Different pairs of degenerate primers were designed on the basis of the conserved motifs present in lcy-e from microalgae. A partial cDNA fragment of 669 bp was isolated by PCR amplification using degenerate primers (lcy-e-1F and lcy-e-1R) (Table 1). A complete BLAST homology searches in the Genbank database showed that this fragment had enough similarity with the lcy-e gene from other species and provided sequence information for designing specific primers for rapid amplification of 5′ and 3′ cDNA ends (RACE-PCR). This analysis generated a full-length cDNA of 2204 bp, which contained an ORF of 1965 bp, 2-nucleotides of 5′-untranslated region (UTR), and a long 3′ UTR of 237 nucleotides. A typical algal polyadenylation signal TGTAAA [26] was present in the 3′ UTR at 82 nucleotides upstream from the beginning of the poly(A) tail. The predicted protein has 654 amino acids residues, with an estimated molecular weight of 71.1 kDa, a theoretical isoelectric point of 8.45 and an instability index of 43.6 (data obtained with ProtParam program, [27]). The differences between the C. zofingiensis lcy-e gene and the cDNA sequence were compared and revealed the presence of 8 exons and 7 introns. Exon size ranged between 124 (exon III) and 488 bp (exon I), and intron size between 159 (intron 1) and 331 bp (intron 3). Approximately 56% (2.2 out of 3.9 kb) of the Czlcy-e gene corresponded to exon sequences. In all introns, the consensus GT donor and AG acceptor sequences at the 5′ and 3′ termini were found (Figure 2).

To determine the copy number of lcy-e gene in the genome of C. zofingiensis, genomic DNA was digested with two different restriction enzymes (either NdeI or PstI) and subjected to Southern blot analysis at two different conditions of stringency, both in hibridization and washing (42 and 65 °C). Using a 1194-bp fragment of Czlcy-e as a probe, strong hybridization signals were obtained with both digestions. The digestion with PstI enzyme, which cuts once inside the probe sequence, showed two bands, while digestion with NdeI, which cuts at one extreme of the probe, exhibited only one band for the two conditions of stringency tested (Figure 3). These results suggested the presence of a single copy of the lcy-e gene in the genome of C. zofingiensis.

The BlastP search results demonstrated that the cloned CzLCYe showed the highest overall homology sequence with other LCYe from green algae, such as Chlamydomonas reinhardtii and Volvox carteri (identity 66%, similarity 77%) and Auxenochlorella protothecoides and Chlorella variabilis (identity 62%, similarity 74%). The GC content of the Czlcy-e coding region was 53%, which was lower than that of C. reinhardtii and V. carteri (63%) or of A. protothecoides (61%). The phylogenetic analysis of lycopene ε- and β-cyclases from green algae, cyanobacteria, plants and bacteria is illustrated in Figure 4. Analysis was conducted in MEGA5 using UPGMA method [28]. The predicted CzLCYe forms a cluster with the LCYe of green algae, which are phylogenetically close to LCYe of plants (~44% of identity, and 61% similarity). The degree of homology was lower with the LCYb of green algae, including C. zofingiensis, and plants (about 39% identity and 55% similarity) and with both cyanobacterial cyclases (CRTLe and CRTLb) (around 35% identity and 51% similarity). As other algal LCYe, CzLCYe was distantly related to bacterial CRTY cyclases, sharing with them only a few conserved motifs and about 23% identity. The characteristic Rossmann or dinucleotide binding fold and two cyclase motifs, present in LCYe and LCYb of green algae and plants, CRTLe and CRTLb of cyanobacteria and CRTY of bacteria, were also identified in the LCYe of C. zofingiensis, between amino acids 187 and 215, 405 and 421, and 480 and 489, respectively. In addition, a leucine in a region near the C-terminus, which was demonstrated to determine ε-monocyclase activity, was also identified in LCYe of C. zofingiensis. One basic amino acid histidine or lysine in that position was found in lycopene ε-cyclases from Lactuca sativa and Prochlorococcus marinus MED4 both of them showing ε-bicyclase activity [29,30].

Since microalgae and plants LCYe is located in the chloroplast membranes, we analyzed the CzLCYe sequence with different programs to determine both the presence of a signal peptide and transmembrane domains. The iPSORT program [31] predicted a putative plastid localization for the CzLCYe, and ChloroP 1.1 server [32] identified a chloroplast transit peptide at N-terminal end. Analysis with ProtScale [27] and TopPred [33] servers identified five deduced transmembrane domains of CzLCYe located between amino acids 1–21, 186–206, 472–492, 576–596 and 633–656 (data not shown). In addition, the predicted CzLCYe was highly hydrophobic. It had 46% of hydrophobic and non polar amino acids and 19% of charged amino acids.

In order to check the functionality of the recently isolated gene, the full-length ORF of Czlcy-e was amplified and cloned into pQE-80L vector under the control of the β-galactosidase promoter. The resulting plasmid (pQE-Czlcy-e) was introduced in E. coli engineered to accumulate either lycopene or δ-carotene as final products, due to the presence of the plasmids pAC-LYC or pAC-DELTA, respectively. HPLC analysis of carotenoids extracted from E. coli showed that cells containing pAC-LYC produced lycopene (Figure 5A), while cells co-transformed with both pAC-LYC and pQE-Czlcy-e accumulated δ-carotene (Figure 5B). δ-carotene was also the only carotenoid synthesized by E. coli cells containing pAC-DELTA or both pAC-DELTA and pQE-Czlcy-e (data not shown). As negative controls, E. coli co-transformed with either pAC-LYC or pAC-DELTA and empty pQE-80L were used, resulting in the accumulation of lycopene or δ-carotene, respectively (data not shown). These results indicated that CzLCYe could catalyze the formation of one ε-ring at one of the ends of lineal lycopene to yield δ-carotene, but not the conversion of δ-carotene into ε-carotene (with one ε-ring in each end of lycopene), exhibiting, therefore, monocyclase activity.

C. zofingiensis cells were grown photoautotrophically at low irradiance (20 μmol photon m⁻² s⁻¹) and a nitrogen replete concentration (nitrate 20 mM), as indicated in Experimental Section, until the middle of the exponential phase. Cells were then kept in the dark for 18 h, in order to make the mRNA levels come down to basal values. After this dark period, cells were subjected to either low (20 μmol photon m⁻² s⁻¹) or high irradiance (300 μmol photon m⁻² s⁻¹), under either nitrogen replete or nitrogen-deprivation conditions. Under conditions of nitrogen replete, nitrate concentration was measured daily (data not shown) and the nitrate consumed by the cells was added to the cultures. The evolution with time of transcriptional expression of the carotenogenic gene lcy-e as well as psy, pds, lcy-b, chyB and bkt as affected by irradiance and nitrogen availability was monitored by qRT-PCR. Changes in content of both the primary carotenoids lutein and α-carotene, β-carotene and violaxanthin and the secondary carotenoids astaxanthin, canthaxanthin and zeaxanthin were also determined in order to correlate transcript levels with the biosynthesis of them. As shown in Figure 6, under nitrogen replete concentrations and both at low and high irradiance, the relative transcript levels of lcy-e increased significantly, attaining 14-fold higher values than basal ones after 48 h, decreasing thereafter. On the contrary, nitrogen starvation did not affect the transcript levels of lcy-e, registering similar values to basal ones at both low and high irradiance. As previously described [25], the transcriptional expression of the Czlcy-b gene increased under nitrogen starvation (4-fold higher than basal levels), attaining a maximum after 24 h at low irradiance and decreasing later, similar maximal values being reached later, at 48 h, at high irradiance and being kept with time. Under nitrogen replete, no difference in maximal transcript levels either at high or low irradiance was observed. In the case of psy, mRNA levels at low irradiance and nitrogen replete were similar to those in the dark, increasing by about 5-fold when irradiance was raised from 20 to 300 µmol photons m⁻² s⁻¹, both under nitrogen starvation and nitrogen replete, attaining a peak after 24 h and decreasing slightly with time. At low irradiance, nitrogen starvation enhanced psy transcript levels by 3-fold with respect to the basal level after 96 h. The pds transcript levels, regardless of nitrogen availability, were two-fold higher at low irradiance after 24 h and three-fold higher at high irradiance after 5 h, when compared to basal levels, decreasing with time in both cases. The relative mRNA levels of chyB were about 4- and 7-fold higher than basal levels at low and high irradiance, respectively, under nitrogen replete, attaining a peak after 24 h. Nitrate deprivation increased slightly the transcript levels of chyB at low irradiance, showing similar values at high light intensity. In the case of bkt, mRNA levels at low irradiance and nitrogen replete were similar to basal values, increasing by about 4-fold when irradiance was raised from low to high light intensity, attaining a peak after 24 h and remaining constant with time. At low irradiance, nitrogen starvation enhanced bkt transcript levels by 3-fold with respect to the basal level after 24–48 h, remaining constant thereafter. The highest values were attained at high irradiance and nitrogen deprivation, reaching mRNA levels of bkt 7-fold higher as compared to the basal ones after 24 h, decreasing with time.

With regard to cell primary carotenoid contents such as lutein, α-carotene, β-carotene and violaxanthin (Figure 7), they accumulated at low irradiance and enough nitrogen availability, cell contents decreasing at high irradiance and more significantly under conditions of both high irradiance and nitrogen starvation. An opposite trend was observed for secondary carotenoids, such as canthaxanthin, zeaxanthin and astaxanthin, their synthesis being triggered at high irradiance and the highest accumulations of these carotenoids being registered under conditions of nitrogen starvation. In addition, cells were lacking these carotenoids at low irradiance, regardless of the nitrogen availability. Total carotenoids and chlorophylls a and b contents showed a similar response than that registered for lutein under the different conditions studied (data not shown).

The green microalga strain Chlorella zofingiensis SAG 211-14 (recently classified as Chromochloris zofingiensis, [50]) was obtained from the Culture Collection of Göttingen University (SAG, Germany). This microalga was maintained and grown photoautotrophically in Arnon medium [51] modified to contain 4 mM K₂HPO₄ and 20 mM NaNO₃, at 25 °C under continuous illumination (50 µmol photons m⁻² s⁻¹), except where indicated. The liquid cultures were continuously bubbled with air, supplemented with 1% (v/v) CO₂ as the only source of carbon. For the expression experiments, cells were grown in Roux flaks of 1 L capacity laterally and continuously illuminated with mercury halide lamps at either 20 (low irradiance) or 300 µmol photons m⁻² s⁻¹ (high irradiance) either in the presence or in the absence of nitrate. To keep constant the nitrate in the medium, the nitrate concentration was measured daily by HPLC according to Cordero et al. [25], and the nitrate consumed by the cells was added to the cultures. The light intensity was measured at the surface of the flasks using a LI-COR quantum sensor (model L1-1905B, Li-Cor, Inc. Lincoln, NE, USA).

Escherichia coli DH5α and BL21 strains were used as the hosts for DNA manipulation and for heterologous expression of lcy-e gene, respectively.

DNA and total RNA were isolated using DNeasy Plant Mini Kit and RNeasy Plant Mini Kit (Qiagen, Düsseldorf, Germany), respectively. For Quantitative Real-Time PCR analysis (qRT-PCR), first-strand cDNA synthesis was obtained from total RNA treated with DNase as recommended by the manufacturer, by using the SuperScript First-Strand Synthesis System (Invitrogen, Barcelona, Spain) primed with oligo(dT)₁₈ according to the manufacturer’s instructions.

For isolating the lcy-e cDNA from C. zofingiensis, degenerate primers were designed from the conserved regions of LCYe from different species of algae. The PCR product was cloned in the pGEM-T vector (Promega, Madison, WI, USA) according to the manufacturer’s manual and then sequenced. The cDNA fragment obtained corresponding to partial lcy-e clone provided sequence information for the designing of gene-specific primers for amplification of 5′ and 3′ cDNA ends by RACE-PCR. All reactions were performed with kits according to the manufacturer’s instructions (Smart RACE cDNA Amplification Kit, Clontech, Mountain View, CA, USA). 5′ and 3′ RACE products were cloned into pGEM-T vector and sequenced. Specific primers were synthesized for genomic DNA amplification based on cDNA sequence. The primers sets used in this study are listed in Table 1.

The Czlcy-e cDNA and genomic DNA sequences have been registered in the EMBL database under the accession numbers HE664109 and HE664108, respectively.

Genomic DNA was digested with NdeI and PstI, which showed one recognition site in the probed region of the lcy-e gene. The probe was prepared by amplifying genomic DNA with the primers Czlcy-e-S-F and Czlcy-e-S-R, resulting in a 1194-bp fragment of Czlcy-e gene. The digested DNA was transferred to a Hybond-N membrane (GE Healthcare, Little Chanfont, UK) by capillary transfer and hybridized with the ³²P labelled DNA probe at both low and high stringency overnight. After hybridization, the radioactivity of the membrane was monitored by the Cyclone Phosphor System (Perkin Elmer, Waltham, MA, USA).

The Czlcy-e ORF was amplified by PCR with the primers pQE-lcy-e-F and pQE-lcy-e-R, which were designed to contain SacI and HindIII restriction sites, respectively, and cloned into pQE-80L expression vector (Qiagen) resulting in plasmid pQE-Czlcy-e, which carries ampicillin resistance. The plasmids pAC-LYC and pAC-DELTA, kindly donated by Prof. Cunningham, carried the carotenoid pathway genes responsible for the synthesis of lycopene (crtE, crtB, and crtI of Erwinia herbicola) and δ-carotene (crtE, crtB, and crtI of E. herbicola + lcy-e of Arabidopsis), respectively [52,53]. Transformation of E. coli BL21 (DE3) with pQE-Czlcy-e and/or one of the two plasmids pAC-LYC or pAC-DELTA (both carrying chloramphenicol resistance) was made by electroporation. Transformed cells were plated on Luria-Bertani (LB) [54], supplemented with 100 μg mL^-1 ampicillin and/or 40 μg mL⁻¹ chloramphenicol, and grown at 37 °C for 1 day. The inducer isopropyl-β-D-1-thiogalactopyranoside (IPTG) was added at a final concentration of 1 mM.

The mRNA relative abundance of psy, pds, lcy-b, lcy-e, chyB and bkt genes of C. zofingiensis was examined by qRT-PCR on an IQ5 Real-Time PCR Detection System (BioRad, Hercules, CA, USA), according to Cordero et al. (2010) [25]. In each experiment, a series of standard dilutions containing a specific concentration of a PCR fragment or a cDNA template was amplified in 20 µL of reaction containing 1× SYBR Green PCR Master Mix (Quantimix Easy SYG kit, BioTools B&M Labs, Madrid, Spain) and corresponding primers (Table 1). After heating at 95 °C for 10 min, cycling parameters were: 40 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. Finally, the specificity of the qRT-PCR products was confirmed by performing a melting temperature analysis at temperatures ranging from 55 to 95 °C at 0.5 °C per min and also by electrophoresis on a 2% agarose gel. Data were captured as amplification plots. Transcription levels of the target genes were calculated from the threshold cycle by interpolation from the standard curve. To standardize the results, the relative abundance of actin gene was also determined and used as the internal standard [45,55]. All calculations and statistical analyses were performed as described in the IQ5 Optical System Software 1.0 (BioRad). The complete experiments (RNA isolation, cDNA synthesis followed with qRT-PCR) were independently repeated twice, and the data were averaged.