The ELOVL6 gene product catalyzes elongation of n-13:0 and n-15:0 odd chain saturated fatty acids in human cells

Abstract

Normal odd chain saturated fatty acids (OCSFA), particularly tridecanoic acid (n-13:0), pentadecanoic acid (n-15:0) and heptadecanoic acid (n-17:0), are normal components of dairy, beef and seafood. The ratio of n-15:0 to n-17:0 in ruminant foods (dairy and beef) is 2:1, while in seafood and human tissues it is 1:2, and their appearance in plasma is often used as a marker for ruminant fat intake. Human elongases encoded by ELOVL1, ELOVL3, ELOVL6, and ELOVL7 catalyze biosynthesis of the dominant even chain saturated fatty acids (ECSFA), however, there are no reports of elongase function on OCSFA. ELOVL transfected MCF7 cells were treated with n-13:0, n-15:0, or n-17:0 (80 μΜ) and products analyzed. ELOVL6 catalyzed elongation of n-13:0 → n-15:0 and n-15:0 → n-17:0; and ELOVL7 had modest activity toward n-15:0 (n-15:0 → n-17:0). No elongation activity was detected for n-17:0 → n- 19:0. Our data expand ELOVL specificity to OCSFA, providing the first molecular evidence demonstrating ELOVL6 as the major elongase acting on OCS n-13:0 and n-15:0 FA. Studies of food intake relying on OCS as biomarker should consider endogenous human metabolism when relying on OCSFA ratios to indicate specific food intake.

Introduction

Normal odd chain saturated fatty acids (OCSFA), particularly tridecanoic acid (n-13:0), pentadecanoic acid (n-15:0) and heptadecanoic acid (n-17:0), are normal components of ruminant products, specifically dairy and beef ^{(1; 2)}. They are also found in non-ruminant sources such as seafood ⁽³⁾. In recent years, n- 15:0 and n-17:0 are considered biomarkers of dairy fat intake, mainly because their concentrations in serum and adipose tissue correspond with dairy intake ^{(4; 5; 6; 7; 8; 9)}. Not only serum and adipose tissue, they are also found to incorporate in other human tissues such as plasma, red blood cells (RBCs) and liver ⁽²⁾. Recent studies showed that n-15:0 and n-17:0 are biomarkers of not only dairy but also seafood ⁽³⁾ and dietary fiber intake ⁽¹⁰⁾, which again indicates OCSFA is applicable in estimation of food intake. Additionally, n-15:0 and n-17:0 are positively associated with insulin sensitivity and inversely associated with Type 2 diabetes in both cohort ^{(6; 11)} and case-control studies ^{(12; 13; 14)}, which contrasts with links between incident diabetes and prevalent even chain saturated (ECS) FA intake such as stearic acid (n- 18:0). Circulating OCSFA and/or their relative concentrations (e.g. ratios) may be useful as a physiological index of health with careful attention to their origin and endogenous metabolism ^{(2; 12; 14; 15)}.

Concentrations of n-13:0, n-15:0 and n-17:0 in general milk products in the U.S. are about 0.1%, 1.2% and 0.6% of total FA, respectively ⁽¹⁾. They are likely to be originated primarily via ruminal bacteria. The ratio of n-15:0 to n-17:0 in U.S. dairy fat is 2:1 ^{(16; 17; 18)}. In contrast, is approximately 1:2 in both freshwater and marine fishes ^{(3; 19; 20; 21)} as well as in human plasma ^{(10; 12; 15; 18)}. Comparable amount of OCSFA are found in vegan RBCs ⁽²²⁾. Because dairy fat is the predominant source of OCSFA in the U.S., it is likely that endogenous fatty acid interconversion alters their ratio via elongation, most importantly as n-15:0→-n-17:0 ^{(2; 18; 23)}; they may also arise by de novo synthesis. Current endogenous synthesis pathways of OCSFA are well summarized in previous reviews ^{(2; 18; 23)}. Among them, one theory proposes an α-oxidation of ECSFA by intermediate hydroxylation/removal of one carbon from carboxylic end ⁽¹⁸⁾. However, few details are available on the gene products responsible for OCSFA biosynthesis, unlike well-known biochemical routes to ECSFA.

In ECSFA metabolism, FA elongation follows a four-step cycle comprised of condensation, reduction, dehydration and second time reduction which occurs in endoplasmic reticulum (ER) ⁽²⁴⁾. Mammalian FA elongases ELOVL1–7 (elongation of very long chain fatty acids) work in the first and rate-limiting condensation step and all have substrate specificities and tissue specific expression distribution. Among them, ELOVL1, 3, 6 and 7 preferentially act on saturated and monounsaturated FA (SFA and MUFA); ELOVL2 and 5 are known to work on polyunsaturated FA (PUFA); while ELOVL4 elongate very long chain FA (VLCFA) with more than 24 carbons regardless of unsaturation ^{(24; 25; 26)}. Currently established substrate specificities of ELOVL towards ECSFA are as follows: ELOVL6, n-12:0→n-14:0→n- 16:0→n-18:0 ⁽²⁵; ²⁷⁾; ELOVL1, 3 and 7, n-18:0→n-20:0→→→n-26:0; ELOVL4, n-26:0→→30:0 ⁽²⁴; ²⁵⁾. ELOVL6 catalyzes n-16:0→n-18:0 in ruminant mammary cells ⁽²⁸⁾ and is expressed ubiquitously in bovine mammary epithelial cells ⁽²⁹⁾. ELOVL6 genetic variants are reportedly associated with insulin sensitivity in a Spanish population ⁽³⁰⁾.

Unlike the well-studied ECSFA, endogenous metabolism of OCSFA once ingested is not well characterized with respect to the relevant genes encoding enzymes that catalyze their interconversion. As circulating n-15:0 and n-17:0 are regarded as biomarkers of dairy, dietary fiber and seafood intake, we aimed to establish specificity of the ELOVL responsible for elongation of OCSFA of quantitative significance in the human diet. We tested the hypothesis that ELOVL6 is specific to n-13:0→n-15:0→n- 17:0 compared to the other ELOVL known to operate on straight chain fatty acids. We adopted an approach analogous to previous successful studies that established numerous novel functions for PUFA biosynthetic genes by transient or stable transfection of the open reading frame (ORF) into MCF7 cells and other models ^{(31; 32; 33; 34; 35; 36; 37; 38)}. To test ELOVLx (ELOVL1, 3, 6 and 7) function we constructed expression vectors and transiently transfected them individually into MCF7 cells as a human cell host.

Materials and Methods

Chemicals and reagents

Fatty acids (n-13:0, n-15:0 and n-17:0) were purchased from Sigma-Aldrich (St. Louis, MO). Solvents are HPLC grade for fatty acid extraction and were purchased from Sigma-Aldrich (St. Louis, MO) and Burdick & Jackson (Muskegon, MI). Cell culture media, fetal bovine serum (FBS) and other cell culture reagents were obtained from Life Technologies (NY), Corning (MA) and Thermo Fisher Scientific (MA).

ELOVL6 sequence and phylogenetic analysis

The amino acid (AA) sequences of ELOVL6 from various vertebrate species are obtained from GenBank accession numbers (Table S1). The AA sequence of human ELOVL6 was aligned with several other vertebrate ELOVL6 sequences using ClustalX2.1 software ⁽³⁹⁾. The phylogenetic tree was constructed using the neighbor-joining method ⁽⁴⁰⁾ with MEGA7 ⁽⁴¹⁾. Confidence in the resulting phylogenetic tree branch topology was measured by bootstrapping test method with 1000 replicates ⁽⁴²⁾.

ELOVL expression vector constructs

The open reading frame of ELOVL transcripts (ELOVL1, ELOVL3, ELOVL6 and ELOVL7) were cloned into a pcDNA3.1(+) expression vector (Thermo Fisher Scientific, MA) containing cytomegalovirus (CMV) promoter. The specific ELOVL gene synthesis and cloning was carried out by GenScript service, NJ. The GenBank accession numbers of ELOVL mRNA (NM) and protein (NP) are provided in Table S1. Plasmid DNA used for transfection assays was extracted and purified using Plasmid Midi Kit (QIAGEN, MD). The extracted DNA was verified by DNA sequencing and stored at - 20°C. DNA sequencing was performed at Cornell University life sciences core laboratories center using the Applied Biosystems automated 3730 DNA analyzer.

Mammalian cell culture, transfection, and fatty acid supplementation

MCF7 human breast cancer cells were grown at 37°C in a humidified environment with 5% CO₂, using minimum essential medium alpha (MEM-a) with 10% FBS and 10mM buffer [4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid; HEPES] as described previously ^{(32; 34)}.

MCF7 cells were seeded at 1× 10⁶ cell density into 60mm cell culture dishes. After 48h, when they reach 60–80% confluency, cells were washed with 1× PBS and ELOVL (ELOVL1, 3, 6 and 7) transcripts were transfected individually using Polyplus (NY) jetPRIME transfection reagent. Empty vector was used as control. According to the jetPRIME reagent kit protocol, 4μg of Vector (control) or ELOVL DNA was transfected into cells along with 200μl jetPRIME buffer, 8μl jetPRIME reagent, and 5ml growth media. After 24h, the transfected MCF7 cells were supplemented with 80 μΜ of bovine serum albumin (BSA) bound OCSFA substrates (n-13:0, n-15:0 and n-17:0). Briefly, to make BSA bound substrates, n-13:0, n-15:0 and n-17:0 were dissolved in absolute ethanol to make 100mM FA stock. FA stock (200 $\mathrm{\mu }$ l) was then mixed with FA free BSA in 1× PBS (4.4% w/w) and incubated overnight at 37°C. BSA bound OCSFA were filtered using 0.22 μm syringe, and diluted to 80 μΜ with non-FBS MCF7 media then added to cells. After additional 24 h incubation, cells were washed twice with 1 × PBS, harvested by trypsinization and supernatant removed after centrifuging.

RNA isolation and cDNA synthesis

RNA was isolated from harvested MCF7 cell pellets using E.Z.N.A. Total RNA Kit I (Omega Bio-tek Inc, GA). The RNA quantity and quality was verified by micro spectrophotometer Nanodrop 2000 (Thermo Scientific). cDNA was synthesized from 1 $\mathrm{\mu }$ g of RNA using High Capacity cDNA Reverse Transcription Kit (Life Technologies, NY). Synthesized cDNA was then used as template for RT-PCR reactions ⁽⁴³⁾.

RT-PCR

Gene specific primers ELOVL1–7 were designed using PrimerQuest software (Integrated DNA Technologies, Coralville, IA), primer sequences and annealing temperatures are shown in Table S2. RT- PCR amplification reactions were performed using EmeraldAmp GT PCR Master Mix (Clontech, CA) using gradient thermal cycler (Eppendorf, NY). PCR products were separated by 2% agarose gel electrophoresis stained with ethidium bromide, and bands were visualized under UV light. GAPDH was used as control gene.

Fatty acid extraction and analysis

Fatty acid methyl esters (FAME) from the harvested MCF7 cell pellets were prepared according to the modified one-step method of Garces and Mancha ⁽⁴⁴⁾. FAME were structurally identified by gas chromatography (GC) - chemical ionization, electron ionization (EI) mass spectrometry (MS) and EIMS/MS using a Saturn 2000 mass spectrometer attached to a Varian Star 3400 gas chromatograph ⁽⁴⁵⁾. FAME were quantified by GC-flame ionization detector (GC-FID) (Hewlett-Packard). An equal weight FAME mixture, GLC462 (Nu-Check Prep, Inc.), was used to calculate response factors of all fatty acids. Percent conversion of substrates (S) to products (P) was calculated as: [(P) / (S + P)] * 100, and normalized to the control group.

Statistical analysis

All treatments were performed using two biological replicates; the mean of three technical replicate GC- FID analyses were used for each biological replicate and no data was excluded. In numerous earlier studies, we used an analogous approach with MCF7 transfection and fatty acid treatment with 2–3 biological replicates for functional characterization of PUFA biosynthetic genes ^{(31; 32; 33; 34)}; this approach is comparable to functional characterization studies conducted by others ^{(35; 36; 37; 38)}. Values generated using averages across the two biological replicates are expressed as mean ± SD. Statistical analysis of comparisons between multiple groups was performed using OriginPro 8 Software (OriginLab Corporation, MA). One-way analysis of variance (ANOVA) with post-hoc Tukey Honest Significant Difference (HSD) test was used to analyze significant differences between groups. Differences when P<0.05 are considered significant.

Results

Amino acid sequence and phylogenetic analysis of ELOVL6

The human ELOVL6 cDNA (NM_024090.2) consist of a 798 base pair (bp) ORF, encoding a protein of 265 AA (NP_076995.1). As shown in Figure 1, human ELOVL6 shares >90% AA sequence identity with other vertebrate ELOVL6 sequences. The human ELOVL6 shared 98.11%, 96.98%, 96.60% and 93.16% identity, with monkey, mouse, rat and cattle sequences respectively. All vertebrate ELOVL6 possessed five transmembrane domains (I-V) found among elongases ⁽⁴⁶⁾ as well as the conserved histidine box HXXHH motif characteristic of elongase families ^{(47; 48; 49)}.

Figure 1: Alignment of amino acid (AA) sequences of ELOVL6 from human and other vertebrates.

The AA sequences of various species obtained from GenBank accession numbers were aligned using ClustalX 2.1. The well conserved histidine motif HXXHH is depicted in the box. The dotted lines with Roman numerals indicate the putative transmembrane regions ⁽⁴⁶⁾.

A phylogenetic tree was constructed by comparing the AA sequences of ELOVL6 from various vertebrates (Figure 2). As expected, the human ELOVL6 grouped with primates, while rodents (rat and mouse) and fish (catfish and zebrafish) grouped together.

Figure 2: Phylogenetic tree of ELOVL6 from human and other organisms.

The tree was constructed using the Neighbor-Joining method with 1000 bootstrap replicates in MEGA7. The numbers represent the frequencies (%); and horizontal branch length is proportional to AA substitution rate per site. Each species was followed by its NCBI Reference Sequence (NP).

Distribution of SFA in MCF7 cells

MCF7 cells are high in ECSFA n-16:0 (16.63 ± 0.43%) and n-18:0 (12.98 ± 0.52%), but contain only trace amount of OCSFA n-17:0 (0.22 ± 0.02%) and n-13:0 and n-15:0 are at undetectable amounts (unpublished data). Table 1 summarizes when MCF7 cells are dosed with 80 μΜ of specific OCSFA, cells readily uptake OCSFA in the order of n-17:0 > n-15:0 > n-13:0.

Table 1.

Uptake Efficiency of OCSFA by MCF7 cells.

OCSFA	n-13:0		n-15:0		n-17:0
	Mean	SD	Mean	SD	Mean	SD
Uptake Efficiency1	3.276	0.002	20.334	0.136	27.733	0.175

¹Uptake Efficiency (%) = wt (BCFA uptake) / wt (total fatty acids excluding OCSFA)×100%

MCF7 cells have trace amount of basal OCSFA; control cells readily uptake OCSFA when treated with 80 μΜ of albumin bound individual OCSFA, and was in the order of n-17:0 > n-15:0 > n-13:0.

Data from two biological replicates.

Transient Transfection ELOVL gene expression

Figure S1 shows ELOVL (ELOVL1 to ELOVL7) gene expression in wild type (WT), control (empty pcDNA3.1(+) vector) and ELOVLx transiently transfected cells. The cells were incubated with A) 80pM n-13:0 B) 80μM n-15:0 and C) 80μM n-17:0. ELOVL1 expression was higher in ELOVL1 transfected cells; similarly ELOVL3, ELOVL6 and ELOVL7 expression are higher in ELOVL3, 6 and 7 transfected cells. These results show transfections were successful.

Elongation, n-13:0→n-15:0

MCF7 cells transiently expressing ELOVL1, ELOVL3, ELOVL6 and ELOVL7 and control (empty vector) were incubated with 80 μΜ of albumin bound n-13:0 OCSFA. MCF7 cells have native elongase activity, so the gain of function is compared to control (Figure 3). The cells expressing ELOVL6 showed significantly increased activity towards n-13:0 (131.19 ± 5.52 (%)). No gain of activity was seen with other elongases.

Figure 3: ELOVLx activity, n-13:0→n-15:0.

Percent conversion of fatty acid substrate n-13:0 into elongation product n-15:0 (ratios shown, calculated as [n-15:0]/[n-13:0 ]+ [n-15:0]) was measured in MCF7 cells and normalized to control group. ELOVL1, 3 and 7 showed no activity towards n-13:0 (n-13:0→-n-15:0). ELOVL6 had significantly higher catalytic activity towards n-13:0 (n-13:0→-n-15:0). Data from two biological replicates. a-b: P<0.05.

Elongation, n-15:0→n-17:0

MCF7 cells transiently expressing ELOVL1, ELOVL3, ELOVL6 and ELOVL7 and control (empty vector) were incubated with 80 μΜ of albumin bound n-15:0 OCSFA. MCF7 cells have native elongase activity, so the gain of function is compared to control (Figure 4). The cells expressing ELOVL6 showed significantly increased activity towards n-15:0 (129.51 ± 1.74 (%)), followed by moderate activity in ELOVL7 cells (113.99 ± 1.30 (%)). No gain of activity was seen with ELOVL1. Elongation activity was lower than control in the ELOVL3 cells though not different than ELOVL1 cells.

Figure 4: ELOVLx activity, n-15:0→n-17:0.

Percent conversion of fatty acid substrate n-15:0 into elongation product n-17:0 (ratios shown, calculated as [n-17:0]/([n-15:0] + [n-17:0]) was measured in MCF7 cells and normalized to control group. ELOVL1 showed no activity towards n-15:0 (n- 15:0→n-17:0). ELOVL3 reduced activity towards n-15:0 (n-15:0→n-17:0) compared to control though it was not different than ELOVL1. ELOVL6 had significantly higher catalytic activity towards n-15:0 (n-15:0→n-17:0). ELOVL7 showed moderate activity towards n-15:0 (n-15:0→n-17:0). Data from two biological replicates. a-d: P<0.05.

Interconversion, n-17:0→n-15:0 but not n-19:0

MCF7 cells transiently expressing ELOVL1, ELOVL3, ELOVL6 and ELOVL7 and control (empty vector) were incubated with 80 μΜ of albumin bound n-17:0 OCSFA. All MCF7 cells (ELOVL1, ELOVL3, ELOVL6 and ELOVL7 and control) readily uptake n-17:0 but none of the cells showed elongation activity towards n-17:0; n-19:0 is not detected. While n-15:0 is not detectable in untreated cells, incubation with n-17:0 results in n-15:0 appearance at approximately equal amounts ([15:0]/([15:0]+[17:0]) ~ 3%) regardless of ELOVLx transfection, suggesting that β-oxidation is a route of n-17:0 conversion.

Discussion

Our results confirm the hypothesis that ELOVL6 is the primary ELOVL with activity toward OCSFA, specifically catalyzing n-13:0→n-15:0 and n-15:0→n-17:0. Further, ELOVL7 has moderate activity toward n-15:0→n-17:0. ELOVL1 had no activity toward any OCSFA, and no ELOVL catalyzed n- 17:0→n-19:0, consistent with the trace amount of n-19:0 in human tissue. ELOVL3 significantly reduced activity compared to control but not compared to ELOVL1 which was not different than control. MCF7 cells have native elongation activity, thus it is plausible that ELOVL3 may have inhibited that native conversion, possibly by non-active substrate binding. The magnitude of the effect is small, however.

Our AA sequence alignments of ELOVL6 showed a high degree of conservation among evolutionarily distant related genomes (human to zebrafish). The histidine rich motif and all five transmembrane regions are conserved from human to zebrafish. The pattern of ELOVL6 homology among several vertebrate species even with distant phylogeny points to similar metabolic conservation.

ELOVL6 belongs to a highly conserved microsomal enzyme family that is involved in fatty acid biosynthesis. ELOVL6 activity characterized by cloning of mammalian ELOVL6 ORF into expression vector and expressing the vector in mammalian cells shows that ELOVL6 specifically catalyzes the elongation of even chain saturated and monounsaturated fatty acids with chain length 12, 14, and 16 carbons ^{(46; 50)}. ELOVL6 is found to be ubiquitously expressed in mice and bovine tissues ^{(29; 46)}, and plays a role in energy metabolism and insulin sensitivity ⁽²⁷⁾, nonalcoholic steatohepatitis ⁽⁵¹⁾, breast cancer ⁽⁵²⁾, pulmonary fibrosis ⁽⁵³⁾ and squamous cell carcinoma of the lung ⁽⁵⁴⁾.

The endogenous synthesis of OCSFA via α-oxidation of ECSFA has been previously reported ^{(18; 55; 56)}. In an adipocyte differentiation study, both kinetic and ¹³C palmitate labelled experiments showed significant increase in the synthesis of OCSFA via α-oxidation ⁽⁵⁵⁾. The differentiating adipocytes converted ¹³C palmitate (n-16:0) to n-15:0, showing that OCSFA were endogenously synthesized from n-16:0 by α-oxidation in this cell type. This conversion happened only in the cells and not in cell culture media ⁽⁵⁵⁾. Similarly, Su et al. showed differentiating adipocytes when incubated with [9,−10-³H]n-16:0 resulted in the production of radiolabeled [³H]n-15:0 via α-oxidation ⁽⁵⁷⁾. Casteels et al. ⁽⁵⁸⁾ proposed α- oxidation mechanism might play a role in the formation of OCSFA in the brain. An in-vivo experiment with rats infused with n-18:0 showed ~70% (p < 0001) increase in the n-17:0 levels in the serum compared to a control rat group ⁽⁵⁹⁾. Our results show that specific ELOVL6 or 7 may operate on these nascent products α-oxidation n-13:0 and n-15:0 but not on n-17:0 to further alter OCSFA profile (Figure 5).

Figure 5: Even chain saturated fatty acid (ECSFA) and Odd chain saturated fatty acid (OCSFA) biosynthesis in vertebrates.

ELOVL6 catalyzes elongation of n-12:0 to n-18:0, whereas ELOVL3 catalyzes elongation of n-18:0 to n-20:0. Depicted in red font color: ELOVL6 elongates OCSFA n- 13:0→n-15:0 and n-15:0→n-17:0. ELOVL7 (shown as 7) has moderate activity (n-15:0→n-17:0). Palmitic acid (n-16:0), the major product of fatty acid synthase (FASN), and α-oxidation of ECSFA to OCSFA are also shown. FASN, fatty acid synthase; ER, endoplasmic reticulum; N.R., no reaction.

Here we found overlapping activity for ELOVL6 and ELOVL7; both had substrate specificity for n-15:0. Elongases in several species are known to share common overlapping functions ^{(49; 60; 61)}. Previously it has been shown that mammalian ELOVL1, ELOVL3 and ELOVL7 share common substrates ^{(25; 49; 62; 63)}. ELOVL1 elongates SFA of chain lengths n-18:0 to n-26:0, with the highest activity toward n-22:0 FA, whereas, ELOVL3 and ELOVL7 were found to elongate n-16:0 to n-22:0 FA, with the highest activity toward n-18:0 FA ^{(62; 63; 64)}.

The relatively small biological replicate size (n=2) in our study has consistently yielded reliable results in our previous work ^{(31; 32; 33; 34)}. For functional characterization studies, others have used a single assay to support conclusions based on identification of a novel fatty acid product by chromatography with subsequent calculation of a conversion ratio ^{(35; 36; 37; 38)}, thus greater confidence in matching biological duplicates of the present results is warranted. Our uses of duplicates with consistently small standard deviations in FAME analysis, as well as validation of ELOVLx transfection efficiency via RT-PCR, support a similar level of confidence in the present results. Finally, we note that our data on normal polyunsaturated fatty acids ^{(31; 32; 33; 34)} and ECSFA (data not shown) faithfully replicates that of others.

In conclusion, we provide the first molecular evidence demonstrating ELOVL6 is the major elongase acting on OCSFA, with specificity toward n-13:0 and n-15:0. Modest activity was found for ELOVL7 toward n-15:0. The present study expands ELOVL substrate specificity range to OCSFA. Nutrition studies considering these fatty acids as markers of specific food intake should consider interconversion of OCSFA in light of these findings, particularly genome-wide association studies (GWAS) and targeted gene studies as they associate circulating OCSFA levels with single nucleotide polymorphisms (SNPs) and other polymorphisms.