Racemic Phospholipids for Origin of Life Studies †

: Although prebiotic condensations of glycerol, phosphate and fatty acids produce phospholipid esters with a racemic backbone, most experimental studies on vesicles intended as protocell models have been carried out by employing commercial enantiopure phospholipids. Current experimental research on realistic protocell models urgently requires racemic phospholipids and e ﬃ cient synthetic routes for their production. Here we propose three synthetic pathways starting from glycerol or from racemic solketal ( α , β -isopropylidene- dl -glycerol) for the gram-scale production (up to 4 g) of racemic phospholipid ester precursors. We describe and compare these synthetic pathways with literature data. Racemic phosphatidylcholines and phosphatidylethanolamines were obtained in good yields and high purity from 1,2-diacylglycerols. Racemic POPC ( rac -POPC, ( R , S )-1-palmitoyl-2-oleoyl-3-phosphocholine), was used as a model compound for the preparation of giant vesicles (GVs). Confocal laser scanning ﬂuorescence microscopy was used to compare GVs prepared from enantiopure ( R )-POPC), racemic POPC ( rac -POPC) and a scalemic mixture ( scal -POPC) of ( R )-POPC enriched with rac -POPC. Vesicle morphology and size distribution were similar among the di ﬀ erent ( R )-POPC, rac -POPC and scal -POPC, while calcein entrapments in ( R )-POPC and in scal -POPC were signiﬁcantly distinct by about 10%. Strazewski); data curation, equal contribution for all authors; writing-original draft preparation, M.F.; ﬁrst review: M.F., P.S.(Pasquale Stano), R.B. and P.S. (Peter Strazewski); Electronic Supporting Information part, M.F. visualization, M.F.; supervision, M.F. and P.S. (Peter Strazewski); project administration, M.F. and P.S. (Peter Strazewski); funding acquisition, P.S. (Peter Strazewski); All authors have read and agreed to the published version of the


Introduction
In a series of papers published in 1848, Louis Pasteur argued that the crystals of both (+)-tartaric and (-)-tartaric acids were composed of the same molecules, albeit bearing different symmetries.
When combined in what is now called a racemic mixture, the different molecules cancelled each other's ability to rotate the direction of uniformly polarized light; he described these samples as "dissymmetric crystals facing one another in a mirror" [1]. At the time, Pasteur probably ignored the fact that he Commercially available achiral or racemic building blocks were used for the large-scale synthesis. In addition, POPC was selected as a model molecule to compare three distinct chemical synthetic pathways named A, B and C for its large-scale preparation (Scheme 1). The enantiopure precursor of (R)-POPC, 1-palmitoyl-2-oleoyl-sn-3-glycerol, was also synthesized in large scale from commercially available (S)-solketal, (S)-3b [37]. The characterizations of all compounds were performed via 1D and 2D 1 H and 13 C NMR spectroscopy and mass spectrometry (cf. Materials and Methods, Appendix A). Comparisons of giant vesicles (GVs) made from commercial enantiopure (R)-POPC with those made from rac-POPC and scalemic mixtures, i.e., rac-POPC enriched with (R)-POPC in a 1:1 molar ratio (scal-POPC, R:S = 2:1), deriving from our synthetic pathways, were performed using confocal laser scanning microscopy analysis.

Results
Racemic phospholipids were produced by first acylating a suitably protected racemic glycerol, then phosphorylating it. This strategy resembles the expected prebiotic formation of phospholipids [18,40]. As a model molecule for our speculation, racemic 1,2-dioleoylglycerol was synthesized under conditions identical to those reported for the naturally occurring enantiopure diacylglycerols [39]. The molecule was prepared at a medium-scale and this was crucial for carrying out a study on the properties of the resulting vesicles with no parsimony of starting material [18]. The three different chemical pathways bear the advantage of using pro-chiral compounds such as glycerol (3a), or racemic mixtures of glycerol derivatives, such as α,β-isopropylidene-dl-glycerol (3b, rac-solketal). Both molecules represent very good sources of starting material for the preparation of racemic phospholipid esters [41].

Synthesis of Racemic Di-Acyl Glycerols 7a-7d
2.1.1. Pathway A Glycerol (3a) served as a common pro-chiral building block for the synthesis of four different racemic diacyl glycerols (rac-DAG, 7a-7d, Scheme 1). The preparation of rac-DAGs bearing two different acyl chains (7a) required one or two additional steps compared with the synthesis of those bearing two identical acyl chains (7b-7d, Scheme 1). The derivative 3a was protected on a 10-gram scale with triphenylmethyl chloride (TrtCl), catalytic 4-dimethylaminopyridine (DMAP) and triethylamine (Et 3 N) to afford the product 4 in very good yields (89%) after precipitation from dichloromethane/pentane (1:10, v/v) [39]. 1-Palmitoyl-3-trityl-glycerol (5) was obtained upon the reaction of 4 with palmitic anhydride in the presence of DMAP. After 48 h, the product was obtained in a modest 45% yield due notably to the difficult removal of residual palmitic acid in the crude mixtures. The insertion of the oleoyl chain was performed by reacting 5 with oleic acid, DMAP and N-(3-dimethylaminopropyl)-N -ethylcarbodiimide hydrochloride (EDC·HCl) in dry CH 2 Cl 2 . The overnight reaction, followed by purification over a silica gel column, yielded 6a in a decent yield (55%). The key compound 7a was obtained from 6a by slight modifications of previously reported conditions [39] and with an improved yield (71%). Dilute HCl was added dropwise over 6 h to a diluted solution of 6a kept at 0 • C (Scheme 1, step d). The reaction was left stirring overnight and allowed the complete deprotection of the trityl group avoiding the acyl migration inconvenience previously observed [18]. The overall yield of 7a using pathway A was 16% as reported in Table 1, entry 1. Racemic 7b-7d bearing, respectively, two oleoyl, two palmitoyl or two myristoyl chains, were obtained by reacting the product 4 with the corresponding acyl chlorides in the presence of DMAP. These reactions were also performed on a gram-scale and the products were obtained with yields ranging from 31 to 58%. The product 6b and 6c were purified, while 6d was recovered in a form sufficiently pure to omit further purification, allowing the overall yield of 7d to be drastically increased, as reported in Table A1 in Appendix A. Table 1. Synthesis of the key compounds 5a for the preparation of racemic compounds 1a-2c.

Entry
Starting Material Scale 1 Pathway Compound Overall Yield 1 with respect to the starting material used; 2 slightly increased yield is due to the higher chemical purity of the starting material used in step m in Scheme 1; 3 yield calculated over the steps e-d, Scheme 1; 4 non purified mixtures of 6d were directly used.

Pathway B
The key compound 7a was prepared alternatively starting from α,β-isopropylidene-dl-glycerol (3b, Scheme 1, pathway B). The main difference with respect to pathway A was the introduction of the palmitoyl residue on the protected glycerol 3b. The reaction was performed using palmitoyl chloride, instead of palmitic anhydride, in the presence of DMAP and in a slight excess with respect to 3b (1.25 eq.). The reaction was left stirring overnight and the desired product 8 was isolated in a good yield (62%) and high purity. Compound 8 [42] was quantitatively deprotected using acidic Amberlyst ® H + resin and the primary alcohol of the resulting compound 9 was tritylated in conditions previously described [43]. The resulting compound 5, obtained with the same chemical purity as that obtained via pathway A, was then treated with oleoyl chloride and DMAP to obtain the product 6a. The deprotection of the trityl group was performed as previously described affording the product 7a in 71% yield and good purity. The overall yield obtained using Pathway B was 21% as reported in Table 1, entry 2.

Pathway C
A third route for obtaining the compound 7a was also optimized. This pathway employed a different protecting group, a tert-butyl-dimethylsilyl ether (TBDMS) to protect the terminal alcohol of the monoacyl glycerol 9 [42]. The first reaction was the Steglich esterification [44] of α,β-isopropylidene-dl-glycerol (3b or commercially enantiopure solketal,(S)-3b) [37], using palmitic acid, EDC as the coupling agent and DMAP, to obtain the desired compound 8 in quantitative yield. The isopropylidene moiety was then removed by treatment of 8 with aqueous acetic acid at 55 • C (Scheme 1, step j). The primary alcohol of resulting product 9 [43] was selectively protected as silyl derivative 10 [45] using tert-butyl-dimethylsilyl chloride and imidazole as a catalyst. The same esterification conditions applied for the synthesis of 8 were applied to obtain compound 10 and 11 [39,46,47] that were isolated almost quantitatively. The deprotection of the silyl group was achieved using a five-fold molar excess of triethylamine tris(hydrofluoride) in a 1:1 mixture of MeCN and THF [48] to yield the final compound rac-POG 7a. As described in the literature [49], this sequence using a silyl protecting group is often more efficient, especially during the deprotection step, than the more common trityl-based strategies (pathway A or B). In fact, the deprotection of the trityl group often leads to undesired products due to the acyl chain migration in acidic medium and suffers from a slow rate of deprotection. In contrast, the hindered chlorosilane reagents enable the protection of primary alcohols with yields varying from 60 to 90% and a clean removal using a source of fluoride (Et 3 N or pyridine HF adducts, or TBAF) with near quantitative yields. Using pathway C, the overall yield of 7a was 60% and the enantiopure (S)-7a following the same reaction pathway was synthesized with an overall yield of 66% (Table 1, entries 3 and 4, respectively).

The Next-To-Last Step: the Introduction of the Phospholipid Headgroup
Compounds 7a-7d were used to prepare the corresponding PCs (1a-1d) and PEs (2a-2c) via cyclic phosphotriester intermediates 13a-13d by nucleophilic addition of dry trimethylamine or ammonia (Scheme 2). The synthetic procedures are well known [9] and the racemic diacyl glycerols 7a-7d obtained following pathways A-C were used. Purifications of the crude materials were carried out using flash chromatography as reported in Appendix A.

Comparison of Giant Vesicles Made from (R)-POPC with Those Made from Racemic and Scalemic Mixtures of Rac-POPC and Scal-POPC
POPC served as a model compound for the preparation of GVs. In particular, we compared enantiopure membranes made of (R)-POPC with racemic membranes made of the synthesized rac-POPC, and with membranes made of a scalemic mixture (R/S molar ratio = 2:1). Typical giant vesicles produced in this work are shown in Figure 2. The scalemic mixture was prepared by adding to rac-POPC the naturally occurring commercially available (R)-POPC, thus, avoiding the eight-step synthesis from (S)-glycidol of non-natural and non-commercial (S)-POPC [50].
GVs are very large vesicles, the diameters of which lie in the micrometer range (1-100 µm). For this reason, they can be directly observed by optical microscopy rather than using indirect methods [51]. In this study we have used the so-called "droplet transfer" method, originally devised by Weitz and collaborators [52,53] and widely used in the community of protocell researchers for constructing solute-filled GVs for the purpose of studying protocell models [34].
The key mechanism of the droplet transfer method is the formation of vesicular bilayer membranes while lipid-stabilized water-in-oil droplets cross a flat interface where other lipid molecules are aligned in a monolayer [52]. In this step, the lipids coating a droplet and the lipids aligned at the flat interface come into close tail-to-tail contact and form the hydrophobic core of the membrane, while the polar headgroups of all lipids are facing the aqueous phases. In other words, the membrane assembly takes into account both lipid/water and lipid/lipid interactions. The droplet transfer is facilitated by centrifugation [30]. The method is sensitive to the type of lipids employed for the aforementioned reasons, i.e., different molecular structures generate different intermolecular interactions, thus affecting the overall droplet transfer efficiency [29,30,52].
The visual inspection of microscopy images ( Figure 2) revealed no major morphological differences between the three samples suggesting that under the experimental conditions tested (low ionic strength, high sugar concentration, 25 • C) the POPC chirality was not critical to the primary goal of forming GVs. Moreover, the green fluorescence detected inside GVs confirmed that in all cases the entrapment of solutes was clearly achievable. These two essential conclusions are key pre-requisites for the future employment of racemic GVs in this field.

Figure 2.
Confocal micrographs of the vesicles obtained from enantiopure (column a), racemic (column b) and scalemic (R:S, 2:1 column c) POPC GVs were imaged with a SP8 X laser scanning confocal microscope (Leica Microsystems, Wetzlar, Germany). Calcein (20 µM) was included in the inner aqueous buffer to mark the vesicle lumen and largely washed out from the outside by replacing with calcein-free outside buffer. Calcein fluorescence is evident in the green channel row, while the vesicle contours are shown in the bright field image row. The green fluorescence confinement is clearly shown in the merged channels row. Scale bar is 30 µm for all the micrographs.
To further evaluate the possible differences among GVs made of homochiral, racemic and scalemic POPC mixtures, quantitative analyses are needed. For example, comparing the size distribution, the concentration of entrapped solutes, and especially the biophysical properties of the membrane will shed light on how phospholipid chirality affects self-assembly, dynamics, and their interactions at the supramolecular level of a bilayer membrane. While detailed investigations on these aspects are postponed to future studies, a preliminary account of vesicle size and content is given in Figure 3 (see legend for the numerical values).  Figure 3, left histogram) revealed that the size distributions of the three samples were not statistically different at the 95% confidence level. The only statistically significant difference was detected between the calcein fluorescence distributions of (R)-POPC and scal-POPC GVs as indicated by Kruskal-Wallis ANOVA (p < 0.01) because their distributions were not normal.
Accordingly, 10% variation in the average fluorescence intensity was observed. Such a variation is consistent with the solute entrapment variability throughout the vesicle formation [54]. However, another intriguing explanation could refer to the mechanism of GV formation. Since a lipid monolayer should be transformed into a lipid bilayer (see original reports for details), supramolecular interactions can play a crucial role during the assembly. The excess of one enantiomer in the scal-POPC could be the origin of membrane defects, possibly located between lipid micro-domains, which would lead to a partial release of calcein. Such defects could play a role in dynamic behavior of the vesicle membrane and in the long-term release of the solutes affecting the permeability of the lipid bilayer.

Discussion
Our findings demonstrate that it is possible to prepare, from cheap and commercially available starting materials, racemic phospholipids in high chemical purity at a ≥1 g scale.
In particular, we obtained four different PCs and three PEs. Since racemic diacyl glycerols are the essential compounds for the preparation of the desired racemic phospholipids, our interest was focused on the synthesis of these derivatives. The key compounds 7a-7d were prepared in moderate yields from commercially available glycerol (3a) using the described Pathway A. The same synthetic route served also to prepare the derivatives bearing two identical acyl chains, 7b-7d, with high yields (50-89%) as described in Table 1, entries 5-8. The synthesis of such compounds presented significant experimental advantages, as it involved a lower number of synthetic steps from the starting material 3a and, in addition, the presence of two identical esters, thus, escaping the problem of acyl migration encountered in the preparation of 7a. Although the chemical purification of intermediate compounds 6b-6d was expected to be quite challenging, the purification proceeded smoothly and with a very low loss of material. The compound 7a was obtained in higher yields when using Pathway B, starting from α,β-isopropylidene-dl-glycerol (3b) protected with a triphenylmethyl group. However, the best results, with excellent overall yield, were obtained via Pathway C that involved the use of a tert-butyl-dimethylsilyl ether in the key synthetic step. This protecting group allowed the overall yields to be increased from 16 to 60%. The synthesis of the enantiopure (S)-7a, prepared from commercial (S)-solketal, further improved the overall yield to 66% as shown in Table 1.
In particular, using sterically hindered chlorosilane reagents allowed for the primary alcohol protection with yields varying from 60 to 90%, as well as their clean removal with a source of fluoride (Et 3 N or pyridine HF adducts, or TBAF) in near quantitative yields. The last steps for the preparation of the target compounds 1a-2c were performed accordingly to the reported synthetic procedures. Yields and purity were higher with respect to any reported data (Appendix A) [39,42,43,45,50,55]. We also produced GVs made of (R)-POPC, rac-POPC and a scalemic mixture (scal-POPC) made of (R:S = 2:1) POPC enantiomers (Figures 2 and 3). The feasibility of building GVs either with enantiomeric (R)-POPC, rac-POPC or scal-POPC made of R:S in 2:1 molar ratio was evident. Their respective morphologies were similar based on observations through confocal laser scanning microscopy. Statistical analyses revealed that the size distribution was not affected by the POPC chirality, whereas a low but statistically significant difference was detected between the calcein entrapment inside (R)-POPC (higher content) and scal-POPC (lower content). Although a possible cause for this difference might lie in the mechanisms of bilayer assembly in the very moment of droplet transfer and GV formation, available data did not allow further discussion. Additional physico-chemical characterizations are needed to ascertain their eventual distinct properties. In this respect, a meaningful example refers to racemic sphingomyelins, which significantly differ in their biophysical properties from the physiologically relevant d-erythro sphingomyelins [56].

Future Research
To address several unanswered questions on lipid synthesis in ancient times, two future research directions are proposed.
First, to further delineate the properties of chiral and racemic phospholipids, a more stringent physico-chemical analysis of vesicles (enantiopure, racemic, scalemic) will be performed: (a) Turbidity variation induced by shrinking and swelling vesicles as measured by UV spectrometry [57]. (b) Dynamic light scattering and/or spectrophotometry, to monitor morphologies as a function of temperature and the addition of cryoprotectants such as trehalose or other chiral mono/disaccharides. Racemic/enantiopure fatty acid vesicles made of chiral α-methyl fatty acids behave differently under temperature stress, the homochiral ones being more stable [58]. (c) Effects of detergents such as β-d-octylglucopyranoside or sodium cholate (both chiral) affecting the membrane permeability to peptides. (d) Vesicles doped with small amounts of anionic surfactants and subjected to the presence of cationic proteins to determine recognition property of vesicles. (e) Osmotic stress, under the above mentioned conditions. For example, the interaction between anionic vesicles and lysozyme revealed specific properties of vesicles [59][60][61]. Likewise, the opposite can be done, i.e., membranes doped with cationic lipids and interactions with polyanions (poly (Glu), nucleic acids, etc.).
Second, to evaluate dynamical properties, such as the growth-division mechanism [32], GVs made of racemic or scalemic mixtures will be treated with enantiopure compounds such as chiral fatty acids [58] acting like detergents. The fatty acid uptake by GVs may destabilize the growth and division of GVs, a process that mimics the growth and division of cells. Parallel experiments can be run together with membranes stressed as described above. Such a combinatorial approach may reveal the stereochemical diversity and selection of modern membranes. In this context, racemic and scalemic mixtures of conveniently temperature-sensitive phospholipids can be used, e.g., 1,2-dipalmitoyl-sn-glycero-3-phosphocholine, DPPC (1c), showing T m values above room temperature.

Conclusions
A series of racemic phosphatidylcholines and racemic phosphatidylethanolamines was synthetized from four different key racemic 1,2-diacylglycerols using different chemical pathways. The use of a silyl ether protecting group drastically increased the overall yields from 16 to 60% for a convenient gram scale scale-up synthesis of POPC. The synthesized racemic POPC was used for the construction of racemic membranes approaching primitive membranes before the onset of the homochiral bio-world that characterizes life as we know it (see next article in this issue). Generally speaking, here we have concluded that racemic and scalemic lipids, in particular POPC, form stable membranes essentially, as well as homochiral lipids. Such a decisive observation will pave the way to deeper investigations on the subject of homo/hetero-chiral primitive membranes. This work represents the first step for a systematic study of phospholipid chirality and its effect on many possible vesicle properties.

Appendix A. Materials and Methods, Preparation of Giant Vesicles and Chemical Characterization of Synthetic Compounds
Appendix A.1. Materials and Methods (R)-POPC was purchased from Avanti Polar Lipids, Alabaster AL (USA). All the other reagents, including calcein, glycerol, (S)-solketal, α,β-isopropylidene-dl-glycerol, oleic acid and oleoyl chloride, palmitic acid, palmitic anhydride and palmitoyl chloride, myristoyl chloride, dimethylaminopyririne (DMAP) were purchased from Sigma-Aldrich (Paris, France), Thermo-Fisher Scientific (Dortmund, Germany) or TCI Europe (Paris Cedex 7, France) and were used without further purification. NH 3  Optical rotations were measured as CHCl 3 solutions (c = 0.01 g/L, unless specified otherwise) on a JASCO P-1010 digital polarimeter and converted to specific rotations [α] D .
HRMS analyses were performed on a Bruker Impact II quadrupole-time of flight mass spectrometer. NMR spectra were recorded in CDCl 3 on a Bruker Avance 300 spectrometer at 300 MHz for 1 H, 75 for 13 C and 121.5 for 31 P and on a Bruker Avance 400 spectrometer at 400 MHz for 1 H and 100 for 13 C. Chemical shifts of CDCl 3 : δ H = 7.26 and δ C = 77.23 served as internal references. Signal shapes and multiplicities are abbreviated as br (broad), s (singlet), d (doublet), t (triplet), q (quartet), quint (quintet) and m (multiplet). Where possible, a scalar coupling constant J is given in Hertz (Hz).
Microscopic images were recorded with a confocal laser scanning microscope Leica SP8 X. Samples (5 µL) were placed in micro-welled plastic slides (ibidi GmbH, Gräfelfing-Münich, Germany, #81821). Images were recorded with a HCX PL APO lambda blue 40.0 N.A. 1.25 oil immersion objective: calcein fluorescence was acquired with excitation source at 488 nm selected wavelength of an argon laser and the emission recorded in the 500-600 nm range.

Appendix A.2. Preparation of Giant Vesicles
Giant vesicles (GVs) were obtained via the droplet transfer method [52], following an optimized procedure [30]. In a 1.5 mL Eppendorf tube (tube A) 300 µL of organic phase consisting of 0.5 mM phospholipids, (R)-POPC for enantiopure sample; rac-POPC for the racemic one; 2:1 ratio of R-POPC and rac-POPC for a scalemic mixture, dispersion in mineral oil (Sigma-Aldrich, #M5904), were gently laid over 500 µL of aqueous O-solution (outer solution, 200 mM glucose, Tris-HCl 20 mM, pH 7.8) In another 1.5 mL Eppendorf tube (tube B), 20 µL of I-solution (inner solution: Tris-HCl 20 mM pH 7.8, 200 mM sucrose, calcein 2 µM) were added to 600 µL of 0.5 mM phospholipids in mineral oil. A water-in-oil (w/o) emulsion was obtained by pipetting repeatedly up and down the mixture for 30 s. Next, the w/o emulsion (tube B) was gently poured on top of the organic phase in tube A. Tube A was centrifuged at 2500 rpm for 10 min at room temperature. After the centrifugation, the mineral oil appeared clear and was removed. The resulting GV pellets were collected from the bottom of the tube by direct aspiration with a polypropylene micropipette tip (50 µL). The vesicles were washed twice by centrifugation, supernatant removal, and re-suspension in fresh O-solution to remove most of the non-entrapped substances. (4). To a stirred solution of glycerol (3a, 10.0 g, 109.6 mmol), DMAP (0.075 g, 0.6 mmol) and trityl chloride (7.5 g, 26.9 mmol) in 20 mL of anhydrous THF at 0 • C were added 4.5 mL of anhydrous triethylamine. The reaction mixture was stirred at r.t. overnight. A solution of NaHCO 3 (2.0 g in 50 mL of H 2 O) was added followed by stirring for 15 min. The product was then extracted with EtOAc (2 × 35 mL). The combined organic phases were washed with brine (2 × 50 mL) and dried over anhydrous Na 2 SO 4 . The crude material obtained after evaporation of the solvent was crystallized from dichloromethane upon addition of pentane (1:10 v/v) to give 32. Compound 5, (Pathway A). To a cold (0 • C) solution of 4 (4 g, 12.2 mmol) in 100 mL of CHCl 3 were added portion wise 6 g palmitic anhydride (12.1 mmol) and 1.6 g DMAP (13.1 mmol). The resulting solution was allowed to return to room temperature and left under vigorous stirring at r.t. for 48 h. The solution was cooled down to 0 • C (ice bath) and 100 mL of saturated NaHCO 3 were slowly added until the excess of palmitic anhydride was hydrolyzed. The phases were separated, and the organic layers were washed with brine (4 × 50 mL) and dried over dry Na 2 SO 4 . The crude material obtained after evaporation of the solvent was purified over freshly activated SiO 2 with PE:EtOAc (10:0 to 8:2 v/v) yielding 5 as a viscous oil (4.96 g, 55%). Compound 6a (Pathway A). 1.96 g of 5 (5.85 mmol) were dissolved in cold CHCl 3 (50 mL, 0 • C) together with DMAP (0.8 g, 6.44 mmol) and oleic acid (3.18 g, 6.44 mmol). The resulting solution was slowly warmed to r.t. and the conversion of 5 into 6a was monitored periodically by TLC (PE:EtOAc 4:1 v/v). The starting material 5 was consumed after 16 h. To the cold mixture, 50 mL of a solution of NaHCO 3 (3% w/w in water) were slowly added and the biphasic mixtures was stirred for 30 min until it went back to r.t. The organic layer was extracted by adding extra volumes of CHCl 3 (3 × 20 mL) and the combined organic layers were washed with saturated solutions of citric acid (pH 6, 2 × 50 mL), NaHCO 3 (3 × 25 mL) and brine (3 × 50 mL) and dried over Na 2 SO 4 . The crude material obtained after evaporation of the solvent was purified over freshly activated SiO 2 eluting with PE:EtOAc (4:1 to 3:1 v/v) yielding 6a as white wax (1.57 g, 55%). 1  6d. The crude mixture containing 6d was directly treated for deprotection. ESI-MS m/z 777 as M+Na + ; Compound 8 (Pathway B). 3.0 g of α,β-isopropylidene-dl-glycerol (3b, 22.7 mmol) were dissolved in dry CH 2 Cl 2 (80.0 mL) and the solution was cooled to 0 • C using a thermostatic bath. Palmitoyl chloride (7.80 g, 28.4 mmol) was added together with DMAP (3.5 g, 28.4 mmol) and the resulting solution was left under vigorous stirring at r.t. for 18 h. 25 mL of saturated NaHCO 3 were added dropwise until the excess of palmitoyl chloride was consumed. The phases were separated, and the organic layers were washed with brine (3 × 25 mL) and dried over dry Na 2 SO 4 . The crude material obtained after evaporation of the solvent was purified over freshly activated SiO 2 with Cy: EtOAc Compound 9. (Pathway B). 1.8 g of 8 (4.9 mmol) were dissolved in 40 mL of dry CH 2 Cl 2 and 5.8 g of Amberlyst ® 15 H + resin were added. The suspension was stirred vigorously at r.t. until complete disappearing of the starting material was observed (4 h). The solution was filtered over a pad of Celite and the solvent evaporated. 1.6 g of 9 (>99.0%) were recovered as a yellowish oil. R f (PE/EtOAc 1:1) 0. 47 (3 × 250 mL), and the combined organic phases was washed with brine (3 × 100 mL) and dried over Na 2 SO 4 . Evaporation of the solvent followed by chromatography over freshly activated SiO 2 with CHCl 3 gave products 7a-7d as pale-yellow oils. Yields are reported in Table A1. Table A1. Data for the preparation of rac diacyl glycerols 6b-6d using 4 as common building block.

Entry
Acylation  Compound 11. A quantity of 5 g of 10 (11.2 mmol) was dissolved in dry CH 2 Cl 2 (50.0 mL) and oleic acid (3.5 g, 12.3 mmol) was added. The solution was cooled to 0 • C using an ice bath before DMAP (0.4 g, 3.4 mmol) and EDC·HCl (2.8 g, 14.6 mmol) were added together. The resulting solution was left under vigorous stirring at r.t. for 18 h. 75 mL of water were added to quench the reaction and the product was then extracted with CH 2 Cl 2 (2 × 50 mL). The combined organic phases were dried over anhydrous MgSO 4 . The crude material obtained after evaporation of the solvent was purified over SiO 2 with PE:EtOAc (99:1 to 9:1, v/v) giving 11 as a colorless oil (7.71 g, 97%).  1 mmol) was dissolved in dry CH 2 Cl 2 (40 mL) and palmitic acid (3.8 g, 15.1 mmol) was added. The solution was cooled to 0 • C using an ice bath, before DMAP (0.5 g, 4.5 mmol) and EDC·HCl (3.8 g, 19.6 mmol) were added together. The resulting solution was left under vigorous stirring at r.t. for 18 h. 75 mL of saturated NaHCO 3 were added to quench the reaction and the product was then extracted with CH 2 Cl 2 (2 × 50 mL). The combined organic phases were dried over anhydrous Compound (R)-10. A quantity of 4.61 g of (S)-9 (13.94 mmol) was dissolved in dry CH 2 Cl 2 (130.0 mL) and imidazole (1.42 g, 20.9 mmol) was added. A solution of TBDMSCl (2.3 g, 15.3 mmol) in dry CH 2 Cl 2 (30 mL) was added dropwise via an addition funnel and the resulting solution was left under vigorous stirring at r.t. for 18 h. The suspension was filtered over a pad of Celite and solvent was evaporated. The crude material obtained was purified over SiO 2 with PE:EtOAc (95:5 to 4:1, v/v) giving (R)-10 as a yellowish oil (4.77 g, 76%).
[α] D 25 = 1.73 (c 0.05, CHCl 3 ) Compound (R)-11. A quantity of 4.77 g of (R)-10 (10.7 mmol) was dissolved in dry CH 2 Cl 2 (50 mL) and oleic acid (3.3 g, 11.7 mmol) was added. The solution was cooled to 0 • C using an ice bath before DMAP (0.4 g, 3.2 mmol) and EDC·HCl (2.6 g, 13.9 mmol) were added. The resulting solution was left under vigorous stirring at r.t. for 18 h 75 mL of water were added to quench the reaction and the product was then extracted with CH 2 Cl 2 (2 × 50 mL). The combined organic phases were dried over anhydrous MgSO 4 . The crude material obtained after evaporation of the solvent was purified over SiO 2 with PE:EtOAc (99:1 to 9:1 v/v) giving (R)-11 as a colorless oil (7.35 g, 96%). (R)-11: [α] D 25 = 2.31 (c 0.05, CHCl 3 ) Compound (S)-7a. A quantity of 7.35 g of (R)-11 (10.3 mmol) was dissolved in a mixture of THF/MeCN (50/50 mL, 1:1 v/v) and Et 3 N·3HF (8.3 g, 51.2 mmol) was added slowly. The resulting solution was left under vigorous stirring at r.t. for 7 h. 150 mL of saturated NaHCO 3 were added dropwise to quench the reaction and the product was then extracted with CH 2 Cl 2 (2 × 100 mL). The combined organic phases were washed with water (100 mL), dried over anhydrous MgSO 4 and concentrated, giving (S)-7a as a colorless oil (6.08 g, 98%). Synthesis of 13a-13d and (R)-13a. General method. 7a-7d and (S)-7a were dissolved in dry toluene (6 mL) with dry Et 3 N (0.025 mL, 0.178 mmol) and cooled (0 • C). To this cold solution, a second solution prepared by dissolving 2-chloro-2-oxo-1,3,2-dioxaphospholane (12, 1.1 equiv.) in dry toluene (4 mL) were slowly added and the resulting mixture was stirred at r.t. The white precipitate obtained after 16 h was filtered off over a Celite pad (2 cm thick, 4 cm Ø, filter porosity n • 4) and the filtrate was evaporated as quickly as possible while keeping the temperature of the water bath below 20 • C. 1 H NMR and 31 P NMR (CDCl 3 ) were used to confirm the formation of 13a-13d that were used without further purifications for the next steps. Yields are reported in Table A2. (R)-13a was not used instead.  Table A2. Data for the synthesis of 1a-1d and 2a-2c from 7a-7d.

Entry
Step Scale 1 Yield Entry Step Scale 1 Yield