Hormonal modulation of Δ6 and Δ5 desaturases: case of diabetes

Abstract

Animal biosynthesis of high polyunsaturated fatty acids from linoleic, α-linolenic and oleic acids is mainly modulated by the Δ6 and Δ5 desaturases through dietary and hormonal stimulated mechanisms. From hormones, only insulin activates both enzymes. In experimental diabetes mellitus type-1, the depressed Δ6 desaturase is restored by insulin stimulation of the gene expression of its mRNA. However, cAMP or cycloheximide injection prevents this effect. The depression of Δ6 and Δ5 desaturases in diabetes is rapidly correlated by lower contents of arachidonic acid and higher contents of linoleic in almost all the tissues except brain. However, docosahexaenoic n-3 acid enhancement, mainly in liver phospholipids, is not explained yet. In experimental non-insulin dependent diabetes, the effect upon the Δ6 and Δ5 desaturases is not clear. From all other hormones glucagon, adrenaline, glucocorticoids, mineralocorticoids, oestriol, oestradiol, testosterone and ACTH depress both desaturases, and a few hormones: progesterone, cortexolone and pregnanediol are inactive.

Introduction

The discovery in 1929 of the essentiality of linoleic acid for the rat [1] and the insinuation of the same property for the α-linolenic acid started the investigation and demonstration of the importance of these acids and derivatives in the animal physiology. The demonstration of the essentiality of α-linolenic acid and its derivatives specially in man started, indeed, much later (as described by Hugh Sinclair [2] in his last lecture 2 months before his death, Horroks and Yen [3] and others).

However, rather soon it began to be known that the main, true physiological effects were due, in the case of linoleic acid to its conversion to the fatty acids, eicosa-8,11,14-trienoic, arachidonic and docosa-4,7,10,13,16-pentaenoic and derived substances (now known as eicosanoids, hepoxilines, lipoxines and anandamida, etc.) whereas in the case of α-linolenic acid to its products, eicosa-5,8,11,14,17-pentaenoic (EPA) and docosa-4,7,10,13,16,19-hexaenoic (DHA) acids.

These high polyunsaturated acids (PUFA) of n-6 and n-3 families were formed identically in the rat by a series of alternative desaturations and elongations from their precursors. Two of the desaturases, first, the then called Δ6 desaturase [4], [5], and later on the then called Δ5 desaturase, were shown to modulate their biosynthesis at two different steps. However, our interest in the possible effect of hormones in the modulation of their activity was started as early as the mid 60s. It was only a few years later when the first in vitro enzymatic measurement of animal desaturases was done in rat liver microsomes, between 1958 and 1966. The stearic acid desaturation in Δ9 was proved by Bernard et al. in 1958 [6], the 20:3n-6 to 20:4n-6 Δ5 desaturation by Stoffel in 1961 [7], 18:2n-6 to 18:3n-6 Δ6 desaturation by Nugteren in 1962 [8], the 18:1n-9 to 18:2n-9 also Δ6 desaturation by Holloway et al. in 1963 [9] and all these reactions besides the 18:3n-3 Δ6 desaturation to 18:4n-3 by us [4] in 1966.

At that time, although it was known that the reaction was produced in the microsomes on the fatty acyl-CoA thioesters in the presence of oxygen and NADH or NADPH, the features of the mechanism like the electron transport by NADH-cytochrome b₅ reductase and cytochrome b₅, and even the independence of the three Δ9, Δ6 and Δ5 rat desaturases was only being envisaged. The independence of the desaturases was definitely proved later, in 1974 [10], and unequivocally confirmed in human [11], [12] and rats [13], [14] by the cloning of the corresponding DNA in 1999. However, now it has been surprisingly found [15], for the first time, in one freshwater fish, the zebrafish (Danio rerio) a peculiar Δ6 desaturase, derived from a cDNA with 64% identity to human Δ6 desaturase and 58% to human Δ5 desaturase, that can perform both the Δ6 desaturation of 18:2n-6 and 18:3n-6 and Δ5 desaturation of 20:3n-6 and 20:4n-3.

My present view of the model of the microsomal bound Δ6 desaturase system adapted to the actual findings [11], [12], [13], [14] and functionality is the one schematized in Fig. 1. In it the two separate hydrophobic domains of the desaturase are specially located in the lipid bilayer. Moreover, according to Aki's¹ finding the possible double contribution of a free Cyt.b₅ and a desaturase fusioned Cyt.b₅ are designed.

The Δ5 desaturase would show a very similar appearance due to the close similarity of the aminoacid sequences and properties of both enzymes.

The alternative contribution of a Δ6 desaturation and elongation and a Δ5 desaturation to the microsomal biosynthesis of 20:4n-6, 20:5n-3 and 20:3n-9 from linoleic, α-linolenic and oleic acids, respectively, was proved rather early [16]. However, the further biosynthesis of 22:5n-6 and 22:6n-3 acids was, without true experimental proof and even negative results [17] attributed stubbornly by lipidologists, we included, until 1991 to be produced by a second elongation and an elusive Δ4 desaturation of 22:4n-6 and 22:5n-3 acyl-CoA, respectively. In 1991, Sprecher's group [18] knocked down this generally adopted hypothesis, demonstrating that the biosynthesis of 22:6n-3 was not produced in the rat by a microsomal Δ4 desaturation. Instead, it was synthesized by a microsomal elongation of 22:5n-3 to 24:5n-3 and farther Δ6 desaturation to 24:6n-3 followed by a peroxisomal β-oxidation to 22:6n-3. This mechanism repeatedly proved farther on [19], [20] was also shown to work for the biosynthesis of 22:5n-6 fatty acid. Notwithstanding, it is interesting to add that now, that the Δ4 desaturation step has been discarded by generally all researchers from the series of reactions involved in 22:6n-3 and 22:5n-6 acids biosynthesis in animals, Infante and Huszagh [21] have proposed a mitochondrial mechanism in which carnitine and a Δ4 desaturase are involved instead of the peroxisomes. However, Ferdinandusse et al. [22] have now proved that this mitochondrial carnitine-dependent mechanism cannot be operative, at least, in human fibroblasts. Moreover, the peroxisomal straight-chain acyl-CoA oxidase, d-bifunctional protein and 3-ketoacyl-CoA thiolase are necessary for the biosynthesis of 22:6n-3 acid.

Notwithstanding, we have to add that Qiu et al. [23] have found that an enzyme with Δ4 desaturase properties is operative in a marine fungoid protist, a marine microheterotroph of Thranstochystriüm sp. Moreover, they could identify its c.DNA that coded for a desaturase that was expressed in yeast and the plant Brassica juncea and this enzyme could introduce a Δ4 double bond into 22:5n-3 and 22:4n-6 acids. Like the Δ6 and Δ5 desaturases, it shares similar N-terminal fusion cytochrome b₅-like domain and three conservative histidine motifs. However, this novel result does not modify Sprecher's group conclusions but opens new lines of research.

The recognition of Sprecher's route for the biosynthesis of 22 carbons PUFA, immediately gave rise to the question “Is the Δ6 desaturase for 18 carbon acyl-CoA the same than that Δ6 desaturase for 24 carbons acyl-CoA? In the case of the rats Δ6 desaturase, Sprecher's group [24] results are consistent with a single Δ6 desaturase. Moreover, we have proved [25] more than two decades ago that rat liver Δ6 desaturase was also able to desaturate actively the synthetic ¹⁴C labeled 20 carbons unsaturated acids eicosa-9,12-dienoic and eicosa-9,12,15-trienoic acids. These acids did not belong, of course, to the n-6 and n-3 series, but had the double bonds appropriately located like linoleic and α-linolenic acids in relation to COOH end. In addition the three double bond unsaturated 20 carbon acid was a better substrate than the two double bond acid, as it happens with linoleic and α-linolenic acids [4].

In the case of humans, the first enzymatic measurement in liver microsomes of the Δ6 desaturation of linoleic and α-linolenic acid as well of stearic Δ9 desaturation was done in 1975 [26]. From that very moment, the existence in human of a Δ6 desaturase able to desaturate linoleoyl, α-linolenoyl and oleoyl-CoA thioesters has been repeatedly confirmed using cell cultures and now by cloning its cDNA [11].

The fact that this same enzyme desaturates in human 18 carbons unsaturated acids and 24 carbon unsaturated acids has been questioned by Marzo et al. [27], showing that all trans linoleic acid inhibited α-linolenic acid desaturation but not in the same way 22:6n-3 formation from 20:5n-3 acid. These experiments were done using human Y-79 retinoblastoma and Juskat T-cells. However, very recently both Williard et al. [28] and de Antueno et al. [29] presented experimental evidence using human fibroblasts and yeast co-expressing human Δ6 desaturase gen, respectively, that in both cases apparently a single Δ6 enzyme desaturates 18:2n-6, 18:3n-3, 24:4n-6 and 24:5n-3 fatty acids.

Therefore, though the problem would be apparently settled, it would be possible that different Δ6 desaturase isoforms exist in different tissues.

High PUFA, in human and animals, except some differences in strictly carnivores (that show a very low Δ6 desaturase expression) may be provided directly by the food or by biosynthesis of the ingested precursors linoleic and α-linolenic acids. In consequence, it is necessary to indicate that the regulatory function of Δ6 and Δ5 desaturases may reach special importance when the biosynthetic passway is the main supply.

Therefore, it is more important in herbivores, less in omnivores and even less in carnivores since the last two ingest PUFA. In the case of humans, the problem would be different for vegetarians, omnivores and fish eaters (high provision of n-3 PUFA). The competition between n-6 and n-3 fatty acids existing at different steps of their biosynthesis, esterification in the lipids, selected incorporation in tissues and functions, has been extensively discussed and will not be treated here. However, it is interesting to indicate that according to Zhou and Nilsson [30] in normal Western diets due to the high percentage of linoleic acid and its high incorporation in depot fats and tissue phospholipids, the amount of arachidonic acid formed by interconversion normally exceeds the dietary intake of arachidonic acid, and therefore the modulation of desaturase activity in n-6 acids becomes very important.

The case for the n-3 PUFA, from our point of view, would not be so fundamental because α-linolenic acid is present in very low proportion in normal human foods, since it is generally provided by the green parts of the vegetals or some special seed oils such as soy seed oil or canola oil, it is stored very little in the human depot fats and tissue phospholipids and it is rapidly oxidized. So, it would very sparingly contribute to the formation of high PUFA in spite that it is a much better substrate for the Δ6 desaturase than linoleic acid [4]. Therefore, the direct provision of EPA and DHA in food becomes important from a nutritional point of view, but anyhow the modulation of the different steps of their biosynthesis must be also taken into account.

Besides we have to bear in mind that arachidonic [31] and docosahexaenoic [32] acids evoke a direct competitive feedback inhibition of their biosynthesis at the level of the Δ6 desaturase, and Clarke's group [12] also found that fish oil administration to rats reduced the abundance of Δ6 desaturase mRNA by 80% and Δ5 desaturase mRNA by 60%. In this way, dietary provision of these PUFA would inhibit in some way their biosynthesis.