Introduction

Chlorella vulgaris NJ-7 was isolated from an algal mat residing on the surface of a rock nearby the Zhongshan Station (China) of Antarctica where the temperature was almost 9.9°C in the summer and can fall to −45.7°C in severe winters, a range of 55°C. Compared with other well studied cold-adapted chlorophytes that failed to grow at temperatures above 20°C (Morgan-Kiss et al. 2005, 2008), C. vulgaris NJ-7 grow well at temperature between 4 and 30°C (Hu et al. 2008). Thus, this newly isolated photopsychrotrophile has retained a higher versatility in response to environmental change.

Cold-adapted organisms are evidently adapted to the extreme environmental conditions found within the ice and have evolved a number of adaptive mechanisms to counteract the effects of temperature change (Stefanidi and Vorgias 2008). But more importantly, they are able to acclimate rapidly to the changing physicochemical conditions within the brine channel systems. Even small temperature changes of 1–2°C greatly influence the structure of these microhabitats as well as the chemical and physical properties of the ice. Furthermore, organisms that inhabit the interstices and underside of sea ice are exposed to wide variations of salinity, particularly during the periods of brine drainage and ice melting. The capability of responding to rapid changes in a host of external stresses, including temperature and salinity, is therefore an intrinsic characteristic of life in sea ice and just as important as being able to survive freezing temperatures (Mock and Thomas 2005). C. vulgaris NJ-7 would make the first candidate for the research on the mechanism of wide acclimation to dramatically changing habitat.

In this study, we analyzed the fatty acid composition of C. vulgaris NJ-7. The extremely high rate of Δ12 unsaturated fatty acids (UFAs) allowed us to isolate a Δ12 desaturase. The function of CvFAD2 was determined by expression in Saccharomyces cerevisiae, in particular, the effects of temperature and salinity on the expression profile of CvFAD2 were investigated. The results provide new insights for the acclimation of organisms to low temperature and high salinity.

Materials and methods

Strains and growth condition

C. vulgaris NJ-7 was presented from Pro. Xu Xudong (Institute of Hydrobiology, Chinese Academy of Sciences). The green alga was grown in batch cultures in a salt medium (BG11) (Stanier et al. 1971), and initially maintained at a temperature of 25°C with 12 h of incandescent light and 12 h of darkness.

Fatty acid analysis

As a first step in defining the acclimation to changing Antarctic ambience, we analyzed the fatty acids composition of C. vugaris NJ-7. Algae were cultivated to logarithmic phase at 25°C, and then some algae were harvested. Others were divided into five treatments: 25°C, 0% (w/v) NaCl (control); 15°C, 0% NaCl; 4°C, 0% NaCl; 25°C, 3% NaCl; 25°C, 6% NaCl. After an interval of 36 h, algae were harvested and the total lipids were extracted and analyzed according to the method described by Hsiao et al. (2007).

Nucleic acid manipulation

C. vulgaris NJ-7 in exponential growth phase were harvested by centrifugation at 5000 rpm for 5 min at 4°C and washed with sterile, distilled, deionized water to remove salts before they were centrifuged again. The dried mass was frozen in liquid nitrogen, ground with mortar and pestled into a fine powder. Total RNA was extracted from the powder according to the method of Li (Li et al. 2008b). The RNA samples were used for RACE and real-time PCR after DNaseI treatment to remove DNA. The first-strand cDNA was synthesized with RT-PCR kit (Qiagen) according to the manufacturer’s instructions. 500 ng of total RNA was used in a 20 μl reaction system. Controls received water instead of reverse transcriptase to assess any contamination from genomic DNA as described by Zhou et al. (2007).

PCR with degenerate primers

First-strand cDNA was used as a template for PCR. Available sequences of Δ12-acyl-lipid desaturases from the GenBank database (Benson et al. 2005) were used to design two degenerate primers, F1/R1 (Table 1). F1 and R1 were corresponding to the highly conserved regions HRRHHSN and DTHVAHH.

Table 1 Primers used in experiment
Full size table

PCR contained 1× PCR Ex-Taq buffer (Mg2+ plus), 0.2 mM of each dNTP, 0.4 μM of each primer, and 1 U of Ex-Taq (Takara) per 25 μl reaction. Amplifications run on a Tpersonal Thermocycler 118 (Biometra, Germany) using the following parameters: an initial 5 min denaturation step at 94°C, followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min, finally extended at 72°C for 7 min. The PCR fragments were cloned into pMD-18-T vector (Takara) and sequenced.

Full-length cDNA and genomic sequences isolation

The nucleotide sequences of the 3′ and 5′-ends of CvFAD2 were amplified by the RACE method (Sambrook et al. 2001). Amplifications were carried out using the SMART RACE cDNA Amplification Kit (BD-Clontech) according to the manual. 3′ and 5′ RACE were performed using gene-specific primers (3GSP1, 3GSP2, 5GSP1 and 5GSP2, Table 1). PCR was carried out using the formulation described in the manual. The program was performed as follows: 5 min at 94°C, followed by 30 s at 94°C, 3 min at 72°C, for 5 cycles, 30 s at 94°C, 30 s at 70°C and 3 min at 72°C, for 5 cycles, 30 s at 94°C, 30 s at 68°C and 3 min at 72°C, for 25 cycles, finally extended at 72°C for 7 min. Nested PCR was carried out using the nested universal primers and gene-specific primers. Secondary PCR fragments were subcloned into pMD-18-T vector and nucleotide sequences were determined subsequently.

Based on the information of 3′ and 5′-ends sequences, two gene-specific primers F2 and R2 (Table 1) were designed to amplify full-length cDNA. PCR was performed with the GC-Rich PCR system (Takara) using 1× GC buffer (Mg2+ plus) 0.2 mM dNTP, 0.5 μM F2 and R2, 1 U of LA-Taq per 25 μl reaction. Reaction conditions were as follows: 5 min denaturation at 94°C, followed by 30 cycles of 94°C for 30 s, 58°C for 30 s and 72°C for 90 s, finally extended at 72°C for 7 min. The PCR fragments of the expected length (1.2 kb) was cloned into pMD-18-T vector and sequenced. For genomic sequence cloning, two sets of primers were designed and named as F3, R3, F4 and R4 (Table 1). An overlapping region was obtained for the splicing of the whole genomic sequence.

Sequence analysis

Sequence analysis was carried out according to the method of Li (Li et al. 2008a). A BLAST search program (http://www.ncbi.nlm.nih.gov/blast/) was used. Multiple sequence alignment of putative CvFAD2 was performed with Clustalx1.81 (Thompson et al. 1997). The distribution of the hydrophobic amino acids was analyzed using Kyte–Doolittle hydropathy scale (http://www.expasy.ch/cgi-bin/protscale.pl) (Kyte and Doolittle 1982). Transmembrane (TM) regions of the protein were predicted with transmembrane hidden Markov model (TMHMM) (http://www.cbs.dtu.dk/services/TMHMM/) (Moller et al. 2001). The signal sequence analysis was done using a signal peptide prediction server (http://www.cbs.dtu.dk/services/SignalP-3.0/) (Bendtsen et al. 2004).

Phylogenetic affiliation of the CvFAD2 was examined by comparing it to the confirmed and putative plastidial and microsomal Δ12 fatty acid desaturases (FAD) obtained from public database. Amino acid sequences were aligned using ClustalX program with the implanted BioEdit (Chenna et al. 2003; Thompson et al. 1997). Phylogenetic trees were constructed using neighbor-joining methods, as implemented in the program MEGA4 (Tamura et al. 2007). Minimum evolution and maximum parsimony analyses were also performed. Bootstrap support was estimated using 1000 replicates for distance analyses. Default program parameters were used.

Quantification real-time PCR

For real-time PCR, algae were cultivated to logarithmic phase at 25°C, and were then divided into five treatments: 25°C, 0% (w/v) NaCl (control); 15°C, 0% NaCl; 4°C, 0% NaCl; 25°C, 3% NaCl; 25°C, 6% NaCl. After an interval of 36 h, algae were harvested.

Two pairs of gene-specific primers (Table 1) were designed according to the CvFAD2 cDNA (F5 and R5) and β-actin sequences (F6 and R6). Standard curves for CvFAD2 and β-actin genes were generated by tenfold serial dilutions (105–1010 copies/μl) of each plasmid DNA. PCR products were quantified continuously with ABI 7900 Real-Time PCR System (Applied Biosystems) using SYBR Green fluorescence (Takara) according to the manufacturer’s instructions. Control reactions were amplified with RNA template but without reverse transcriptase. The relative amounts of CvFAD2 RNA were normalized to the respective β-actin transcripts. The 2−△△CT method (Livak and Schmittgen 2001) was used to analyze quantitative real-time PCR data.

Expression in S. cerevisiae

Two specific primers were used to obtain the full-length open reading frames. F7 and R7 (Table 1) corresponds to the nucleotide sequences of start and stop codons (in bold face), respectively. The 5′ end of the F7 and R7 contains an EcoRI or an XhoI restriction site (italicized) to facilitate subsequent manipulation. The amplified cDNA was digested and subcloned into the EcoRI–XhoI site of the expression vector pYES2.0 (Invitrogen) to generate a plasmid designated pYFAD2. The sequence orientation and identity were confirmed by sequencing. S. cerevisiae INVSc1 was transformed and selected according to the manual.

Heterologous expression of CvFAD2 was induced under transcriptional control of the yeast GAL1 promoter. Yeast cultures were grown to logarithmic phase at 30°C in synthetic minimal medium (SC-Ura). The cells were induced as described by Robert et al. (2009). Cultures were incubated at 15°C for a further 48 h prior to harvesting. The cells were harvested by centrifugation, and washed three times with sterile distilled water and then dried by lyophilization. The total lipids were extracted and analyzed according to the method described by Hsiao et al. (2007).

Nucleotide sequencing and accession number

The nucleotide sequences of CvFAD2 genomic and cDNA have been deposited and assigned accession number GQ175318 and EU596474 in the EMBL/GenBank/DDBJ database.

Results

Fatty acid distribution

The amount of 18:2 and 18:3 UFAs were 33.08 and 24.99% in C. vulgaris NJ-7 (Table 2). On the other side, they were 5.88 and 21.59% in Chlamydomonas raudensis Ettl (UWO241) and 19.25 and 11.50% in Chlorella BI, respectively (Morgan-Kiss et al. 2008). The total amount of the Δ12 UFAs of C. vulgaris NJ-7 (58.07%) is 2.11- and 1.89-fold compared with C. raudensis Ettl (UWO241) and Chlorella BI, respectively.

Table 2 Comparision of fatty acid distribution of C. vulgaris NJ-7, C. raudensis and Chlorella BI
Full size table

The major composition was 16:0 which shared half of fatty acids in Chlorella BI. On the other hand, the main components were 14:0 and 18:3 in C. raudensis and their amounts were closely equal to each other. Although all three microalgae were Antarctic, the difference of fatty acid composition was impressive. A question must be explicitly addressed here in understanding of the role of lipids in fine tuning membrane function. At low temperatures, a more fluid membrane can usually be attained by increasing the degree of fatty acid unsaturation or branching and by decreasing the fatty acid chain length (Feller 2007). These low melting-point fatty acids lower the gel–liquid crystalline-phase transition temperature of the lipid and therefore maintain proper fluidity at low temperatures. For the sake of maintaining fluidity of the membrane, different strategies were adopted. The strategy for C. vulgaris NJ-7 is primarily to increase the content of straight-chain UFAs. In contrast, the strategy for Chlorella BI appears to be mainly dependent upon increasing the amount of short chain fatty acids. C. raudensis may regulate membrane fluidity by both strategies.

Fatty acid distributions under different treatments were studied in order to know the linkage between fatty acids alteration and stresses. Total amount of Δ12 fatty acids (16:2, 16:3, 18:2, 18:3) increased at lower temperature. On the other hand, less Δ12 fatty acids were observed in C. vulgaris NJ-7 at 3 and 6% NaCl (Table 3). However, 18:2 and 16:2 increased in response to exposure to lower temperature and higher salinity (Fig. 1).

Table 3 Comparision of fatty acid distribution of C. vulgaris NJ-7 at different temperature or salt concentration
Full size table
Fig. 1
figure 1

Comparisons of fatty acid distributions of C. vulgaris NJ-7 at different temperature or salt concentration. 1 16:2, 2 18:2, 3 16:2 + 18:2

Full size image

Based on these observations, we tentatively put forward that Δ12 UFAs were important players in the rapid acclimation of C. vulgaris NJ-7 to the Antarctic environment. Thus, a research on its Δ12 FADs expression profile will help to illuminate the mechanism of the process.

cDNA and genomic DNA sequences identification

A 542 bp cDNA fragment was generated by RT-PCR with degenerate primers. The fragment shared a 73% sequence similarity to Δ12 FAD genes from Chlorella C-27 and 57% to Chlamydomonas reinhardtii. It indicated a partial putative FAD2 was isolated from Antarctic C. vulgaris NJ-7. Gene-specific primers were further designed to obtain the full-length CvFAD2. A 517 bp 5′-RACE product and a 1068 bp 3′-RACE product were amplified. The nucleotide sequences of both products from RACE experiments shared an identical 35 and 60 bp sequence overlap on flanking regions of the partial putative CvFAD2 cDNA fragment, suggesting that these fragments are portions of the same gene. Sequence analysis revealed that the cloned CvFAD2 cDNA was 2032 bp in length, which contained a 1158 bp open reading frame, a 50 bp 5′-untranslated region, and a 824 bp 3′-untranslated region with the characteristics of the poly (A) tail. An ATG translation initiation codon was identified in the sequence of the 5′ terminus (51–53 bp), and a TAA termination codon was found 1158 bp downstream of the initiation site.

To elucidate the genomic organization, the genomic DNA sequence of CvFAD2 was cloned and aligned with cDNA sequence. The result indicated that CvFAD2 consisted of eight exons. The sizes of the exons and introns are summarized in (Table 4). Seven are canonical GT/AG splicing signals at the ends of the putative introns. However, the sixth intron does not conform to that pattern. Rare splice donor consensus (GC) had been found, instead of (GT) after exon ending at position 2542. Spliceosomal introns generally begin with GT and end with AG dinucleotide motifs that are referred to as donor and acceptor splice sites, respectively. However, introns with noncanonical splice sites have been identified in some organisms (Burset et al. 2000; Davis et al. 2000; Bon et al. 2003).

Table 4 Gene structures of C. vulgaris NJ-7 FAD2
Full size table

Comparison with other desaturases

CvFAD2 encoded 385 amino acid residues with an estimated molecular mass of 45.5 kDa and a theoretical isoelectric point of 6.87. A Blast search revealed that the primary structure of CvFAD2 showed high similarity to known microsomal Δ12 FAD, with a 75% identity to Chlorella C-27 (BAB78716, Suga et al. 2002), 57% to C. reinhardtii (XP_001691669), 53% to Punica granatum (AAO37754), 52% to Jatropha curcas (ABA41034) and 52% to Arachis hypogaea (AAY67653).

The distribution of the hydrophobic amino acids of CvFAD2 was typical for a membrane protein, as identified using the Kyte–Doolittle hydropathy scale (Kyte and Doolittle 1982). The prediction of TM helices by TMHMM indicated that the deduced protein contained four hydrophobic domains between amino acids 54–76, 86–108, 178–200 and 250–272, which would be long enough to span the membrane bilayer twice, with the N-terminus facing the cytosol (Moller et al. 2001).

A multiple sequence alignment of CvFAD2 against similar Δ12 desaturase highlighted conserved motifs including HRRHHSN and DTHVAHH against which the degenerate primers were designed (Fig. 2). Otherwise three conserved histidine boxes, HXCGH (109–113) and HXXHH (145–149, 321–325), were found to be located in hydrophilic regions separated by long hydrophobic or membrane spanning regions. The histidine-rich motifs, highly conserved among membrane-bound acyl-CoA and acyl-lipid desaturases, were proposed to form the potential diiron active site (Shanklin et al. 1994).

Fig. 2
figure 2

Multiple alignments of amino acid sequences of C. vulgaris NJ-7 Δ12-desaturase and other Δ12 desaturase. The deduced amino acid sequence of the C. vulgaris NJ-7 Δ12-desaturase was subjected to BLAST analysis. The closest relatives were then selected and aligned using ClustalW. The conserved motifs HRRHHSN and DTHVAHH are highlighted in boxes and asterisks denote the histidine motifs. Sequences shown are Δ12 FADs: BAB78716 Chlorell C-27, EU596474-CvFAD2, XP_001691669 C. reinhardtii, AAO37754 P. granatum, ABA41034 J. curcas

Full size image

Phylogenetic analysis using deduced amino acid sequences of several other Δ12 FADs is shown (Fig. 3). All tree topologies are highly congruent. Two major clades are very well supported based on bootstrap values. All of the plastidial Δ12 FADs are grouped. CvFAD2 is clustered within microsomal Δ12 FAD group. Based upon its sequence similarity and phylogenetic analysis, it implies that CvFAD2 encodes a microsomal Δ12 FAD. Further evidence to support this proposal is that CvFAD2 has a truncated N-terminal (data not shown) whereas the plastidial Δ12 FADs have redundant N-terminals signal peptides for subcellular location predicted by hidden Markov models (Bendtsen et al. 2004).

Fig. 3
figure 3

Phylogenetic tree for Δ12 FAD and related desaturases based on amino acid sequences. Bootstrap analysis of 1000 randomized sequence replicates was performed. Sequences shown are: NP_194824 Arabidopsis thaliana, ABI73993 Descurainia Sophia, AAW63039 Olea europaea, ABD58898 Mesostigma viride, XP_001693068, XP_001691669 C. reinhardtii, ABA41034 J. curcas, AAO37754 P. granatum, EU596474-CvFAD2, BAB78716 Chlorell C-27

Full size image

Functional analysis in S. cerevisiae

Heterologous expression in yeast was used to confirm Δ12 regioselectivity and function of CvFAD2. Both pYFAD2 and empty vector, pYES2 (control), were transformed into the S. cerevisiae INVSc1. Total lipids of the transformants were put through GC–MS analysis. The results showed a novel fatty acid peak from pYFAD2, which was absent from the control (Table 5). The novel fatty acid was 18:2, by comparison of the retention time to FAME standard mixtures (Sigma). The result indicates that CvFAD2 encodes a Δ12-FAD, which can convert 18:1 into 18:2 in yeast. This is consistent with the recent reports where Δ12-FADs from the Lentinula edodes (Sakai and Kajiwara 2005) and Rhizopus arrhizus (Wei et al. 2004) recognize only one substrate (18:1) in yeast. Conversely, Δ12 FAD from Phaeodactylum tricornutum (Domergue et al. 2003), Sapium sebiferum (Niu et al. 2007) and Chlorella C-27 (Suga et al. 2002) were found to recognize two substrates (16:1 and 18:1) though 18:1 was the preferred substrate. It may indicate different TM topologies of these Δ12 FADs (Wei et al. 2004).

Table 5 Composition of the major fatty acids (% w/w) of pYES2 and pYFAD2 yeast transformants by GC–MS analysis
Full size table

Effects of temperature and salt stress on CvFAD2 expression

The gene expression patterns of CvFAD2 in Antarctic C. vulgaris NJ-7 were analyzed by quantification real-time PCR (qRT-PCR). β-actin was used as a reference for total RNA. β-actin PCR product was not detected when reverse transcriptase was omitted, indicating that the RNA template was free of genomic DNA. The data reveal that temperature and salt concentration regulate the accumulation of CvFAD2 mRNA. Transcripts accumulation increased 2.2- and 3.9-fold when the algae at 25°C were transferred to 15 or 4°C (Fig. 4a). It is different from plastidial Δ12 FAD (CvFAD6) which remained constantly when the algae at 25°C were transferred to 15°C (unpublished results). It signifies that they play different roles in temperature acclimation. Microsomal Δ12 FAD initiated transcripts accumulation in response to major or intermediate temperature shifts whereas plastidial one was only upregulated in response to major shifts. On the other hand, CvFAD2 was expressed abundantly under salt stress, 1.7-fold increase at 3% NaCl and 8.5-fold at 6% NaCl (Fig. 4b).

Fig. 4
figure 4

Expression of CvFAD2 at different temperature (a) or salt concentration (b). The qRT-PCRs were employed to quantify the relative amounts of CvFAD2 transcripts in 25°C (1), 15°C (2), 4°C (3), 3% NaCl (4) and 6% NaCl (5). Data are presented as the mean ± SD (n = 3)

Full size image

Discussion

The Antarctic C. vugaris NJ-7 grew well when exposed to wide variations of temperature compared with other characterized Antarctic microalgae (Hu et al. 2008; Morgan-Kiss et al. 2005, 2008). In order to realize the mechanism, we analyzed its fatty acid composition as the first step on account of the close correlation between fatty acid and temperature. As expected, C. vugaris NJ-7 overwhelmingly exceeds others in Δ12 UFAs constitution. Thus, we proposed that C. vulgaris NJ-7 take advantage of the increased content of straight-chain UFAs instead of short chain fatty acids to acclimate the extreme habitats. Therefore, Δ12 FAD may be an important player in the psychology modulation to counteract the effects of the cold. So we isolated CvFAD2 from C. vulgaris NJ-7 using RT-PCR with degenerate primers designed from conserved motifs and RACE method. The isolated protein was found to contain features characterizing Δ12 desaturases. It was grouped with the microsomal Δ12 desaturases according to the phylogenetic analysis. The function of CvFAD2 was validated by the expression in S. cerevisiae. 18:1 was utilized as precursor and converted to 18:2 by CvFAD2 in vivo. These results suggested that this gene encoded microsomal Δ12 FAD. CvFAD2 genomic organization suggested an alternative splicing since the rare splice sites (GC) in the sixth intron was found. The clarification to noncanonical splicing has the potential to facilitate accurate gene prediction further.

Based on the qRT-PCR analysis, the expression profile of CvFAD2 was characterized. Transcript was increased when the alga was transfer from 25 to 15°C or 4°C after 36 h. It was different from the cold response of CvFAD6 which remained a constant accumulation from 25 to 15°C (unpublished results). It suggested that CvFAD2 and CvFAD6 responded differently to changing temperature. CvFAD2 expression was activated in intermediate and major temperature shifts whereas CvFAD6 was only activated in major shifts. On the other hand, CvFAD2 also engaged in the adaption of high salinity stress. Changes in cellular membranes are adaptations to hypersalinity, that have been previously reported in yeasts (Turk et al. 2007; Gostincar et al. 2009), fungi (Turk et al. 2004, 2007) and cyanobacteria (Allakhverdiev et al. 2001). The restructuring of membrane lipid composition is one of the adaptations to high concentration of salt, which is mainly achieved by increasing the unsaturation of its phospholipid fatty acids. The possible reason was as that increase in unsaturation of fatty acids in membrane lipids significantly enhanced the tolerance of the photosynthetic machinery to salt stress. The enhanced tolerance was due both to the increased resistance of the photosynthetic machinery to the salt-induced damage and to the increased ability of cells to repair the photosynthetic and Na+/H+ antiport systems (Allakhverdiev et al. 2001).

Although temperature-dependent expression of FAD2 genes have been previously reported (Suga et al. 2002), to our knowledge, this is the first characterized FAD2 with the response to both low temperature and high salinity. It should be noted that this study has only examined the accumulation of CvFAD2 in transcription level, further study should be focus on the post-transcription and translation levels of global differentially expressed proteins, and stability of fatty acid biosynthesis genes to demonstrate the detailed involvement of FAD profiles in the process of acclimation. Nevertheless, this desaturase along with the recently characterized genes from the same organism represents an attractive biotechnological resource for development of an agricultural viable breed as highlighted in recent publications (Tran et al. 2009; Sengupta and Majumder 2009; Yokotani et al. 2009). In addition, it can be used for conferring high oleic acid profile on the seed oil and for producing high oleic plant lines. Further mining of the genome of Antarctic Chlorella NJ-7 should lead to identification of gene encoding these additional enzyme activities (de Pascale et al. 2008) and shed light on the acclimation of organisms to low temperature and high salinity.