Integrated analyses of rice dark response and leaf color regulation reveal links with porphyrin and chlorophyll metabolism

Tianxingzi Wang Hebei Agricultural University Yue Chen Hebei Agricultural University Zheng Zhu Hebei Agricultural University Yuqing Liu Hebei Agricultural University Gaowei Yan Hebei Agricultural University Shan Xu Hebei Agricultural University Haiqing Xu Hebei Agricultural University Ming Yang Hebei Agricultural University Shijuan Dou Hebei Agricultural University Liyun Li Hebei Agricultural University Guozhen Liu (  gzhliu@hebau.edu.cn ) Hebei Agricultural University https://orcid.org/0000-0002-7072-3357


Background
Light is the essential environmental factor for normal plant growth and development, both as an energy source processed via photosynthesis and as a signal for many biological processes. When cultivated in regular day-night cycles (diurnal cycle), rice grows with green leaves. In contrast, rice seedlings grown under dark treatment elongate more quickly and have yellow or pale colored leaves. Dark-stress is a severe stress, producing serious physiological damage that affects plant growth and development.
Through the course of evolution, plants have established e cient mechanisms for responding to light signals (Demarsy et al., 2017). Photoreceptors play essential roles, sensing light signals and initiating signal transduction. Three types of photoreceptors have been identi ed in Arabidopsis: ve subtypes of red / far red light receptor phytochromes (PhyA-PhyE), blue light receptor cryptochromes (Cry1 and Cry2) and phototropins (Phot1 and Phot2) and UV-B speci c photoreceptors (UVR8) (Demarsy et al., 2017;Yang et al., 2018). In light signal transduction pathways, dozens of elements that are downstream of photoreceptors, such as chromatin remodeling factors, histone modifying enzymes, transcription factors (such as PIFS, HY5 and FHY3), splice factors, protein degradation factors (such as COP1, SPAs and LR8), factors interacting with exogenous environmental signals and endogenous hormones have been identi ed. These elements are coordinated to transmit light signals accurately and quantitatively, thereby regulating seed germination, photomorphogenesis, chloroplast development, shade response, stomatal opening and closing, owering, biological rhythm and aging to ensure normal plant growth and development (Demarsy et al., 2017). On the basis of sequence similarity, homolog photoreceptor genes for Phy (Kay et al., 1989), Cry (Zhang et al., 2006b), Phot (Kanegae et al., 2000) and UVR8 (Fernández et al., 2016) have been identi ed in rice. However, limited information has been reported about regulatory elements downstream of these photoreceptors.
Whole-genome transcriptome analysis, involving deep sequencing or DNA microarrays, is a method used to reveal differential gene transcription at varying locations and times or under various stresses. It is useful for identifying clues that enable subsequent investigations of molecular mechanisms. Using oligonucleotide arrays, transcriptional pro ling was performed under four light treatments (blue, green, red and white), as well as dark treatment, in rice. The results showed that the expression of transcription factors, such as bHLH, MYB, C2H2, ERF, NAC and WRKY, changed signi cantly, and carbohydrate degradation decreased under dark treatment (Lakshmanan et al., 2015). It was also revealed that both Phy A and Phy C cooperatively regulate transient gene expression in red-light treated rice seedlings (Kiyota et al., 2012). It has been observed that middle mesocotyl elongation was almost completely inhibited when germinating seeds were exposed to low-intensity light. RNA-Seq analysis revealed that most of the differential expressed genes (DEGs) were associated with hormone changes that occur in response to light exposure (Feng et al., 2017). Comparative transcriptome pro ling for low-light tolerant and sensitive rice varieties was carried out via RNA-Seq analysis, and DEGs induced by low-intensity light at the tillering stage were identi ed (Sekhar et al., 2019). Rice leaf color mutant accessions were analyzed by wholegenome resequencing and transcriptomic approaches, and the identi ed DEGs were classi ed into different categories, including genes related to macronutrient (e.g. magnesium and sulfur) transport and genes related to avonoid biosynthesis . RNA-seq analysis was performed for darktreated rice Hwaseong and a backcross inbred line-CR2002, the results suggested that OsGW2 controls chlorophyll content and dark-induced senescence (Shim et al., 2020).
In this study, transcriptomic analysis using an RNA-Seq approach was carried out for yellow leaves of dark-treated rice seedlings. DEGs were identi ed and DEG-enriched metabolic pathway analysis was performed with KEGG. Furthermore, data for previously reported LCR genes were collected for KEGG analysis, and metabolic pathways that include both LCRs and dark-response DEGs were identi ed. It was found that the transcript abundance of most LCR genes changed under dark treatment, suggesting an overlap between leaf color regulation and dark response. Therefore, integrated analyses were carried out, producing supporting, supplementary evidence for this hypothesis. This study demonstrated the evidence to support the overlap and described the overlap in a quantitative manner. It also provides important clues for identifying additional LCR genes, improving mechanistic understanding of dark response and leaf color regulation.

Results
Phenotypic characterization and transcriptomic analysis of dark-treated rice seedlings Rice seedlings grown in normal light condition (12-h light/12-h dark) for 5 days were switched to dark treatment, and samples were collected at 0 (CK), 3 (CD_3d) and 6 days (CD_6d). The seedling height and chlorophyll a, chlorophyll b and carotenoid content were measured. Under CD stress, rice seedlings showed more leaf yellowing and taller plant height than CK (Fig. 1A). The plant height increased 19.2% and 11.5% at CD_3d and CD_6d, respectively, compared with control (Fig. 1B). The chlorophyll a, chlorophyll b, carotenoid and total chlorophyll contents under 6d of CD signi cantly decreased by 82.0%, 99.4%, 70.4% and 85.0%, respectively (Fig. 1C, 1D, 1E and 1F). Additionally, the ratio of chlorophyll a/b under dark stress was signi cantly increased in comparison with the control (Fig. 1G). Taken together, yellowed leaves, increased plant height and decreased chlorophyll content are typical characteristics of plant skotomorphogenesis, suggesting that CD stress resulted in de cient rice growth and development.
RNA-Seq transcriptome analysis was carried out for CK, CD_3d and CD_6d rice seedlings. The results revealed 33,755 non-redundant transcripts, accounting for 59.2% of all predicted rice genes (about 57,023). A total of 13,115 DEGs were identi ed, which accounts for 38.9% of all non-redundant transcripts (Additional le: Table S1). Approximately 8,242 DEGs were detected between CK and CD_3d, of which 4,467 were up-regulated and 3,775 were down-regulated. Additionally, there were 8,266 DEGs between CK and CD_6d, of which 3,834 were up-regulated and 4,432 were down-regulated. There were more upregulated DEGs than down-regulated DEGs for CD_3d, while there were fewer up-regulated DEGs than down-regulated DEGs for CD_6d ( Fig. 2A). There were 4,849 CD_3d only and 4,873 CD_6d only DEGs. A total of 3,393 genes were DEGs in both CD_3d and CD_6d (Fig. 2B). It can be inferred that the early stages of dark treatment induced the transcription of many genes, while longer periods of dark treatment led to increased numbers of down-regulated genes. This switch from up-regulation to down-regulation is an indication that as the CD treatment is extended, rice seedling growth becomes more inhibited and the expression of additional genes is down regulated.
To verify the transcription data and determine protein abundance, we conducted WB analysis for nine DEGs under dark-stress (Additional le: Figure S1). The results showed that changes in the expression of seven tested proteins (OsCAB1, OsCAB2, OsDGW10, OsGST, OsLhcb4, OsLIR1, and OsPAO5) were consistent with the direction of differential transcript expression, supporting the fact that consistent changes were seen for most DEGs at both transcriptional and translational levels.

DEG-enriched KEGG metabolic pathway analysis
KEGG metabolic pathway analysis of dark-response DEGs was performed by ClusterPro ler R. DEGenriched pathways and corresponding rich factors (RF) for CD_3d and CD_6d are listed in Additional le: Table S2. The 20 most signi cantly enriched pathways are shown in Fig. 3. In CD_3d, the term with the highest RF was cutin, suberine and wax biosynthesis. There were 11 up-regulated DEGs and 6 downregulated DEGs. The second highest RF term was glyoxylate and dicarboxylate metabolism with 13 upregulated DEGs and 22 down-regulated DEGs. The third highest RF term was amino sugar and nucleotide sugar metabolism with 49 up-regulated DEGs and 13 down-regulated DEGs. In CD_6d, the pathway with highest RF was photosynthesis-antenna proteins with 13 down-regulated DEGs, the second highest RF term was monobactam biosynthesis with 7 down-regulated DEGs, and the third highest RF term was photosynthesis with 1 up-regulated DEG and 28 down-regulated DEGs. Comparison the terms from 3d and 6d, 7 out of 20 were overlapped, which supported the unique dark-stress response at different time points.
Next, the CD_3d and CD_6d DEGs were combined, and RF was recalculated for each pathway using nonredundant DEGs. The 20 pathways with the highest RF are shown in Fig. 4, and the highest RF pathways include photosynthesis-antenna proteins; carotenoid biosynthesis; cutin, suberine and wax biosynthesis; monobactam biosynthesis; porphyrin and chlorophyll metabolism (PCM); other glycan degradation; one carbon pool by folate; glyoxylate and dicarboxylate metabolism; and photosynthesis and biotin metabolism. Data for additional pathways are listed in Additional le: Table S3.

Identi cation of transcription factors among dark-response DEGs
Because transcription factors (TFs) play important roles in many biological processes, TFs among the dark-response DEGs were identi ed from matches with the Oryza sativa subsp. japonica transcription factor database in PlantTFDB (planttfdb.cbi.pku.edu.cn). Of the 8,242 DEGs in CD_3d, 449 (5.4%) were annotated as TFs belonging to 48 TF families, of which 296 were up-regulated and 153 were downregulated. Of the 8,266 DEGs from CD_6d, 394 (4.8%) were annotated as TFs belonging to 47 families, of which 262 were up-regulated and 132 were down-regulated. Combined, 661 non-redundant TFs, representing 52 TF families, were identi ed as DEGs at the two dark treatment time points. Note, as shown in Table 1, the TF families that were most affected by CD were the bHLH, MYB, C2H2 and WRKY TF families. In particular, there were 47 bHLH among 449 TFs (10.5%) in CD_3d and 36 bHLH among 394 TFs Collection of rice leaf color-related (LCR) genes Rice leaves showed green color under normal growth condition. Numerous leaf color mutants have been identi ed via traditional forward genetics and breeding programs in rice. Information on 150 LCR genes was collected from literature, and loc#, gene name and references are summarized in Table 2 and Additional le: Table S6. As shown in Table 2, the phenotypes of known LCR mutants are albino (26), yellow (52), temperature-sensitive (14), stay green (16), stripe (15), spot (9), purple (1) and turning green (17). On the basis of their annotation, LCR genes mainly participate in chlorophyll synthesis, chloroplast development and photosynthesis.
Transcription analysis of LCR genes in dark-treated rice RNA-Seq measurements of the transcript levels under CD stress for 150 LCR genes were surveyed. This analysis revealed that 102 of 150 LCR genes (68.0%) were also dark-response DEGs, of which 18 were upregulated at both 3d and 6d, 77 were down-regulated at both 3d and 6d, and 7 had opposing up-and down-regulation at the two time points (Additional le: Table S6). Thus, the transcript abundance of most LCR genes changed in response to dark stress.
The 102 LCR DEGs were functionally classi ed according to their annotations. A heat map was drawn showing log2 (fold change) values under 3 d and 6 d dark treatments (Fig. 5). In Fig. 5, there are 50 chloroplast development related genes, including three that were up-regulated and 46 that were downregulated at both time points. There were 15 chlorophyll synthesis-related LCR genes, including three that were up-regulated and eight that were down-regulated at both time points. There were 10 photosynthesisrelated LCR genes and ve chlorophyll degradation-related LCR genes.
These data support that chloroplast development, chlorophyll synthesis and photosynthesis LCR genes play important roles in leaf yellowing under CD treatment and, particularly those down-regulated genes in these pathways. Moreover, aging, active oxygen scavenging, light signal transduction and carotenoid synthesis are also involved in the yellowing of leaves. Signi cant down-regulated expression was observed for most LCR genes under constant dark treatment, which implies that these LCR genes play a positive role in light signal responses.

KEGG analysis of LCR genes
Although hundreds of LCR genes have been cloned through traditional genetic approaches, larger-scale understanding of molecular mechanisms controlling leaf color remains poor. Thus, LCR genes were functionally classi ed and assigned to KEGG Oryza sativa japonica (Japanese rice) pathways. Similar to the calculation of rich factor in KEGG enrichment analysis, the number of LCR genes as a proportion of the total number of genes in a speci c metabolic pathway (#GMP) was calculated as mutant rich factor (mRF). Fig. 6 shows pathway names and mRF values. The highest mRF pathway was PCM, accounting for 40.9% (18/44), followed by betalain biosynthesis (16.7%; 1/6) and terpenoid backbone biosynthesis (7.7%; 4/52). The mRF for other metabolic pathways was less than 6% (Additional le: Table S7). Based on this result, it can be postulated that the PCM pathway plays important roles in regulating rice leaf color. KEGG analysis of the collection of LCR genes brought focus to speci c metabolic pathways, providing a broader understanding of leaf color mechanisms and clues for identifying novel LCR genes.

Integrated analysis of dark-response DEGs and LCR genes
Among the 150 LCR genes, the transcript abundance of 102 genes (68.0%) changed under CD treatment, indicating an overlap between dark-response and leaf color regulation. Thus, it is reasonable and necessary to carry out integrated analysis of the two biological networks.
A correlation between the two sets of data can be seen in Fig. 7. As shown in Fig. 7, overlap between darkresponse DEGs and LCRs were indicated. In the PCM pathway (Dosa00860), there are 27 dark response DEGs and 18 are LCR genes. The number of LCR genes overlapped with dark-response DEGs normalized to the total number of dark-response DEGs is 55.6% (15/27) and it is 83.3% when normalized to total number of LCR genes (15/18), which provides strong evidence supporting PCM pathway as a core component of overlap between dark response and LCR. Fig. 8 highlighted the pathways that include both dark-enriched DEGs and LCR genes, and the name or locus number of each gene is shown. Fifteen genes in PCM pathway are both LCR genes and dark-response DEGs: OsNYC1, OsGGR2, OsCRD1, OsHY2, OsLYL1, OsDVR, OsChlI, OsChlD, OsPORA, OsSGRL, OsYGL1, OsYGL18, OsSGR, OsRCCR1, and OsPGL. Taken together, integrated analysis provided a tool for speci c examination of genome-wide transcriptome data within the context of metabolic pathways established using traditional forward genetics data. The complementary sets of data mutually support one another.
The chlorophyll biosynthesis process in the PCM pathway plays important roles in rice leaf color regulation To further examine the function of LCR genes and DEGs in PCM, a ow chart was drawn based on the KEGG data, to highlight the position and distribution of the 15 LCR dark-response DEGs (Fig.9). Fig. 9 and Additional le: Table S8 shows that the LCR overlapped dark-response DEGs were neither evenly nor randomly distributed throughout the metabolic pathway. Instead, almost all of them were concentrated to chlorophyll biosynthesis, which is conserved in higher plants. Thus, it can be postulated that this process is the key part of the PCM pathway responsible for both LCR and dark response, and additional genes that participate in this process are potential LCR genes.

Discussion
To understand the effect of light on rice growth, RNA-Seq analysis was carried out for rice leaves treated with constant dark, and DEG-enriched KEGG pathways with high rich factors were identi ed. Meanwhile, a collection of rice LCR genes identi ed and cloned by traditional genetics were analyzed for mutantenriched pathways with KEGG. It was found that 102 of 150 LCRs (68.0%) were DEGs under CD treatment, suggesting an overlap between dark response and leaf color regulation networks. An integrated analysis of the two sets of data found that 83.3% of the LCRs in PCM pathways were also dark-response DEGs.
More importantly, most of the LCR genes participate in chlorophyll synthesis, which suggests that chlorophyll synthesis, as a central part of the PCM pathway, plays an important role in both leaf color regulation and dark response.

The mechanisms of dark response in rice
Darkness is a severe form of stress, that causes rice leaves to become yellow and unhealthy. Results obtained in this study and reported literature have shown that chlorophyll a, chlorophyll b and carotenoid contents are signi cantly reduced under CD treatment. RNA-seq analysis of dark-treated leaves revealed the down-regulation of genes involved in the synthesis of chlorophyll, the light-harvesting complex (LHC), and carotenoids (Additional le: Table S9). Speci cally, the transcript abundance of OsChlI, OsChlH, and OsYGL1 decreased 13.3-, 5.5-and 5.30-fold, respectively. OsChlI and OsChlH were reported to encode the CHLI and CHLH subunits of Mg 2+ -protoporphyrin IX chelatase (Mg 2+ -chelatase) (Zhang et al., 2006a;Inagaki et al., 2015), while OsYGL1 encodes chlorophyll synthase . These are key enzymes for chlorophyll synthesis and leaf color regulation. We found that LHC synthesis genes OsCAB1, OsDGW10, OsLhca4, and OsLhcb4 were down-regulated and that degradation genes OsSGR and OsSGRL were increased 121.6-and 74.5-fold, respectively, which is consistent with previous ndings (Rong et al. 2013;Jiang et al. 2017). Additionally, the expression of β-OsLCY, encoding lycopene β-cyclase, decreased by 2.4-fold, that of OsPDS, encoding phytoene desaturase decreased by 3.5-fold, and the expression of OsZDS, encoding ζ-carotene desaturase, decreased by 2.9-fold. These three genes encode key enzymes involved in carotenoid synthesis, with β-OsLCY is also known to be a LCR gene (Fang et al., 2008). Taken together, the down-regulation of chlorophyll-related genes provide an explanation for the decreased chlorophyll content observed in dark-treated rice leave.
To verify the results obtained from TP309 in our experiment, transcriptional signals obtained from Hwaseong and CR2002 were analyzed (Additional le: Table S9). This showed that most of the chlorophyll synthesis genes were down-regulated, consistent with the performance at 6 d in TP309.
CD-enriched DEGs in photosynthesis-related pathways were detected by RNA-Seq analysis in this study.
Four of the 10 most highly DEG-enriched metabolic pathways in dark response are associated with photosynthesis, they are photosynthesis-antenna proteins; photosynthesis; carotenoid biosynthesis; glyoxylate and dicarboxylate metabolism (Fig. 4). In the photosynthesis-antenna proteins pathway, all 13 DEGs were down-regulated. In glyoxylate and dicarboxylate metabolism, OsRBCS4, which encodes a small subunit of rubisco, was down-regulated 3001-fold and is also an LCR gene (OGAWA et al., 2012). In the photosynthesis pathway, 30 out of 33 DEGs were down-regulated, among which, LOC_Os12g10570 was down-regulated 90-fold. RNA-Seq analysis has previously been carried out for different light treatments (white, red, blue and green). It was found that photosynthesis-related genes were signi cantly downregulated and carbohydrate degradation was pronounced in darkness (Lakshmanan et al., 2015), consistent with the results obtained in this study. Furthermore, it was found that the four photosynthesisrelated pathways were among the 10 most highly DEG-enriched metabolic pathways in dark treated Hwaseong and CR2002 rice lines (Additional le: Table S10). These results supported that dark treatment affected the photosynthesis process dramatically.

The mechanisms of leaf color regulation
Leaf color is an important agronomic trait that is directly related to rice growth and grain yield. In this study, data on 150 LCR genes were consolidated from the literature. According to LCR gene transcription data, most LCR genes were down-regulated under constant dark treatment, which demonstrates that light signals can play positive roles in regulating LCR gene expression. Light is the upstream initiator of signal transduction in plants. PIFs (phytochrome interacting factors) are a class of bHLH transcription factors that can interact with phytochrome (Phy). Light signals regulate PIFs protein stability; that is, the protein is stable in the dark and degrades in the light (Demarsy et al., 2017). In rice, six PIF genes have been identi ed and designated as OsPIL11 to OsPIL16 (Nakamura et al., 2007;Piao et al., 2015). It was reported that OsPIL13 is an LCR gene, with expression controlled by circadian rhythms (Nakamura et al., 2007). OsPIL13 binds to the promoters of two Chl biosynthetic genes, OsPORB and OsCAO1, and induces the transcription of downstream genes (Sakuraba et al., 2017). OsPIL15 is responsible for regulating rice tiller angle in response to light and gravity, and OsPIL15 expression is negatively regulated in etiolated seedlings exposed to light (Xie et al., 2019). In this study, up-regulated expression for OsPIL11, OsPIL13, and OsPIL16 was detected, and more bHLH TFs were found to be CD-response DEGs than any other TF family, suggesting that bHLH TFs play important roles to propagating light signal transduction in rice.
Darkness can lead to extensive stress responses, including loss of green leaf color. Direct correlations between light and leaf color mutation have been reported. For example, OsLYL1, encoding a geranylgeranyl reductase, was induced by light and suppressed by dark (Zhou et al., 2013b). OsYGL18 encodes a putative magnesium protoporphyrin IX methyltransferase (ChlM). When an OsYGL18 deletion mutant (ygl18) was transferred from dark to light, chlorophyll content increased and its expression was up-regulated (Wang et al., 2017c). OsZN encodes a thylakoid-bound protein of unknown function, and its mRNA level in constant light is higher than that in CD, indicating that OsZN transcription is controlled by light (Li et al., 2010). In addition, an OsOTP51 mutant, encoding a pentatricopeptide repeat protein, showed dramatic changes in PSI structure and function, which led to severe photoinhibition (Ye et al., 2012). In this study, it was found that expression of these genes was down-regulated after CD treatment, which suggests that light signals positively regulate LCR genes.
Although more than 100 LCR genes have been cloned by traditional genetic methods, it has been di cult to assess the role of speci c LCR genes in the context of metabolic pathways. In this study, LCR genes were surveyed with KEGG metabolic pathway analysis. LCR gene-enriched pathways were identi ed by calculating the mutant rich factor (mRF). It was found that the mRF was highest for PCM, which suggested that PCM plays an important role in leaf color regulation, and additional genes from the PCM pathway may be LCR genes.
The feasibility and importance of integrated analysis of transcriptome and genetic data Genome-wide transcriptome analysis can detect stress-related DEGs, and KEGG enrichment analysis can further identify DEG-enriched metabolic pathways. However, clues obtained by transcriptomic analysis require veri cation from genetic analyses and functional studies. Traditional forward genetics approaches can e ciently identify genes related to speci c mutant phenotypes, but it is di cult to integrate a group of mutant-related genes into speci c metabolic pathways. In this circumstance, it was possible to obtain mutual supporting, complementary evidence through integration of two sets of data. The integrated analysis can get focused view from transcriptomic data and a magni ed view from genetic data.
In this study, RNA-Seq analysis of CD-treated rice seedlings was carried out and DEGs were identi ed. Data on 150 LCR genes previously cloned by traditional genetic means were collected, and 102 LCR genes (68.0%) were found to be dark-response DEGs, which suggests an overlap between dark response and leaf color regulation networks. Furthermore, KEGG analysis of LCR genes showed that the mRF of LCR genes were the highest for PCM, and most PCM LCR genes (83.3%,15/18) were dark-response DEGs.
DEGs identi ed from Hwaseong and CR2002 lines were downloaded from Shim et al. (2020) and parallel analysis were carried out for the overlap with LCR genes (Additional le: Figure S2). As shown in the gure, 91 and 81 LCR genes that were DEGs in Hwaseong and CR2002 account for 60.7% (91/150) and 54.0% (81/150) of LCRs, respectively. These results support the existence of an overlap between dark response DEGs and LCRs. In the PCM pathway, 14 genes overlap between DEGs and LCRs, accounting for 77.8% (14/18) of LCR genes in both rice lines, indicating that the PCM was is the core component of the overlap.
This independent evidence provides further support for the ndings from TP309. Moreover, most of the genes overlapping between LCR and DEG in the PCM pathway are involved in chlorophyll biosynthesis.
RNA-Seq data provided understanding of leaf color mutations, and traditional genetic research provided complementary data towards clarifying the mechanism of dark-stress response. Genome-wide analysis is aimed at investigating all genes. When facing different biological questions in the same organism, data from different sources can be integrated for coordinated analysis. With the rapid accumulation of transcriptome data, which can be combined with genetic data that has been collected over many years, integrated analysis can now be carried out to a greater extent, which will contribute to improved understanding of rice biology.

Dissection of the PCM pathway
DEGs can be allocated to speci c metabolic pathways via KEGG analysis; however, a metabolic pathway may contain dozens or even hundreds of genes that are connected with multiple interrelated processes.
Analysis of the location and distribution of DEGs is helpful for making functional associations with speci c processes. In other words, more accurate and precise results can be obtained if metabolic pathways are dissected. In this study, DEGs derived from whole-genome transcriptomic analysis and LCR genes collected from traditional genetics literature were associated with the PCM pathway. Further, the DEGs and LCR genes were neither randomly nor evenly distributed throughout the pathway. Instead, they were concentrated in the chlorophyll biosynthesis part of the pathway (Nagata et al., 2005). On this basis, it can be postulated that chlorophyll biosynthesis is the key process for leaf color regulation. Analyzing dissected metabolic pathways is important because it helps narrow the range of target genes and leads to comprehensive understanding of gene functions.

Conclusions
In this study, transcriptomic analysis using an RNA-Seq approach was carried out for yellow leaves of dark-treated rice seedlings. DEGs and DEG-enriched metabolic pathways were analyzed. Data for reported LCR genes were collected. Metabolic pathways that included both LCRs and dark-response DEGs were identi ed. It was found that the transcript abundance of most LCR genes changed under dark treatment, suggesting an overlap between leaf color regulation and dark response. KEGG analysis revealed enrichment of LCR genes in porphyrin and chlorophyll metabolism (PCM). Interestingly, most of the overlapped LCR genes and DEGs were concentrated at chlorophyll biosynthesis in the central of PCM, indicating that PCM pathway, particularly chlorophyll biosynthesis process, is the core component of the overlap between rice LCR and dark stress-response. This study provides important clues for identifying additional LCR genes, understanding the mechanisms of dark response and leaf color regulation.

Materials And Methods
Plant materials, cultivation and dark treatment Rice variety TP309 was used in this study. After being soaked in water for 3 days, seeds were surface sterilized and then germinated in soil (soil: vermiculite = 1:1). Rice seedlings were grown for 5 days in a 30°C growth chamber under a 12-h light/12-h dark cycle (60 µmol·m -2 s -1 ). Then, seedlings were transferred to 24-h dark (constant dark, CD) conditions until sample collection.

Measurement of plant height and chlorophyll content
Phenotypic photos were taken at different time points during constant dark treatment. The plant height was measured for more than 10 randomly selected seedlings. The mean and standard derivation were calculated. Rice leaves were harvested, sliced and soaked in 80% acetone for 48 h in the dark at 4℃. The optical density was recorded at 470 nm, 646 nm and 663 nm with an ultraviolet spectrophotometer. The chlorophyll a, chlorophyll b, carotenoids, and total chlorophyll contents, chlorophyll a/b ratio were calculated in accordance with Lichtener and Wellburn's method (Lichtenthaler and Wellburn, 1983). Three replicates were measured.
Total RNA extraction from rice seedlings and RNA-Seq analysis Rice seedlings under CD treatment for 0, 3 and 6 days were collected for total RNA extraction and library construction. Transcriptomic analysis was performed using Illumina Novaseq 6000 by Novegene Co., Ltd (Beijing, China).

Identi cation of differentially expressed genes (DEGs) and KEGG-enrichment analysis
Genes with log2 fold change (absolute value) >1 and P-value <0.05 were de ned as differentially expressed genes. The Kyoto Encyclopedia of Genes and Genomes (KEGG, https://www.genome.jp/kegg/) pathway network was used to identify DEG-enriched metabolic pathways under dark treatment. The DEGenriched KEGG analysis was performed by ClusterPro lerR software. Rich factor (RF), the number of DEGs normalized to the total number of genes in a speci c metabolic pathway (#GMP), was calculated for all metabolic pathways. Bubble charts were drawn for the 20 pathways with the highest P-value.

Identi cation of transcription factors among DEGs
Transcription factors among DEGs were identi ed with PlantTFDB v5.0 (planttfdb.cbi.pku.edu.cn), and their relative expression levels were obtained from the RNA-Seq data.

Collection of leaf color-related (LCR) genes
Rice LCR genes were collected by literature mining. The gene name, locus number, annotation and references for cloned LCR genes are summarized in Table 2. KEGG analysis of LCR genes was carried out as DEGs. The number of LCR genes measured as a proportion of the #GMP is referred to as the mutant rich factor (mRF).

Heat Map
LCR genes were functionally classi ed by their protein's annotated function, and different colors were assigned to indicate expression on the basis of log2 (fold change) values. Red and green colors depict upand down-regulated genes, respectively. The LCR heat map was made with EXCEL software (Microsoft Corporation, USA).
Metabolic pathway chart LCR gene locus numbers were input to obtain corresponding EC (Enzyme Commission) numbers from the KEGG database (https://www.genome.jp/kegg/pathway.html). KEGG Mapper software was used to assign a color to each gene based on the transcript abundance, as determined by RNA-Seq analysis. On the basis of log2 (fold change) values, red lines represent up-regulated DEGs, green lines represent down-regulated DEGs, and yellow lines represent DEGs with discrepant up-and down-regulation at 3 d and 6 d under constant dark.

Supplementary Information
The datasets supporting the conclusions of this article are included within the article and its additional le.
Additional le: Figure S1. Comparative analysis for the expression of 9 DEGs under dark stress. Figure S2. The integrated analysis between LCRs and DEGs identi ed from rice Hwaseong and CR2002 varieties. Table S1. Transcriptional data for rice under constant dark stress. Table S2. DEG-enriched metabolic pathway analysis by KEGG. Table S3. List of metabolic pathways containing dark-response DEGs. Table S4. List of TFs among dark-response DEGs. Table S5. Transcript abundance of TFs that are dark-response DEGs. Table S6. List of rice leaf color-related (LCR) genes. Table S7. KEGG analysis of rice leaf color-related (LCR) genes. Table S8. Integrated analysis of dark response and leaf color regulation networks. Table S9. List of DEGs related with chlorophyll synthesis and degradation Table S10. KEGG analysis for DEGs identi ed in Hwaseong and CR2002 (Shim et al. 2020) Table S11. Integrated analysis between LCR genes and DEGs identi ed in Hwaseong and CR2002 (Shim et al. 2020) Abbreviations CD: constant dark; Chl: chlorophyll; DEG: differentially expressed gene; GMP: genes in a speci c metabolic pathway; KEGG: Kyoto encyclopedia of genes and genomes; LCR: leaf color-related; LHC: light-harvesting complex; mRF: mutant rich factor; PCM: porphyrin and chlorophyll metabolism; Phy: phytochrome; PIF: phytochrome interacting factor; PIL: rice phytochrome-interacting factor-like; RF: rich factor.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
No applicable.

Availability of data and material
All data generated or analyzed during this study are included in this published article or its supplementary information les or are available from the corresponding authors on reasonable request.  Rich factors for DEG-enriched metabolic pathways identi ed by KEGG analysis of dark-treated rice. Twenty pathways with highest rich factor value were listed. More pathways were listed in additional le: Table S3.

Figure 5
The expression pro le of LCR genes that are dark-response DEGs. 3d and 6d refer to 3 and 6 days under constant dark treatment, respectively. Red and green colors depict up-and down-regulation, respectively. The scale shows log2 (fold change) values.  Integrated analysis of dark response and leaf color regulation. Gene names in red (italic) are LCR genes that were also dark-response DEGs. Dosa# indicated the number of pathways in KEGG database. The percentages at the bottom represent the number of LCR genes overlapped with dark-response DEGs normalized to the total number of dark-response DEGs. Ten pathways are listed here, and detailed information for additional pathways is listed in additional le: Table S8.