Life demands membranes. Biological membranes are dynamic structures that generally consist of amphipathic molecules bilayers held together by non-covalent bonds ^[1][2][1,2]. In eukaryotic cells, phospholipids are the predominant membrane lipids and are, from a topographic point of view, asymmetrically distributed across the bilayer ^[3][4][5][3–5]. Phospholipids consist of a hydrophilic head group with attached hydrophobic acyl chains. The variation in head groups and aliphatic chains allows the existence of more than 1000 different phospholipid species in any eukaryotic cell ^[6][7][6,7]. Phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylinositol (PI) and cardiolipin (CL), as well as sphingomyelin (SM) and glycosphingolipids (GS) are the major phospholipid classes ^[1][2][1,2]. In most eukaryotic membranes, PC and PE represent together around 60%–85% of the phospholipid fraction, while the fraction of other phospholipids depends on the cell membrane and even on animal species ^[1][2][8][9][1,2,8,9]. Phospholipids play multiple roles. They constitute a permeability barrier, modulate the functional properties of membrane-associated activities, provide a matrix for the assembly and function of a wide variety of catalytic processes, and act as donors during the synthesis of macromolecules. The wide range of processes in which phospholipids are specifically involved explains the need for diversity in phospholipid structures and fatty acid composition ^[6][10][6,10]. This diversity requires complex metabolic and regulatory pathways ^[1][2][1,2]. Therefore, for example, eukaryotic cells invest around 5% of their genes to synthesize all of these lipids [9].

The various phospholipid classes that comprise cell membranes are distributed over both leaflets of the bilayer in a non-random fashion ^[3][4][5][3–5]. This is especially evident for the aminophospholipids, PS and PE, which preferentially reside in the plasma membranes’ inner leaflet ^[11][12][11,12]. Where does lipid asymmetry originate, and how is it maintained and regulated? What is the functional role of aminophospholipid asymmetry, and what are the consequences of the breakdown of regulatory processes that result in the exposure of aminophospholipids in the cells’ outer leaflet? Although these key questions mainly remain unanswered, recent studies have established that cells have developed a number of mechanisms to deal with this issue. Targeting phospholipids to specific membrane sites is essential for maintaining critical signal transduction cascades, cell shape, hemostasis, and homeostasis [12]. In particular, aminophospholipids have been implicated in a diverse array of processes ranging from cell proliferation to cell death, from catabolism to inflammation [12]. In this scenario, asymmetry is maintained by active ATP-dependent processes, suggesting that is critical to normal cell function. Specifically, aminophospholipid asymmetry is controlled by one or more specific mechanisms, which involve selective interactions between lipids and cytoskeletal proteins and an aminophospholipid-specific active transport system ^[11][12][11,12].

The acyl chains are composed of either saturated, monounsaturated or polyunsaturated hydrocarbon chains that normally vary from 14 to 22 carbons in length [13]. In eukaryotic cells from vertebrate species, the average chain length of a biological membrane is strictly maintained by around 18 carbon atoms, and the relative distribution between saturated and unsaturated fatty acids follows the ratio 40:60 [14]. Polyunsaturated fatty acids (PUFAs) are essential components of cellular membranes in higher eukaryotes that strongly affect their fluidity, flexibility and selective permeability. Additionally, PUFAs affect many cellular and physiological processes in animals, including modulation of ion channels and carriers, activities of membrane-associated enzymes, and regulation of gene expression, among others [13].

2. Membrane Unsaturation and Lipid Peroxidation

As a principle, chemical reactions in living cells are under strict enzyme control and are tightly regulated by the metabolic program. One of the attractors involved in biomolecular evolution is the minimizing of unnecessary side reactions. Nevertheless, uncontrolled and potentially deleterious reactions occur, even under physiological conditions. Oxidative damage is a broad term used to cover the attack upon biological molecules by free radicals—chemical species with one unpaired electron. Free radicals attack/damage all cellular constituents [14]. In this context, the susceptibility of membrane phospholipids to oxidative damage is related to two inherent traits, the physico-chemical properties of the membrane bilayer and the chemical reactivity of the fatty acids composing the membrane [15]. The first property is related to the fact that oxygen and free radicals are more soluble in the fluid lipid bilayer than in the aqueous solution. Thus, membranes contain an interior organic phase, in which oxygen may tend to concentrate. Therefore, these differences in solubility are important when considering the availability of oxygen/free radicals for chemical reactions inside living systems: Organic regions may contain more free radicals than aqueous regions ^[16][17][16,17] and, consequently, membrane lipids become primary targets of oxidative damage. The second property is related to the fact that PUFA residues of phospholipids are extremely sensitive to oxidation. Every membrane phospholipid contains an unsaturated fatty acid residue esterified to the 2-hydroxyl group of its glycerol moiety. Many of these are polyunsaturated and the presence of a methylene group between two double bonds renders the fatty acid sensitive to free radical-induced damage, their sensitivity to oxidation increasing exponentially as a function of the number of double bonds per fatty acid molecule ^[18][19][18,19]. Consequently, the high concentration of PUFAs in phospholipids not only makes them prime targets for reaction with oxidizing agents but also enables them to participate in long free radical chain reactions. Lipid peroxidation generates hydroperoxides as well as endoperoxides, which undergo fragmentation to produce a broad range of reactive intermediates called reactive carbonyl species (RCS) with three to nine carbons in length (see Figure 1), the most reactive being α,β-unsaturated aldehydes (4-hydroxy-trans-2-nonenal (HNE) and acrolein), di-aldehydes (malondialdehyde (MDA) and glyoxal), and keto-aldehydes (4-oxo-trans-2-nonenal (ONE) and isoketals) ^[20][21][20,21]. 2-Hydroxyheptanal (2-HH) is another major aldehydic product of lipid peroxidation of PUFAn-6, while 4-hydroxyhexenal (4-HHE) is generated in a lower yield. Additionally, a number of other short chain aldehydes are produced during lipid peroxidation through poorly understood mechanisms. These carbonyl compounds, ubiquitously generated in biological systems, have unique properties contrasted with free radicals [14]. For instance, compared with free radicals, reactive aldehydes have a much longer half-life (i.e., minutes to hours instead of microseconds to nanoseconds for most free radicals). Further, the non-charged structure of aldehydes allows them to migrate with relative ease through hydrophobic membranes and hydrophilic cytosolic media, thereby extending the migration distance far from the production site. Based on these features alone, these carbonyl compounds can be more destructive than free radicals and may have far-reaching damaging effects on target sites within or outside membranes. Therefore, the highly unsaturated fatty acids present in cellular membranes are the most susceptible macromolecules to oxidative damage in cells, and this sensitivity increases as a function of their number of double bonds. In addition, the carbonyl compounds generated as lipid peroxidation-derived end products extend the membrane damage to other cellular constituents.

3. Non-Enzymatic Modification of Cellular Components: The Maillard Reaction-Derived Molecular Damage

Carbonyl compounds react with nucleophilic groups in macromolecules like proteins and DNA, resulting in their chemical, non-enzymatic, and irreversible modification and finally in the formation of a variety of adducts and cross-links collectively named Advanced Lipoxidation Endproducts (ALEs) ^[22][23][22,23]. Thus, by reacting with nucleophilic sites in proteins (belonging basically to Cys, Lys, Arg, and His residues), carbonyl compounds generate ALE adducts such as MDA-Lys, HNE-Lys, FDP-Lys, and S-carboxymethyl-cysteine; and the cross-links glyoxal-lysine dimmer, and lysine-MDA-lysine, among several others. The accumulation of MDA adducts on proteins is also involved in the formation of lipofuscin. Thus, lipofuscin becomes a nondegradable intralysosomal fluorescent pigment formed through lipoxidative reactions [24]. In addition to proteins, lipid peroxidation-derived endproducts can also react with the exocyclic amino groups of deoxyguanosine, deoxyadenosine, and deoxycytosine to form various alkylated products [25]. Guanine is, however, the most commonly modified DNA base because of its high nucleophilicity. Some common enals that cause DNA damage, analogously to proteins, are MDA, HNE, and acrolein, among others. Thus, the most common adducts arising from enals are exocyclic adducts such as etheno adducts, and MDA-deoxyguanosine (M1dG). In addition to lipid peroxidation derived carbonyl compounds, reducing sugars and carbonyl compounds derived from carbohydrate oxidation can also react with the primary amino groups of macromolecules such as proteins and DNA, following the principles of the carbonyl-amine reaction (also named Maillard reaction) [22]. The early Schiff base and Amadori adducts (glycation reaction), which form subsequently, slowly undergo a succession of intramolecular rearrangements, dehydration, and oxidation-reduction reactions to produce the terminal products termed advanced glycation end products (AGEs), which are often chemically irreversible, thus persisting throughout the life of the affected macromolecule ^[26][27][26,27]. More important, a major spin-off of studies on glycation during the 1980s was the recognition that oxidative reactions, and by inference, oxidative stress, catalyzed the chemical modification of proteins and DNA by Maillard reactions in vivo [28]. In this scenario, it is likely that the amino group of aminophospholipids will also react with carbonyl compounds and initiates some of the reactions occurring in proteins and DNA, expanding the biological effects of the carbonyl-amine or Maillard reaction (see Figure 2). Hence, Maillard reaction-derived molecular damage is a natural consequence of aerobic life. Maillard reaction products (MRPs, including AGEs and ALEs) induce the chemical, non-enzymatic and irreversible modification of cellular constituents, and they are part of the evidence for the existence of an “oxidative stress” in vivo. Little is known about molecular modification by MRPs, particularly regarding aminophospholipids. A current challenge is to establish the chemical structure of these modifications and the mechanisms for their formation, and to identify which factors control the nature, selectivity, extent, irreversibility and biological-pathological consequences of the molecular modification occurring in vivo.

4. Chemical Modification of Aminophospholipids by Carbonyl-Amine Reactions

Recent reports indicate that, like proteins and DNA, aminophospholipids are also targets of Maillard-type reactions. In early studies, the formation of products resulting from the reaction between aminophospholipids and lipoperoxidation-derived aldehydes such as MDA and HNE were described, but the exact structure of the products was not established. In these works, it was established that the amount of free amino groups significantly decreased during oxidation of phospholipids, developing a brownish-yellow color attributed to a Maillard-type reaction [29]. It was also reported that free amino groups of PE disappeared during oxidation in proportion to the oxygen absorbed [30]. Probably, both findings can be attributed to the reaction of the PE amino group with carbonyls, mainly MDA produced by lipid oxidation, leading to Schiff base formation as assessed by fluorescence. So, during peroxidation, PE and PS formed fluorescent chromophores with maximum emission ranging from 440 to 490 nm and maximum excitation between 360 and 400 nm ^[31][32][31,32]. Fluorescence development was related to (i) formation of thiobarbituric acid reactive substances (TBARS), especially MDA; (ii) reaction time; (iii) availability of reactive amino groups on the aminophospholipids, and (iv) antioxidant (alpha-tocopherol) content in an inverse fashion. These chromophores showed similarities to those formed in model membranes [33] or in rat liver mitochondrial and microsomal fractions peroxidized in vitro ^[31][32][31,32]. Furthermore, lipid extracts isolated from lipofuscin [34] and from tissues of lipid peroxidation-experimental models such as old, vitamin E deficient animals or animals stressed with highly unsaturated lipid diets, showed similar fluorescence properties ^[31][32][31,32].  Using thin-layer chromatography (TLC), non-enzymatic modification of aminophospholipids by lipoperoxidation derived products were also detected in red blood cells (RBC) ^{[35][36][37][38][39][40][41][42]}[35–42] and in eye lens membranes ^[43][44][43,44]. These modifications corresponded to a Schiff’s base adduct formed by cross-linking the PE and PS amino with MDA aldehyde groups, based on the following evidence: (i) A new lipid spot appeared between PS and PE; (ii) its intensity was proportional to MDA concentration both in vivo and in vitro; (iii) in selective staining procedures, it was phosphorus positive and ninhydrin negative; (iv) when this compound was exposed to acid vapors and then developed in a second direction, the “adduct” was resolved into two equimolar spots of PS and PE which were ninhydrin positive; (v) other non-amino phospholipids were ineffective in the formation of this compound; (vi) its fluorescence characteristics were compatible with a Schiff base conjugate formed between MDA and aminophospholipids; and finally, (vii) added antioxidants, blocking MDA formation, avoided its appearance. Lipid peroxidation leads to the formation of other aldehydes, such as HNE, able to react with aminophospholipids. Accordingly, the formation of fluorescent chromolipids was detected when HNE was incubated with aminophospholipids, microsomes and mitochondrial fractions ^[45][46][45,46]. Spectral characteristics of these chromolipids showed excitation maxima at 350–360 nm and emission at 430 nm, with the fluorescence intensity being linearly related to the number of HNE molecules reacting with aminophospholipids [45]. More recently, HNE potential capacity to react with aminophospholipids has been reported. Thus, by using TLC-high performance liquid chromatography (HPLC)-liquid chromatography (LC)-mass spectrometry (MS) techniques, it has been found that the main resulting compounds were a Michael adduct plus a minor Schiff base adduct, partly cyclized as a pyrrole derivative via a loss of water, with PE being more reactive than PS [46]. Much evidence demonstrates the in vitro and in vivo occurrence of the Schiff base, Amadori and AGEs-lipid products resulting from the Maillard reaction (see also Figure 3). The Schiff base formation between glucose and aminophospholipids was documented in experimental models and in human RBC membranes, plasma, and low-density lipoproteins (LDL) ^{[47][48][49][50]}[53,77–79]. The existence of glycated aminophospholipids in its Schiff base form was confirmed by using HPLC LC-electrospray ionization (ESI)-MS. Reduction with NaBH₃CN, shifting the retention time and increasing the detected mass of glycated lipids by two units, confirmed the identity of the major analyte as a Schiff base. Surprisingly, only the diacyl species became glycated and neither the alkylacyl nor the alkenylacyl were modified; furthermore, in contrast with in vitro experiments, PS glycation was not detected. In vitro model reactions of d-glucose and PE demonstrated the formation of Amadori products, which were also detected in vivo ^{[51][52][53][54][55]}[55–58,73]. Chromatographic and spectroscopic characterization unequivocally proved the existence of deoxy-d-fructosyl PE ^[56][57][55,56]. The formation of the Amadori product has also been demonstrated through the synthesis and characterization of N-(glucitol)ethanolamine (GE), a stable reduction product of glycated PE, detected in RBC membranes [58]; and by 5-(hydroxymethyl)-2-furfuraldehyde formation from the acid-treated phospholipid fraction as a stable derivative of an Amadori product, in this case being detected in rat liver lipid extracts ^[59][73]. Early immunological studies indicated the presence of MRPs covalently attached to aminophospholipids such as PE in LDL, although their structures remain unknown ^[60][72]. In addition, by using an ELISA assay, it was concluded that the majority of MRPs present in LDL were localized in the lipid phase. Nevertheless, subsequent analyses identified carboxymethyl-lysine (CML) as the major AGE antigen in tissue proteins, and it was proposed that AGE-lipids could be actually immunologically cross-reactive carboxymethyl (CM) derivatives of PE and PS formed by phospholipid glyco- and lipoxidation, similarly to proteins ^[61][80]. So, by using a selected selective ion monitoring-gas chromatography-Mass spectrometry (SIM-GC-MS) assay for carboxymethylethanolamine (CME) ^{[62][63][64][65]}[54,74,75,81] and carboxymethylserine (CMS) [58], the hydrolysis products of CM-PE and CM-PS, respectively, it was possible to demonstrate CME and CMS formation in vitro during the glycation of dioleoyl-PE under air and from linoleoylpalmitoyl-PE, but not from dioleoyl-PE, in absence of glucose. These data seem to indicate that carboxymethylation may proceed either from glucose or PUFA under oxidizing conditions. In experimental models, glucose, which is more resistant to autoxidation than PUFAs, was present at a 33-fold molar excess over free amino groups, whereas PUFA (linoleic acid) in PE was present in equimolar amounts with the amino group. So, it is difficult to anticipate which one of the routes, glycoxidation or lipoxidation, predominates in vivo, being CME and CMS, like CML ^[66][82], mixed AGEs and ALEs. The possible contribution of myeloperoxidase-catalyzed formation of CM groups from serine interaction with hypochlorite should not be dismissed ^[67][83]. In this line, it the possible contribution of PS moieties to glycoaldehyde generation, as precursor of CM groups, might be hypothesized. An alternative source of these compounds could be the oxidative cleavage of the Amadori compounds that would result from glycation of PE or PS. However, glycated PS has not been detected in vivo. CME and CMS were detected in RBC and liver mitochondria membranes, and also in urine ^{[62][63][64][65]}[54,74,75,81]. The Maillard reaction can generate free radicals through autoxidative processes that, in turn, lead to protein damage [28]. Similarly, these processes, initially described for proteins, may be expanded to aminophospholipids, representing another mechanism for the initiation of lipid oxidation. Accordingly, much evidence indicates that the Maillard reaction on aminophospholipids enhances lipid peroxidation: (i) Autoxidation is usually accelerated by primary amines ^[68][52]; (ii) in liposomes, inclusion of PE was found to accelerate the Fe²⁺-dependent peroxidation ^[68][52]; (iii) in vitro, highest levels of fluorescence intensity were obtained by incubating dilinoleoyl-PE with glucose in oxidative conditions; dipalmitoyl-PE and dilinoleoyl- or dipalmitoyl-PC were ineffective in fluorescence generation independently of the presence/absence of glucose and oxidative conditions ^[69][84]; (iv) the presence of a carboxyl group on PS, or two in the case of carboxymethylated serine, may enhance its metal binding capacity, leading to increased metal-catalyzed lipid peroxidation [58]; (v) even in the absence of exogenously added transition metals or free radical generating systems, the formation of AGE-lipids is accompanied by oxidation of the unsaturated fatty acid side chains, suggesting that AGEs formation is a source of free radicals that leads to lipid oxidation ^[70][85]; and finally, (vi) the incorporation of glycated PE into LDL facilitates its peroxidation ^[71][79]. Despite the fact that non-enzymatic modification of aminophospholipids by glycation, carboxymethylation and lipid peroxidation has been described, they represent only a limited range of the possible products that can likely be formed by the Maillard reaction.