Identification and characterization of a novel type-2 DGAT homologue gene from the green microalga Haematococcus pluvialis

Background: The unicellular green microalga Haematococcus pluvialis is an ideal source of astaxanthin (AST) which is stored at oil bodies contain both esterified astaxanthin (EAST) and triacylglycerol (TAG). Diacylglycerol acyltransferase (DGAT) catalyzes the last step of the acyl-CoA-dependent TAG biosynthesis and is considered as the crucial enzyme involving in EAST biosynthesis in H. pluvialis. However, the function of DGAT in H. pluvialis has not been reported. Results: A full-length cDNA sequences encoding a putative DGAT2 (HaeDGAT2E) was obtained from H. pluvialis . It contained an open reading frame (ORF) of 1,017-bp encoding a protein of 338 amino acid residues. The isolated HaeDGAT2E protein shared high identity of 57.6% and 54.1% with DGAT2E from Chlamydomonas reinhardtii and Chromochloris zofingiensis respectively. There were 7 conserved motifs and 3 trans-membrane regions in HaeDGAT2E. The phylogenetic analysis suggested that HaeDGAT2E belonged to DGAT2E subfamily. HaeDGAT2E activity was confirmed in the TAG deficient yeast strain (H1246) by restoring its ability to produce TAG. Upon expression of HaeDGAT2E , C16:0 and C18:1 fatty acid contents were 190.2% and 132.4% higher respectively than that of the H1246 strain. In addition, over-expression of HaeDGAT2E in transgenic Nicotiana Benthamiana resulted in increased contents of C16:0 (113.5%) and C18:1 (234.5%). Conclusions: A novel gene encoding HaeDGAT2E was identified in H. pluvialis . This is the first functional analysis of DGAT2 in Haematococcus . This information is important for understanding TAG accumulation and for further elucidating EAST biosynthesis in H. pluvialis .

into acyl-CoA-dependent and acyl-CoA independent pathways [6]. Diacylglycerol acyltransferase (DGAT) catalyzes the final acylation of sn-1,2-diacylglycerol (DAG) to form TAG, which is the last and limiting step in the acyl-CoA dependent TAG formation [7]. This enzyme represents a bottleneck in TAG biosynthesis in some oilseed crops and algal species, and thus has been regarded as a key target in manipulating oil production [7]. In higher plants and microalgae, there are three major groups of DGATs: (1) membrane bound forms of DGAT1 and DGAT2 which shares no sequence similarity; (2) soluble type of DGAT3 which is localized in the cytosol; and (3) dual functional of WS/DGAT which possess both wax ester and DGAT biosynthesis activities [8][9][10][11]. DGAT1 is considered to play a critical role in TAG accumulation in many higher plants and microalgae, whereas DGAT2 appears to have an important role in the formation of TAG containing unusual fatty acids. There is strong evidence support the involvement of DGAT3 and WS/DGAT in TAG biosynthesis in microalgae [12][13][14].
Interesting, only one copy of DGAT1 has been identified in a number of microalgae, whereas multiple copies of DGAT2 genes are typically present, suggesting that DGAT2 may play an important function in TAG biosynthesis and algal growth [8].
Haematococcus pluvialis is a green microalga widely known for its ability to synthesize the highest amount of astaxanthin (4% dry weight) under stress conditions [15][16][17]. Natural astaxanthin (AST) is a red-colored carotenoid with strong antioxidant ability and important commercial value [18][19][20]. This microalga also represents a potential source of TAG, since a considerable increase in TAG content accompanies the accumulation of AST [21][22][23][24]. Moreover, the previous studies have indicated that the main form of AST is esterified astaxanthin (EAST), which includes astaxanthin monoester and diester, and which is stored in TAG rich cytosolic oil bodies (OBs) [26]. Although the exact mechanisms of stress-induced TAG and AST accumulation in H. pluvialis are not well understood, several lines of evidence have suggested that the biosynthesis of both compounds appears to be linked through the regulation of oil biosynthetic enzymes [26]. Indeed, the accumulation of AST appears to be dependent on the accumulation of TAG. In addition, it has been speculated that certain DGAT is the candidate enzyme catalyzing the esterification of AST in H. pluvialis [26]. Although DGATs from different microalgae had been confirmed that have the ability to catalyze TAG biosynthesis [27][28][29], there are few studies on the cloning and functional identification of DGAT from the green alga H. pluvialis.
In the present study, homologous cloning coupled with the rapid amplification of complementary DNA ends (RACEs) was applied to clone the full-length cDNA sequences of HaeDGAT2E. The sequence analysis for DGAT from green algae and higher plants were finished, focusing on their phylogeny, evolution, and conserved domains. The function of HaeDGAT2E in TAG biosynthesis was confirmed in the TAG deficient yeast strain (H1246). The total lipids accumulation and fatty acid composition of yeast and tobacco were studied according to the over expression of HaeDGAT2E respectively. These results lay the fundament for elucidating TAG biosynthesis and provide evidence for illuminating EAST biosynthesis in future in H. pluvialis.
HaeDGAT2E from H. pluvialis was not. The first two amino acids (YF) were highly conserved among all DGAT2s examined, while the third residue was variable in microalgae. Interestingly, in the PHG block, the first two continuous residues Pro-His (PH) were obviously conserved among all DGAT2s examined, whereas the third residue was Gly (G) or Ser (S). It was worth mentioning that the Tyr-Ile-Phe (YIF) motif was conserved in this region and was replaced by Leu-Val-Met (LVM) in HaeDGAT2E. In the following PR block, HaeDGAT2E had a conserved PxxR motif as well as other DGAT2s examined.
Similar with PR block, the core GGxxE motif in the GGE block was highly conserved. Basing on this block, CzDGAT2D and CrDGAT2D proteins belonged to plant-type DGAT2. The RGFA block and VPFG block were also conserved among all DGAT2s. In the last block, the alignment result showed that NoDGAT2B was special because it had no G block.
In order to further investigate the evolutionary relationship of HaeDGAT2E, phylogenetic analysis was performed using proteins of DGATs (DGAT1, DGAT2, DGAT3, and WS/DGAT) from different higher plants and microalgae (Fig. 3). Four groups including DGAT1, DGAT2, DGAT3, and WS/DGAT were clustered in this tree. As expected, HaDGAT2E was separated into DGAT2 subgroup with other DGAT2 from algae, and it was clearly separated with those from fungal and higher plants. In addition, HaeDGAT2E also had a close evolutionary relationship with CrDGAT2B, CrDGAT2C, and CzDGAT2E, which implied that they have the same origin and function.

Recovery the TAG synthesis in quadruple mutant yeast strain H1246 with HaeDGAT2E
To verify the function of the putative HaeDGAT2E enzyme, the ORF encoding sequences was heterologously expressed in the quadruple mutant yeast strain S. cerevisiae H1246 (∆dga1∆lro1∆are1∆are2) that lacks the activity of TAG synthesis. This yeast mutant strain contains knockout mutations in four TAG biosynthesis-related genes (dga1, lro1, are1, and are2) and is unable to synthesize TAG. The mutant type (MT) yeast can formed TAG when at least one of these four genes was expressed. Furthermore, WT (INVSc1) and H-EV (H1246 harboring empty vector pYES2.0) yeast strains were used as positive and negative controls, respectively.
As shown in Fig. 4, HaeDGAT2E was able to restore yeast TAG biosynthesis. There was a prominent TAG spot on the TLC plate from WT and the H-HaeDGAT2E strains respectively, but no TAG spot was detected in both H-EV and MT strains (Fig. 4a). Nile red can specifically stain the lipid molecule in cells, resulting in an orange fluorescence that can be used to quantify the lipid accumulation. As shown in Fig. 4b, the fluorescence in the cells of H-EV and MT strains was difficult to observe and exhibited a shaded orange. However, the lipid droplets were easier to observe, and they were large, clear, and bright in the MT and H-HaeDGAT2E strains. These results suggested that expression of HaeDAGT2 in the quadruple mutant strain H1246 can recover its ability to form neutral lipids through interaction with yeast lipids biosynthesis pathway and confirm that HaeDGAT2E indeed encoded a functional protein with DGAT activity.

Analysis of total lipids and fatty acid composition in H-HaeDGAT2E yeast strain
The changes of total lipids content and fatty acid composition were studied in different yeast strains.
As shown in Fig. 5a, the total lipids content of MT strain still remained as low as that in H-EV strain, whereas the total lipids content in the yeast transformed with HaeDGAT2E significantly increased and was 404.5% higher than that of the H-EV or MT strain. However, the total lipids content of H-HaeDGAT2E strain is still lower (77.6%) than that of the WT yeast INVSc1.
To further test the change of fatty acid composition in different yeast strains, the TAG extracted from

Transient expression of HaeDGAT2E in Nicotiana benthamiana
To explore HaeDGAT2E as a tool to manipulate acyl-CoA pools and to engineer TAG in higher plants, HaeDGAT2E was over-expressed in the leaves of Nicotiana benthamiana by injecting the Agrobacterium (GV3101) strain harboring binary vector (pCAMBIA1303) with the targeted gene. RT-PCR results showed that the HaeDGAT2E transcript was expressed (6.7 fold) in transgenic lines (Fig.   6a). The total lipids content significantly increased and was 138.9% higher than that in WT (Fig. 6b).
Transgenic N. benthamiana lines did not show any visible difference on the total starch and protein contents from wild-type plants ( Fig. 6c and 6d). In addition, the overe-xpression of HaeDGAT2E resulted in increased C16:0 and C18:1 fatty acid content, which was consistent with results from yeast strain (Fig. 6e).

Discussion
H. pluvialis is not only able to produce a substantial amount of TAG but also the highest content of AST under stress conditions, which has great potential as an alternative source of industrial oil production [21-24]. In addition, the previous studies have implied that the main form of AST is esterified astaxanthin (EAST), which includes astaxanthin monoester and diester, and which is stored in TAG-rich cytosolic oil bodies (OBs) [25]. Although the exact mechanisms of stress-induced TAG and AST accumulation in H. pluvialis are largely unknown, the AST accumulation is found to be dependent by the TAG biosynthesis [26]. Therefore, elucidating the stress induced TAG and EAST biosynthetic pathway is crucial to the improvement in the production of both TAG and AST in H. pluvialis. DGAT catalyzes the terminal step in acyl-CoA-dependent TAG production, and the expression of four DGAT2 isoforms is found to be up-regulated in H. pluvialis under stress conditions [21], which indicates that DGAT2 is the possible candidate enzyme involvement in TAG and EAST accumulation [26]. Recently, DGATs from different oil-accumulating species have been widely reported [29][30][31][32][33]. To our knowledge, there is no report about the function of DGAT2 in H. pluvialis. In the current study, a novel DGAT2 cDNA was isolated from the green algae H. pluvialis and its function was characterized using a yeast system. In addition, the potential of HaeDGAT2E as a tool to manipulate acyl-CoA pools for improving TAG accumulation was also explored.
One DGAT2 isoform (HaeDGAT2E) was identified from H. pluvialis (Fig. 1a). According to the sequences analysis result, the HaeDGAT2E protein shared high identity with DGAT2E of C. zofingiensis and C. reinhardtii, which implied that the isolated gene belonged to DGAT2E subfamily and might have the same function. Trans-membrane (TM) represents the typical property of DGAT1 and DGAT2.
Generally, DGAT1 contains 8-10 predicted TMs, while DGAT2 contains 2 TMs [34,35]. HaeDGAT2E protein had 3 putative TMs which was consistent with previous study (Fig. 1b). The conserved motifs determine the potentially important functions, which has been previously identified in DGAT2 enzymes from plants, animals, fungi, and microalgae [34,35]. These typical conserved motifs were also present in HaeDGAT2E but with varying degrees of conservation (Fig. 2). The origin and evolution of multi-copies DGAT2 members is still an interesting and puzzling topic [8]. The HaeDGAT2E was found to be clustered into DGAT2 subfamily and constructed a monophyletic subgroup with CrDGAT2B, CrDGAT2C, and CzDGAT2E (Fig. 1), which was different from the classification of multiple sequences alignment. Together, these results indicate that this isolated DGAT2 might have important physiological functions in H. pluvialis.
Heterologous expression in S. cerevisiae H1246 strain, which lacks the activity of neutral lipid biosynthesis, can intuitively detect the function of target gene in TAG biosynthesis [28]. This method has been widely used in DGATs from higher plants, fungi, and microalgae [34][35][36][37][38][39][40]. In this study, we also introduced HaeDGAT2E gene into this mutant type H1246 strain. Nile red staining and TLC analysis results indicated that HaeDGAT2E encode a protein with DGAT activity (Fig. 4). Recently, the function of different members of DGAT2 remains unclear since the seed oil content of type DGAT2 gene mutant has no significant decrease compared with that of wild type [36]. Moreover, CrDGAT2D failed to accumulate TAG in H1246 yeast [36,37]. The CzDGAT2D, which was highly close to CrDGAT2D, produced a trace amount of TAG in H1246 yeast [38]. All these previous studies have indicated that different members of DGAT2 family in distinct organisms even in same organism are various. It was further noted that HaeDGAT2E has a close evolutionary relationship with CrDGAT2B which has the ability of restored TAG biosynthesis in H1246 yeast strains [37]. As expected, the HaeDGAT2E was able to restore the TAG biosynthesis ability of yeast mutant H1246 (Fig. 4). Two possible reasons are responsible for the failure to restore yeast TAG biosynthesis by expressing an algal or plant DGAT2 [41]. The first is the differences in codon usage between yeast and algae or plants, and the second is the limited fatty acid composition of yeast, which does not contain the diverse species of fatty acids as algae and thus cannot provide appropriate substrates for DGAT [41].
Therefore, it will be interesting to further test the function of HaeDGAT2E in yeast strain feeding with exogenous fatty acid. Unfortunately, the substrates specificity of HaeDGAT2E to polyunsaturated fatty acids (C18:2 and C18:3) was weak (Fig. 5b). The C18:2 and C18:3 polyunsaturated fatty acids are rich in H. pluvialis.
The fatty acid composition of the resulting yeast TAG appears to be related to the substrate specificity of the introduced DGAT2 toward the four dominant fatty acids in yeast [34][35][36][37][38][39][40]. As such, the introduction of DGAT gene from microalgae into yeast cells has also caused changes in fatty acid composition [5,30,34,38,41]. For instance, yeast cells harboring DGAT2A from Nannochloropsis oceanica led to the accumulation of about 40% of C16:0 in TAG, which may derive from its preference for 16:0-CoA [34]. In the present study, the TAG isolated from yeast producing HaeDGAT2E contained 23% of C16:0 and ~ 35% C18:1 (Fig. 5b) suggesting that HaeDGAT2E may have a lower preference for C16:0 containing substrate than the two DGAT2 from N. oceanica, but a higher preference for monounsaturated fatty acid C18:1. Indeed, DGAT2 form C. reinhardtii have been demonstrated to contribute to the synthesis of diverse TAG species in algal cells by displaying distinct specificities toward acyl-CoA and DAG [36]. Therefore, it is interesting to conduct a comprehensive in vitro characterization of HaeDGAT2E with different substrates including both acyl-CoA and DAG in future.
Algal DGAT is a potential target to engineer improved oil rich biomass accumulation [33]. For instance, genetic engineering of Arabidopsis thaliana and Brassica napus by expressing a DGAT1 cDNA from Chlorella ellipsoidea led to increases in the contents of lipids and polyunsaturated fatty acid [33]. In this present work, HaeDGAT2E was over-expressed in the leaves of Nicotiana benthamiana. The over-expression of HaeDGAT2E significantly increased the total lipids content and did not show any visible difference in total starch and protein contents (Fig. 6a-6d). Similarly, the over-expression of DGAT gene in tobacco increased the total lipids accumulation [42][43][44][45][46]. In addition, the change of fatty acid composition from Nb-HaeDGAT2E tobacco leaves indicated that HaeDGAT2E has a higher preference for fatty acid C16:0 and C18:1, which was consistent with results from yeast strain.

Conclusion 11
A novel HaeDGAT2E gene was obtained from H. pluvialis. HaeDGAT2E activity was confirmed in the TAG deficient yeast strain (H1246) by restoring its ability to produce TAG. HaeDGAT2E has a higher preference for fatty acid C16:0 and C18:1 in both yeast strain and higher plants. This is the first functional analysis of DGAT2 in Haematococcus. This information is important for understanding TAG accumulation and for further elucidating EAST biosynthesis in H. pluvialis.

Extraction of total RNA and synthesis of cDNA and RACEs templates
The total RNA was extracted using the H. pluvialis cells at the phase of exponential growth according to the EasySpin RNA Extraction Kit (Aidlab Biotech, Beijing, China). The total RNA concentration was quantified by NanoDrop 2000c (Thermo Scientific, USA). Totally, 2 μg RNA was used to synthesize the first-strand cDNAs by the PrimeScript® RT Enzyme Mix I (TaKaRa DRR047A, China) Kit. It is worth to note that RNA solution should be store at -80 °C if not use immediately. The RACE cDNA template was made using the SMARTer TM RACE cDNA Amplication Kit (Clontech) according to the manufacturer's instructions.

Gene cloning of HaeDGAT2E
The core sequence design was based on the highly conserved regions of DGAT2 from some green algae (Chlamydomonas reinhardtii, Chlorella zofingiensis, and Phaeodactylum tricornutum). Two pair homologous cloning degenerate primers (F1/R1 and F2/R2) were designed by CODEHOP software [47]. Moreover, the 5'-and 3'-RACEs gene specific primers were designed based on the homologous cloning sequences obtained in the previous step (5'RACE R3, R4 and 3'RACE F3, F4). All primers were showed in Table 1 and were synthesized by Sangon Biotech (China) company.
First-strand cDNA was used as template and PCR amplification was conducted with TaKaRa LATaq® (TaKaRa DRR002A, China) according to the manufacturer's instructions. The PCR was processed with the following parameters: initial denaturation at 94 °C for 5 min followed by 35 cycles of 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 1 min (according to the length of product, 1, 000 bp min -1 ), with a final extension at 72 °C for 7 min and cooling to 4 °C. Then the RACEs template was used as template and nested PCR was carried out using the nested universal primers and gene specific primers for 5'-and 3'-RACE reaction. The PCR products were resolved by electrophoresis on 1 % agarose gel. Then, the fragment of interest was excised and purified using an agarose gel DNA fragment recovery kit (TaKaRa D823A, China). Finally, the fragment was cloned into pMD-18T vector (TaKaRa D101A, China) and sequenced (Invitrogen, China).

Heterologous expression in yeast
The S. cerevisiae wild type strain INVSc1 and quadruple mutant H1246 strain, which lacks the ability of TAG biosynthesis, were used to determine the function of HaeDGAT2E by heterologous expression.
The yeast expression vector pYES2.0, which is controlled by the inducible promoter GAL1, was selected to complete this test. The primers information of pYES2.0 vector was showed in Table 1 (pYES2-F/pYES2-R). The mutant type yeast strain H1246 was cultivated on YPD medium until the OD value reached to 0.6-0.8, and then the construction of HaeDGAT2E-pYES2.0 was transformed using the method of LiAc [52]. Selection of the transformants was finished by using the synthetic medium without uracil (SC-ura). For the transformed strains, we selected individual colonies from petri dishes and added them into a 250-mL conical flask with 100-mL liquid SC-ura medium containing 2% (w/v) glucose. Then it was cultured at 30 °C with the shaker speed of 150 RPM for 24 hours. The yeast cells were collected by centrifugation and imported them into a 500-mL conical flask with 200-mL liquid SC-ura medium containing 2% (w/v) galactose to induce. After 72 hours cultivation, yeast cells were collected again and freeze-dried for subsequent experiments. The expression of HaeDAGT2 in H1246 yeast strain was verified at the transcript level by qRT-PCR method.

Nile Red staining and microscopy
Nile red fluorescent staining was used for the lipid qualitative analysis. Firstly, the collected yeast cells were diluted with sterile water to OD 600 =0.

Total lipids extraction and fatty acids analysis
Extraction of total lipids was finished according to the method of Bligh and Dyer [53]. The TAG was separated and recovered from the total lipid extracts by thin-layer chromatography (TLC) on Silica Gel plates (Merck, Darmstadt, Germany) in petroleum ether/diethyl ether/glacial acetic acid (80: 20: 2, v/v). The liquid volume ratio was 16 mL: 4 mL: 0.4 mL, and the total volume was 20.4 mL. After mixing, the liquid was added to the 200 × 100 TLC expansion cylinder. Then, after the solvent on the silica gel plate had completely evaporated, put it in an iodine tank for staining and analyze it. The TAG was recovered from the TLC plate and was transfered into a straw with glass fiber. Then, the TAG was trans-esterified with 10% H 2 SO 4 in methanol at 80 °C for 2 hour. Lastly, fatty acid methyl esters (FAMEs) were extracted with hexane and analyzed by GC according to the previous method [54]. The total fatty acid content was quantified with the intimal standard of heptadecanoic acid (C17:0).

Transient expression of HaeDGAT2E in Nicotiana benthamiana
The complete HaeDGAT2E was cloned into the plant expression vector pCAMBIA1301 under the control of the nopaline synthase (NOS) promoter and nos terminator, yielding pCAMBIA1301-HaeDGAT2E. The final binary vector was verified and then transferred into Agrobacterium tumefaciens strain GV3101 by the freeze-thaw method [55]. The strong tobacco plants cultured in an artificial climate incubator at 24 °C with a diurnal cycle of 16 h light (light intensity is 4000 lx) and 8 h dark were selected. The tobacco leaves were infected by the Agrobacterium strain which harbors binary vector pCAMBIA1303-HaeDGAT2E according to the method [56]. The transformants were selected according to the expression of target gene. The total RNA was extracted from tobacco leaves 2 days after infiltration according to the above method. The qRT-PCR technology was used to analyze the HaeDGAT2E gene expression. The NbActin gene was used as internal reference. Five days after infiltration, the tobacco leaves were harvested and subsequent extracted for total lipids (TL), total protein (TP), and total starch (TS). All primers used were showed in Table 1.

Statistical analysis
All experiments were performed three times, and the data were analyzed using a one-way analysis of variance (ANOVA) with the Social Sciences (SPSS) software, and the statistically significant at P values of < 0.05. Furthermore, the software of GraphPad Prism 8 was used to draw charts. Partial sequence alignment of HaeDGAT2E with DGAT2s from other species. The GenBank accession numbers were showed in Table 2.