Colorimetric changes during the leaf color transformation
By constant observation, we found part of C. camphora leaves would change their color from green to red, and this phenotype can be observed all the year around (Fig. 1A). After counted 492 WRL on 335 branches and record their locations, we found that all WRL were distributed in the last 5 leaves at the end of the branch with 44.72%, 22.97%, 22.97%, 6.91% and 2.44%, respectively (Fig. 1B). Colorimetric measurements were performed on both the adaxial (Ad) surface (Fig. 1C) and abaxial (Ab) surface (Fig. 1D) of the leaves. On the Ad surface, the L* values were no significant difference among GL, RL and WRL. The L* values on the Ab surface of WRL were significant lower than GL and RL. For the a* values, both Ad and Ab surfaces shown the significant difference among GL, RL and WRL. Whether Ad or Ab surface, GL were negative means green color, while RL and WRL were positive means red color. a* values of WRL were higher than RL on both surfaces. These results suggested that the WRL were redder than RL. b* values of WRL exhibited lowest level compared with GL and RL on both surface.
Major pigments in the red leaves of C. camphora
To determine the mainly pigments responsible for the leaf color transformation, the chlorophylls, carotenoids and anthocyanins content were measured. The results showed that color profiles were differently between GL, RL and WRL. The contents of chlorophyll a (Fig. 2A), chlorophyll b (Fig. 2B) and total chlorophyll (Fig. 2C) were significant decrease from the GL to WRL. When the leaf color completed turn to red, chlorophyll a, chlorophyll b and total chlorophyll only remain 1.79%, 9.47% and 3.98% of the GL, respectively. The contents of carotenoid were similar between RL and WRL, and significant decrease compared to GL (Fig. 2D). Compared to the GL and RL, WRL own the highest anthocyanins content, reached up to 532.13 ug·g− 1 (Fig. 2E). By the HPLC MS/MS, 27 anthocyanins were detected in GL and WRL. Among them, 11 anthocyanins were identified as differentially accumulated metabolites, including 3 cyanidins (Cyanidin-3,5-O-diglucoside, Cyanidin 3-O-sophoroside and Cyanidin 3-O-xyloside), 1 peonidin (Peonidin 3-O-galactoside), 1 pelargonidin (Pelargonidin 3-O-beta-D-glucopyranoside), 3 delphinidins [Delphinidin 3-O-sophoroside, Delphinidin 3-(6-rhamnosylgalactoside) and Delphinidin 3-rutinoside-5-glucoside] and 3 other anthocyanidins (Aurantinidin, 6-Hydroxycyanidin and 5-Carboxypyranocyanidin 3-O-beta-glucopyranoside) (Fig. 2F). The WRL accumulated more anthocyanins than the GL, which are sources of red color. Among them, Cyanidin-3,5-O-diglucoside, Peonidin 3-O-galactoside, and Pelargonidin 3-O-beta-D-glucopyranoside showed the top three significance difference.
RNA library construction and sequencing
In order to characterize the difference of gene expression levels between GL and RL, 6 cDNA libraries were constructed, including 3 GL (named GL1, GL2 and GL3) and 3 RL (named RL1, RL2 and RL3), respectively. By the way, we found that the RNA of WRL was degraded lead to unable construct cDNA library. A total of 299,566,040 reads (44,934,906,000 bp) were generated in raw data. After sweep away the low-quality reads, a total of 297,256,232 reads (44,163,916,131 bp) were filtered out. On average have 92.14% of the reads were mapped to the camphor tree reference genome (Supplementary Table S1). A total of 22,948 genes were sequenced compared to reference genes, and 544 new genes were identified (Supplementary Table S2). Among these new genes, 231 unnamed, 212 belonged to Cinnamomum micranthum f. kanehirae, and other 101 distributed to 63 plants.
Contrastive analysis of transcriptome data between GL and RL
The leaf color transformation is a complex process. High transcriptome variation was observed between GL and RL, as evidenced by a large number of differentially expressed genes (DEGs). In total, 12,217 DEGs were filter out (Supplementary Table S3), including 4,065 up-regulated genes and 8,152 down-regulated genes (Fig. 3A). We selected 30 DEGs with the greatest gene expression difference, 24 DEGs were down-regulated, 5 DEGs were new genes (Fig. 3B). Ccam02G002785 was the maximum positive value with log2 (fold change) = 14.928, while Ccam01G002624 was the maximum negative value with log2 (fold change) = -16.493. To confirm the accuracy and reliability of the RNA-seq data, 6 DEGs were randomly chosen and validation by qRT-PCR. The candidate genes expression showed large consistence between RNA-seq and qRT-PCR (Fig. 3C).
GO and KEGG analysis of the DEGs
To further elucidate the functional roles of DEGs, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis through AgriGO and GeneCodis. All DEGs were assigned to three main categories of GO classification, including biological process (BP), cell components (CC) and molecular functions (MF) (Fig. 4A). For BP, the dominant subcategories were ‘cellular process’ and ‘metabolic process’. For CC, ‘cellular anatomical entity’ and ‘protein-containing complex’ were highly represented. In MF, ‘binding’ and ‘catalytic activity’ were the most represented.
All DEGs were further performed by KEGG pathway enrichment analysis. DEGs were mainly classified into 137 pathways. Only 23 pathways with the P-value ≤ 0.05, including ‘ribosome’, ‘pentose phosphate pathway’, ‘porphyrin metabolism’, ‘photosynthesis’, ‘carbon metabolism’, ‘carbon fixation in photosynthetic organisms’ and ‘photosynthesis-antenna proteins’. ‘Ribosome’ (Ko03010) contained the largest number of DEGs (247 DEGs, accounts for 8.99% of the total DEGs), and most of the DEGs were down-regulated in RL. Ribosome was the cellular factory responsible for synthesize proteins, most of DEGs were down-regulation could affect other biological process. List the top 30 most enriched pathways (Fig. 4B). ‘Porphyrin metabolism’ (Ko00860), ‘carotenoid biosynthesis’ (Ko00906) and ‘anthocyanin biosynthesis’ (Ko00942) were related to plant pigment chlorophylls, anthocyanins and carotenoids, respectively.
Ribosome related DEGs between GL and RL
Ribosome is composed of two subunits, which are built from RNA and protein. In cells, ribosome perform translation function and play vital role in cell division, growth and development25. In this study, we found that the ‘ribosome’ was the richest KEGG term. Total of 247 DEGs belonged to this pathway, but only 3 DEGs (Ccam04G001485, Ccam02G003266, and MSTRG.15482) were up-regulated. 24 DEGs with |log2 (Fold change)| ≥ 3 were belong to small subunit and only PR-S23e (MSTRG.15482) is up-regulated (Fig. 5A). 61 DEGs with |log2 (Fold change)| ≥ 3 were belong to large subunit and only PR-L31 (Ccam02G003266) is up-regulated (Fig. 5B). Most of genes were down-regulated would affect the ribosome generation and protein synthesis. Thus, we measured the change of protein concentration during leaf color transformation. The results shown that protein contents were downtrend, the GL own the highest protein contention than RL and WRL (Fig. 5C).
DEGs related to chlorophyll biosynthesis and degradation
Chlorophylls biosynthesis and degradation is belonged to ‘porphyrin metabolism’ (Ko00860). Based on the KEGG pathway analysis, a total of 26 DEGs were identified, including 2 up-regulated DEGs and 24 down-regulated DEGs (Fig. 6). Among them, 22 DEGs related to chlorophylls biosynthesis, including OVA3 (Ccam05G002255), HEMA1 (Ccam01G002476 and Ccam04G001696), GSA (Ccam01G002922 and Ccam06G000281), HEMB (Ccam11G001295), HEMC (Ccam07G001413), HEME1 (Ccam10G000902), CPX1 (Ccam10G000600), PPOX1 (Ccam10G001810), CHLH (Ccam10G001073), CHLD (Ccam02G001092), CHLI (Ccam04G000078), CHLM (Ccam02G002041), CRD1 (Ccam01G003276), DVR (Ccam02G001588), PORA (Ccam12G000060), CHLG (Ccam03G003207), CAO (Ccam01G000545), NYC1 (Ccam02G002556), NOL (Ccam03G002262 ) and HCAR (Ccam12G001422). Almost all of the genes were down-regulated, except for NYC1 was up-regulated in RL. 4 DEGs related to chlorophyll degradation, including CLH2 (Ccam02G002637), RCCR (Ccam08G000276), SGR (Ccam07G000443), SGRL (Ccam09G000375). These results indicated that chlorophylls biosynthesis was severely decreased in RL. The genes expression result was consistent with the result of chlorophylls concentration.
DEGs related to anthocyanin biosynthesis
Anthocyanin is associated with three pathways, including ‘phenylpropanoid biosynthesis’ (Ko00940), ‘flavonoid biosynthesis’ (Ko00941) and ‘anthocyanin biosynthesis’ (Ko00942). In this study, 31 DEGs were identified belong to four PAL, one C4H, seven 4CL, four CCoAOMT, five CHS, one CHI, one F3H, one FLS, one DFR, one ANS and four UFGT (Fig. 7). 18 DEGs were significantly up-regulated, including three PAL, three CCoAOMT, one C4H, four 4CL, three CHS, one FLS, one ANS and two UFGT. Among these up-regulated genes, 4CL (Ccam06G000458, Ccam06G002027, Ccam08G001811 and Ccam03G001045), ANS (Ccam10G000921) and UFGT (Ccam02G000147 and Ccam02G000149) may result in increased anthocyanin synthesis in RL.
TFs analysis
TFs analysis was performed in this study, total of 1691 TFs have been identified into 58 families. 657 DEGs belongs to 51 different TF families were identified (Fig. 8, Supplementary Table S4). These DEGs are mainly in the bHLH, MYB, ERF, NAC, WRKY, C2H2 families (Fig. 8). TFs could affect the pigment biosynthesis by regulate the expression of structural genes. MYB and bHLH are the critical TFs in flavonoid and anthocyanin biosynthesis. We found 64 bHLH, 54 MYB, and 21 MYB-like genes were difference expression. Among the most 10 significantly up-regulated genes were three MYB, three LBD, one bHLH, one WRKY, one ERF, and one G2-like, while the most 10 significantly down-regulated genes were three ERF, one MYB, one bHLH, one bZIP, one LBD, one GATA, one C2H2, and one TCP.