2019 Volume 67 Issue 4 Pages 327-332
Δ9-Fatty acid desaturase (Δ9-desaturase) is a rate-limiting enzyme of unsaturated fatty acid biosynthesis in animal cells and specifically introduces a cis-double bond at the Δ9-position of acyl-CoA. Since the chemical structure of fatty acids determines the physicochemical properties of cellular membrane and modulates a broad range of cellular functions, double bond introduction into a fatty acid by Δ9-desaturase should be specifically carried out. Reported crystal structures of stearoyl-CoA desaturase (SCD)1, one of the most studied Δ9-desaturases, have revealed the mechanism underlying the determination of substrate preference, as well as the position (Δ9) and conformation (cis) of double bond introduction. The crystal structures of SCD1 have also provided insights into the function of other Δ9-desaturases, including Drosophila homologs. Moreover, the amino-terminal sequences of Δ9-desaturases are shown to have unique roles in protein degradation. In this review, we introduce recent advances in the understanding of the function and regulation of Δ9-desaturase from the standpoint of protein structure.
Fatty acid desaturases are a family of enzymes that introduce cis-double bonds into the acyl chain of acyl-CoA. There are several types of fatty acid desaturases having a variety of substrate preferences and specificities for the position of double bond introduction. Among them, Δ9-fatty acid desaturase (hereafter referred to as Δ9-desaturase), which is embedded in the membrane of the endoplasmic reticulum (ER), introduces a cis-double bond exclusively at the Δ9 position of acyl-CoA1) (Fig. 1). This reaction requires molecular oxygen and electrons derived from electron relay systems via cytochrome b5, cytochrome b5 reductase, and nicotinamide adenine dinucleotide (phosphate) (NAD(P)H)2) (Fig. 1).
Δ9-Desaturase introduces a cis-double bond at the Δ9 position of acyl-CoA. The reaction requires molecular oxygen and electrons derived from electron relay systems via cytochrome b5, cytochrome b5 reductase, and NAD(P)H.
Δ9-Desaturase is a rate-limiting enzyme involved in the biosynthesis of monounsaturated fatty acids that are used to synthesize polyunsaturated fatty acids, phospholipids, triacylglycerols, cholesteryl esters, and wax esters. The fatty acid double bond affects the properties and functions of fatty acid-containing lipids.3) For example, the number and position of the double bonds in the fatty acid moieties of phospholipids determine the physicochemical parameters of cellular membranes: unsaturated fatty acid-containing phospholipids have lower phase transition temperatures and tend to form membranes with a liquid-disordered phase.4) Recently, Budin et al. demonstrated that the regulation of unsaturated fatty acid biosynthesis in Escherichia coli and budding yeast modulates membrane viscosity, as well as the activity of electron transport chains that feature diffusion-coupled reactions between enzymes and electron carriers.5) Therefore, the reaction of double bond introduction should be specifically carried out to maintain the appropriate properties and functions of cellular lipids.
Recently reported crystal structures of mammalian stearoyl-CoA desaturase (SCD)1, one of the most studied Δ9-desaturases, demonstrated the mechanism underlying the determination of substrate preference, as well as the position (Δ9) and conformation (cis) of double bond introduction. Furthermore, the amino acid sequences of Δ9-desaturase that are responsible for the regulation of protein degradation have been identified. In this review, we introduce recent advances in the understanding of the function and regulation of Δ9-desaturase from the standpoint of protein structure.
Biochemical studies have predicted that the amino- (N) and carboxy- (C) termini of SCD1 are oriented toward the cytosol, with four transmembrane helices (TMs) separated by two short hydrophilic loops in the ER lumen and one large hydrophilic loop in the cytosol.6) Consistent with these biochemical results, the reported crystal structures of human and mouse SCD1 have four TMs arranged in a cone-like shape7,8) (Fig. 2A). The crystal structure has a narrow tunnel extending approximately 24 Å in which the acyl chain of acyl-CoA is enclosed. The substrate-binding tunnel has a hydrophobic interior with a sharp kink around the 9th and 10th carbons of the acyl chain of the bound acyl-CoA. Furthermore, the carbonyl group of the acyl chain and CoA moiety can be specifically recognized via hydrogen bonds and electrostatic interactions. There is a large positively charged surface on the CoA moiety-binding site, which is well suited to the recognition of a CoA moiety containing negatively charged phosphate groups. This specific recognition of acyl-CoA enables the enzyme to determine the arrangement of acyl chain of bound acyl-CoA. It is easy to assume that this narrow tunnel with its sharp kink is well suited to the crucial determination of the positon (Δ9) and conformation (cis) of the double bond introduced by Δ9-desaturase.
(A) Structure of mouse SCD1 (PDB ID: 4YMK).7) Two metal ions are shown as black spheres. The structure of bound acyl-CoA is also shown. (B) Residues of mouse SCD1 involved in the recognition of acyl-chain and the coordination of metal ions.7) Two metal ions are shown as black spheres. The structure of bound acyl-CoA and the side chain of Tyr104 and Ala108 are shown. The side chains of conserved histidine and asparagine residues in the dimetal center are also shown.
There are nine conserved histidine residues (His120, His125, H157, His160, His161, His269, His298, His301, and His302 in human SCD1; His116, His121, His153, His156, His157, His265, His294, His297, and His298 in mouse SCD1) and one conserved asparagine residue (Asn265 in human SCD1; Asn261 in mouse SCD1) in TM2, TM4, the cytosolic loop between TM2 and TM3, and the C-terminal domain. It has been reported that substitution of a single histidine residue among conserved eight histidine residues in rat SCD1 (corresponding to His116, His121, His153, His156, His157, His294, His297, and His298 in mouse SCD1) eliminates the enzyme’s ability to complement the growth defects of a Δ9-desaturase-deficient yeast strain.9) These histidine and asparagine residues compose the dimetal center, which is adjacent to the kink of the substrate-binding tunnel (Fig. 2B). In the dimetal center, two metals are coordinated by the nine nitrogen atoms on the side chains of the histidine residues and one water molecule that interacts with carbonyl group on the side chain of the asparagine residue (Fig. 2B). Although two zinc ions are detected in the reported crystal structures of human and mouse SCD1, this is expected to be an artifact of protein overexpression.7,8) There are several reasons why the coordinated ions in native SCD1 protein are expected to be iron ions rather than zinc ions: i) although zinc ion normally has a tetrahedral coordination, coordinated ions in SCD1 structures have octahedral coordination, which is the typical form for the coordination of iron ion; ii) Fe2+ (0.92 Å) and Zn2+ (0.88 Å) have similar ionic radii; and iii) the diiron center is widely observed in a variety of oxidase enzymes including soluble Δ9-stearoyl-acyl carrier protein (ACP) desaturase.10–12)
To introduce a cis-double bond into the acyl chain of acyl-CoA, Δ9-desaturase accepts electrons from cytochrome b5 (Fig. 1). It is likely that cytochrome b5 binds to SCD1 in the vicinity of the dimetal center to effectively transfer electrons because the dimetal center is accessible from the cytoplasmic side. It is estimated that the positively charged surface of SCD1 and negatively charged surface of cytochrome b5 complement each other.7) To reveal the exact mechanism of electron transfer and double bond formation, an understanding of the complex structure of Δ9-desaturase and cytochrome b5 is required.
The mechanism underlying the determination of acyl-chain length of the substrate acyl-CoA is also revealed by SCD1 crystal structures. It has been reported that SCD1 prefers acyl-CoA with lengths of 17-, 18-, and 19-carbons.2) Mouse SCD1 has Tyr104 and Ala108 residues at the end of the substrate-binding tunnel (Fig. 2B). Notably, the hydroxyl group of Tyr104 is located close (4.1 Å) to the methyl group of the acyl chain of stearoyl-CoA in mouse SCD1. This tyrosine residue is widely conserved in human, mouse, and Drosophila Δ9-desaturases (Fig. 3). Mouse SCD3 is also called palmitoyl-CoA-desaturase, and prefers palmitoyl-CoA (C16-carbon length acyl-CoA) rather than stearoyl-CoA (C18-carbon length acyl-CoA) as a substrate.13) Mouse SCD3 has Tyr108 and Ile112 at the sites corresponding to Tyr104 and Ala108 of mouse SCD1, respectively (Fig. 3). Substitution of alanine for isoleucine at position 112 together with three (Glu113Leu, Asp281Gln, and Pro282Ser) or four (Glu113Leu, Val119Ala, Asp281Gln, and Pro282Ser) other mutations convert the mouse SCD3 to a stearoyl-CoA-preferring Δ9-desaturase.7) Because the side chain of isoleucine (–CH(CH3)–CH2–CH3) is larger than that of alanine (-CH3), it is likely that the large hydrophobic side chain of isoleucine hinders the binding of stearoyl-CoA, but not that of palmitoyl-CoA in mouse SCD3.
Residues comprising the end of the substrate-binding tunnel in Δ9-desaturases (highlighted in black) from various species were compared: Homo sapiens SCD1 (NP_005054.3), H. sapiens SCD5 (NP_001032671.2), Mus musculus SCD1 (NP_033153.2), M. musculus SCD2 (NP_033154.2), M. musculus SCD3 (NP_077770.1), M. musculus SCD4 (NP_899039.2), Drosophila melanogaster DESAT1 (NP_652731.1), and D. melanogaster DESAT2 (NP_650201.1).
These substrate-determining residues at the end of the substrate-binding tunnel may also play a role in the unique substrate preference of Drosophila Δ9-desaturaes. The Drosophila genome contains the Δ9-desaturase-encoding genes Desat1 and Desat2. DESAT1 and DESAT2 comprise 383 and 361 amino acids with four TMs, and show structural features similar to those of SCD1. DESAT1 was identified by its homology to vertebrate fatty acid desaturases14) and subsequent genetic studies have revealed the role of DESAT1 in the control of sensory communications via pheromone production, as well as regulation of the double bond contents in the acyl chains of phospholipids.15–17) DESAT2 was identified through genomic screening for the enzyme which determines the population-specific composition of female cuticular pheromone.17) There is a considerable variety in the composition of Drosophila melanogaster female cuticular pheromones, which are composed of cis-double bond-containing hydrocarbons and are produced from unsaturated fatty acids. Female cuticular pheromones from African and Caribbean populations have high ratios of 5,9-heptacosadiene/7,11-heptacosadiene, whereas this ratio is low in populations from other areas. Genomic analysis revealed that DESAT2 is not expressed in populations other than those of Africa and the Caribbean because of a 16-bp deletion in the 5′ region of the Desat2 gene.18) When DESAT1 is expressed in a Δ9-desaturase-deficient yeast strain, palmitoleic acid (C16:1) and oleic acid (C18:1) are the monounsaturated fatty acids that are primarily produced.17) On the other hand, DESAT2-expressing yeast produces mainly myristoleic acid (C14:1), with only trace amounts of C16:1 and C18:1.17) Because C14:1 and C16:1 are required for the production of 5,9-heptacosadiene and 7,11-heptacosadiene, respectively, functional expression of DESAT2 is required for the production of female cuticular pheromone with a high ratio of 5,9-heptacosadiene/7,11-heptacosadiene. Interestingly, DESAT2 has a Met91 residue at the site corresponding to that of the Ala108 residue in mouse SCD1 (Fig. 3). Reflecting on the similarity of the substrate preference between SCD1 and DESAT1, DESAT1 has a Gly112 residue at the site corresponding to that of the Ala108 residue of mouse SCD1 (Fig. 3). Because the side chain of methionine (–CH2–CH2–S–CH3; DESAT2) is larger than that of alanine (–CH3; mouse SCD1), glycine (–H; DESAT1), or isoleucine (–CH(CH3)–CH2–CH3; mouse SCD3), differences in the substrate preference of Drosophila Δ9-desaturases may be explained by the side chain of the amino acid residue at the end of the substrate-binding tunnel. In fact, a DESAT2-Met91Gly mutant effectively produces C16:1 and C18:1 (Miyamoto K., Nagao K., Umeda M. unpublished observation). Unsaturated fatty acids with different chain lengths are reported to have different roles not only in Drosophila, but also in mammalian models (i.e., C16:1 is called lipokine, a lipid hormone linking adipose tissue to systemic metabolism).19,20) Comparison of the substrate preference of a wide range of Δ9-desatuarses based on the amino acid residues at the end of the substrate-binding tunnel will facilitate our understanding of the physiological roles and the synthesis pathway of each unsaturated fatty acid molecule with a different chemical structure.
Expression levels of mammalian SCD1 are regulated at the transcriptional level by sterol regulatory element-binding protein (SREBP), carbohydrate-responsive element-binding protein, liver X receptor, and insulin signaling.21) Similar to SCD1, expression levels of Drosophila DESAT1 are also regulated at the transcriptional level.22) There are five splicing variants that are transcribed from different upstream regions of the Desat1 gene, but encode the same amino acid sequence. Tissue-specific expression of the Desat1 gene is precisely regulated by distinct putative regulatory regions targeting either pheromone biosynthetic cells, neurons involved in pheromone perception, or non-neuronal cells.23) It has also been reported that the expression of Desat1 gene in pheromone biosynthetic cells called oenocytes is under the control of circadian clock genes, which affects pheromone production and mating behavior in Drosophila.24)
Although the expression levels of Δ9-desaturases could be intensively regulated at the transcriptional level, Δ9-desaturases have intrinsic motifs regulating their degradation in their N-terminal domains. In contrast to the highly conserved TMs, the cytosolic loop, and the C-terminal domain, the N-terminal cytosolic domain has low homology between human, mouse, and Drosophila Δ9-desaturses (Fig. 4). Secondary structure prediction program25) suggests that the N-terminal domains of DESAT1 and mouse SCD1 do not have apparent secondary structures. Furthermore, the structure of the N-terminal domain of human and mouse SCD1 is not completely solved in reported crystal structures.7,8) Therefore, it is thought that the N-terminal domains of Δ9-desaturases have a variety of regulatory motifs for post-translational regulation such as protein degradation.
The N-terminal amino acid sequences of Δ9-desaturases from various species were compared: H. sapiens SCD1 (NP_005054.3), M. musculus SCD1 (NP_033153.2), D. melanogaster DESAT1 (NP_652731.1), and D. melanogaster DESAT2 (NP_650201.1). Proline, glutamic acid, serine, and threonine are highlighted in gray. The di-proline motif of DESAT1 is enclosed.
Since the degradation rate of SCD1 in the microsome fraction is fast, with a half-life of only a few hours, SCD1 is regarded as a short-lived protein.26–28) Because the transcription of SCD1 could be strictly regulated, the short life of protein is well suited to the finely tuned regulation of SCD1 expression. It has been reported that C-terminally green fluorescent protein (GFP)-tagged rat SCD1 protein expressed in CHO-K1 cells has a half-life of a few hours.29) However, N-terminal truncated SCD1 proteins consisting of residues 27–358 or 45–358 of SCD1 are stable.29) Furthermore, a fusion protein of the 33 N-terminal residues of SCD1 and GFP has a short half-life when expressed in the cytosol, but not in the ER lumen, demonstrating that the N-terminal cytosolic domain of SCD1 has a rapid degradation signal for the cytosol proteases.29) Using a biochemical approach, Heinemann et al. identified a plasminogen-like protein as a protease for the microsomal degradation of SCD1 protein.30,31) They also demonstrated that expression level of SCD1 in liver microsomes is decreased in plasminogen-deficient mice.30)
The N-terminal sequence of SCD1 contains a tag for rapid protein degradation, which is called PEST sequence and enriched in proline, glutamic acid, serine, and threonine32,33) (Fig. 4). Kato et al. reported that degradation of endogenously and exogenously expressed SCD1 protein is suppressed by proteasome inhibitors (MG132 and epoxomicin).33) Furthermore, in MG132-treated cells, SCD1 protein is poly-ubiquitinated and interacts with AAA-ATPase p97, indicating that SCD1 is degraded via the ER-associated degradation pathway.33) Moreover, they demonstrated that the 66 N-terminal residues containing the PEST sequence are important for the proteasomal degradation of ER-localizing proteins.33)
The expression level of Δ9-desaturase should be strictly regulated to maintain the appropriate physicochemical properties of cellular membranes. Although the regulatory mechanisms of the expression of several fatty acid desaturases are reported, it is unclear how changes in the level of cellular unsaturated fatty acids are recognized by intracellular machinery to regulate the expression of fatty acid desaturase. While the degradation rate of SCD1 protein is irrespective of the cellular levels of unsaturated fatty acids,33) we recently reported that the expression level of Δ9-desaturase is post-translationally regulated by the cellular fatty acid composition in Drosophila melanogaster.34) Drosophila melanogaster is a model organism that provides advantages for the study of mechanisms underlying the expression and function of Δ9-desaturase.34) First, DESAT1 is the sole fatty acid desaturase that introduces a cis double bond into acyl chain of acyl-CoA in typical Drosophila cell lines such as S2 cells because another fatty acid desaturase, DESAT2, is not expressed.18) Second, because Drosophila cannot synthesize sterols, and only a trace amount of polyunsaturated fatty acids are detected in cellular phospholipids,34–36) changes in Δ9-desaturase activity are expected to directly affect the physicochemical properties of the cellular membrane. Therefore, Drosophila provides a useful model to study how cells recognize changes in the level of cellular unsaturated fatty acids and regulate the expression of Δ9-desaturase.
DESAT1 protein is rapidly degraded in Drosophila S2 cells, with a half-life of approximately 2 h,34) which is comparable to those observed for mammalian SCD1 proteins. The degradation of DESAT1 protein is significantly enhanced by supplementation of culture medium with unsaturated fatty acids but not saturated fatty acids, in which exogenously added fatty acids are rapidly incorporated into cellular phospholipids.34) Furthermore, Δ9-desaturase inhibitors37,38) decrease the amount of unsaturated fatty acyl chain of phospholipids, and significantly suppress the degradation of DESAT1,34) suggesting that changes in the composition of the acyl chains of phospholipids are responsible for regulation of the expression level of DESAT1.
We found that the two sequential proline residues in the N-terminal sequence Met1-Pro2-Pro3-Asn4-Ala5-Gln6 are responsible for unsaturated fatty acid-dependent degradation of DESAT1; single mutations (Pro2Ala and Pro3Ala), as well as a double mutation (Pro2,3Ala) in DESAT1 remarkably abolish the responsiveness of DESAT1 protein degradation to the level of fatty acid desaturation.34) In the light of these findings, we designated the sequential prolines (Pro2–Pro3) of DESAT1 as a di-proline motif, which is crucial for the regulation of DESAT1 expression in response to changes in the level of cellular unsaturated fatty acids.34) By comparing the N-terminal sequences of Δ9-desaturases, we found that most Δ9-desaturases have a single proline residue at N-terminal position 2 or 3, except DESAT1 with its di-proline motif34) (Fig. 4). Interestingly, DESAT2, which is different from DESAT1 in substrate preference and physiological function, does not have a di-proline motif. Introduction of Pro2–Pro3 residues into the N-terminus of mouse SCD1 significantly enhances the unsaturated fatty acid-dependent degradation of mouse SCD1 in S2 cells,34) demonstrating that the di-proline motif works in the context of mammalian Δ9-desaturase protein. Furthermore, genetics and pharmacological analyses demonstrated that calpain A and calpain B, typical calpains containing motifs for Ca2+ and lipid binding, are involved in the di-proline motif-mediated degradation of Δ9-desaturase protein.34)
Structural and functional analyses of Δ9-desaturases will provide opportunities in several fields, including pharmaceutical research examining metabolic disorders and cancer.39,40) It has been reported that SCD1-deficient mice show decreased hepatic triacylglycerol content, reduced body adiposity, increased insulin sensitivity, and resistance to diet-induced weight gain.41,42) However, liver-specific SCD1-deficient mice are not protected from high-fat diet-induced adiposity and hepatic steatosis, indicating that SCD1 inhibition in the extrahepatic tissues is required for protection from high-fat diet-induced obesity and insulin resistance.43) In contrast, skin-specific deletion of SCD1 causes increased energy expenditure and protection from high-fat diet-induced obesity.44) Skin-specific SCD1-deficient mice are reported to show sebaceous gland hypoplasia and depletion of sebaceous lipids.44) Furthermore, expression of genes related to fat oxidation, lipolysis, and thermogenesis is increased by skin-specific depletion of SCD1.44) Therefore, the protective effects of SCD1 depletion against high-fat diet-induced obesity may be explained by its effect on the skin. In contrast, high-carbohydrate diet-induced adiposity is prevented in liver-specific SCD1-deficient mice.43) Moreover, hepatic lipogenesis and levels of active SREBP1 are decreased in liver-specific SCD1-deficent mice on a high-sucrose very low-fat diet,43) indicating that SCD1 plays a pivotal role in hepatic lipogenesis.
Inhibition of Δ9-desaturase is an attractive target for therapeutic intervention in cancer because the production of monounsaturated fatty acids is required for the replication and survival of mammalian cells.39) In proliferating cells, the expression of SCD1 is upregulated by the activation of SREBP1, a target of the phosphatidylinositol 3-kinase (PI3K), Akt, and mechanistic target of rapamycin (mTOR) pathways.45) Furthermore, inhibition of SCD1 causes the suppression of proliferation and survival signaling in cancer cells.46,47) However, because SCD1 is required for a wide range of physiological functions, including lipogenesis in liver and skin, it should be noted that SCD1 inhibitors may have harmful side effects. Recently, tumor-specific irreversible inhibitors of SCD1 were discovered.48) Sensitive cell lines against these compounds express CYP4F11 and metabolize these compounds into irreversible inhibitors of SCD1.48) Because these compounds are not activated in sebocytes, which do not express CYP4F11, toxicity in the skin can be avoided.48) Although specific interaction between SCD1 and these inhibitors is reported, the binding site and mechanism for SCD1 inhibition have not been elucidated. Structure-based evaluation and development of SCD1 inhibitors will be facilitated by reported crystal structures of SCD1, as well as further studies of Δ9-desaturases using biophysical, biochemical, and biological approaches.
This work was supported by Grant-in-Aid for Scientific research 15H05930 (to M. U.), 15K21744 (to M. U.), 18K19296 (to M. U.), 17H03805 (to M. U.), and 18K05433 (to K. N.) from Japan Society for the Promotion of Science (JSPS) and Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan.
The authors declare no conflict of interest.
