A Desaturase Gene Involved in the Formation of 1,14-Nonadecadiene in Synechococcus sp. Strain PCC 7002

Abstract

The marine cyanobacterium Synechococcus sp. strain PCC 7002 synthesizes two alkenes, 1-nonadecene and 1,14-nonadecadiene. Whereas the genetic basis for the biosynthesis of the terminal double bond in both alkenes has been characterized, the origin of the internal double bond in 1,14-nonadecadiene has not. In this study, we demonstrate that a gene encoding an uncharacterized desaturase is involved in the formation of the internal double bond of 1,14-nonadecadiene. Further, at low temperatures, the desaturase gene is essential for growth, and in wild-type cells the levels of 1,14-nonadecadiene increase relative to that of cells grown at 38°C. These data suggest that 1,14-nonadecadiene plays a role in responding to cold stress.

INTRODUCTION

Hydrocarbon biosynthesis is a common trait among both eukaryotes and prokaryotes that has gained significant attention for use in developing sustainable alternatives to petroleum (1–3). Although it has been known for decades that cyanobacteria accumulate hydrocarbons with C₁₇ to C₂₀ chains (4, 5), the cellular function of these molecules and the biosynthetic pathways responsible for their production remain poorly understood. A recent survey of sequenced cyanobacteria discovered that all strains possess one, but not both, of two pathways for producing saturated or unsaturated straight-chain hydrocarbons (6), indicating that hydrocarbons play an important role in cyanobacterial physiology. The majority of strains harbor the FAR/ADO pathway that was first characterized in Synechococcus elongatus PCC 7942 (7). The pathway synthesizes saturated alkanes (heptadecane and pentadecane) from acyl-acyl carrier proteins (acyl-ACP) via the action of an acyl-ACP reductase and an aldehyde-deformylating oxygenase. Conversely, a smaller percentage of cyanobacteria produce long-chain alpha-olefins via the Ols pathway, first characterized in Synechococcus sp. strain PCC 7002 (8). The Ols pathway is comprised of a single, multidomain enzyme (Ols) possessing modular organization similar to that of a polyketide synthase. Based on the protein's domain architecture and experimental feeding studies, Ols was predicted to produce alpha-olefins via elongation and decarboxylation of C₁₈ acyl-ACP, the terminal product of fatty acid biosynthesis and substrate for phospholipid synthesis (8). Unlike other cyanobacterial strains that possess a more complex hydrocarbon profile in terms of length and degree of unsaturation (4, 5, 9), Synechococcus sp. strain PCC 7002 synthesizes only two alkenes: 1-nonadecene (C_19:1) and 1,14-nonadecadiene (C_19:2). The only difference between these compounds is the presence of an internal double bond at position 14 of the C_19:2 hydrocarbon. Whereas the biosynthesis of the terminal double bond in both hydrocarbons was explained by the decarboxylation and dehydration reactions performed by the Ols domains, the biosynthesis of the internal double bond in 1,14-nonadecadiene was not. In this study, we linked the presence of this internal double bond to a gene predicted to encode a desaturase. Further, we demonstrated an increase in C_19:2 abundance as an inverse function of temperature. This finding suggests that the hydrocarbons play a role in responding to cold stress similar to the way unsaturated fatty acid content in the cell membrane is modulated in many bacteria (10).

MATERIALS AND METHODS

Reagents, media, and growth conditions.

Enzymes and reagents were purchased from New England BioLabs or Fisher Scientific unless otherwise noted. Oligonucleotides used in this study were purchased from Integrated DNA Technologies, Inc., and are listed in Table 1. Isolation of genomic DNA and purification of DNA fragments were performed with commercial reagents (Promega, Qiagen).

TABLE 1.

Oligonucleotides used in this study^a

No.	Oligonucleotide	Sequence (5′→3′)
1	desA-a2	CCGTTCTGCAGGTTTCTTGGCGCAAGGGTTACAGCTTCC
2	desA-a1	GCAACCATGGGAAACCCAACGCAAGG
3	desA-b2	CCGTTGGATCCGTACGCTTCCATACCATGTTCACCAATATCG
4	desA-b1	CCTCACAGGTTCGGCCTACAGTGG
5	desB-a2	CCGTTCTGCAGCTTTACAACCCCTAATCCGCCTTTATTCATTTCC
6	desB-a1	CAGATCGAGGGGAACCTGGTTTGCG
7	desB-b2	CCGTTGGATCCCAACAACGCCTTGCAGAAAATCCCCAGC
8	desB-b1	CCAGTTTTAACAGACCTTGGGTAAAGGCTTC
9	desC-a2	CCGTTCTGCAGCCTTGTCACCTACGGCGAAGGTTGG
10	desC-a1	GGAATACACTGACGAATACCGCGATGGG
11	desC-b2	CCGTTGGATCCGGTAATTCCAATGCCGCCGGTAATCCAGT
12	desC-b1	GGTTTCTGGGCGAATTGCGATCTTGAGG
13	desE-a2	CCGTTCTGCAGATTGTTGTCCCCAAAGGGAAATATCTATTCCG
14	desE-a1	GACCCAGAAAACCCCGTAGAATAATCACG
15	desE-b2	CCGTTGGATCCAATGCTTTCCTAACGAGTTGAGAATATCTTCTATG
16	desE-b1	CGTCGATTTTGCCTCATTAATTTAGTTAAAGCAGC
17	desF-a2	CCGTTCTGCAGGAATAAATTTGCGTTTGATCATTACCGCCAATTC
18	desF-a1	GGTTGAAACCATTTAGGGAAACCCATACTG
19	desF-b2	CCGTTGGATCCGAATTCTCAAACAATAGAACAAGACAAAGGGGAATATC
20	desF-b1	CGAGGCAGGTTTTGAGAGCGTCAAC
21	desE-US-Fw	TATGCACATATGCGTCGATTTTGCCTCATTAATTTAGTTAAAGCAGC
22	desE-US-Rv	GCAGTTTCATTTGATGCTCGATGAGTTTTTCTAAGCTTTCCTAACGAGTTGAGAATATCTTCTATGAAACCG
23	KM-Fw	TTAGAAAAACTCATCGAGCATCAAATGAAACTGC
24	KM-Rv	GGACTCTTCTCTACAGGTGGGTATAGATTTGTTAAGCTTTGGCAGGATCCGGCTGCTAACAAA
25	Cpc-prom-Fw	CTTAACAAATCTATACCCACCTGTAGAGAAGAGTCC
26	Cpc-prom-Rv	GGGTCAAGAACGTTGCTGTAATGCGTCATGGAATTAATCTCCTACTTGACTTTATGAGTTGGG
27	desE-DS-Fw	ACGCATTACAGCAACGTTCTTGACCC
28	desE-DS-Rv	CTGCCGCCAGGCAAATTCTGTTTTATCCATATGCTAGGGATTGGCCGCGTTTTGTAGATC
29	pBAD18-Fw	GATAAAACAGAATTTGCCTGGCGGCAG
30	pBAD18-desE-USRv	GCTGCTTTAACTAAATTAATGAGGCAAAATCGACGCATATGTGCATAGGAGAAACAGTAGAGAGTTGCGATAAAAAGCG
31	pBAD18-NdeI-Fw	CATATGGATAAAACAGAATTTGCCTGGCGGCAG
32	pBAD18-A-acsRv	TAGGAGGTTACGGGGAAAAGCCAATAGGCATATGTGCATAGGAGAAACAGTAGAGAGTTGCGATAAAAAGCG
33	A-acsA-Fw	CCTATTGGCTTTTCCCCGTAACCTCCTA
34	A-acsA-Rv	GCAGTTTCATTTGATGCTCGATGAGTTTTTCTAACCTCGGCAGCAAAGTCTGGTG
35	KM-desE-Rv	GTTTCGGTGGTGACAGTTTCTGGGCTTTGGCAGGATCCGGCTGCTAACAAA
36	desE-comp-Fw	CCCAGAAACTGTCACCACCGAAAC
37	desE-comp-Rv	GTGTCGCCCACAATTTCCTGACCCCCAGGGCATCGTTTTAGCAACG
38	B-acs-Fw	GGTCAGGAAATTGTGGGCGACAC
39	B-acs-Rv	CTGCCGCCAGGCAAATTCTGTTTTATCCATATGCCAACAAGCCTTTGCCGCTGATC
40	AAS675-a1	TATGCACATATGCCTTCCATCACGTCGGCAGTAATTTC
41	AAS675-a2	GCAGTTTCATTTGATGCTCGATGAGTTTTTCTAACAAGCCGAAATCATGGCTACAATCCTAC
42	KM-675-Rv	GAAGATTCCGCCATTCGGATCGCCTTTGGCAGGATCCGGCTGCTAACAAA
43	AAS675-b1	GCGATCCGAATGGCGGAATCTTC
44	AAS675-b2	CTGCCGCCAGGCAAATTCTGTTTTATCCATATGCACTGAGGCCACATCCGTCAAAATC
45	pBAD18-675Rv	GAAATTACTGCCGACGTGATGGAAGGCATATGTGCATAGGAGAAACAGTAGAGAGTTGCGATAAAAAGCG
46	AAS1977-a1	TATGCACATATGGCAGTTCTGTAAGGCCCTACTAGAGG
47	AAS1977-a2	GCAGTTTCATTTGATGCTCGATGAGTTTTTCTAAGTTTTCGCAGAATGGGTCATGGTGG
48	KM-1977-Rv	GCAATTTTCATCGCCACCCTTTAGAGGCTTTGGCAGGATCCGGCTGCTAACAAA
49	AAS1977-b1	CCTCTAAAGGGTGGCGATGAAAATTGC
50	AAS1977-b2	CTGCCGCCAGGCAAATTCTGTTTTATCCATATGCATTGACCCCAGGCTCAATCAGATTCC
51	pBAD18-1977Rv	CCTCTAGTAGGGCCTTACAGAACTGCCATATGTGCATAGGAGAAACAGTAGAGAGTTGCGATAAAAAGCG

^aRestriction sites are underlined. Oligonucleotides 1 to 20 were used to construct integration cassettes for creating the desaturase knockouts (a1 and a2 primers were used for the upstream region of each gene, b1 and b2 for the downstream region). Oligonucleotides 21 to 30 were used to construct integration cassettes for creating the desE-up strain. Oligonucleotides 31 to 39 and 23 were used to amplify DNA fragments for subsequent Gibson assembly of an integration cassette used to create the ΔdesE+ complementation strain. Oligonocleotides 40 to 51 as well as 24 and 29 were used to construct integration cassettes for creating Δaas strains.

A wild-type strain of Synechococcus sp. strain PCC 7002 obtained from the Pasteur Culture Collection was grown photoautotrophically (140 μE m⁻² s⁻¹) on solid medium A (11) (supplemented with 1 mg/ml NaNO₃) agar plates or in 20-ml liquid cultures sparged with air under constant illumination from cool-white lamps at the specified temperatures. When required, cultures were supplemented with streptomycin and/or kanamycin (final concentration, 100 μg/ml).

Strain construction.

Knockout mutants were constructed by homologous recombination using a linear DNA fragment containing a resistance marker flanked by ∼1,000 bases homologous to the regions flanking the corresponding gene. For the construction of desaturase knockouts (Table 2), the upstream and downstream flanking sequences of the desaturase genes (desA, SYNPCC7002_A2756; desB, SYNPCC7002_A0159; desC, SYNPCC7002_A2198; desE, SYNPCC7002_A2833; desF, SYNPCC7002_A1989) were amplified by PCR (Phusion polymerase) from genomic DNA. PCR products were digested with BamHI and PstI and gel purified. The aadA gene (Sm^r) was excised from plasmid pSRA81 (12) with PstI and BamHI. The fragments were mixed in a 3:1:3 ratio (left flank:Sm^r:right flank) and ligated with T4 DNA ligase. Ligation products were gel extracted, reamplified, and transformed into Synechococcus sp. strain PCC 7002 as described by Frigaard et al. (12).

TABLE 2.

Cyanobacterial strains used in this study

Strain	Phenotype	Reference or source
Wild type	Wild-type Synechococcus sp. strain PCC 7002	Pasteur culture collection
ΔdesA	ΔdesA::aadA	This study
ΔdesB	ΔdesB::aadA	This study
ΔdesE	ΔdesE::aadA	This study
ΔdesF	ΔdesF::aadA	This study
desE-up	Φ(PcpcAB-desE)	This study
ΔdesE+	ΔdesE::aadA ΔacsA::desE-aphII	This study
Δols	Δols::aadA	8
Δols-desE-up	Φ(PcpcAB-desE) Δols::aadA	This study
Δaas675	ΔSYNPCC7002_A0675::aphII	This study
Δaas1977	ΔSYNPCC7002_A1977::aphII	This study

For the construction of the desE-up strain, the upstream region of desE was replaced with the phycocyanin (cpcBA) promoter from Synechocystis sp. strain PCC 6803, a promoter that has been reported to have high activity in Synechococcus sp. strain PCC 7002 (13). The wild-type and Δols strains (8) were transformed with a DNA cassette assembled using the in vitro enzymatic assembly method described by Gibson et al. (14). The cassette contained an upstream flanking sequence of desE, the aphII gene that confers Km^r (from pJ206 plasmid; DNA 2.0, Menlo Park, CA), the cpcBA promoter, and the 5′ end of desE.

The complementation strain (ΔdesE+) was constructed by homologous recombination using a DNA cassette containing the native desE locus (including 390 bases immediately 5′ of the desE coding sequence predicted to encode the desE promoter), the aphII gene that confers Km^r, and the upstream and downstream flanking sequences of acsA (locus SYNPCC7002_A1838). The acsA locus (target for integration of desE) encodes an acetyl coenzyme A (CoA) ligase that confers sensitivity to exogenous acrylic acid. The conditional acrylic acid sensitivity has been used as a counterselection method to facilitate the integration of heterologous genes and segregation of the resulting mutants (15). Here, the heterologous DNA was assembled from four PCR products (primers are listed in Table 2) using the in vitro enzymatic assembly method described by Gibson et al. (14). The resulting DNA cassette was transformed into the ΔdesE strain. Recombinants were selected on Km. Complete segregation of the mutants was verified by colony PCR.

The aas knockouts were created by homologous recombination of a linear targeting cassette. Here, plasmids (pBAD18 backbone) containing a kanamycin resistance cassette (aphII gene) and the upstream and downstream regions of the corresponding genes (SYNPCC7002_A0675 and SYNPCC7002_A1977) were assembled from PCR products (primers are listed in Table 2) using the in vitro enzymatic assembly method described by Gibson et al. (14). Each plasmid was linearized before transformation. Complete segregation of the mutants was verified by colony PCR. Mutant strains were grown in air-sparged, liquid cultures supplemented with pentadecanoic acid (final concentration, 20 mg/liter) at 38°C.

Lipid analysis (GC-MS).

Cultures were grown to an optical density at 730 nm (OD₇₃₀) of ∼1.0, centrifuged, resuspended in 3 ml of water, and extracted and analyzed by following previously described protocols (8, 16). Samples were analyzed using a Shimadzu GCMP QP2010S gas chromatograph mass spectrometer (GC-MS) equipped with an AOC-20i autoinjector and a Restek Rxi-5ms column (catalog no. 13423). The temperature program was a 100°C hold for 2 min, ramping up from 100°C to 150°C at 80°C per min, a hold for 4 min, ramping up from 150°C to 218°C at 4°C per min, ramping up from 218°C to 325°C at 80°C/min, and a hold at 325°C for 2.5 min. A sample injection temperature of 250°C and a volume of 1 μl was used, along with a 1:10 split ration. The MS was operated in scanning mode between 50 and 350 m/z. Quantification was achieved by comparison of integrated peaks with calibration curves of fatty acid methyl ester (FAME; Sigma) and 1-nonadecene standards (Fluka).

RESULTS AND DISCUSSION

Identification of a desaturase gene involved in alkene unsaturation.

Given the lack of functional handles on the olefins, we hypothesized that the internal bond of 1,14-nonadecadiene would be present in the substrate of the Ols-catalyzed elongation-decarboxylation mechanism, i.e., an unsaturated C₁₈ acyl-ACP. Synechococcus sp. strain PCC 7002 synthesizes lipids that incorporate 18-carbon fatty acids with zero, one, two, or three double bonds at the Δ9, Δ12, and Δ15 (or ω3) positions at the sn-1 position of lipids and C₁₆ fatty acids containing zero or one double bond at the Δ9 position at the sn-2 position of lipids (17–19). These unsaturated and polyunsaturated fatty acids are essential constituents of polar glycerolipids and are used to control the fluidity of membranes in response to changes in temperature (10). Three acyl-lipid desaturases, encoded by desA (Δ12), desB (Δ15), and desC (Δ9), have been shown to be involved in the biosynthesis of the unsaturated fatty acids observed in Synechococcus sp. strain PCC 7002 (20–22). Two additional genes are predicted to encode uncharacterized desaturases, desE (SYNPCC7002_A2833) and desF (SYNPCC7002_A1989). We hypothesized that one of the five desaturases was responsible for the internal double bond in 1,14-nonadecadiene.

To test this hypothesis, a disruption mutant of each desaturase was constructed by homologous recombination using a knockout cassette consisting of an antibiotic resistance gene flanked by homology targeting sequences. After multiple attempts, we were unable to obtain a ΔdesC mutant, suggesting that this mutation is lethal to the cells. The same problem was reported when trying to disrupt desC in Synechocystis sp. strain PCC 6803 (23). Conversely, fully segregated knockouts of desA, desB, desE, and desF were obtained after transformation and plating on the appropriate antibiotics. The observed fatty acid profiles of ΔdesA and ΔdesB mutants were consistent with past reports (20). Cells of the desA mutant contained 18:1(Δ9) but no 18:2(Δ9, Δ12) or 18:3(Δ9, Δ12, Δ15) fatty acids, and the cells of the desB mutant contained 18:1(Δ9) and 18:2(Δ9, Δ12) but no 18:3(Δ9, Δ12, Δ15) fatty acids. The hydrocarbon compositions of the ΔdesA, ΔdesB, and ΔdesF mutants were indistinguishable from that of wild-type Synechococcus sp. strain PCC 7002 (data not shown). Conversely, the hydrocarbon extract of the ΔdesE mutant contained no detectable C_19:2 alkene (Fig. 1). When the deletion was complemented by inserting desE under its native promoter in the acsA locus (a location used previously for making chromosomal insertions [15]), 1,14-nonadecadiene production was restored. These results suggest that the desaturase encoded by desE is responsible for the internal double bond in C_19:2. It should be noted that desF recently was reported to be essential for growth on plates under microoxic conditions (24), conditions that were not tested here.

FIG 1.

Comparison of the hydrocarbon composition from the wild-type, desE knockout (ΔdesE), ΔdesE complemented (ΔdesE+), and desE upregulated (desE-up) strains. The deletion of the desE gene eliminated only the production of the hydrocarbon with the internal double bond (C_19:2). Cultures were grown autotrophically at 36°C with bubbling of air, and error bars represent the standard deviations from three biological replicates.

Precursors for 1,14-nonadecadiene biosynthesis.

The putative Ols pathway (8) calls for C₁₈ acyl-ACP precursors to be processed via an elongation-decarboxylation mechanism catalyzed by a multimodular megasynthase. In the case of the C_19:1 hydrocarbon, a fully saturated acyl-ACP (18:0) would be the precursor. However, for the C_19:2 hydrocarbon to have an internal double bond at position 14, a C₁₈ acyl-ACP with a double bond at position 13 would be required. This C_18:1(Δ13) acyl-ACP could be directly synthesized by DesE acting on C₁₈ acyl-ACP or could be the elongation product of a shorter unsaturated acyl-ACP. Potential DesE products consistent with this scheme include C_16:1(Δ11), C_14:1(Δ9), or C_12:1(Δ7) acyl-ACPs. Compared to the ΔdesE mutant extract, the lipid profiles of the wild-type strain contained no additional fatty acids, providing no assistance in determining the DesE substrate.

The amino acid sequence of DesE is similar to the Δ9 desaturases of Rattus norvegicus (rat) and Saccharomyces cerevisiae (34 to 38%) (25). DesE contains the conserved three-histidine cluster motifs observed in Δ9 desaturases: the HXXXXH motifs (residues 79 to 84) and two HXXHH motifs (residues 116 to 120 and 242 to 246). These histidine motifs are thought to bind iron atoms and play an important role in the introduction of the double bond in the hydrocarbon chains of fatty acids (19). The high degree of similarity of DesE to Δ9 desaturases suggested that a C₁₄ acyl-ACP was the most likely substrate to be consistent with the formation of 1,14-nonadecadiene.

If true, the C_14:1(Δ9) acyl-ACP synthesized by DesE would be elongated to a C_16:1(Δ11) and a C_18:1(Δ13) acyl-ACP that ultimately would be the substrate for Ols (Fig. 2a). The lack of these intermediates in the lipid extracts of the wild-type strain suggested that they are not accumulating to a detectable extent. Therefore, we increased the expression of desE by replacing its promoter with the strong P_cpcBA promoter from Synechocystis sp. strain PCC 6803 (13) (desE-up strain). The fatty acid profile of this strain contained no additional fatty acids compared to the wild type. Conversely, the hydrocarbon profile of desE-up was dramatically shifted from 35% C_19:2 hydrocarbon in the wild type to nearly 100% C_19:2 in the desE-up mutant (Fig. 1). These results suggest that the hydrocarbon profile is controlled by the relative level of Ols substrates and that the products of DesE are used exclusively in the formation of olefins. If this hypothesis is true, knocking out ols in the desE-up strain should result in the accumulation of unsaturated acyl-ACPs that ordinarily serve as intermediates in C_19:2 biosynthesis. A Δols-desE-up strain was constructed, and its fatty acid profile was analyzed. Compared to the wild-type strain, the extract of the Δols-desE-up mutant contained three additional peaks in its chromatogram at retention times of 15.2, 19.6, and 19.9 min. One (15.2 min) was identified as 11-hexadecenoic acid, C_16:1(Δ11), by comparison to a commercial standard (Fig. 2b). The mass spectra for the other two peaks were consistent with C₁₈ unsaturated fatty acids, but no commercial standard for the C_18:1(Δ13) fatty acid was available for structural confirmation; in order to corroborate the identity of one of these peaks as the C_18:1(Δ13) fatty acid, a different approach was used.

FIG 2.

(a) Proposed route for alkene biosynthesis in Synechococcus sp. strain PCC 7002. DesE is required to place a double bond at the Δ9 position of the C_14:0 acyl-ACP substrate. The product subsequently is elongated to a C_18:1(Δ13) acyl-ACP that would serve as the precursor for 1,14-nonadecadiene biosynthesis by Ols. (b) Comparison of the fatty acid profiles of Synechococcus sp. strain PCC 7002 and mutants. Feeding of the C_14:1(Δ9) fatty acid to the wild-type strain and overexpression of desE in the Δols strain resulted in the formation of the C_16:1(Δ11) fatty acid (retention time, 15.2 min) and the C_18:1(Δ13) fatty acid (retention time, 19.9 min). Attempts to chemically complement the ΔdesE mutant with C_14:1(Δ9) fatty acid failed, as did attempts to chemically complement the Δols mutant with 1-nonadecene.

Exogenous fatty acids that are transported across the outer membrane can be incorporated into cyanobacterial lipid metabolism after being activated to the acyl-ACP form by an acyl-ACP synthetase (Aas). The Aas involved in this activation have been identified in Synechocystis sp. strain PCC 6803 and Synechococcus elongatus PCC 7942 (26). Although no Aas has been characterized in Synechococcus sp. strain PCC 7002, when pentadecanoic acid (C₁₅) was fed to this strain, we detected the formation of heptadecenoic acid (C₁₇) and 1-octadecene, showing that Synechococcus sp. strain PCC 7002 has the ability to elongate and incorporate exogenous fatty acids into lipid metabolism (8). Two enzymes with homology to the Aas from Synechocystis sp. strain PCC 6803 and Synechococcus elongatus PCC 7942 were identified in the genome of Synechococcus sp. strain PCC 7002 (SYNPCC7002_A0675 and SYNPCC7002_A1977). Knockout mutants were constructed for each gene (Δaas₆₇₅ and Δaas₁₉₇₇ strains). After feeding C₁₅ free fatty acid to each mutant, C₁₇ fatty acid and 1-octadecene were not detected in the Δaas₆₇₅ strain (Fig. 3) but were in the Δaas₁₉₇₇ strain, suggesting that the Aas encoded by SYNPCC7002_A0675 is responsible for the activation of exogenous fatty acids to acyl-ACPs, and its activity is required for the formation of the alkenes from exogenous fatty acids.

FIG 3.

Heptadecanoic acid (C₁₇) and 1-octadecene (C_18:1) formation after the addition of pentadecanoic acid (C₁₅). The formation of C₁₇ and C_18:1 was observed for the wild-type (wt) and Δaa₁₉₇₇ strains but not for the Δaas₆₇₅ strain. Peak areas obtained from GC-MS analysis were normalized to the volume, OD₇₃₀, and the area of an internal standard; they are presented as a fraction of the wild-type-normalized area for the corresponding compound. Error bars represent the standard deviations from three biological replicates. n.d., not detected.

Therefore, to corroborate the identity of one of the peaks observed in the Δols-desE-up strain as C_18:1(Δ13) fatty acid, 9-tetradecenoic acid, C_14:1(Δ9), was fed to wild-type Synechococcus sp. strain PCC 7002 (Fig. 2b). Once activated by the Aas, the elongation of the C_14:1(Δ9) fatty acid should result in the formation of the C_16:1(Δ11) and C_18:1(Δ13) acyl-ACPs. After analysis, feeding Synechococcus sp. strain PCC7002 with C_14:1(Δ9) resulted in the formation of two additional peaks with retention times of 15.2 min, corresponding to C_16:1(Δ11), and 19.9 min (Fig. 2b); these peaks were not observed when C_14:1(Δ9) was fed to the Δaas₆₇₅ strain. These data suggest that the peak at 19.9 min observed in the Δols-desE-up strain is indeed the C_18:1(Δ13) fatty acid and that DesE is acting on C₁₄ acyl-ACP.

Efforts to heterologously express DesE in Escherichia coli (Mistic tagged [27] and His tagged under various conditions) failed to generate soluble protein or changes in the fatty acid profile (data not shown). The hydropathy profile of DesE shows two highly hydrophobic regions, suggesting that it is an integral membrane protein (25), and further optimization of the expression construct will be needed to confirm the biochemical activity of DesE.

Effect of temperature in alkene unsaturation.

It is known that upon a downshift in temperature, expression of some desaturases is upregulated and the mRNA stability of desaturase genes is altered (22), resulting in an increase of the unsaturated lipids that are used to maintain membrane fluidity at low temperatures. In the case of desE, global transcriptome profiling via RNA-seq showed that its transcript levels increased approximately 2-fold in cells grown at 22°C compared to cells grown at 38°C (28). Also, it has been shown that desE transcripts had an estimated half-life of only 1 min at 38°C and 21 min at 22°C (25), suggesting that desE is regulated to respond to changes in temperature. To test if desE upregulation at low temperatures would affect the hydrocarbon composition of Synechococcus sp. strain PCC 7002, the wild-type strain was grown at different temperatures. At 38°C, the C_19:2 hydrocarbon represented only 11% of the total hydrocarbon pool, whereas at 22°C it was 96% (Fig. 4a). Except for 38°C, most of these differences were generated by a decrease in C_19:1 hydrocarbon rather than an increase of C_19:2. For the ΔdesE strain, there was a slight decrease in C_19:1 at low temperatures; however, on average, hydrocarbon levels remained constant across all temperatures (Fig. 4b). Interestingly, no C_18:1(Δ13) fatty acid was detected at any of the temperatures tested, suggesting that this intermediate is not incorporated into lipids but rather is exclusively converted to hydrocarbon by Ols. Given that Synechococcus sp. strain PCC7002 does not contain a pathway for fatty acid catabolism (e.g., beta-oxidation), other uses of the C_18:1(Δ13) acyl chain are not obvious.

FIG 4.

Comparison of the hydrocarbon composition from cultures grown at different temperatures. 1-Nonadecene (C_19:1) levels decrease at low temperatures for the wild-type strain (a) but not for the ΔdesE strain (b). Error bars represent the standard deviations from three biological replicates.

Although the physiological functions of hydrocarbons in cyanobacteria remain poorly understood, several hypotheses have been proposed, including intra- or interspecies chemical signaling, prevention of desiccation, enhanced buoyancy, and membrane fluidity/stability (6). The alkenes synthesized by Synechococcus sp. strain PCC 7002 are a significant portion of the total lipid content in this strain (for example, at 36°C the total alkene content is ∼5.5 μg/ml-OD₇₃₀ and the total fatty acid content is ∼35 μg/ml-OD₇₃₀). Furthermore, olefins are found exclusively in the cell pellet and not in the spent media, indicating that olefins are not secreted outside the cell (results not shown). Therefore, the change in olefin composition with temperature suggested that these compounds are important for growth at low temperatures. To test this hypothesis, the growth of the wild-type, ΔdesE, and Δols strains was compared at 38°C and 22°C on solid media. As can be seen in Fig. 5a, the three strains grew at 38°C after 4 days; at 22°C, however, only the wild-type strain was able to grow to comparable levels, suggesting that the alkenes are important for growth at low temperatures. Similar results were observed when the strains were grown in liquid cultures (Fig. 5b).

FIG 5.

Effect of temperature on growth. (a) Cells were harvested from liquid cultures and suspended to the same initial cell concentration (OD₇₃₀, ∼0.05) and then serially diluted (1:10). Seven μl of each dilution was plated on solid medium A (supplemented with 1 mg/ml NaNO₃), and cells were grown for 4 days at the specified temperatures. (b) Growth curves for the wild-type (wt), ΔdesE, and Δols strains in liquid cultures.

Conclusions.

In summary, we demonstrated the involvement of a desaturase gene (desE) in the formation of the internal double bond in the 1,14-nonadecadiene synthesized by the cyanobacterium Synechococcus sp. strain PCC 7002. The amino acid sequence encoded by the desE gene shows a high degree of similarity to Δ9 desaturases, suggesting that its most likely substrate is a C₁₄ fatty acid, which, after elongation to a C_18:1(Δ13) fatty acid, would serve as the precursor for the formation of the hydrocarbon with the internal double bond at position 14. When expression of desE was increased by replacing its promoter, we observed the formation of only the C_19:2 (and not the C_19:1) hydrocarbon, suggesting that the products of DesE are earmarked for the hydrocarbon pathway. Moreover, since no hydrocarbons coming from the other unsaturated C₁₈ fatty acids synthesized by Synechococcus sp. strain PCC 7002 have been detected, such as C_18:1(Δ9) or C_{18:2(Δ9, Δ11)}, it seems that not only the presence of the additional unsaturation is important but also its location. Since the hydrocarbon content in Synechococcus sp. strain PCC 7002 represents about 10 to 15% of the total lipid content, the increase in C_19:2 (diunsaturated hydrocarbon) production and the essentiality of DesE at low temperatures suggests that hydrocarbons play a role in responding to cold, possibly in maintaining membrane fluidity.