Transcriptomics reveal how circadian regulation contributes to starch hyperaccumulation in marine alga Tetraselmis helgolandica

Tetraselmis helgolandica var. Tsingtaoensis is a marine microalga. It can produce a large amount of starch, especially amylose, with addition of carbon source and specific circadian rhythm. The mechanism behind this phenomenon is still unclear. Analysis of this mechanism can help to develop T. helgolandica into a new green bioengineering chassis organism. We explained how circadian rhythm and glucose affect the rate of starch accumulation and starch structure in T. helgolandica based on the transcriptome. The glucose inhibited the photosynthetic system of T. helgolandica, while the circadian rhythm can alleviate the inhibition. Circadian rhythm induced the upregulation of Embden–Meyerhof–Parnas pathway and pentose phosphate pathway (PPP) in T. helgolandica, but had little effect on the tricarboxylic acid cycle. PPP pathway provides Ribulose‐1,5‐bisphosphate, which may be beneficial for dark reactions and nucleotide synthesis. And PPP pathway provides Nicotinamide adenine dinucleotide phosphate, which facilitates energy substance synthesis. This will further upregulate the starch metabolic pathway. The transcript level of the key gene ADP‐Glucose pyrophosphorylase is mainly regulated by glucose. The granule‐bound starch synthase (gbss), a key gene for amylose synthesis, is mainly influenced by circadian rhythm. In general, the increase of starch synthesis and amylose ratio requires both glucose addition and circadian rhythm. We report the first referenced transcriptome of T. helgolandica. Differences between transcripts reveal how circadian rhythm and glucose addition affected the rate of starch synthesis and structural variation. It provides a reference for an in‐depth study of starch synthesis in green algae.


| INTRODUCTION
Microalgae, with thousands of species and widely distributed in various ecosystems, have the most efficient photosynthetic system in the nature, consuming CO 2 10 times more efficiently than higher plants (Singh & Ahluwalia, 2013).Under the current background of carbon neutrality in the world, the application of microalgae has attracted much attention and has carried out extensive applications in the fields of biofuel production, high value molecule production, food, wastewater treatment, and CO 2 absorption (Chen et al., 2022;Levasseur et al., 2020).Microalgae are efficient starch producers and can accumulate more than 50% dry weight of starch with nutrient stress, carbon source addition, or other methods (Morales-Sánchez et al., 2017;Yao et al., 2018).Microalgae starch is mainly amylopectin with a structure similar to plant starch and is considered as an excellent alternative to it.
The starch synthesis pathway has been adequately studied in land plants and model algae.In Arabidopsis thaliana, 30%-50% assimilates will enter starch synthesis pathway.The core starch synthesis pathway in plant: The precursor ADP-Glucose can be catalyzed to α-1,4-glucan, and to amylopectin by starch synthase (SS), and to amylose by granule-bound starch synthase (GBSS; Smith & Zeeman, 2020;Stitt & Zeeman, 2012;Streb et al., 2008).The key enzymes for starch synthesis are mostly regulated at transcriptional level, such as ADP-Glucose pyrophosphorylase (AGPase).AGPase catalyzes the formation of starch synthesis precursor ADP-Glucose from 1-phosphate glucose (Glc-1-P) and ATP.In Zea mays (Li et al., 2011), Triticum aestivum (Kang et al., 2013), and Oryza sativa (Oiestad et al., 2016), the overexpression of AGPase promoted biomass accumulation and increased starch content.And the expression of AGPase is affected by environment.For example, the AGPase transcription level and starch content decreased with the addition of nitrate in Dioscorea esculenta roots (Kim et al., 2002); the AGPase transcription level of Musa nana was significantly changed under the temperature, salt, and osmotic stress (Miao et al., 2017).
Other key enzymes are also similar to AGPase, such as overexpression of SBEII in Solanum tuberosum resulted in higher amylopectin branching degree, and the post-transcriptional silencing of gbss reduced the amylose content (Brummell et al., 2015).In addition, GBSS is also regulated by circadian rhythm and sugar, sucrose in Dioscorea esculenta stimulates GBSS mRNA accumulation (Wang et al., 2001).It has been confirmed that circadian rhythm has a significant effect on the expression of gbss in Chlamydomonas reinhardtii, Antirrhinum majus, and Oryza sativa, and affects starch content and structure (Dian et al., 2003;Mérida et al., 1999;Ral et al., 2006).
Omics analyses such as genomics, transcriptomics, proteomics, and metabolomics have gradually become common tools for exploring starch metabolic pathways and the roles of enzymes at different levels.For instance, Liu et al. (2016) identified candidate genes regulating starch content in maize kernels by genomics.Koo et al. (2017) fingered out the starch accumulation mechanism in C. reinhardtii high-yield starch mutant by transcriptomics, and found that the increased expression of pgm (phosphoglucomutase gene) and the accumulation of glucose-6-phosphate may lead to the large starch accumulation in mutant.Willamme et al. (2015) analyze the effects of circadian rhythms on starch accumulation and chlorophyll accumulation in C. reinhardtii using transcriptomics and metabolomics.Jia et al. (2013) developed transcriptome and proteome analyses to search genetic changes when mutant opaque2 maize produced high crystallinity starch.Zhen et al. (2017) used phosphoproteomics to study the phosphorylated proteins in wheat grains under high nitrogen environment, and found lots of enzymes associating with starch synthesis phosphorylated, including starch branching enzyme (SBE), AGPase, glucose phosphate isomerase, phosphoglucomutase (PGM).Phosphorylation increased their activity and promoted starch accumulation.In general, omics analysis provides a better platform to study starch metabolic mechanism.
T. helgolandica is a unique species of Chlorophyta that grows in the coastal waters of China.In our previous study, we have revealed a special phenomenon that carbon addition and regulating circadian rhythm could promote starch synthesis, especially amylose, in T. helgolandica (Shi et al., 2022).The special starch structure has aroused our interest.Amylose is also ideal for the manufacture of bioplastics (Ren et al., 2021).We have preliminarily glucose addition affected the rate of starch synthesis and structural variation.It provides a reference for an in-depth study of starch synthesis in green algae.

K E Y W O R D S
circadian rhythm, gene expression, microalgae, starch, synthesis, transcriptome revealed the mechanism by single gene transcription analysis (Ren et al., 2021).In this study, we report the first referenced transcriptome of T. helgolandica and reveal how circadian rhythm and glucose addition affect the rate of starch synthesis and structural variation according to the transcriptional differences.It provides a reference for an in-depth study of starch synthesis in green algae.

| Strain and culture conditions
T. helgolandica var.Tsingtaoensis HL-1 (were called Platymonas helgolandica once), a marine green microalga, was isolated from the Donghai Sea near Yancheng, Jiangsu Province, P.R. China, streaking on the plate repeatedly to purify to sterility in the laboratory.The culture conditions was described in our previous study (Shi et al., 2022).

| Biomass and starch assay
The starch concentration and amylose/amylopectin ratio (Am/Ap) was measured by the methods described by Hovenkamp-Hermelink et al. (1988).The dry weight was measured by the same way described in our previous study (Shi et al., 2022).The changes of starch during 1 day to night were carried out in a total of four groups of ±glucose and ±circadian.Samples were taken at 0, 2, 4, 6, 12, 15, 18, 21, and 24 h after the "sunrise" for analysis.

| RNA isolation, library preparation, and sequencing
Totally six groups of samples were used for transcriptome analysis, including autotrophy cultured group (L:D = 24:0 h, −Glc), mixotrophy cultured group (L:D = 24:0 h, +Glc), and Circadian cultured group (L:D = 6:18 h, +Glc), on the 4th and 8th day separately.After 4 and 8 days of cultivation, cells were centrifuged at 8000 rpm, 5 min (4°C) into Eppendorf tubes for RNA extraction.Each sample group includes two biological replicates.All samples were collected at 1 h after shifting from dark to light.The total RNA was isolated using the TRIzol method.The extracted samples were tested for concentration and purity by Nanodrop, and integrity by Caliper GX.Use NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, USA) for library construction.The library fragments were purified using the AMPure XP system (Beckman Coulter, Beverly, USA).Evaluate the quality of the library on the Agilent Bioanalyzer 2100 system, and sequence it using the Illumina platform.

| Transcriptome mapping, annotation, and differential expression analysis
HISAT2 (version 2.0.4) was used for efficient comparison of transcriptome results and reference genomes.Assemble the reads on the comparison pair using StringTie, and after the comparison analysis is completed, assemble and quantify the reads on the comparison pair using StringTie.Normalize the number of mapped reads and transcript length in the sample.Using StringTie and the maximum flow algorithm, FPKM (Fragments Per Kilobase of script per Million fragments mapped) is used as an indicator to measure transcript or gene expression levels.DESeq2 was used for differential expression analysis between sample groups to obtain a set of differentially expressed genes between two biological conditions.In the process of differential expression gene (DEG) detection, genes with an intergroup expression level of Fold Change ≥ 2 and |log2FC| ≥ 1 are defined as DEG.The functional annotation of DEGs was completed based on the annotation results of the reference genome.Refer to the GO database for DEG annotation classification and enrichment analysis.Refer to the KEGG database to query the integrated metabolic pathway.Classify, enrich, and visualize the KEGG annotation results of DEG, and apply hypergeometric tests to identify significantly enriched pathways in DEG compared to the entire genome background.

| Data availability
The T. helgolandica transcriptome data are available through the NCBI under project number PRJNA970912.The T. helgolandica genome data are available through NCBI under project number PRJNA971153 (Unreleased).The genome datasets used during the current study are available from the corresponding author on reasonable request.

| Statistical analysis
All the presented data are average values of three biological replications.Error bars indicate the standard deviation.Statistical analysis was performed using SPSS 26.0 for Windows (SPSS Inc., USA), and a value of p < 0.05 was considered statistically significant.

| Synergistic promotion of starch synthesis in T. helgolandica by circadian rhythm and glucose
We identified that glucose addition and circadian rhythm could effectively promote the growth of T. helgolandica (Figure 1A).Under 10 g/L glucose (Glc) and L:D (L:D) = 6:18 (h), T. helgolandica could accumulate the highest starch density (3.88 g/L) and the highest proportion of amylose (Am/Ap = 1.87; Figure 1B).In previous studies, we also detected changes in cellular physiological parameters and the ratio of amylose/amylopectin at different time points (Shi et al., 2022).The intracellular starch content of T. helgolandica was then monitored with time for 1 day (Figure 1C), and the trend of starch content was similar between the two groups under continuous light (L:D = 24:0 h, ±Glc).The mixotrophy group (L:D = 24:0 h, +Glc) consistently have a higher starch content than the autotrophy group (L:D = 24:0 h, −Glc).A classical trend of linear increase in starch content by day and linear decrease by night was observed in the autotrophy group (L:D = 24:0 h, −Glc), similar to the phenomenon observed in A. thaliana (Feugier & Satake, 2014).However, this group is continuously illuminated and there is no darkness, so degradation occurs in the subjective darkness.This indicates that the molecular clock in algal cells continues to work regardless of the presence of objective circadian variation.The starch content of the circadian group fluctuated widely.The circadian group without glucose (L:D = 6:18 h, −Glc) accumulated starch rapidly from 2 to 6 h, reaching a maximum at 6 h, then gradually decreasing to a low point at 15 h, and then stabilizing at about 10% DW accumulation.The accumulation did not significantly differ from that of the autotrophy group.Starch accumulation in the circadian group with glucose (L:D = 6:18 h, +Glc) occurred continuously from 2 to 15 h, reached a maximum at 15 h, and then dropped rapidly, and then maintained at about 55% DW accumulation, which increased compared to 0 h.The circadian rhythm caused the starch content in T. helgolandica to fluctuate with the light-dark cycle, but after the addition of glucose, the starch synthesis pathway was not closed at night and starch synthesis continued.The amount of starch accumulated in the last hours of the day was less variable and increased compared to the beginning hours, similar to that of C. reinhardtii (Vitova et al., 2011).

| Transcriptome sequencing and gene enrichment
Six groups of samples were prepared for transcriptome sequencing: on the 4th day of culture (mid-logarithmic growth phase), the autotrophic group T4A (L:D = 24:0 h, − Glc), the mixotrophic group T4M (L:D = 24:0 h, +Glc), and the circadian group T4C (L:D = 6:18, +Glc) were taken; on the 8th day of culture (late logarithmic growth phase), the autotrophic group T8A, the mixotrophic group T8M, and the circadian group T8C were taken.Each group included two biological replicates.
Transcriptome sequencing of all simples obtained 82.98 Gb clean data, the Q30 bases percentage was 93.85% and above, and the GC bases was about 59%.The clean reads of each sample were compared with the reference genome, with 91.31%-96.40%comparing efficiency.Mapped reads had 81.11% alignment to exons and 14.01% alignment to introns on the reference genome, which supposed to be the intron retention by mRNA precursor and variable splicing.We identified and sequenced 22,531 genes in total, and functionally enriched these genes by GO and KEGG databases.There were 12,125 genes enriched to multiple categories by searching and comparing in GO database (Figure 2a).The most genes were enriched to the BPcellular process and BP-metabolic process, each with 5303 and 4948 in total, which proved up to 43.73% and 40.81% genes were involved in cell growth, maintenance, and metabolism.It showed that the T. helgolandica grew vigorously and metabolized actively.In addition, other 1374 genes were enriched to BP-response to stimulus, which declared that T. helgolandica need mass genes to grow in such complex environment.The membrane is the main gene functional structure, with 4012 genes, other 3178 and 1671 genes were enriched to organelle and organelle part, including nucleus, plasmid, plastid, etc.The main molecular functions include catalytic activity, binding, and transporter activity, with 6110, 5603, and 782 genes, respectively.In general, genes in T. helgolandica exhibit many functions, including growth, metabolism, environmental stimuli, and immune responses.However, there are almost no genes enriched in nutrient storage, translation regulation, and metal ion-binding activity.

| Light-dark cycle and glucose addition lead to differentially expressed genes
To investigate the effects of the light-dark cycle and glucose addition on the transcriptional level of flatworms, we counted the number of differentially expressed genes (Table 1).The group with the most differentially expressed genes was T8A versus T8C, with 3999 DEGs, of which 58% (2333 genes) were upregulated.
On Day 4, there were the most endemic DEGs (1027 genes) in T4A versus T4M and the least in T4M versus T4C.On Day 8, there were the most endemic DEGs with 1286 in T8A versus T8C and the least in T8A versus T8M.It can be speculated that the effect of glucose addition has an effect mainly in the early culture period, while the circadian rhythm has an effect in the middle and late culture period.There were also differences in DEGs between the same groups at days 4 and 8.The effect of glucose on T. helgolandica diminished in the later culture period, producing fewer DEGs, while the effect of circadian rhythm was more pronounced in the later period.We analyzed the effects of glucose and light-dark cycle on various aspects of metabolism in T. helgolandica based on DEGs.

| The light-dark cycle facilitates synthesis of the light harvesting complex to influence photosynthesis
Photosynthesis is one of the most important physiological functions of green algae.The photosynthetic system of green algae mainly includes the parts of light-harvesting protein complex, PSI (photosystem I), PSII (photosystem II), and cytochrome complex.Photosynthesis fixes CO 2 and degrades water through two steps: light reaction and dark reaction.Photosynthesis is an important step for green algal cells to obtain energy and synthesize precursors of organic compounds, and the efficiency of the light system often reflects the living state of green algae cells.Light-harvesting protein complexes (LHCs) are membrane protein complexes in plants and green algae that are mainly responsible for the capture and transmission of light.Sugar signal has been confirmed to be an important physiological signal in plants such as A. thaliana, Nicotiana tabacum, and Zea mays.It mainly reduces photosynthesis in high sugar environments, regulates phytohormone production, and reduces Rubisco concentration (Bonardi et al., 2005;Dodd et al., 2014;Farré & Weise, 2012;Mulo et al., 2012).In T4A versus T4M, we found that genes related to the photosynthetic system were downregulated with the addition of glucose (Table 2), including light harvesting complex genes lhca1, lhcb1, and lhcb2; the PSII subunit synthesis gene psbP; ferredoxin genes petF, petH, and petJ.This is similar to A. thaliana and Chromochloris zofingiensis (Yang et al., 2000;Zhen et al., 2017).In T4M versus T4C, enrichment results of different databases showed that DEGs between two groups were mainly related to the photosynthetic system.This reflects that glucose inhibition of photosynthetic system was relieved by circadian rhythm.And some light-harvesting complex genes, including lhca4, lhca5, lhcb1, lhcb2, lhcb4, and lhcb5, were upregulated.In T4A versus T4C, LHC subunits mentioned above also were upregulated, which suggesting that circadian rhythm leads to an increase of LHC synthesis.Since in the circadian group, cells have to complete light-demanding activities in shorter periods, the increase in the synthesis of light-harvesting antennas not only facilitates the rapid collection of light energy but also can be used for repairing light harvesting systems at night.And we observed an upregulation of the genes psbP, psbQ, psbY, psb27, psb28, petC, petN, petE, petF, and petH.It is worth noting that only psbP, psbQ, psbY, and petC are upregulated in T4A versus T4C, and petF even shows downregulation (Figure 3; Table 2).We infer that the upregulation of photosynthetic system-related genes in T4M versus T4C is partly due to the downregulation in T4M, while promotion of photosynthetic systems in T4C is limited.It can be assumed that the circadian rhythm counteracts part of the glucose inhibition.Compared to T4A, the photosystem efficiency of T4C still decreased, and the decrease may be related to the long darkness (Shi et al., 2022).And the downregulation of petF in T4C compared to T4A and upregulation compared to T4M can explain the lower performance index based on absorption (PI abs ) in T4C compared to TA4 and higher compared to T4M (Shi et al., 2022).
We identified that the circadian rhythm facilitates the photosynthetic system of T. helgolandica in multiple ways.
In previous studies, light-harvesting systems in marine algae receive circadian rhythm regulation.In plants, it was found that the photosynthetic system electron transfer efficiency and PSII quantum yield are also regulated by light-dark cycle.In general, circadian rhythm is an important factor affecting photosynthetic system (Dodd et al., 2014).Bonardi et al. (2005) indicated that photosynthetic system can adapt to light changes through the regulation of the proteins of PsbA, PsbD, and PsbC by two kinases, STN7 and STN8.The circadian rhythm was found to have an effect on the transcriptional level of STN7 and it may play a major role in circadian rhythm regulation of photosynthetic system efficiency (Dodd et al., 2014).Only STN8 was annotated in the transcriptome of T. helgolandica.We found that the gene was upregulated in T8C compared to T8A and T8M.STN8 may also have a vital role in the regulation of photosynthesis by the circadian rhythm.

| Effects of glucose addition and circadian rhythm on carbon metabolism
The effect on carbon metabolism was greater due to the addition of glucose.By comparing the circadian group with the parthenogenic and autotrophic groups, it can be found that circadian rhythm affected carbon metabolism more significantly (Figure 4; Table 3).
The Embden-Meyerhof-Parnas (EMP) pathway lies at the core of carbon metabolism.We started with a detailed analysis of the EMP pathway.On days 4 and 8, EMP pathway was less altered in the mixotrophy group compared to the autotrophic group.On day 4, there were no significant changes in all three EMP pathway rate-limiting enzymes.On day 8, hexokinase (EC: 2.7.1.1)was downregulated, while phosphofructokinase and pyruvate kinase showed no significant changes.The downregulation of hexokinase may be associated with glucose inhibition of the photosynthetic system (Roth et al., 2019).Tricarboxylic acid cycle (TCA cycle) was also not significantly upregulated.There are two possible reasons: (i) It has been reported that the effect of glucose addition on the central carbon metabolism pathway appears very rapidly, with transcriptional upregulation occurring within 1-4 h.In our previous study, glucose was almost depleted on day 4, and the low concentration of remaining glucose may fail to stimulate EMP pathway.(ii) On day 4, upregulation of pentose phosphate pathway (PPP) pathway occurred.T. helgolandica are in the middle of logarithmic growth, and the exuberant cell division requires nucleic acid, so the extra glucose enters PPP pathway for ribose synthesis.In circadian groups, glucokinase (EC: 2.7.1.2) showed significant upregulation on 4 (T4A versus T4C, T4M versus T4C) and day 8 (T8A versus T8C); pyruvate kinase (EC: 2.7.1.40)was also upregulated in T8A versus T8C on day 8. Overall, circadian rhythm and glucose addition resulted in the upregulation of the EMP pathway in T. helgolandica.Transcripts of TCA cycle-related enzymes were less variable.Pyruvate carboxylase (EC: 6.4.1.1) of both T4C and T8C has a significant upregulation compared to control groups.Pyruvate carboxylase catalyzes the production of oxaloacetate from pyruvate, which is an important complementary reaction to oxaloacetate.Upregulation of pyruvate carboxylase may allow rapid conversion of pyruvate to oxaloacetate, and the increase of oxaloacetate in circadian groups in the metabolite assay results confirmed this hypothesis.In circadian groups, TCA cycle rate-limiting enzymes were not significantly changed at day 4 compared to control groups.In T8C, citrate synthase (EC: 2.3.3.1) was upregulated compared to T8A and downregulated compared to T8M.In general, the circadian rhythm has a smaller effect on TCA cycle, probably because: (i) The genes related to TCA cycle are constitutively expressed genes, so the transcripts fluctuated slightly; (ii) the acetyl-CoA may enter the fatty acid synthesis pathway, resulting in no significant upregulation of TCA cycle.Combined with the fatty acid metabolism-related pathways in the transcriptome, several enzyme-related fatty acid synthesis pathways in the circadian groups showed significant upregulation, such as acetyl-CoA carboxylase (EC: 6.4.1.2;T4A vs. T4C, T8A vs. T8C, T8M vs. T8C), fatty acid synthase (EC: 2.3.1.85;T4A vs. T4C, T4M vs. T4C, T8A vs. T8C), ACP-synthase (EC: 2.3.1.179;T4M vs. T4C, T8A vs. T8C, T8M vs. T8C), ACP-reductase (EC: 1.1.1.100;T4A vs. T4C, T4M vs. T4C, T8A vs. T8C, T8M vs. T8C), and several other enzymes.Acetyl-CoA is the initiator of fatty acid synthesis, and upregulation of the fatty acid synthesis pathway requires more acetyl-CoA as the precursor.Therefore, it can be assumed that acetyl-CoA may tend to enter the fatty acid synthesis pathway, and there is a deficiency of acetyl CoA in TCA cycle.That resulting in no upregulation of TCA cycle.Since TCA cycle produces precursors for the synthesis of many substances, including many amino acids, the above changes limit the production of amino acid precursors, resulting in a decrease of intracellular soluble protein.
The PPP pathway is another pathway of glucose metabolism.It mainly produces nicotinamide adenine dinucleotide phosphate (NADPH), ribose 5-phosphate, and erythrose 4-phosphate, which can be used to synthesize fatty acids, nucleotides, aromatic amino acids and participate in photosynthetic dark reactions.The rate-limiting enzyme of the pentose phosphate pathway, glucose 6-phosphate dehydrogenase (G6PD), was significantly upregulated on days 4 and 8 compared to the autotrophic group, but did not change significantly compared to the parthenogenic group.This indicates that glucose addition led to the change of G6PD.The effect of circadian rhythm on PPP pathway was mainly in the upregulation of enzymes related to the synthesis of ribose 5-phosphate and fructose 6-phosphate.Upregulation of ribose 5-phosphate synthesis can affect nucleotide synthesis and the dark reaction.It has been demonstrated that glucose metabolism in microalgae mainly through the EMP pathway in light, while in darkness, it mainly through the PPP pathway (Li et al., 2020;Yang et al., 2000).The upregulation of the PPP pathway in the cultivation environment with dark periods compared to pure light cultivation confirms the above conclusion to some extent.

| Circadian rhythm promotes starch accumulation
We found increased starch accumulation in T8M compared to T8A, but lower than T8C, and without increased amylose starch accumulation.In T8A versus T8M, it showed that DEGs were significantly enriched in starch and sucrose metabolism pathways.The key gene of starch synthesis, AGPase (EC: 2.7.7.27), was upregulated, while starch synthase (ss, EC: 2.4.1.21)and gbss (EC: 2.4.1.242)showed no significant changes.This indicates that the upregulation of the starch synthesis pathway is limited.Meanwhile, glycogen phosphorylase (EC: 2.4.1.1)was significantly upregulated.This indicates that starch synthesis and degradation pathways were upregulated at the same time with glucose addition.This is the reason why T8M has less starch accumulation compared to T8C (Figure 1).T. helgolandica accumulate large amounts of starch, which is mainly amylose under specific circadian rhythm.The transcriptome results on day 4 showed that the key enzyme genes of starch metabolism were not significantly changed in T4C compared to T4A.The key genes for starch synthesis, AGPase and ss, were upregulated; the key genes for starch degradation, βAMY, were upregulated, while αAMY was downregulated in T4C compared to T4M.This indicated that the transcripts of starch metabolism-related enzymes were less changed in T4C.We speculate that the starch degradation pathway is also upregulated in the mixotrophy group, and starch acted only as a transient energy storer.The circadian group acquired more energy due to higher photosynthetic efficiency and respiration rate (Shi et al., 2022), and thus started starch storage in the long term.
On the 8th day, starch metabolism in T8C was upregulated compared to T8A.AGPase, ss, gbss, and gbe (EC: 2.4.1.18)in the starch synthesis pathway were significantly upregulated, and αAMY and βAMY in starch degradation were upregulated.In T8C versus T8M, gbss and gbe were upregulated; AGPase was not significantly changed.The changes in multiple transcripts of SS were different, and αAMY and βAMY in the degradation pathway were similar to SS (Figure 5; Table 4).Scheibe (1991) proposed that the regulation of AGPase by light is mainly achieved through allosteric interaction.AGPase was upregulated in T8M and T8C compared to T8A, while AGPase expression levels were similar in T8M and T8C.This indicates that the addition of glucose may affect AGPase at the transcription level, while the addition of circadian rhythm has a relatively small impact on AGPase.GBSS is a key enzyme for amylose synthesis.According to the transcriptome results, gbss in T. helgolandica was insensitive to glucose.However, gbss was significantly upregulated after 8 days of cultivation under circadian rhythm, resulting in the accumulation of amylose.With circadian rhythm but without the glucose addition, T. helgolandica grew slowly, with no increase in starch accumulation and no significant  increase in proportion of (Figure 1).Based on this phenomenon, it can be inferred that the simultaneous addition of glucose and the application of circadian rhythm could lead to a large accumulation of starch and an increase in the proportion of amylose.Key enzyme expression genes, such as GBSS, continue to be highly expressed under the stimulation of both sugar and circadian signals, significantly increasing the rate  of starch (Table 4; Figure 5).The above results also validate the conclusion that gbss is regulated by circadian signals (Morell et al., 1997).It used to be thought that in microalgae and plants, starch plays an important role as an energy buffer in the day-night transition.In nature, starch accumulates during the day degrades at night to maintain the physiological activities.The shorter the daytime, the faster starch is synthesized.However, it was found that the starch synthesis pathway is not closed at night with sucrose addition, leading to a continuous starch accumulation in plants (Lunn et al., 2006).We propose that shorter day led to upregulation of the starch synthesis pathway in response to long nights, while glucose leads to the nonclosure of the starch synthesis pathway at night.T. helgolandica continued to synthesize starch at a high rate.The increase in amylose synthesis is mainly associated with the upregulation of key enzymes.

| Molecule clock of circadian rhythm
Circadian rhythm in microalgae is influenced by a combination of genes.These genes together form a transcriptional feedback loop ("molecular clock").Multiple related physiological processes are initiated at specific times in the day-night cycle.The molecular clocks of the circadian rhythms of microalgae have been discovered mainly from the model organism C. reinhardtii, revealing multiple clock genes and their relationships (Ryo et al., 2016).
The molecular clock model of the circadian rhythms that has been validated in C. reinhardtii is shown in Figure 6.This model is now widely accepted and similar models exist in both green algae and A. thaliana.Timing of cab expression 1 (toc1), Late Elongated Hypocotoyl (lhy) and Circadian Clock-Associated 1 (cca1) are conserved components of the molecular clock found in green algae and plants.LUX (LUX Arrhythmo), ELF3 (Early Flowering 3), and ELF4 (Early Flowering 4) form the evening complex (EC).Toc1 belongs to the family of pseudoresponse regulators (PRRs, members of which are all associated with circadian rhythms) and is mainly expressed at the beginning of the night, while lhy/cca1 belongs to the family of MYB transcription factors and is mainly expressed at the beginning of the day.These three genes are at the core of the molecular clock transcriptional feedback loop and repress each other to keep the molecular running.The molecular clock also several important genes, such as GIGANTEA (gi) and Clock-associated PAS protein ZTL (ztl), which constitute transcriptional regulation of toc1 and control nighttime activity.prr7/prr9 and lhy/cca1 control daytime activity.
In T. helgolandica, several homologous genes of the above molecular clock key genes that were annotated, such as lhy, cry (Cryptochrome 1), cop1, ztl, and toc1 (Table 5).Some other important genes such as cca1 and gi were not annotated.On the one hand, it may be due to incomplete annotation, and on the other hand, the possibility of a special molecular clock structure in T. helgolandica cannot be excluded.The study about Euglena (Carre & Edmunds, 1993) also revealed its unique rhythmic regulation mechanism, which is controlled by oscillation conditions at cAMP levels.The circadian rhythm regulation mechanism that is not completely similar to C. reinhardtii has also been found in Ostreococcus tauri (Corellou et al., 2009).This proves that the molecular clock composition in different microalgae is not completely conservative.de Winter et al. ( 2014) indicated that the circadian clock was not affected by N-limitation, and cell division was timed during the natural night under constant light conditions.
Cop1 was downregulated in T4C compared to T4M.Cop1 plays an important role in light regulator gene in green algae and plants (Tilbrook et al., 2016;Zhang et al., 2023).The downregulation of cop1 relieved its inhibition of growth-related genes, which is beneficial for cell growth and reproduction, and this result is consistent with the expressed traits.Since we did not obtain transcriptomes in different circadian periods, it is difficult to assume whether cop1 has the same functions and trends as other species based on the current results.In the future, cop1 inhibitors can be added to further investigate whether cop1 plays a key role in circadian rhythm of T. helgolandica to promote growth and starch accumulation.

| CONCLUSION
Up to now, there is still a gap of the research about T. helgolandica.The genome of T. helgolandica is still unreported.And very few research have reported the transcriptome of T. helgolandica.Chu et al. (2023) reported the transcriptome and indicate that circadian rhythms altered the growth and wastewater treatment efficiency of T. helgolandica.We explain the mechanisms by which circadian rhythm and glucose affect the rate of starch accumulation and starch structure in T. helgolandica based on the transcriptome.The glucose inhibited the photosynthetic of T. helgolandica, while light-dark cycle can alleviate the inhibition.Circadian rhythm induced upregulation of EMP pathway and PPP pathway in T. helgolandica, but had less effect on TCA cycle.PPP pathway provides Ribulose-1,5bisphosphate, which may beneficial for dark reactions and nucleotide synthesis.And it also provides NADPH, which facilitates energy substance synthesis.This will further upregulate the starch metabolic pathway.The transcript level of the key gene AGPase is mainly regulated by glucose.The gbss, a key gene for amylose synthesis, is mainly influenced by circadian rhythm.
In general, the increase in starch synthesis and amylose ratio requires both glucose addition and circadian rhythm.Due to the imperfect genome annotation, some questions remain to be fully elucidated.In the future, we need to promote genomic progress to better analyze the causes of physiological changes in microalgae at the genetic level.It can also be supplemented with emerging genomic tools such as metabolomics to reveal the mechanisms of physiological changes in microalgae more comprehensively.

F
I G U R E 1 Effects of glucose and circadian rhythm on the growth and starch accumulation in Tetraselmis helgolandica.(A) Dry weight accumulation at Day 12 under different circadian rhythm with/without glucose; (B) different types starch accumulation under different circadian rhythm with glucose.Different English letters represent the significant differences in total starch concentration; (C) starch proportion change under L:D = 6:18 h with/without glucose, L:D = 24:0 h with/without glucose.
Transcript enrichment results: (a) enrichment results of transcripts in GO database; (b) enrichment results of transcripts in KEGG database.
Differential expression of key enzymes in central carbon metabolism.

F
I G U R E 6 Putative pathway of molecule clock in green algae.The red-filled boxes indicate genes upregulated at different time.The blue-filled boxes represent genes downregulation at different time.The black solid line border in the picture represents the annotated genes in Tetraselmis helgolandica, while the black dashed line represents potential molecular clock-related genes in T. helgolandica.The blue background represents that the relevant genes can form complexes or perform similar functions (CCA1, Circadian Clock Associated; COP, Constitutive Photomorphogenic; CRY, Cryptochromes; ELF, Early Flowering; GDH, Glutamate Dehydrogenase; GI, Gigantea; GLN, Glutamine synthetase; HY5, Elongated Hypocotyl 5; LHY, Late Elongated Hypocotyl; PRR, Pseudoresponse Regulator; TOC1, Timing of Chlorophyll A/B-Binding Protein 1).

T A B L E 5
Differential expression of key enzymes in circadian clock.
Differential expression of key proteins in photosynthesis and light-harvesting antennae.
T A B L E 2Note: Bold markings indicate significant differences in expression.
Differential expression of key enzymes in starch metabolism.
T A B L E 4