Analysis of relationship between key genes and astaxanthin biosynthesis in Phaffia rhodozyma by transcript level and gene knock-out

Background Phaffia rhodozyma is a potential industrial source for production of natural astaxanthin. The synthetic mechanism of astaxanthin in P. rhodozyma is complex and unclear that blocked its development. Results In this study, eight genes related to dicyclic and monocyclic pathway in three different strains of P. rhodozyma were analyzed, and the relationship between the expression and astaxanthin biosynthesis was explored. Among these genes, crt YB (R=0.75, P<0.05) and asy genes (R=0.74, P<0.05) showed the most closely correlation with astaxanthin biosynthesis. In order to further study exact relationship, crt YB and asy genes were knocked out by homologous recombination. After crt YB knock-out, astaxanthin was decreased to be under detected line. It suggested crt YB played a role in dicyclic and monocyclic pathway. Meanwhile, the asy gene was in dicyclic pathway of astaxanthin biosynthesis, and its knock-out would promote the astaxanthin biosynthesis in monocyclic pathway, resulting in a 25.04% increase in astaxanthin production. Conclusion The possible rate-limiting enzymes were asy gene and crt YB illustrated by analysis of regression. Knock-out of asy and crt YB gene was great helpful to understand the synthetic pathway of astaxanthin, and significant to the industrial application of producing astaxanthin.

Therefore, we investigated expression levels of the genes related to dicyclic and monocyclic pathway in three P. rhodozyma of high-yield, low-yield and wild strains, and their relationship with astaxanthin and carotene synthesis. Meanwhile, the possible rate-limiting enzymes in astaxanthin synthesis were speculated by regression analysis. Moreover, the mechanism of action of the speculated ratelimiting enzymes were further validated and analyzed by gene knock-out. It provided support for understanding the relationship between these key genes and astaxanthin synthesis process in P. rhodozyma.

Results
Relationship between total carotene content and genes expression level We detected the relative expression of eight genes in three different yields of P. rhodozyma and analyzed their correlation with total carotene, β-carotene and astaxanthin. Figure 2 shows that four genes such as asy, mature mRNA of crtYB (mm crtYB ), alternatively spliced mRNA of crtYB (am crtYB) and crtE were significantly correlated with total carotene content, while other genes showed a low correlation.
The results indicated that the four genes have an impact on carotene synthesis.
The asy is presumed to be ending gene for the conversion of carotene to astaxanthin biosynthesis in the dicyclic pathway [6,21]. Figure 2A shows that it had the highest positive correlation with total carotene content (R = 0.68, P < 0.05). It further confirmed that the high expression of asy gene promoted the accumulation of carotene.
In this study, am crtYB gene showed a positive linear correlation (R = 0.675, P = 0.0011, Fig. 2B) with total carotene content. The results suggested that the gene played an important role in the whole process of carotene synthesis. Interestingly, unlike am crtYB, mm crtYB showed negatively correlated (R = 0.475, P < 0.05, Fig. 2C). This was related to the functional complementarity between the two genes [16].
The crtE gene encodes GGPP synthase that catalyzes the addition of three-molecule IPP to a DMAPP in turn to form GGPP. We have confirmed that there was a positive correlation between crtE and total carotene synthesis (R = 0.44, P < 0.05), that GGPP synthase could play a role in regulating carbon flow [22].
On the other hand, there were different trends of these genes for β-carotene ( Fig.S1). The significant correlation genes with β-carotene were the am crtYB, crtE and the alternatively spliced mRNA of crtI (am crtI), while other genes had lower correlation and regression coefficient. The crtYB catalyzes the synthesis of βcarotene from lycopene [21]. It was a direct gene involved in β-carotene synthesis.
The correlation (R = 0.61, P < 0.05) between am crtYB and β-carotene further confirmed this result. The crtE could provide precursor GGPP for the synthesis of βcarotene. That was why its expression showed positively correlated. The crtI was an octahydrogen tomato dehydrogenase and a key synthetase for γ-carotene synthesis [23]. Fig. S1G shows that the expression of am crtI was positively correlated with βcarotene (R = 0.57, P < 0.05). Some studies have pointed out that there is a certain positive correlation between γ-carotene and β-carotene synthesis [22]. Relationship between astaxanthin content and genes expression level Astaxanthin synthesis might be regulated by a variety of genes [24]. In this study, we focused on the relationship between astaxanthin production and these genes. Figure 3 shows that there were significant correlations of am crtYB, asy, am crtI and crtE gene with astaxanthin production (p < 0.05). There was a high correlation (R = 0.74, P = 0.0042, Fig. 3A) between asy gene expression and astaxanthin production.
Overall, there were great different affection on carotene synthesis among different genes, and the most significant correlation for total carotene was asy gene (R = 0.69, P = 0.003). Moreover, the genes of the most significant correlation with astaxanthin were am crtYB (R = 0.75, P = 0.0005) and asy (R = 0.74, P = 0.0042).
Verification of genes knock-out in P. rhodozyma In the above studies, we confirmed that am crtYB and asy genes were highly correlated with astaxanthin synthesis. Therefore, the two genes were selected for deletion in high-yield and wild strain P. rhodozyma in order to further analyze their mechanism of action in astaxanthin synthesis.
Single colonies grown on the transformed culture plate were randomly selected to form cell solution. The genome was identified by PCR using primers M-F/M-R and K-F/K-R designed according to exogenous gene fragments. Fig. S2A, B shows that exogenous gene fragments could not be amplified in the original strain JMU-MVP14 and JMU-VDL668, while the crtYB gene knock-out strains could be used as template to amplify the products of about 1200 bp. It suggested that the crtYB gene knockout strains were successfully screened. The result of electrophoresis of asy gene was consistent with the expectation (Fig. S2C) that indicated asy gene also was knocked out. The strains JMU-MVP14 and JMU-VDL668 after knocking out the asy gene were named JMU-MVP△asy and JMU-VDL △asy, respectively. So was crtYB gene.
From the colony morphology, the strains with the crtYB gene knock-out could be preliminarily screened out (Fig. S3). The colony of the original strain JMU-MVP14 was dark red. The colony of the strain JMU-VDL668 was pink. The colony of both strains turned white after the crtYB gene was knocked out. However, the color of asy knockout strain did not change significantly.
Effect of genes knock-out on growth By comparing the total carotenoid production of strains JMU-MVP14 and JMU-MVP △asy (Fig. 5C, D), it can be found that the trends of total carotenoid accumulation of the two strains were similar in the whole fermentation process. The total carotenoid increased slowly in 0 ~ 48 h period with low yield, it showed rapid accumulation in 48 ~ 80 h period, and then the yield increased slowly and tended to increase gradually. However, the total carotenoid accumulation of the original strain JMU-MVP14 was higher than that of the strain JMU-MVP △asy (Fig. 5C, D). It further indicated that asy gene had a significant effect on the carotenoid accumulation.
Furthermore, the production of β-carotene and astaxanthin in different strains of P. rhodozyma was determined by HPLC. The curve of astaxanthin production in the strain of asy gene knock-out was drawn according to the experimental results, as shown in Fig. 6. Because the knock-out of crtYB led to the total carotene content under the detected line (Fig. 5A, B), astaxanthin was not detected in them.
The astaxanthin production of strains JMU-MVP 14 and JMU-MVP △ asy were relatively low and no significant difference was found during 0 ~ 24 h. At 24 ~ 56 h, astaxanthin accumulation of both strains were relatively slow, and astaxanthin production of strain JMU-MVP 14 was higher than that of strain JMU-MVP △asy. After 56 h of fermentation, the astaxanthin production of strain JMU-MVP △asy was higher than that of original strain. The astaxanthin production of strain JMU-MVP △asy was slightly higher than that of strain JMU-MVP (12.32 ~ 15.12%) at most (Fig. 6A).
The trend of astaxanthin production of strains JMU-VDL668 and JMU-VDL △ asy with time was similar to that of strains JMU-MVP14 and JMU-MVP △asy with time.
Astaxanthin could not be detected in the cell extracts of both strains at 0 ~ 8 h, and the astaxanthin yield of strain JMU-VDL668 was higher than that of strain JMU-VDL △ asy at 8 ~ 40 h. After 40 h, the astaxanthin production of strain JMU-VDL △ asy exceeded that of strain JMU-VDL 668, and was 25.04% higher than that of strain JMU-VDL 668 (Fig. 6B).

Discussion
The crtYB was considered to be a variable expression protein that belongs to a class of regulatory proteins [16]. It can be transcripted to produce two kinds of products: mature mRNA (mmRNA) and alternatively spliced mRNA (amRNA) [16]. In the nematode C. elegans, splicing of the RNA encoding the ribosomal protein was a regulated process for productive or unproductive transcripts [25]. In this study, when the synthesis of mm crtYB was reduced, am crtYB was accumulated. These results indicated that a similar regulatory pattern of functional complementarity might exist between the two splicing products in P. rhodozyma. Moreover, the knock-out result further confirmed the important role of crtYB in astaxanthin and total carotene synthesis. Therefore, its positive effect on the synthesis of total carotene was beyond doubt.
The direct gene for astaxanthin synthesis is asy gene (Fig. 7). The previous studies of Lodato, Miao and Marcoleta et al. pointed that the expression of asy gene in highyielding astaxanthin-producing strains was higher than that in low-yielding strains [17,26,27]. Contreras et al. also found that asy gene was overexpressed in P. rhodozyma [ 28], the total carotenoid production of the strain did not change significantly, and the astaxanthin ratio increased. In this study, the asy gene expression showed high correlation with astaxanthin content and total carotene content. Therefore, asy was an important gene to regulate the accumulation of pigments, especially astaxanthin.
As secondary metabolites [29], carotene and astaxanthin have no significant effect on biomass of primary production in P. rhodozyma. That was why the biomass of the knock-out strain did not decrease. Due to substance conservation, secondary metabolism was weakened, while primary metabolism would be strengthened. The knock-out of crtYB gene promoted the growth of JMU-VDL668. Compared with asy gene, crtYB gene played a more significant role as a key gene in the synthesis of secondary metabolites such as astaxanthin and carotene. The JMU-VDL668 strain was a low-yield carotene strain (Table 1), and the asy gene presumed to be the terminal pathway of its carotene [18,19], the effect of asy gene knock-out was lower than that of crtYB. The astaxanthin biosynthesis system in P. rhodozyma has dicyclic and monocyclic pathways (Fig. 7) [5]. The crtYB gene knock-out could interrupt the whole carotene pathway, resulting in the absence of carotene (Fig. 5A, B). The crtYB is distributed in both pathways. Meanwhile, asy gene was considered as terminal enzyme of forming astaxanthin from β-carotene [11,18] in the dicyclic pathway (Fig. 7). When asy in the dicyclic pathway was disrupted, the efficiency of the monocyclic pathway might be enhanced. Under some specific conditions, the monocyclic pathway is more efficient than the dicyclic pathway for astaxanthin synthesis [20]. Figure 6 also showed that P. rhodozyma could synthesize more effectively astaxanthin after asy gene knock-out. Previous studies have also confirmed that there is no necessary relationship between asy and astaxanthin synthesis [30]. In addition, β-carotene was significantly reduced in asy knock-out strains (Fig. S4). Lei et al. pointed that carotene synthetase in Haematococcus pluvialis was restrained by increased product such as polyunsaturated fatty acid [31]. These results suggested that astaxanthin as terminal product might also restrain β-carotene synthetase. It further illustrated knock-out of asy gene might regulate down the dicyclic pathway.
Knock-out of asy and crtYB gene was great helpful to understand the synthetic pathway of astaxanthin, and significant to the industrial application of producing astaxanthin.

Conclusions
We analyzed expression of the genes related to astaxanthin biosynthesis in P. and JMU-N3 were established through ethyl methylsulfonate mutagenesis from JMU-VDL668. P. rhodozyma strain JMU-N3 was a strain of low astaxanthin production established by our laboratory, while P. rhodozyma JMU-MVP14 was a strain with high astaxanthin production. The profiles of the three strains were shown in Table 1.
Culture conditions P. rhodozyma were grown in YPD culture at 22 °C. 2% of the cells were inoculated in 250 mL flasks with 30 mL medium and cultivated for 48 h. After two generations of cultivation, the YPD culture was then transferred to fermentation medium. The fermentation medium and culture condition were the same as our previous study [32] with 12-hour intervals sampling. Three paralleled samples were performed and 5 mL fermentation broths were sampled and stored at -20 °C until detection.
Cell of dry weight, total carotenoid extraction and analysis of carotenoids For dry weight determinations, 2 mL samples were centrifuged for 5 min at 3500 × g and cells were washed twice by distilled water. Then the washed cells were dried to constant weight at 105 °C to analyze the total weight.
Carotenoid was extracted by the method of dimethysulfoxide [33]: 2 mL of samples were centrifuged for 5 min at 3500 × g, and the cells were washed twice with distilled water. After decanting the water, the cells were then treated with 2 mL of dimethylsulfoxide at 75 °C. Next, 5 mL of ethanol was used to extract carotenoid.
Total carotenoid concentration was detected by spectrophotometer at a wavelength of 474 nm [34]. The analyses of total carotenoid were performed in triplicate, and pigments were normalized by the dry weight of the yeast. Profiles of carotenoid family were analyzed by RP-HPLC using a reverse phase Nova-Pak C18 column. The gradient conditions were shown in Table S1. The elusion spectra were recuperated using a diode array detector. Peaks were identified by comparison to the prepared standards (Sigma, USA) and integration of the peak areas was used to quantify carotenoids from obtained standard curves.

Total RNA extraction
The cellular pellets obtained from 40 mL of culture were utilized for the extraction of total RNA by a modified protocol of TRI REAGENT RNA extraction kit (TRI, Sigma, USA). The total RNA concentration was quantified at 260 nm and the purity was

Real-time quantified PCR
RT-qPCR analyses were performed using cDNA samples as a template, with an ABI 7900HT apparatus (Applied Biosys-tems, Norwalk, CT). Dissociation curves were constructed to test amplification validity. All sequences of target genes were obtained from NCBI. Database accession numbers and the corresponding primer sets in RT-QPCR are shown in Table S2. The β-actin was used as the control gene. Preparation of competent cells: The P. rhodozyma strains named JMU-MVP14 and JMU-VDL668 were transferred to seed YPD culture medium at 22℃, 180 rpm/min for two days. After three generations of cultivation, 5 mL of culture in exponential growth was collected in a sterile centrifugal tube, centrifuged at 22℃ for 5 min at 3000 × g, and the supernatant was discarded as far as possible to collect cells. 25

Ethics approval and consent to participate
This article does not contain any studies with human participants or animals performed by any of the authors.

Consent for publication
Not applicable.

Data and materials availability
All data needed to evaluate the conclusions are present in the paper and/or Supplementary file.

Funding
This work was financially supported by National Natural Science Foundation of China  The regression correlation of total carotene content with expression of gene asy (A), am crtY The regression correlation of astaxanthin with expression of gene asy (A), am crtYB (B), mm Figure 5 The curves of total carotenoids in original strains and the mutants with crtYB (A, B) and asy Figure 6 The curves of astaxanthin in original strains and the mutants with asy gene knocking-out Supplementary material2.pdf