Differential gene expression for carotenoid biosynthesis in a green alga Ulva prolifera based on transcriptome analysis

Background Carotenoids are widely distributed in plants and algae, and their biosynthesis has attracted widespread interest. Carotenoid-related research has mostly focused on model species, and there is a lack of data on the carotenoid biosynthetic pathway in U. prolifera that is the main species leading to green tide, a harmful plague of floating green algae. Results The carotenoid content of U. prolifera samples, that is the main species leading to green tide, a harmful plague of floating green algae at different temperatures revealed that its terpenoid was highest in the samples subjected to high temperature at 28 °C (H), followed by the samples subjected to low temperature at 12 °C (L). Its terpenoid was lowest in the samples subjected to medium temperature at 20 °C (M). We conducted transcriptome sequencing (148.5 million raw reads and 49,676 unigenes in total) of samples that were subjected to different temperatures to study the carotenoid biosynthesis of U. prolifera. There were 1125–3164 significant differentially expressed genes between L, M and H incubation temperatures, of which 11–672 genes were upregulated and 453–3102 genes were downregulated. A total of 3164 genes were significantly differentially expressed between H and M, of which 62 genes were upregulated and 3102 genes were downregulated. A total of 2669 significant differentially expressed genes were observed between L and H, of which 11 genes were upregulated and 2658 genes were downregulated. A total of 13 genes were identified to be involved in carotenoid biosynthesis in U. prolifera, and the expression levels of the majority were highest at H and lowest at M of incubation temperature. Both the carotenoid concentrations and the expression of the analysed genes were lowest in the normal temperature group, while low temperature and high temperature seemed to activate the biosynthesis of carotenoids in U. prolifera. Conclusions In this study, transcriptome sequencing provided critical information for understanding the accumulation of carotenoids and will serve as an important reference for the study of other metabolic pathways in U. prolifera. Electronic supplementary material The online version of this article (10.1186/s12864-018-5337-y) contains supplementary material, which is available to authorized users.


Background
Terpenoids are the most widely distributed natural compounds in nature, playing an important role in growth, development, light absorption, hormone synthesis, photoprotection and stress resistance in plants [1]. Terpenoids are the largest and most diverse volatiles released by plants [2]. Isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) are the precursors for terpenoid synthesis. Two pathways participate in the biosynthesis of IPP and DMAPP in living organisms: the MVA pathway and the MEP pathway. In recent years, terpenoid metabolism in algae has been intensively studied, and there is adequate evidence that in green algae, the MVA pathway has been abandoned, and only the MEP pathway supplies the precursors of all cellular terpenoids [3]. The MEP pathway is not dependent on mevalonic acid (MVA) for IPP and DMAPP synthesis. There are seven enzymes involved in the pathway, including DXS, DXR, CMS, CMK, MCS, HDS and HDR ( Fig. 1) [4]. These enzymes play an important role in regulating the pathway.
Carotenoids are mostly C40 compounds consisting of eight isopentenyl-pyrophosphate units and comprise a class of natural pigments that belong to the tetraterpenoid group [5]. Carotenoids are generally synthesized through the terpenoid biosynthetic pathway in plant plastids. There are six enzymes involved in the pathway, including IPI, GGPS, PSY, PDS, ZDS and CRTL-B (Fig. 2) [6]. The formation of monoterpenes, sesquiterpenes, and diterpenes is activated by terpene synthases (TPS) in plants [7], however, scientists have found almost no TPS in algae [8], so it is critical to study the genes of the MEP pathway and the downstream pathway to understand carotenoid biosynthesis in algae.
Ulva prolifera (Ulvaceae, Chlorophyta), that is the main species leading to green tide, a harmful plague of floating green algae [9]. The macroalgae usually live in intertidal zones and have complex life histories and multiple reproductive events [10]. These features have been considered critical strategies to grow rapidly. Since 2007, U. prolifera blooms have occurred in the southern Yellow Sea continuously and led to a green tide disaster along the coast of the Shandong and Jiangsu provinces of China. This phenomenon had negative impacts on the local economy and environment [11][12][13]. Therefore, the study of U. prolifera is an increasing focus of scientists and has drawn considerable attention from the Chinese government [14,15].
With the development of sequencing technology, transcriptome sequencing has been widely used in the study of various species. Transcriptomics revealed dynamics of photopigment [16], flavonoid [17] or terpenoid [18] synthesis in algae or land plants. Stress response of algae or cyanobacteria to desiccation [19], deprivation of essential element [20] or temperature [21] was portrayed by this technique. In green algae, transcriptome sequencing has been widely applied to study various biological processes [22][23][24].
Temperature is a key factor for terpenoid biosynthesis in algae [25]. In this study, we aimed to better understand the expression of key genes related to carotenoid biosynthesis in U. prolifera under different temperatures. All of the genes involved in the MEP pathway and the downstream pathway were screened, and their expression patterns were analysed through transcription data. Our results not only provide insight into carotenoid biosynthesis in U. prolifera but also provide an important reference for further study in U. prolifera. Terpenoids metabolism is an important metabolism of U. prolifera, we made a significative preliminary work for exploring the reason of the rapid formation of green tide in terms of the metabolic activities in the future.

Results
The carotenoid, chlorophyll (Chl) a and Chl b content of the L, M and H samples The quantitative determination of carotenoid, Chl a and Chl b content was performed using a spectrophotometer. The concentrations of carotenoids, Chl a and Chl b in the L group were 3.34 ± 0.03 μg/ml, 2.20 ± 0.07 μg/ml and 1.67 ± 0.01 μg/ml, respectively. The concentrations of carotenoids, Chl a and Chl b in the M group were 2.80 ± 0.03 μg/ml, 1.78 ± 0.05 μg/ml and 1.34 ± 0.01 μg/ ml, respectively. The concentrations of carotenoids, Chl a and Chl b in the H group were 4.72 ± 0.03 μg/ml, 3.2 ± 0.04 μg/ml and 1.90 ± 0.06 μg/ml, respectively (Fig. 3). The concentrations of terpenoids were highest in the samples that were subjected to high temperature, followed by the samples that were subjected to low temperature; the samples that were subjected to the medium temperature had the lowest terpenoid concentrations.
GO functional annotations consist of three ontologies: cellular component, molecular function, and biological process. A total of 18,027 annotated unigenes were categorized into three ontologies with 56 GO terms (Additional file 1: Figure S1). The 'cellular component' category had the most unigenes (16183), followed by 'molecular function' (15,886 unigenes) and 'biological process' (15,205 unigenes). For the cellular component category, 'cell' and 'cell part' dominated in the ontology; 'binding' and 'catalytic activity' were the two most abundant terms in the 'molecular function' category, and the most highly represented terms in the 'biological process' category were 'cellular process' and 'metabolic process'.
Differentially expressed genes (DEGs) among the L, M and H samples Differential expression analysis was performed for the L, M and H samples (Fig. 5). The results showed that there were 1125 significant differentially expressed genes between L and M, of which 672 genes were upregulated and 453 genes were downregulated in L compared to M. There were 3164 significant differentially expressed genes between H and M, of which 62 genes were upregulated and 3102 genes were downregulated in H compared to M. There were 2669 significant differentially expressed genes between L and H, of which 11 genes were upregulated and 2658 genes were downregulated L compared to H (Fig. 5a).
Venn diagrams of the differentially expressed unigenes among the L, M and H samples showed that 220 genes were significantly differentially expressed both between L and M and between H and M. These genes may play a key role in responses to temperature stress in U. prolifera. The number of genes that were differentially expressed between L and M only was 763; these genes The concentrations of and total carotenoids, Chl a and Chl b in samples subjected to different temperatures. The absorbances at wavelengths of 470 nm, 646.8 nm and 663.2 nm were measured using a spectrophotometer, and the results were obtained using formulas [62]  may play a key role in responses to low-temperature stress in U. prolifera. The number of genes that were differentially expressed between H and M only was 2340; these genes may play a key role in responses to high-temperature stress in U. prolifera (Fig. 5b). KEGG analysis was performed to further explore the functional classification and pathway assignment of the DEGs in the L vs M, H vs M and L vs H comparisons. Many genes were enriched for the gap junction, apoptosis, carbon fixation in photosynthetic organisms, photosynthesis, pentose phosphate pathway, carotenoid biosynthesis, linoleic acid metabolism, carbon metabolism, carotenoid biosynthesis, plant hormone signal transduction, and citrate cycle terms (Fig. 6). We used GO assignments to classify the functions of DEGs in pairwise comparisons of the cDNA libraries of samples subjected to different temperatures. The GO enrichment revealed 10 biological process categories, 10 cellular component categories and 10 molecular function categories that were enriched for the L vs M, H vs M and L vs H comparisons (Additional file 2: Figure S2).

Genes involved in carotenoid biosynthesis
All carotenoid biosynthetic pathway genes were identified in U. prolifera by analysing 56 unigenes that participate in terpenoid backbone biosynthesis and 22 unigenes that participate in carotenoid biosynthesis. The carotenoid biosynthesis genes in Chlamydomonas reinhardtii, Chlamydomonas eustigma, Chlorella sorokiniana, Chlorella variabilis and Volvox carteri were subjected to BLAST analysis with the annotated carotenoid biosynthesis genes in U. prolifera. The genes of the MEP pathway were divided into 7 types: DXS, DXR, CMS, CMK, MCS, HDS, and HDR (Fig. 7a); the genes of the carotenoid biosynthetic pathway were divided into 6 types: IPI, GGPS, PSY, PDS, ZDS and CRTL-B (Fig. 7b). Heatmaps were constructed using the program MeV to show the differentially expressed transcripts among L, M and H (Fig. 8). Hierarchical clustering analysis was performed with seven genes of the MEP pathway for L, M and H (Fig. 8a). This analysis indicated two gene clusters. One cluster included the genes in L or H that were upregulated compared with the levels in M, namely, DXS, DXR, CMS, CMK, MCS and HDR; these genes had the highest expression levels in H. HDS grouped into the other cluster and had the highest expression level in L. In addition, hierarchical clustering analysis was performed with six genes of the carotenoid biosynthetic pathway (Fig. 8b). The results showed two gene clusters. One of the clusters, ZDS, this gene had the highest expression level in L. The other genes, GGPS, IPI, PSY, PDS, CRTL-B, grouped into the other cluster, and these genes had the highest expression level in H.

qRT-PCR verification of changes in gene expression from the RNA-Seq analysis
To confirm the RNA-Seq results, the expression levels of all seven genes of the MEP pathway were quantified via qRT-PCR, and the expression levels (FPKM) among L, M and H were compared. There was a high degree of correlation between the qRT-PCR results and the RNA-Seq results (Fig. 9) indicating credibility of the transcriptomic profiling data. The FPKM values of GGPS, IPI, PSY, PDS, CRTL-B among L, M and H were compared simultaneously (Fig. 10). The expression levels of all the genes selected increased under low or high temperature, and our findings regarding the high expression of genes in the carotenoid biosynthetic pathway under low or high temperature were consistent with the high concentrations of carotenoids, Chl a and Chl b in the samples that were subjected to temperature stress.

Discussion
Carotenoids have been studied for more than 100 years. They play important roles in the structure and function of the photosynthetic apparatus of living organisms, including bacteria, algae and higher plants [26]. More than 750 different carotenoids have been reported in nature [27]. The majority of carotenoids exist in the photosynthetic tissues of plants and algae, and the green colour of chlorophyll masks various carotenoids, carotenoids that are produced in photosynthetic tissues are not well known. Carotenoids are also widespread in photosynthetic bacteria and in microorganisms such as non-photosynthetic bacteria and yeast [28,29]. Carotenoids have multiple functions, including enhancing immunity, inhibiting bacterial growth and exerting antioxidative activity [30,31]. Many types of carotenoids have been identified from different plants, and they play an important role as antioxidants [32,33]. Carotenoids participate in photoprotection in plants [34,35]. In the past, carotenoid-related research has mostly focused on model species, such as maize, tomato, rice and Arabidopsis [36]. Many microalgae and macroalgae are rich in carotenoids, therefore, carotenoids extracted from algae may be the main natural resource for studying potential functional components [37]. However, genes for carotenogenesis in algae are not yet known [38]. There is also a lack of data on the expression patterns of genes related to carotenoid metabolism in U. prolifera and there have been limited attempts to understand carotenoid biosynthesis in this species. In this work, we analysed the carotenoid biosynthetic pathway in U. prolifera by transcriptome sequencing.
Functionally confirmed enzymes of carotenoid biosynthesis have been found in algae species such as Chlorella, Chlamydomonas, Dunaliella and Haematococcus [39][40][41][42]. Isopentenyl pyrophosphate (IPP), a C 5 -compound, is the source of chlorophylls and carotenoids. There are two pathways of synthesis of this precursor: the MVA pathway and the MEP pathway [43]. The pathway of carotenoid biosynthesis in algae is similar to that in plants and is dependent on the MVA or MEP pathway for precursor production. In green algae, some biochemical and genomic evidence has proven that the MVA pathway has been lost and that the MEP pathway is the sole pathway [44][45][46][47]. In our study, we found that some of the genes of the MVA pathway exist in U. prolifera, such as AACT and HMGR. The other genes, HMGS, MVK, PMK and MVD, were absent in the transcriptome data. This suggests that a complete MVA pathway to synthesize terpenoids is lacking in U. prolifera. The results were in accordance with the results of studies on Porphyra umbilicalis, Cyanidioschyzon merolae 10D, and Chlorella zofingiensis [48][49][50]. The MEP pathway exists in the plastid, which is the only source of terpenoids in green algae [51]. Some green algae regulate the flux of terpenoid metabolism by differential expression of the gene families of enzymes in the MEP pathway [52,53].
Carotenoids are derived from the plastid-localized MEP pathway [54], for which pyruvate and glyceraldehyde 3-P act as initial substrates leading to the synthesis of GGPP [55]. Two GGPPs are catalysed by PSY to form phytoene [36]. Subsequently, carotenoids are produced from phytoene through a complex set of reactions requiring PDS, ZDS and CRTL-B [56]. We analysed the expression profiles of all the genes related to carotenoid biosynthesis and identified orthologues of previously known carotenoid genes in U. perolifera.
Secondary metabolites are the result of biological and non-biological interactions between organisms and the environment throughout evolution, and secondary metabolites play a critical role in improving the ability of organisms to survive and coordinate with the environment [25]. The production of and changes in secondary metabolites are influenced by the environment [57]. Plants have developed many modes for adaptation to temperature variations [58], and temperature is a main environmental factor that affects carotenoid biosynthesis and metabolism in plants [59][60][61] suggests different level of carotenoid synthesis and concentration according to change in temperature conditions. Both the carotenoid concentration and the expression of related genes were lowest under the normal temperature, while low temperature and high temperature seemed to activate the biosynthesis of carotenoids in U. prolifera. This finding revealed that carotenoids are involved in the response to temperature stress; in other words, carotenoids might have a protective function.

Conclusion
In this study, we conducted transcriptomic and carotenoid biosynthetic analysis on samples of U. prolifera that were subjected to different temperatures. The results provided a comprehensive explanation of carotenoid biosynthesis in U. prolifera. The MEP pathway was detected from the transcriptome data. However, the MVA pathway was absent in terpenoid metabolism. Temperature is a key environmental factor affecting carotenoid biosynthesis. The production and concentrations of carotenoids were susceptible to temperature, and carotenoid concentrations were upregulated when U. prolifera were subjected to temperature stress. The data reported in this study provide critical information for understanding the accumulation of carotenoids and will serve as an important reference for the study of other metabolic pathways in U. prolifera that is the main species making green tide.

Plant materials
U. prolifera samples were collected in March 2018 from Pyropia rafts (32°26'N, 121°25′E) in Nantong, Jiangsu, China. The samples were cultivated in seawater medium, and cool-white fluorescent light was provided on a 12:12 L:D cycle. The cultivation environment of the samples was as follows: 120 μmol photons m − 2 ·s − 1 in seawater with a salinity of 30. There were three temperature regimes: 12°C was set as the low temperature (L), 28°C was set as the high temperature (H), and 20°C was set as the medium temperature (M).
The medium temperature group served as a control group. The samples, which were cultivated at the three different temperatures for 7 days, were frozen in liquid nitrogen until they were used for RNA extraction.

Measurement of total carotenoids, chlorophyll (Chl) a and Chl b
Samples (0.05 g) from the three different temperatures were weighed, ground into powder, and placed into 5 mL of an 80% acetone solution.

RNA extraction
Total RNA from all samples was extracted using an E.Z.N.A.® Plant RNA Kit Omega Bio-tek, USA), and an aliquot of total RNA was treated with DNase I (Takara, China) to remove DNA. RNA integrity was measured by 1% agarose gel electrophoresis, and RNA purity was detected with a NanoDrop spectrophotometer (Thermo Fisher, USA).

Illumina HiSeq library preparation and sequencing
These RNA samples were reverse transcribed into cDNA using a SMARTer™ PCR cDNA Synthesis Kit (Takara, China), and cDNA libraries were created. After the quality of the cDNA libraries was assessed with an Agilent Bioanalyzer 2100 system (Agilent Technologies, USA), sequencing was conducted using an Illumina HiSeq X Ten sequencer by Shanghai OE Biotech. Co., Ltd. The raw data were uploaded to the NCBI Sequence Read Archive (SRA, https://trace.ncbi.nlm.nih.gov/Traces/sra/ sra.cgi). The data and scripts used for the analysis is available with accession number SRP157932.

Data analysis and annotation
Raw reads were quality filtered using the Trimmomatic 0.36 [63], and the process consisted of four stages: removal of the adaptor; removal of low-quality reads for which the number of N bases exceeded 10% of the total read length or for which the number of error-prone bases (quality score ≤ 5) exceeded 50% of the total read length; removal of low-quality bases from the 3′ end and the 5′ end in different ways; and statistical analysis of the raw reads and clean reads. The data were assessed for contamination before subsequent analysis; 250,000 pairs of reads (500,000 reads) from the data were extracted randomly, and then the data were aligned using BLAST (E value < 10 − 10 , coverage > 80%) with sequences in the National Center for Biotechnology Information (NCBI) non-redundant nucleotide sequence (Nt) database (ftp://ftp.ncbi.nih.gov/ blast/db). The best dataset was selected.

Differential expression analysis
Analysis of the differential expression of unigenes among different samples was conducted using DESeq [66], a method based on the negative binomial distribution, in the R statistical environment [67]. The number of unigenes in each sample was normalized using the base-Mean value to estimate the expression, the fold change was calculated, and the significance of the difference in the number of reads was tested with an NB test (negative binomial distribution test). Finally, we screened for the differential expression of unigenes based on the fold changes and the results of the significance test.

Phylogenetic analysis of carotenoid biosynthesis genes
The expression profiles of genes involved in terpenoid biosynthesis were analysed using KEGG pathway annotation. A total of 56 expressed unigenes encoding terpenoid biosynthesis enzymes were found in U. prolifera. But most of the genes encoding key enzymes in the MVA pathway were not found, except for AACT and HMGR. In addition, a total of 22 expressed unigenes encoding carotenoid biosynthesis enzymes were found, and six genes encoding key enzymes in carotenoid biosynthesis were identified in U. prolifera. Functional genes that participate in terpenoid biosynthesis in Chlamydomonas reinhardtii and Chlorella variabilis were selected from NCBI and aligned with the sequences of U. prolifera. The phylogenetic analysis was carried out in MEGA 5.1 software [68] with the neighbour-joining (NJ) analysis option based on the amino acid sequences.

Expressional validation of carotenoid biosynthesis genes with qRT-PCR
The expression levels of the seven unigenes that participate in the MEP pathway were examined by qRT-PCR. The primers for qRT-PCR are shown in Table 2. The expression levels were measured with a Baiyuan ASA-4800 Real Time PCR System using SYBR Green fluorescent dye (Takara, China) according to the manufacturer's instructions. The cycling profile included a step at 95°C for 30 s, followed by 40 cycles of amplification (95°C for 5 s and 60°C for 34 s). The relative gene expression was calculated using the 2 -ΔΔCt relative quantitative method. All MEP pathway genes were chosen for analysis of expression patterns in different samples that were subjected to the three different temperatures (12°C, 20°C, and 28°C). The expression levels of these selected genes from qRT-PCR analyses were assessed in comparison to the differentially expressed genes (DEGs) from RNA-Seq.