Evolutionary origins, molecular cloning and expression of carotenoid hydroxylases in eukaryotic photosynthetic algae

Identification, classification, and distribution of CHY proteins

18 algal nuclear genomes were examined for putative genes of CHY proteins. A summary of algal genes putatively encoding CHYs is shown in Table 1, and the classification and distribution of candidate genes, and hypothesized xanthophylls biosynthetic pathways across organisms are given in Figure 1. A total of 11 and 49 putative genes encoding BCH and CYP97, respectively, were predicted and annotated from 18 complete or incomplete eukaryotic algal genomes. Deduced protein sequences of genes encoding CHYs from the 18 algal genomes are shown in the Additional file [see Additional file 2]. Among the 18 eukaryotic photosynthetic algae, red algae possessed a bch homolog (CrtR-type) only. Hydroxylation of β-carotene in different organisms is primarily carried out by three gene subfamilies of the non-heme/di-iron monooxygenase superfamily: bch genes of higher plants and green algae; bch genes of non-photosynthetic bacteria; and CrtR genes of cyanobacteria [37]. Green algae inhabiting freshwater (Chlorophyceae: C. reinhardtii and V. carteri; Trebouxiophyceae: C. sp. NC64A, C. vulgaris, and C. sp. C-169) and marine environments (Prasinophyceae: M. pusilla, M. sp. RCC299, O. sp. RCC809, O. tauri, and O. lucimarinus) each possessed a bch homolog. No CrtR- or bch-homolog was discovered in those algal strains from red algal secondary endosymbiotic event. Previous studies, however, demonstrated a xanthophyll cycle (zeaxanthin-antheraxanthin-violaxanthin), with violaxanthin the putative precursor of both diadinoxanthin and fucoxanthin in the diatom Phaeodactylum tricornutum[49,50], implying that this enzymatic reaction may therefore be catalyzed by other unrelated enzymes, such as LUT-like P450 proteins in those algal strains from Chromalveolates [51]. Therefore, isolation and characterization of novel enzymes involved in β-xanthophylls (zeaxanthin) biosynthesis in Chromalveolates are of going research.

Homologs of cyp97a and cyp97c genes are found in green algae, but absent in other algae, including red algae (C. merolae), Heterokontophyta (P. tricornutum, F. cylindrus and T. pseudonana), Haptophyta (E. huxleyi), Cryptophyta (G. theta), and Stramenopiles (A. anophagefferens). Interestingly, two cyp97a genes were predicted in C. reinhardtii and O. lucimarinus, indicating that lineage-specific gene duplications occurred during the evolution of these algae. Due to gene loss or the incompletely sequenced genome, only genes encoding CYP97B and CYP97C homologs and none gene encoding CYP97A homolog were predicted in B. natans CCMP2755. Moreover, none gene encoding CYP97A homolog has been discovered in Chlorella sp. NC64A. In this study, at least one copy of the gene encoding CYP97B was found to be widely distributed in most of algae, except for red algae, indicating that the originations of this gene is not due to secondary endosymbiosis.

Domain structures of BCH-type CHYs in algae

β-Xanthophylls (zeaxanthin) are widely distributed in nature, and they are common in all photosynthetic eukaryotes, and many photosynthetic, or non-photosynthetic prokaryotes [26,52]. Non-heme/di-iron β-ring carotenoid hydroxylase (BCH) has long been regarded as the only enzyme involved in zeaxanthin synthesis by hydroxylating β–carotene. Genes encoding β-carotene hydroxylases have been identified and characterized from bacteria [53,54], cyanobacteria [55], algae [41,42] and higher plants [34]. The bch genes of algae and higher plants encode the members of the fatty acid hydroxylase superfamily [PF04116] that also includes fatty acid and carotene hydroxylases and sterol desaturases. Members of this superfamily are integral membrane proteins and contain two copies of histidine-rich iron-binding motifs (HXXHH), needed for binding this cofactor with enzymes [34,56].

An alignment of the deduced amino acid sequences of CrtRs of cyanobacteria, and BCHs of red algae and green algae is displayed in Figure 2. There are few gaps in the alignment and a number of highly conserved regions. Four transmembrane segments and four histidine-rich boxes (HKXLWH, HXSHH, HDGLVH, and HXXHH) are distributed among the amino acid residues of BCHs of green algae (Figure 2). It is interesting that the first transmembrane segment was absent in two BCHs of green algae O. sp. RCC809 and O. lucimarinus, and replaced instead by another histidine-rich box in the N-terminal (HKHLWH). Three conserved histidine-rich boxes (HDASHXXAH, HXQHHXX, and HLIHH) and five transmembrane segments are generally well-conserved between red algae and cyanobacteria, indicating that red algae acquired this gene from cyanobacteria during primary endosymbiosis (Figure 2). However, previous studies have demonstrated that the protein sequence of the C. merolae BCH (CrtR-type) is truncated at the N-terminus by at least 22 amino acids relative to the sequences of all cyanobacterial CrtRs, and that the activity of this encoded enzyme could not be shown when expressed in E. coli[57,58]. Two HXXHH boxes are widespread across the BCHs of algae and CrtRs of cyanobacteria, indicating that their catalytic mechanisms are highly similar in these proteins [34,56].

Origins and evolution of BCHs in algae

β–Carotene hydroxylation in different organisms is primarily carried out by members of three subfamilies of the non-heme/di-iron monooxygenase superfamily: BCH-type enzyme of green algae and higher plants; BCH-type enzyme of non-photosynthetic bacteria; and CrtR-type enzyme of cyanobacteria. The former shares sequence homology with non-photosynthetic bacterial-type enzyme, and the latter (CrtR-type) is more closely related to bacterial carotenoid ketolases [59]. Identification of plant-like BCHs in green algae and cyanobacteria-like CrtRs in red algae is the key indicating complicated origination of BCH in algae, for which we constructed a phylogenetic tree of CrtRs of cyanobacteria and BCHs of green algae, higher plants, and bacteria (Figure 3) (Additional file 3: Table S2).

As shown in Figure 3, difference between CrtRs of cyanobacteria and BCHs of algae, bacteria, and higher plants is very clear. Genes encoding BCHs from bacteria, green algae, and higher plants build a monophyletic group (Bootstrap [BS]: 100%), and the phylogenetic relationship present here indicates that they share a common ancestor and strongly supports a non-cyanobacterial origin. In contrast to the apparently widespread retention of BCH paralogs in higher plants, the gene encoding BCH is a single copy in each algal genome sequences (Figure 3). In higher plants, duplication of the ancestral bch gene took place most probably via whole-genome or segmental genome duplication [26] and the duplicates seem to have functionally diverged primarily at the gene expression level [60-62]. It is worth mentioned, however, that the gene encoding BCH (CrtR-type) of red algae forms another monophyletic group (BS: 97%) with CrtRs of cyanobacteria, suggesting that the red alga obtained this gene via the primary endosymbiotic event from cyanobacteria and retained it over the course of evolution persistently. However, the activity of this encoded enzyme was not demonstrated when expressed in E. coli[57,58]. Therefore, hydroxylation of β-carotene in red algae remains unclear, calling for more red-algal genomes sequences to elucidate. In this study, no BCH- or CrtR-type CHY-encoding gene was discovered in algae of Chromalveolates, except for a partial sequence of bch-type gene in the genome sequences of T. pseudonana. This result is consistent with observations from a previous study, in which only a partial sequence of BCH-type gene was found in the genome sequences of T. pseudonana and no one was detected from the P. tricornutum genome [51]. Unfortunately, the enzymatic activity of this putative carotene hydroxylases (BCH) from T. pseudonana has not yet been reported anywhere to the best knowledge of authors. Therefore, the hydroxylation of β-carotene in Chromalveolates remains unknown, resulting from genomes of the few organisms currently sequenced lack entire hydroxylation families and, enzymatic activity of putative carotene hydroxylases has not yet been reported at present [51,57,63-65].

As mentioned above, it is difficult to fully understand the evolution of β-xanthophylls biosynthesis with limited available data, but some hypotheses were proposed based on the tree topology. Firstly, it is presumed that CrtR-type CHY was originated in all algae as the result of primary endosymbiosis. In such a scenario, CrtR-type CHY proteins were lost in extant green algae and higher plants, and these organisms acquiring another BCH from bacteria by lateral gene transfer or from the host during the primary endosymbiosis event. Secondly, red algae have CrtR-type CHY-encoding gene, while Chromalveolates do not have them due possibly to gene lost during the secondary endosymbiosis event. In addition, BCH-type CHYs are also absent in Chromalveolates, which may be resulted from the replacement of algal CrtR- and BCH-type CHYs by another unidentified novel lineage-specific CHY in strains that derived from the red algal secondary endosymbiosis. Previous studies of the diatom P. tricornutum[43], however, have demonstrated the presence of a xanthophyll cycle (zeaxanthin-antheraxanthin-violaxanthin), with violaxanthin the putative precursor of both diadinoxanthin and fucoxanthin, implying this enzymatic reaction may therefore be catalyzed by other unrelated enzymes, such as LUT-like P450 proteins [51]. Therefore, we speculate that the bch genes in Chromalveolates have not been under a strong pressure of natural selection like in green algae and higher plant lineages, and they may have instead evolved different ways for β-xanthophylls biosynthesis.

Domain structures of CYP97s in algae

Cytochrome P450s are defined by the 450 nm light absorption of their heme cofactors. They oxidize various arrays of metabolic intermediates and environmental compounds [66], and participate in many primary, secondary, and xenobiotic metabolic reactions [67]. The CYP97-encoding genes are members of P450s family. These members share a common catalytic center in heme by iron coordination to the thiolate of a conserved cysteine [68]. Despite low sequence identity at amino acid level, P450s display a common overall topology in three-dimensional folding pattern [69,70]. Genes encoding members of CYP97 family have been isolated and functionally investigated from higher plants [8,26,33,37,59,71,72]. There are three CYP97 family members (CYP97A3, CYP97B3, and CYP97C1) in Arabidopsis genome, of which CYP97A3 and CYP97C1 are predicted to be chloroplast-targeted [59,73]. The localization of predicted chloroplast CYP97A/C from algae and higher plants is coincident with the sub-cellular of carotenoids biosynthesis. To further reveal domain structure characteristics of CYP97 proteins, an alignment was constructed using selected CYP97A/B/C protein sequences from different algae (Figure 4) and an additional file shows this in more detail [see Additional file 4: Figure S1].

The alignment (Figure 4) reveals 11 domains in strongly conserved amino acid sequences from all CYP97 protein sequences. All conserved domains are highly similar between CYP97As and CYP97Cs of green algae except for CD1 (QPVFVPLYKLPLXYG) (Figure 4), indicating that they share a common ancestor. In addition, the conserved domains in CYP97Bs of distinct algae phyla are also similar, implying that this gene occurred before the formation of extant diverse algal groups. Degenerate primers for cloning genes encoding CYP97 homologs were designed from these conserved amino acid residues. P450s catalytic motifs were found as expected in all CYP97A/B/C protein sequences including I-helix involved in oxygen binding (CD6 in CYP97B and CD5 in CYP97A and CYP97C), an ERR triad (CD7 in CYP97A/B/C) responsible for locking the heme pockets into position and assuring stabilization of the conserved core structure, and CD10 associated with the conserved heme-binding cysteine. In addition, many amino acids were conserved in each CYP97 subfamily such as CD4 (LLRFLVDXR) and CD9 (LYPXE) were conserved in CYP97B subfamily only. The conserved domains associated with active sites suggest that these protein sequences are members of the P450 family, and that conserved domains within each CYP97 subfamily proteins are responsible for the specificity with respect to the β- or ϵ-ring of different carotenes.

Origins and evolution of CYP97s in algae

Production of α-xanthophylls (lutein) in higher plants requires four reaction steps: β- and ϵ-ring formation from lycopene by the action of β- and ϵ-cyclases (LYCB and LYCE); and subsequently, hydroxylation of each ring of α-carotene by β- and ϵ-ring hydroxylases (CYP97A and CYP97C) [26]. Ample evidence shows that genes encoding CYP97A and CYP97C homologs are only present in green algae and higher plants, but not in Chromalveolates, indicating that α-xanthophylls biosynthesis occurs only in green algae and higher plant lineages. In our results, this biosynthetic pathway is also absent in red algae. In contrast to a previous study postulated that synthesis of α-xanthophylls occurs in only a few lineages of photosynthetic eukaryotes, namely, some red algae, all green algae, and higher plants [26]. An insufficient number of red algal species (only a red alga used in this study) may be responsible for these different conclusions. In addition, our previous and other studies have demonstrated that β-cyclases (LYCBs) genes are widely distributed in nature, and ϵ-cyclase (LYCEs) genes were identified only in green algae, higher plants, and some cyanobacteria (e.g. Prochlorococcus marinus MED4). LYCE genes seems to come from β-cyclases by gene duplication and subsequently functional divergence [30,57,74-77], which indirectly manifests that α-xanthophylls are synthesized only in green algae and higher plants. No CYP97A and CYP97C homologs are detected in algal genome sequences from Chromalveolates, and these algae cannot synthesize α-xanthophylls (lutein), which is consistent with previous studies [51].

At present, no CYP97 protein homolog has been found in cyanobacteria, suggesting that these proteins are an ancient eukaryotic innovation. All three CYP97 subfamilies are represented in Arabidopsis and other land plants, often in a single copy per subfamily, indicating their critical functions [26]. In contrast, some paralogs genes encoding CYP97A in a few green algae and CYP97B in most algae, especially in Chromalveolates, indicating recent gene duplication events occurred. For deep understanding the internal phylogenetic relationships among CYP97 proteins from higher plants and green algae, we constructed a maximum likelihood tree using our CYP97 database and some other CYP97s from higher plants (Figure 5). CYP86A1 from Arabidopsis thaliana was selected as outgroup because its substrates, fatty acids with chain lengths from C₁₂ to C₁₈[78] are mostly molecules similar to carotenoids, and the CYP86 clade is the most closely related to the CYP97 clade [26].

As it shows in maximum likelihood tree, CYP97 proteins constitute three distinct monophyletic groups: CYP97A, CYP97B, and CYP97C. Two of these groups (CYP97A and CYP97C) form a sister group that are composed of proteins from green algae and higher plants. CYP97A and CYP97C sequences in Chromalveolates, including Heterokontophyta (P. tricornutum, F. cylindrus, and T. pseudonana), Haptophyta (E. huxleyi), Cryptophyta (G. theta), and Stramenopiles (A. anophagefferens) are absent in the cluster. Genes encoding CYP97B homologs from all organisms form a monophyletic group (BS: 92%). Surprisingly, algal CYP97Bs from Chromalveolates and green algae (Prasinophyceae) build another monophyletic group (BS: 95%). This phylogenetic relationship and the lack of CYP97B homologs in C. merolae point to a “green” origin of this gene in Chromalveolates, similar to the origin and evolution of hydroxypyruvate reductase [79]. Alternatively, this gene represents early eukaryotic innovations in the Plantae although lacking CYP97B homolog in C. merolae. Further investigations are needed to support this hypothesis.

The topology of the phylogenetic tree shows that the CYP97B is an ancient gene emerging before the divergence of extant algae groups during evolution, and represents an ancient eukaryotic innovation. Our results indicate that CYP97A is evolutionarily more closely to CYP97C than to CYP97B (Figure 5). Therefore, we believe that the CYP97A and CYP97C genes were originated by gene duplication before the split between green algae and higher plants, and were subjected to purifying selection in a lineage-specific fashion (Figure 5). Alternatively, these genes may have been originally present in all algae, and then they were lost gradully from red algae and Chromalveolates. This scenario is similarity to that of Arabidopsis whose evolution and functional divergence of two duplicate gene pairs, CYP97A3/C1 and BCH1/2 involved in carotenoid hydroxylation occurred [26]. Additional lineage- or organism-specific gene duplications have occurred during the evolution of CYP97A in green algae (C. reinhardtii and O. lucimarinus), and CYP97B in most algae, except for C. reinhardtii, C. sp. NC64A, C. vulgaris, C. sp. C-169, V. carteri, M. pusilla, E. huxleyi, and B. natans CCMP2755. Duplication and subsequent functional divergence of genes have been recognized increasingly as an important mechanism of evolution [80-83].

Isolation and characterization of cyp97 genes from H. pluvialis

To further study genes encoding CYP97 homologs in green algae, three full-length cDNA sequences of cyp97 homologs, including Haecyp97a, Haecyp97b, and Haecyp97c[84] were isolated from commercial green alga H. pluvialis strain Flotow 1844. Briefly, 1,017- and 984-bp cDNA fragments encoding HaeCYP97A/B were generated by RT-PCR with degenerate primers [see Additional file 5: Table S3]. The Haecyp97a fragment shared 77% sequence similarity with cyp97a5 gene of C. reinhardtii and 74% similarity with cyp97a3 gene of A. thaliana. The Haecyp97b fragment shared 61% sequence similarity with A. thaliana cyp97b3 gene. The results indicate that two partial putative cyp97a/b genes were isolated from H. pluvialis. Gene-specific primers were then designed to obtain full-length sequences of Haecyp97a/b using RACE methods. Information regarding the full-length sequences of Haecyp97a/b/c is summarized in Table 2. The full-length cDNA sequences of Haecyp97a comprised 1,872-bp with an open reading frame (ORF) of 1,593-bp encoding a 530 amino acid protein, and it was flanked by a 159-bp of 5′-untranslated region (UTR) and a 120-bp of 3′-UTR including the poly-A tail [see Additional file 6: Figure S2]. The deduced protein had a calculated molecular weight of 59.03 kDa and a predicted isoelectric point (pI) of 7.81. The full-length cDNA sequences of Haecyp97b contained an ORF of 1,620-bp, a 2-bp 5′-UTR, and a 249-bp 3′-UTR, and it encoded a new putative carotenoid hydroxylase protein of 539 amino acid protein with a deduced molecular weight of 58.72 kDa and pI of 6.26 [see Additional file 7: Figure S3]. Sequence analysis revealed that the cloned Haecyp97c cDNA was 1,995-bp in length, and contained a 1,620-bp ORF, a 46-bp 5′-UTR, and a 329-bp 3′-UTR in characteristic of a poly (A) tail. An ATG translation initiation codon was identified in the 46-bp terminal sequence (47-49 bp), and a TAA termination codon was found in the 1,620 bp downstream of the initiation site [see Additional file 8: Figure S4].

The ChloroP and TargetP servers [85,86] were used to predict the sub-cellular location of the respective deduced proteins from isolated genes (Table 2). The results indicate that CYP97A and CYP97C in H. pluvialis are probably located in the chloroplast, same as in higher plants Arabidopsis thaliana[59,73]. We also calculated degrees of identity and similarity between predicted amino acid sequences of each isolated CYP97 and corresponding CYP97s of other eukaryotes (Table 3). Amino acid sequences from each of the isolated CYP97 genes shared high similarities (72%-76%) with known eukaryotic proteins, indicating that the three novel cyp97 genes of green alga H. pluvialis Flotow 1844 has been successfully identified using RT-PCR with degenerate primers designed from conserved motifs and RACE methods. To our knowledge, this is the first time these three cyp97 gene homologs have been isolated from commercial green algae.

Expression analysis of four chy genes in H. pluvialis under high light stress

Studies showed that high light (HL) effectively induces carotenoid biosynthesis-related gene expression and astaxanthin accumulation in H. pluvialis[47,48], indicating that light plays an important role in controlling green algal carotenoid biosynthesis. Although the regulatory role of light in the expression of nuclear-encoded plastid-targeted proteins has been studied for decades in green algae and higher plants, elucidation of effects of light on transcriptional levels of three novel cyp97 genes in green algae is still in its infancy. In a previous study on diatom P. tricornutum, a blue-light library was found to be the most enriched in carotenogenesis-related ESTs [51]. For more detail in the transcriptional regulatory role of light on green algae, we studied the gene transcriptional expression profiles of bch, Haecyp97a, Haecyp97b, and Haecyp97c genes in response to white and blue HL conditions.

As shown in Figure 6, transcriptional levels of four chy genes (bch, Haecyp97a, Haecyp97b, and Haecyp97c) were increased throughout the course of HL illumination treatments. Starting from relatively low levels, bch expression level was slowly increased under blue HL treatment and reached a maximum transcriptional level at 24 h exposure that was 5.0-fold higher than that of the control. It then declined sharply after 72 h of exposure (Figure 6A). A similar, although less pronounced trend was observed under white HL: the highest bch transcriptional level occurred at 54 h of exposure, with 3.3-fold higher compared with the control. Our previous study has demonstrated that zeaxanthin concentrations under blue and white HL treatments were increased markedly and reached their highest levels at 34 h (blue) and 48 h (white) of exposure [84], which is listed in Figure S5 for details [see Additional file 9: Figure S5]. BCH is the mainly enzyme catalyzing the hydroxylation of β-carotene, which produces zeaxanthin associated with the xanthophylls cycle [40,87]. The contradiction between low bch expression levels and marked increase of zeaxanthin concentration during the early stage of treatments was explained by the fact that when photoprotection is required, violaxanthin is rapidly converted via antheraxanthin to zeaxanthin by violaxanthin de-epoxidase [88]. The astaxanthin concentration under blue or white HL stress was higher than that of the control, reaching a maximum level at 34 h and 54 h of exposure, respectively [84], which is listed in Figure S5 for details [see Additional file 9: Figure S5]. Accumulation of astaxanthin may be responsible for the rapidly increasing levels of bch and sharply decreasing concentrations of zeaxanthin at later treatment stages. Previous studies have demonstrated that bch and bkt are the main genes involved in astaxanthin biosynthesis [89].

Expression levels of Haecyp97a, Haecyp97b, and Haecyp97c began with a slight decrease followed by a dramatic increase (Figure 6B-D). Strongly steady increase in transcriptional levels of Haecyp97a and Haecyp97b were observed during 4–28 h of both HL (white and blue) treatments. Then, they dropped sharply after 28 h of exposure. It is also intriguing that different conditions had various impacts on Haecyp97a and Haecyp97b mRNA levels. For instance, transcriptional levels of Haecyp97a and Haecyp97b under blue HL were higher than that of under white HL during 4–10 h of exposure, and the contrary tendency was observed during 13–28 h of exposure (Figure 6B and C). The maximum transcriptional levels of Haecyp97c under both the blue and white HL treatments occurred at hour 13 and 10, and transcriptional levels were 4.5- and 2.8-fold higher than that of control, respectively (Figure 6D). Blue HL appeared to have stronger effects on Haecyp97c transcriptional level than that of white HL (Figure 6D). Studies on mutant of higher plants showed that CYP97A3 and CYP97C1 are the enzymes primarily responsible for catalyzing hydroxylation of β- and ϵ-ring of α-carotene, respectively, producing α-branch xanthophylls (lutein), the most abundant carotenoids in light-harvesting complexes (LHCs) which is key structural and functional components of light harvesting [6,8,26,33,37,39,59,72]. During early stages (0–4 h) of HL (blue and white) treatments, lutein concentrations were lower than that of the control [84], which is listed in Figure S5 for details [see Additional file 9: Figure S5]. This result is consistent with the decreased transcriptional levels of Haecyp97a and Haecyp97c observed in H. pluvialis (Figure 6B and D). Although Haecyp97a and Haecyp97c transcriptional levels increased over time, lutein concentration level remained rather stable but was lower than the control. When considering that CYP97A is also involved in hydroxylation of β-carotene into zeaxanthin, the level of Haecyp97a increase is understandable [26]. Previous studies have demonstrated there was a cytochrome P450 involved in astaxanthin biosynthesis in H. pluvialis by the use of ellipticine [90]. Therefore, we speculate that CYP97A may be the cytochrome P450 involved in astaxanthin biosynthesis in H. pluvialis. According to our results, however, it is unclear why and how the levels of Haecyp97b and Haecyp97c increased. The function of CYP97B is currently unknown and they (CYP97B and CYP97C) may play some additional roles (aside from CHY ϵ–ring) in algae under adverse conditions. An earlier study reported that cytochrome P450 reductase is also co-up-regulated with enzymes involved in DNA repair under light/dark cycles [91].

Our data (Figure 6) reveal that mRNA levels of different chy genes increase rapidly under HL exposure and varied with light wavelength. For instance, our results indicate that blue HL very significantly increased Haebch and Haecyp97c genes expression, and produced similar effects on Haecyp97a and Haecyp97b genes expression compare with white HL treatment. Young et al. [40] have demonstrated that different carotenoid metabolic gene isoforms typically had distinctly different expression patterns. Previous study has demonstrated that zeaxanthin and astaxanthin concentrations were more efficiently enhanced for H. pluvialis under blue HL treatment than white [84]. Therefore, we propose that Haematococcus cells are more sensitive to blue induction compare with white. Similar phenomena were reported red Light Emitting Diodes (LEDs) operated at a relatively low light intensity were found to be suitable for cell growth; and LEDs emitting short wavelength (380–470 nm) can induce morphological changes in H. pluvialis and enhance astaxanthin accumulation [92-94]. It is well known that strong light (blue) enhances astaxanthin accumulation, because astaxanthin is produced by H. pluvialis cells to protect cells from the strong intensity of the light [93]. Although blue light contains more energy than red light, according to Planck’s law [95], in fact only energy associated to the S1 transition level of chlorophyll (Chl) can be used for photosynthesis [96]. On this bottom line, the fact that the efficiency of energy transfer from carotenoid to Chl a is far lower (40%) with respect to Chl b to Chl a or Chl a to Chl a. Thus it can easily be concluded that less energy used for photosynthesis is available from blue light than from red or white light under the same light intensity. According to above conclusions, we speculate that astaxanthin accumulation for H. pluvialis depends on not only light intensity, but also light quality (i.e. blue light might be a more useful wavelength to enhance astaxanthin accumulation in H. pluvialis under HL intensity). Therefore, we believe that the response of H. pluvialis Flotow 1844 to high light stress is a complicated process involving bch, Haecyp97a, Haecyp97b, and Haecyp97c.

Identification of chy genes encoding CHY proteins

The genomes of 18 eukaryotic photosynthetic algae included Chlamydomonas reinhardtii, Chlorella sp. NC64A, Chlorella vulgaris, Coccomyxa sp. C-169, Volvox carteri, Micromonas pusilla, Micromonas sp. RCC299, Ostreococcus sp. RCC809, Ostreococcus tauri, Ostreococcus lucimarinus, Phaeodactylum tricornutum, Thalassiosira pseudonana, Fragilariopsis cylindrus, Aureococcus anophagefferens, Emiliania huxleyi, Guillardia theta and Bigelowiella natans CCMP2755 were obtained from the website of the DOE Joint Genome Institute (Walnut Creek, CA, USA; http://genome.jgi.doe.gov/). The genome of the red alga Cyanidioschyzon merolae was obtained from the C. merolae Genome Project (http://merolae.biol.s.u-tokyo.ac.jp). The protein coding sequences of each genome was fed into the program makeblastdb to create an organism-species database [97].

Two methods were applied to identify the putative CHY homologs genes. Firstly, we followed JGI’s or the C. merolae Genome Project’s annotation to determine the number of chy present in the algal genomes. Then, eight previously characterized CHYs from Haematococcus pluvialis [GenBank: BCH, ABB70496.1], Chlamydomonas reinhardtii [GenBank: BCH, AAX54907.1], Synechocystis sp. PCC 6803 [GenBank: CrtR, BAA17468.1] and Arabidopsis thaliana [GenBank: BCH1, sp|Q9SZZ8.1; BCH2, sp|Q9LTG0.1; CYP97A3, NP_564384.1; CYP97B3, NP_193247.2, and CYP97C1, NP_190881.2] were used to construct a query protein set. BLASTp [97,98] and HMMER [99] programs were then conducted locally to identify all chy genes in all 18 algal genomes using a threshold e-value of 1e-10. Finally, we manually checked the extracted proteins by SMART and Pfam analyses to avoid false positive hits that commonly arise during large-scale automated analyses. Putative chy genes found by this method were added to the query set for another round of BLASTp searches. This procedure was iterated until no newly retrieved sequences that belonged to chy homologs. Moreover, in order to check for false negatives, two HMM models [Pfam: PF04116] and [Pfam: PF00067] derived from known bch and cyp97 genes were applied to search for genes encoding chy on all proteins encoded in the 18 algal genomes [100,101]. All translated protein sequences of CHYs-encoding genes used in this paper were listed in more detail [see Additional file 2]. CHY proteins from higher plants and cyanobacteria were also downloaded from National Center for Biotechnology Information GenBank database and Cyanobase (http://bacteria.kazusa.or.jp/cyanobase/).

Multiple sequence alignment and phylogenetic analysis

Proteins identified by the BLAST and HMM searches were aligned using ClustalW [102,103] with a gap opening penalty of 10, a gap extension penalty of 0.2, and Gonnet as the weight matrix. The SMART [104] and Pfam 26.0 [105] databases were applied to delete false positives. Phylogenetic trees were constructed using Maximum likelihood (ML) method that implemented in PhyML [106]. ML trees were built in particular models of amino acid substitutions chosen according to PROTTEST AIC results [107]. The Le and Gascuel evolutionary model [108] was selected to analyze the protein phylogenies by assuming an estimated proportion of invariant sites and a gamma correction (four categories). Bootstrap values (BS) were inferred from 400 replicates. Graphical representation and edition of the phylogenetic tree were performed with TreeDyn (v198.3) [109].

Algal strains and culture conditions

H. pluvialis strain Flotow 1844 was obtained from Culture Collection of Algae and Protozoa, Dunstaffnage Marine Laboratory, and maintained at the Biological Resources Laboratory, Yantai Institute of Costal Zone Research, Chinese Academy of Science. Algae were incubated in 250-ml erlenmeyer flasks, each containing 100-ml BBM, and placed in an illuminating incubator (Ningbo Jiangnan Instrument Factory, GXZ-380, Ningbo, China) under a light intensity of 25 μmol photons m^–2 s^–1 in 14 h:10 h diurnal scheme at temperature of 22 ± 1°C without aeration.

For high light stress conditions, exponentially growing cultures (cell density approximately 5 × 10⁷ cells ml^–1) were harvested, and then transferred to 500-ml erlenmeyer flasks (named, 1–9), each containing 250-ml BBM (fresh medium) under continuous exposure of white light (390–770 nm) or blue light (420–500 nm) in light intensity of 1,000 μmol photons m^–2 s^–1 without day/night cycle. Algal cells collected (20-mL sample at some selected time) were rinsed with PBS, stored at −80°C if not immediately used. All experimental chemicals and reagents were of analytical grade.

Cloning and characterization of the CYP97 genes

H. pluvialis strain flotow 1844 in the exponential growth phase were harvested. Total RNA was extracted from fresh cells of H. pluvialis using Trizol reagent (TaKaRa D9108B, Dalian, China) according to the user manual. RNA solutions were stored at −80°C, if not immediately used. First-strand cDNAs were synthesized from 2 μg total RNA with PrimeScript® RT Enzyme Mix I (TaKaRa DRR047A, Dalian, China) according to the manufacturer′s instructions.

The CODEHOP (Consensus-degenerate hybrid oligo-nucleotide primers) strategy (http://blocks.fhcrc.org/blocks/codehop.html) was used to generate degenerate primers [110,111] for the cloning of core partial sequences of HaeCYP97A/B/C basing on the highly conserved regions of predicted putative genes encoding CYP97 homologs from some green algae. The nucleotide sequences of the 3’- and 5’-ends of HaeCYP97A/B/C were amplified by the RACE method [112,113]. Gene-specific primers were designed from the amplified core cDNA sequence of HaeCYP97A/B/C and their 3’- and 5’-ends were obtained using SMARTTM RACE cDNA Amplification Kit (Clontech) according to the manual. All primers used in this study were listed in more [see Additional file 5: Table S3]. The PCR products were resolved by electrophoresis on 1% agarose gel. The fragment of interest was excised and purified using agarose gel DNA fragment recovery kit (TaKaRa D823A, Dalian, China). Finally, the fragment was cloned into PMD-18T vector (TaKaRa D101A, Dalian, China) and sequenced (Invitrogen, Beijing, China).

The full-length cDNA sequence of HaeCYP97A/B/C was spliced according to the RACE-PCR results by the SeqMan software of DNAStar 7.1 (DNASTAR Inc., USA). The theoretical molecular weight (Mw) and isoelectric point (pI) of HaeCYP97A/B/C protein were computed by ExPASy Compute pI/Mw tool [114]. Transmembrane regions were predicted by "DAS"-Transmembrane Prediction server [115]. Prediction of sub-cellular localization of the deduced amino acids was conducted using ChloroP and TargetP Servers [82,83].

Gene expression profiling: Real-time RT-PCR

Total RNA from different samples was extracted using Trizol reagent (TaKaRa D9108B, Dalian, China) according to the user manual. Total RNA was treated with RNase-free DNase I (Fermentas, Glen Burnie, MD) to remove any residual genomic DNA that might be carried through the extraction process. Nuclear acids were quantified by NanoDrop 2000c (Thermo Scientific, USA). The first-strand cDNA synthesis for quantitative Real-time RT-PCR (qRT-PCR) was obtained from 1 μg total RNA with PrimeScript® RT Enzyme Mix I (TaKaRa DRR047A, Dalian, China) according to the manufacturer’s instructions.

Gene-specific primers were designed with Primer5 and are listed in more [see Additional file 5: Table S3]. All primer pairs were initially tested by standard RT-PCR and the amplification of single products of the correct size was verified on 2% (w/v) agarose gels (data not shown). The actin gene [116] has been proven experimentally in this study [see Additional file 10: Figure S6]. The qRT-PCR amplifications were carried out in triplicate in a total volume of 20 μL according to the manufacturer’s instructions of SYBR® Premix Ex Taq™ (Tli RNaseH Plus) (TaKaRa DRR420A, Dalian, China). The qRT-PCR program was holding stage, 50°C for 20 s and 95°C for 10 min, followed by 40 cycles of 95°C for 15 s, 60°C for 1 min, and melt curve stage, 95°C for 15 s, 60°C for 1 min, 95°C for 30 s, and 60°C for 15 s. Real-time RT-PCR analysis was performed on an ABI fast 7500 Sequence Detection System (Applied Biosystems) following the protocol previously described using actin gene as the internal control. The relative steady state mRNA transcript levels were normalized to the respective actin transcripts. The 2^–ΔΔCT method [117] was used to analyze quantitative real-time PCR data based on the cycle threshold (C_T) values. The ΔΔC_T is represented as the following formula: ΔΔC_T = (C_T, target gene (test group) - C_T, actin gene (test group)) time x - (C_T, target gene (control group) - C_T, actin gene (control group)) times x, where x is the time of selected sample. Therefore, the relative expression level of chy genes has been normalized to control group (normal light regime).

Statistical analyses

All exposure experiments were repeated three times independently, and data were recorded as the mean with standard deviation (SD). For gene expression experiments, quantitative real-time PCR analysis was performed using software BioRAD iQ5. For each gene, the expressed as the mean ± SD (% control) was calculated using the (standard curve) approximation corrected for primer efficiency and normalized to housekeeping gene actin expression values.

Nucleotide sequencing and accession numbers

The cDNA nucleotide sequences of HaeCYP97A/B/C have been deposited and assigned the accession number AFR31786, AFR36909, and AFQ31612, respectively, in the EMB/GenBank/DDBJ database.