA than other radiation (Shikazono et al., 2005), sequence analyses of radiation-induced mutations have been widely carried out in plants (Bruggenmann et al., 1996;Shikazono et al., 2000;Shikazono et al., 2003).These genetic variations could directly induce expression differentiation of the plenty of genes which involved in the biological processes.Nowadays, Digital gene expression (DGE) tag profiling has been widely utilized to monitor the differences in transcriptional level to elucidate the genome-wide expression profiling among different tissues and organs.It directly quantify the transcript abundance of the uniquely tagged corresponding genes with ultra-high-throughput sequencing of cDNA fragments (Hong et al., 2011), which could be conveniently detected for the organisms without prior annotations of genomic information, such as cotton (Wei et al., 2013), spruce (Albouyeh et al., 2010), Brassica napus (Jiang et al., 2013), and moss (Nishiyama et al., 2012).
This study aimed to gain comprehensive understanding of DGEs in KFJT-1 compared to KFJT-CK at seedling stage and improve our current understanding of the molecular mechanism of KFJT-1 induced by carbon ions beam.

Plant Materials and Growth Conditions
The dry seeds of equal size from KFJT-1 and KFJT-CK, which showed no moldy and lesion, were selected and placed in a 90 mm Petri dish containing double-layer wet filter paper, respectively.The seeds were germinated at 25±2 o C in a growth chamber under a 16 h light photoperiod provided by fluorescent light tubes (50 μmol m -2 s -1 ).Each genotype was replicated three times and 100 seeds were employed for each replication.After 30 days, the samples of roots, stems and leaves were harvested, respectively and quickly frozen in liquid nitrogen for RNA isolation.

RNA Isolation and Library Preparation for DGE
RNA extraction was performed according to the manufacturer's instructions of TRIzol reagent (Invitrogen, USA), followed by RNase-free DNase treatment (TaKaRa, Dalian, China).The total RNA was checked for quality and quantity using a Biophotometer Plus (Multiskan Spectrum, German), and a minimum of 6 ug of total RNA was used for Illumina sequencing.The total RNA samples isolated from the three parallels tissues were pooled for libraries preparation, in which RC and RF, SC and SF, LC and LF represented the transcripts of roots, stems and leaves from control KFJT-CK ("C" characterized) and mutant KFJT-1 ("F" characterized), respectively.Over 6 μg from each total RNA samples were constructed as the DGE libraries using Illumina gene expression kit (IllumingaInc; San Diego, CA, USA) according to the manufacturer's protocol (version 2.1B), mRNA was purified using biotin-Oligo (dT) magnetic bead adsorption.The first-and second-strand cDNA synthesis was performed after the RNA was bound to the beads.The double stranded cDNA were digested with NlaIII to produce cohesive end.After purification with Dynabeads, and the digestion was ligated to GEX adapter 1 which contains MmeI restriction CATG site, and downstream 17 bp then cut with the NlaIII.The 21 bp tags containing adapter I were ligated to GEX adapter 2 to generate a tag library.These tag fragments were amplified by liner PCR for 15 cycles using PCR primers anneral to the adapter ends.The 85 bp amplicons were seperated on 6% TBE PAGE gel, purified and denatured to produce single strand molecules.These molecules were anchored to Solexa sequencing array and sequenced on Illuminga GA II at BGI-Shenzhen, Shenzhen, China.Raw sequence data were generated by Illuminga pipeline.

Sequence Annotation and DGEs Pathways Identification
Raw sequences were transformed into clean tags by filtering off adapter-only tags and low-quality tags as described (Li et al., 2013).All the clean tags were mapped to the reference sequences of Sorghum bicolor and only 1 bp of mismatch was considered.The remaining clean tags were designed as unambiguous clean tags.In order to compare the expression abundance among the samples, the number of unambiguous clean tags for each gene was calculated and then normalized to TPM (number of transcripts per million clean tags).The final assembled transcripts (≥ 100 bp) were submitted for homology and annotation searches using Blast2GO software v2.4.4 (Wei et al., 2013).For BLASTX against the NR database, the threshold was set to E-value lower than 10 -5 .However, most of the gene information of sorghum was hypothetical or putative.Therefore, all the putative sorghum genes were BlastX with was performed against Sorghum genes.GO classification was achieved using WEGO software together with David Bioinformatics Resources 6.7 (http://david.abcc.ncifcrf.gov/home.jsp).Enzyme codes were extracted and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were retrieved from KEGG web server (http://www.genome.jp/kegg/).We used a rigorous algorithm to identify differentially expressed genes between the KFJT-1 and KFJT-CK in this study.FDR ranking, FDR (False Discovery Rate) was used applied to adjust the p-value in multiple tests and analyses (Qin et al., 2011).The transcripts with at least two-fold differences (absolute values of log2 (Ratio) ≥ 1 with FDR < 0.001) were regarded as significantly different expressed genes.

Real-Time Quantitative RT-PCR (q RT-PCR) Analysis
Real-time quantitative RT-PCR (qRT-PCR) analysis was used to verify the DGE results.The RNA samples used for the qRT-PCR assays were the same as in the DGE experiments.Gene specific primers were designed according to the reference unigene sequences using Primer Premier 5.0.Seven genes were selected from the DEGs for quantitative qRT-PCR assays.QRT-PCR was performed according to the manufacturer's specifications ().The following SYBR Green PCR cycling conditions were used: denaturation at 95 C for 10 s, followed by 40 cycles of 94 C for 5 s and 60 C for 20 s.The PCR experiments were performed using an iQ 5 Multicolor real-time PCR detection system (BioRAD, USA).Sorghum actin gene (forward: GCCGAGCGAGAAATTGTAAG and reverse: ATCATGGATGGCTGGAAGAG) was used as a normalizer.The relative gene expression levels were calculated using 2 -△△ CT .

Construction of Digital Gene Expression (DGE) Library for KFJT-1 and KFJT-CK at Seedling Stage
To obtain a global view of the tissue specific characteristics at the transcriptional level between KFJT-1 and KFJT-CK at seedling stage, total six DGE libraries from roots, stems and leaves were sequenced with Solexa/Illumina DGE analysis, respectively.Among the libraries, we got the total numbers of tags ranging from 5.8 to 6.7 million, which composing distinct tags with 125840 and 131973,140208 and 156400, 147227 and 137977 in young roots, stems and leaves libraries for KFJT-CK and KFJT-1, respectively (Table 1).The number of the tags and unambiguous tags mapping to genes was almost the same for about 70%, for example in RC, 62208 distinct tags (72.07% of clean tag), 62089 unambiguous tags (71.99% of clean tag) was mapping to gene, which means that the tags was well matched to the specific genes.Those distinct tags matched to the genes occupied about 50% of the clean tags.The distribution of total clean tags and distinct clean tags over different tag-abundance categories were shown in Figures 1A and 1B.In terms of the total clean tags, the percentage of 2-5 copies ranged from 3.31-3.89%,6-10 copies from 2.3-2.71%,11-20 copies from 3.20-3.81%,21-50 copies from 6.44-7.47%,51-100 copies from 7.7-8.83%.The largest percentage was constituted by the copies over 100 which ranged from 73.58-76.90%.However, compared to the total clean reads, the distinct clean tags displayed different distribution among all the six DGE libraries.The largest proportion was constituted by the 2-5 copies for about 54%.The second largest part was the 6-10 copies which occupied about 14%, approximately 10% had copy numbers higher than 100.The smallest proportion was constituted by 51-100 copies for about 7%.The numbers of the tag-mapped genes or unambiguous tag-mapped gene were decreased sharply compared to the distinct tags.For example, the number of matched gene in RC was 15768, however, the distinct tags was 62208.Finally, 15786 (RC), 16007 (RF), 16179 (SC), 16780 (SF), 15378 (LC) and 15200 (LF) tag-mapped genes were generated against sorghum reference genome between KFJT-1 and KFJT-CK.The saturation analysis showed that the number of the genes was not increased proportionally with the number of sequences (total tag number) when the sequencing counts reached 4M (Figure S1).Thus, these tag-mapped genes were completely satisfying the further analysis.

Tissue-Specific Gene Expression in the Development of the Seedling between KFJT-1 and KFJT-CK
To compare differential expression patterns between KFJT-1 and KFJT-CK, we normalized tag distribution for gene expression level in each library to make an effective library size and extracted significance of differentially expressed transcripts (DETs) with FDR ≤ 0.05 and log2 fold-change ≥ 1 by edgeR (Empirical analysis of Digital Gene Expression in R).The regulated genes were shown in Figure 2. The red dots and green dots represent transcripts higher or lower in abundance for more than two fold, and the blue dots represented the transcripts that differed less than two fold between the KFJT-1 and the wild type.In root, a total of 557 genes were up-regulated (77.7%, red dot in Figure 2A) and 160 genes were down-regulated (22.3%, green dot in Figure 2A).In stem, a total of 1,232 genes were up-regulated (57.0%, red dot in Figure 2B) and 928 genes were down-regulated (43.0%, green dot in Figure 2B).In leaf, total 1,577 genes were up-regulated (67.7%, red dot in Figure 2C) and 754 genes were down-regulated (32.3%, green dot in Figure 2C).An increasing trend in the number of differently expressed genes was observed in young stem and leaf compared to young root.The total numbers of the tags-mapped genes were 717, 2160 and 2331 in root, stem and leaf, respectively (Figure 3A).Venn analysis revealed that 66 genes were differently expressed in the all young root, stem and leaf.In addition, 219 genes were differently expressed both in leaf and root, 489 genes in leaf and stem, and 148 genes in root and stem.Numbers of 284, 1457 and 1557 genes were differentially expressed specific to the root, stem and leaf, respectively (Figure 3B).To be mentioned, about 99% unique tags were expressed within five-fold difference (red bar in Figure 4) between KFJT-1 and KFJT-CK, covered 99.85% in young root, 99.39% in young stem and 98.77% in young leaf, respectively.Only 0.1-1.01% of the DEGs over five folds was up-regulated (green bar in Figure 4), while 0.05-0.23%was down-regulated.ntially Express e used to classi ording to biol r than 0.05.Ac genes cluster in ulated genes wa riteria to assig genes were re ponse to stimul e term "fatty a that KFJT-1 n biosynthetic p ng time were port", which c "polysaccharid nded which su metabolic proc nitrogen fixatio sugar mediated ith the high su s "binding", " ess above.dge which path 05, Table 2C).lpha-Linolenic nthesis", which h and developm   (Mas et al., 2003), mostly it was negatively regulated by CCA1/LHY, and positively regulated by the WNK1 gene (Wang & Tobin, 1998).Interestingly, the genes all above were assigned to late flower, which was contradicted against the early flower phenotypic of KFJT-1.This might be contributed to the complex regulatory module for photoperiodic flowering signaling.We proposed that the earlier flower characteristic of KFTJ-1 was related to the up-regulation of PHYA in young root and leaf.Previous study indicated that PHYA are the major day-length sensors in Arabidopsis (Mockler et al., 2003) and thought to promote flowering.The PHYA mutant flowers significantly later than the wild type in response to day-length extensions with a far-red-enriched white light (Johnson et al., 1994), differed between long-days and short-days (Weller et al., 1997).Interestingly, the PHYA is tissue-specific expressed in KFJT-1.Previous study on Arabidopsis suggested that sucrose could depress expression of PHYA (Dijkwel et al., 1997).Thus, we conferred that the down-regulated PHYA in young stem in KFJT-1 was due to high content of sucrose which mutual adjusted by the gene 3.2.1.26and 2.4.1.13.

Conclusion
To our best knowledge, this study gained comprehensive understanding of DGEs between KFJT-1 and KFJT-CK, which was the first genome-wide effort to investigate the transcription dynamics of sweet sorghum induced by carbon ion beam at seedling stage.Furthermore, this work provides some useful information to develop function genes for the industry process of energy crop by carbon ion beam.
Figure Distributio mutant, re o

Figure
Figure 2. C significant mutant and > 1" were

Figure
Figure 4. T the d repre Figure 5. bio Figure 6 stage betw red bor Figure 7 between borders w

Table 1 .
Statistics of DGE sequencing

Table 2 .
List of first twenty pathways for DEGs

Table 2A .
List of first twenty pathways in root

Table 2B .
List of first twenty pathways in stem