Measurement alliin content
To determine the differences in alliin content among the garlic cloves, inner buds, roots, and sprouts, a post-column derivatization method using ninhydrin was established. The findings revealed that the bulbs contain the highest level of alliin (1.474 mg.g-1, DW), while the roots contain the lowest (0.019 mg.g-1, DW).
Next, to determine the alliin content of the garlic leaves after wounding, damaged leaves were sampled at different time points after injury then analysed using the same method. Three biological replicates were sampled at each time point (T1/3, T4/6, T7/9 and T10/12, at 0, 3, 6 and 12 hours after wounding, 3 biological replicates, respectively). As shown in Fig. 1c, the content of alliin decreased over time in the wounded leaves.
Transcriptome analysis of the garlic leaves after wounding
Approximately, 0.03% of the raw reads, including low quality and short reads, were removed post filtering of the adapter sequences, resulting in 34,855,947 clean reads for T1/3, 30,830,543 for T4/6, 29,001,419 for T7/9 and 34,278,922.33 for T10/12 (Additional file 1). Pair-wise Pearson’s correlation coefficients of the three replicates × four garlic samples indicated high repeatability of the sequencing data (Fig. 2a). At the same time, comparisons of individual genes between the control group and each treatment group revealed significant differences in reliability among genes.
All reads were assembled using Trinity [45], resulting in 194,627 transcripts (N50:1,667) with a mean length of 1157.22 bp (Fig. 2b), and 94,144 unigenes (N50:1,394) with a mean length of 933.08 bp. The size distributions of the unigenes and transcripts are shown in Fig. 2b. Overall, 15.38% of the reads were > 2000 bp in length, while 28.78% were < 1000 bp in length. A minimum read length of 200-300 bp occurred in 11.56% of the transcripts, and only 10.36% of the unigenes were > 2000 bp in length, the majority being between 200 and 500bp. A total of 16,882 unigenes were annotated using the Swiss-Prot database, 26,662 unigenes were identified using the Nr database, 19,206 displayed significant similarities with known proteins in the Pfam database, and 9677 and 16,197 were annotated using the KEGG and KOG databases, respectively (Fig. 2c). The distribution of unigene species annotation is shown in Fig. 2d.
Identification of differentially expressed unigenes (DEGs) in the garlic leaf samples
A number of the unigenes were classified into KEGG metabolic and signalling pathways. Three KEGG pathways, ‘Protein processing in endoplasmic reticulum’ (75 unigenes), ‘Plant hormone signal transduction’ (92 unigenes), and ‘Photosynthesis’ (56 unigenes), were the most enriched (Fig. 3a; Additional file 2). Venn diagrams were also used to represent the number of differentially expressed genes under each treatment compared with the control (Fig. 3c). By this standard of (P < 0.05) in the pathways. Accordingly, 1714 unigenes were found to be up-regulated and 1135 were down-regulated in T1/3 - T4/6, while in T1/3 - T10/12 688 highly-expressed unigenes were up-regulated and 1375 were down-regulated (Fig. 3b; Additional file 3).
Of the above pathways, four were significantly enriched in T1/3 compared to T4/6, while in T7/9, these four pathways were also significantly enriched compared with T1/3. Meanwhile, only three of the four pathways were significantly enriched in T1/3 compared to T10/12 (Fig. 3d).
To predict possible functions and orthology classifications, the unigenes were also compared using the COG database. As a result, 12,375 sequences in T1/3 compared with T4/6 were assigned to 25 COG categories. There is usually only one category of functional prediction. Here, the general functional prediction (2191; 17.71%) represented the largest group followed by replication, recombination and repair (1350; 10.91%); transcription (1099; 8.88%), and signal transduction mechanisms (909; 7.35%). Additional samples used for COG classification are detailed in Additional file 4.
DEGs associated with the alliin biosynthesis pathway
Detailed steps of the alliin biosynthesis pathway are currently unknown[46]. However, some of the corresponding molecules have been determined, including its precursors: glutathione, glycine, serine[47], cysteine[48] and sulfur[49], which feed a series of hydrocarbylation, alkylation and oxidation reactions (Fig.3e). In this study, transcriptome analysis revealed five alliin biosynthesis-related GO terms: ‘sulfur compound biosynthetic process’ (GO:0044272), ‘sulfur amino acid metabolic process’ (GO:0000096), ‘cysteine biosynthetic process’ (GO:0019344), ‘L-serine biosynthetic process’ (GO:0006564) and ‘glutathione peroxidase activity’(GO:0004602), providing a basis for further analysis of the differential expression of alliin biosynthesis-related genes during leaf wounding. After wounding, genes in the GO:0044272 pathway, such as c119107.graph_c0 and c101182.graph_c0, significantly changed gradually over time. Moreover, genes encoding sulfur compounds (c49698.graph_c0, c86971.graph_c0, and c114534.graph_c0) were expressed at high levels. The most highly expressed were involved in cysteine biosynthesis: c107612.graph_c1 and c95022.graph_c0.
Transcriptome analysis of the differences in cysteine pathway-related genes
Based on transcriptome analysis, eight cysteine-related GO terms, ‘D-cysteine catabolic process’ (GO:0019447), ‘cysteine biosynthetic process’ (GO:0019344), ‘cysteine biosynthetic process from serine’ (GO:0006535), ‘cysteine desulfurase activity’ (GO:0031071), ‘D-cysteine desulfhydrase activity’ (GO:0019148), ‘glutamate-cysteine ligase activity’ (GO:0004357), ‘peptidyl-cysteine modification’ GO:0018198, and ‘cysteine metabolic process’ (GO:0006534) were identified. Significant differences in GO:0019344 were observed in the T1/3 vs. T4/6, T1/3 vs. T7/9 and T1/3 vs. T10/12 comparisons at 0.00168, 0.0013 and 0.00012, respectively. Meanwhile, significant differences in with GO:0006535 were also observed between the T1/3 vs. T4/6, T1/3 vs. T7/9 and T1/3 vs. T10/12 comparisons at 0.46558, 0.45829 and 0.23682, respectively. Differences between T1/3, T4/6, T1/3, T7/9, T1/3 and T10/12 in terms of cysteine related genes are shown in Fig.4a and Additional file 5.
The alliin synthesis pathway is thought to involve the cysteine pathway, hence the above difference in related differential genes. With time after wounding, the enrichment of different genes in the cysteine synthesis pathway became increasingly significant, further confirming the involvement of cysteine in alliin synthesis. We also created a heat map based on the differential genes related to cysteine biosynthesis, further highlighting the differences between samples (Fig.4b, Additional file 6).
Transcriptome analysis of the differences in serine pathway-related genes
Serine is also thought to play an important role in the alliin synthesis pathway[32]. Based on transcriptome analysis, seven serine-related GO terms were identified: ‘protein serine/threonine/tyrosine kinase activity’ (GO:0004712), ‘protein tyrosine/serine/threonine phosphatase activity’ (GO:0008138), ‘L-serine biosynthetic process’ (GO:0006564), ‘L-serine metabolic process’ (GO:0006563), ‘serine-type endopeptidase inhibitor activity’ (GO:0004867), ‘serine family amino acid biosynthetic process’ (GO:0009070), and ‘D-serine metabolic process’ (GO:0070178). Significant differences in GO:0008138 were observed in all comparisons (Fig. 4c), while significant differences in GO:0004712 were observed between T1/3 vs. T4/6 and T1/3 vs. T7/9. Significant differences in GO:0008138 were also observed between T1/3 vs. T4/6, T1/3 vs. T7/9, and T1/3 vs. T10/12 at 0.01758 and 0.01673 and 0.01689, respectively. In contrast, significant differences in GO:0004712 were observed only between T1/3 vs. T4/6 and T1/3 vs. T7/9 at 0.01939 and 0.01911, respectively. Differences between T1/3, T4/6, T1/3, T7/9, T1/3, and T10/12 in terms of serine-related genes are shown in Additional file 7. As above, heat maps of the serine metabolism-related genes were also created (Fig. 4d; Additional file 8).
Transcriptome analysis of the differences in sulfur-related diffeences genes
In garlic, sulfur compounds are the most important organic compound. As described earlier, they include alliin, and many other sulfur compounds involved in alliin biosynthesis. Based on transcriptome analysis, six sulfur-related GO terms were identified: ‘sulfate transmembrane transporter activity’ (GO:0015116), ‘sulfate transport’ (GO:0008272), ‘sulfur compound transport’ (GO:0072348), ‘ligase activity, forming carbon-sulfur bonds’ (GO:0016877), ‘sulfur compound transmembrane transporter activity’ (GO:1901682), and ‘sulfur compound metabolic process’ (GO:0006790). Significant differences in GO:0015116 were observed between T1/3 vs. T4/6, T1/3 vs. T7/9, and T1/3 vs. T10/12 at 0.13427 and 0.13008 and 0.03515, respectively (Fig. 4e). In contrast, significant differences in GO:0016877 were only observed between T1/3 vs. T4/6 and T1/3 vs. T7/9 at 0.02281 and 0.02219, respectively. Differences between T1/3, T4/6, T1/3, T7/9, T1/3 and T10/12 in terms of sulfur-related genes are shown in Additional file 9. Heat maps of the sulfur-related genes were also constructed as shown in Fig. 4f and Additional file 10.
Identification of transcription factor (TF) families in alliin
A number of TFs are thought to play important roles in alliin biosynthesis. In this study, 452 putative TF-encoding genes belonging to 47 major TF families were analysed. Of these, 40 were included in the BHLH family, 36 in the FAR1 family, and 36 in the NAC family. In order to screen key regulators of alliin biosynthesis, eight of these TF were selected for further analysis (Additional file 11).