Chemical Synthesis of Marine-Derived Sulfoglycolipids, a New Class of Molecular Adjuvants

Vaccines play a primary role in the protection of human health by preventing infectious and chronic diseases. Recently we have reported 1,2-O-distearoyl-3-O-β-d-sulfoquinovosylglycerol (β-SQDG18), here named Sulfavant A (1), which shows promising properties as a new molecular adjuvant in in vitro and in vivo tests. In the present manuscript, we provide full details about a synthetic strategy for the preparation of 1, including a discussion of chemical determinants of the activity and the major technical hurdles we faced during the study. Synthesis of Sulfavant A (1) is achieved by a versatile procedure based on a trichloroacetimidate methodology and peracetate sugar precursors. The final design opens possibilities for the preparation of a series of interesting analogs for further pharmacological optimization and development, including derivatives containing different saturated and polyunsaturated fatty acids (e.g., 17 and 22).


Introduction
Vaccination consists of the stimulation of immune response by vaccines, historically composed of attenuated or inactivated biological agents [1].However, most recent vaccines are constituted by peptides or recombinant DNA, which are safer, but often insufficiently immunogenic [1].To overcome this problem, modern vaccines contain adjuvants that are necessary to boost the immune response.To date, except for the monophosphoryl lipid A (MPLA), clinically approved formulations of adjuvants are restricted to aluminum salts and emulsions of lipids in water [2].
Dendritic cells (DCs) are the most important antigen-presenting cells (APCs), and play a pivotal role in starting immunogenic reactions and priming a long-lasting and antigen-specific response [3].The ability to activate DCs is a key tool for improving the efficacy of vaccination [3].
Sulfavant A (1) activates human DCs and triggers an efficient immune response in vivo [4].After treatment, DCs show up-regulation of MHC II molecules and other co-stimulatory proteins (CD83, CD86), as well as production of pro-inflammatory cytokines (IL-12 and INF-γ).Mice immunized with OVA associated to Sulfavant A produced a release of anti-OVA Ig comparable to traditional adjuvants, while, in an experimental model of melanoma, vaccination of C57BL/6 mice with Sulfavant A-adjuvanted hgp10 peptide elicited a protective response leading to a promising reduction in tumor growth and increase of mice survival [4].
Here we discuss in detail the strategy for the synthesis of Sulfavant A (1), including an overview of the main technical hurdles we have faced, together with the considerations for the scale-up of the process and the possibility to expand the class of synthesizable analogs.sulfoglycoglycerolipids are marked by the α-configuration of the anomeric carbon and the βconfiguration type is not present in nature.Sulfavant A (1) activates human DCs and triggers an efficient immune response in vivo [4].After treatment, DCs show up-regulation of MHC II molecules and other co-stimulatory proteins (CD83, CD86), as well as production of pro-inflammatory cytokines (IL-12 and INF-γ).Mice immunized with OVA associated to Sulfavant A produced a release of anti-OVA Ig comparable to traditional adjuvants, while, in an experimental model of melanoma, vaccination of C57BL/6 mice with Sulfavant A-adjuvanted hgp10 peptide elicited a protective response leading to a promising reduction in tumor growth and increase of mice survival [4].
Here we discuss in detail the strategy for the synthesis of Sulfavant A (1), including an overview of the main technical hurdles we have faced, together with the considerations for the scale-up of the process and the possibility to expand the class of synthesizable analogs.

Results and Discussion
For the preparation of Sulfavant A, the main challenges lay in the requirement of planning a chemical strategy comprising the introduction of the β-glycosidic bond at the anomeric carbon and the insertion of the sulfonic function on C6 of glucose (Scheme 1).At the same time, the possibility of introducing different and unsaturated fatty acyl residues on the glycerol moiety would make the strategy expandable to a wide class of analogs.Furthermore, the process had to be suitable to scaleup for further pharmacological development.The first crucial point was the choice of orthogonal protecting groups on sugar, whose deprotection should not affect the diacylglycerol aglycon part and the glycosidic linkage.In this frame, tert-butyldimethylsilyl and benzyl groups to which the synthesis of α-and β-SQDGs [14,21,22] had already been applied, were initially tested for the preparation of the sulfolipid.Unfortunately, the tert-butyldimethylsilyl group proved to be too labile in the experimental conditions and this, together with its low deprotection efficiency, led to a considerable decrease in yields; whereas the use of benzyl excluded the possibility of introducing unsaturated acyl chains to the glycerol, and greatly reduced the possibility of extending the synthesis to other analogs.For these reasons, we decided to explore the use of acetate as transient protecting group for the hydroxy functions of the sugar moiety.The acetoxy derivatives are stable under a wide range of conditions and, most importantly, the presence of the acetate residue at C-2′ is capable of ensuring an excellent stereoselectivity in the

Results and Discussion
For the preparation of Sulfavant A, the main challenges lay in the requirement of planning a chemical strategy comprising the introduction of the β-glycosidic bond at the anomeric carbon and the insertion of the sulfonic function on C 6 of glucose (Scheme 1).At the same time, the possibility of introducing different and unsaturated fatty acyl residues on the glycerol moiety would make the strategy expandable to a wide class of analogs.Furthermore, the process had to be suitable to scale-up for further pharmacological development.sulfoglycoglycerolipids are marked by the α-configuration of the anomeric carbon and the βconfiguration type is not present in nature.Sulfavant A (1) activates human DCs and triggers an efficient immune response in vivo [4].After treatment, DCs show up-regulation of MHC II molecules and other co-stimulatory proteins (CD83, CD86), as well as production of pro-inflammatory cytokines (IL-12 and INF-γ).Mice immunized with OVA associated to Sulfavant A produced a release of anti-OVA Ig comparable to traditional adjuvants, while, in an experimental model of melanoma, vaccination of C57BL/6 mice with Sulfavant A-adjuvanted hgp10 peptide elicited a protective response leading to a promising reduction in tumor growth and increase of mice survival [4].
Here we discuss in detail the strategy for the synthesis of Sulfavant A (1), including an overview of the main technical hurdles we have faced, together with the considerations for the scale-up of the process and the possibility to expand the class of synthesizable analogs.

Results and Discussion
For the preparation of Sulfavant A, the main challenges lay in the requirement of planning a chemical strategy comprising the introduction of the β-glycosidic bond at the anomeric carbon and the insertion of the sulfonic function on C6 of glucose (Scheme 1).At the same time, the possibility of introducing different and unsaturated fatty acyl residues on the glycerol moiety would make the strategy expandable to a wide class of analogs.Furthermore, the process had to be suitable to scaleup for further pharmacological development.The first crucial point was the choice of orthogonal protecting groups on sugar, whose deprotection should not affect the diacylglycerol aglycon part and the glycosidic linkage.In this frame, tert-butyldimethylsilyl and benzyl groups to which the synthesis of α-and β-SQDGs [14,21,22] had already been applied, were initially tested for the preparation of the sulfolipid.Unfortunately, the tert-butyldimethylsilyl group proved to be too labile in the experimental conditions and this, together with its low deprotection efficiency, led to a considerable decrease in yields; whereas the use of benzyl excluded the possibility of introducing unsaturated acyl chains to the glycerol, and greatly reduced the possibility of extending the synthesis to other analogs.For these reasons, we decided to explore the use of acetate as transient protecting group for the hydroxy functions of the sugar moiety.The acetoxy derivatives are stable under a wide range of conditions and, most importantly, the presence of the acetate residue at C-2′ is capable of ensuring an excellent stereoselectivity in the Scheme 1. Retrosynthetic scheme for preparation of Sulfavant A (1) and its analogs.
The first crucial point was the choice of orthogonal protecting groups on sugar, whose deprotection should not affect the diacylglycerol aglycon part and the glycosidic linkage.In this frame, tert-butyldimethylsilyl and benzyl groups to which the synthesis of αand β-SQDGs [14,21,22] had already been applied, were initially tested for the preparation of the sulfolipid.Unfortunately, the tert-butyldimethylsilyl group proved to be too labile in the experimental conditions and this, together with its low deprotection efficiency, led to a considerable decrease in yields; whereas the use of benzyl excluded the possibility of introducing unsaturated acyl chains to the glycerol, and greatly reduced the possibility of extending the synthesis to other analogs.For these reasons, we decided to explore the use of acetate as transient protecting group for the hydroxy functions of the sugar moiety.The acetoxy derivatives are stable under a wide range of conditions and, most importantly, the presence of the acetate residue at C-2 is capable of ensuring an excellent stereoselectivity in the formation of the β glycosidic bond by anchimeric assistance [23,24].On the other hand, acetate removal implies basic conditions that are not suitable for preserving the ester linkage of glycerol with the acyl chains.To this end, a careful tuning of hydrolysis by hydrazine was crucial to completing the synthesis of the sulfolipid [23,24].
As reported in Scheme 2, the total synthesis of Sulfavant A (1) started with acetylation of D-glucose, followed by selective deacetylation at the anomeric position with benzylamine and derivatization with trichloroacetimidate.The subsequent coupling with 1,2-O-isopropylidene glycerol by Schmidt methodology [25,26] gave the derivative 5 in high yields (about 80%).As expected, the acetate groups determined the high stereoselectivity of the reaction (β/α ratio 95:5) with the β-orientation of the glycosidic bond by neighboring group participation.
Mar. Drugs 2017, 15, 288 3 of 10 formation of the β glycosidic bond by anchimeric assistance [23,24].On the other hand, acetate removal implies basic conditions that are not suitable for preserving the ester linkage of glycerol with the acyl chains.To this end, a careful tuning of hydrolysis by hydrazine was crucial to completing the synthesis of the sulfolipid [23,24].As reported in Scheme 2, the total synthesis of Sulfavant A (1) started with acetylation of D- glucose, followed by selective deacetylation at the anomeric position with benzylamine and derivatization with trichloroacetimidate.The subsequent coupling with 1,2-O-isopropylidene glycerol by Schmidt methodology [25,26] gave the derivative 5 in high yields (about 80%).As expected, the acetate groups determined the high stereoselectivity of the reaction (β/α ratio 95:5) with the β-orientation of the glycosidic bond by neighboring group participation.The next step was the removal of the isopropylidene residue under soft acidic conditions to preserve the glycosidic linkage.Use of zinc nitrate hexahydrate in acetonitrile allowed hydrolysis in quantitative yields [27], whereas N,N-dicyclohexylcarbodiimide (DCC)-mediated condensation with two equivalents of stearic acid yielded the diacyl derivative 7.At this point, with the aim of introducing the sulfonic function on carbon 6′, a selective removal of sugar acetyl groups without affecting the acyls on glycerol was necessary to obtain β-glucosyl-distearoyl-glycerol 8.This was achieved by an accurate setting of the basic conditions that drove us to use monohydrate hydrazine The next step was the removal of the isopropylidene residue under soft acidic conditions to preserve the glycosidic linkage.Use of zinc nitrate hexahydrate in acetonitrile allowed hydrolysis in quantitative yields [27], whereas N,N-dicyclohexylcarbodiimide (DCC)-mediated condensation with two equivalents of stearic acid yielded the diacyl derivative 7.At this point, with the aim of introducing the sulfonic function on carbon 6 , a selective removal of sugar acetyl groups without affecting the acyls on glycerol was necessary to obtain β-glucosyl-distearoyl-glycerol 8.This was achieved by an accurate setting of the basic conditions that drove us to use monohydrate hydrazine in an amount of 2.4 moles per acetate unit at a temperature below 45 • C [23,24].After trytilation of the primary alcohol of glucose (compound 9, 70 %), acetylation (compound 10, 89%), detrytilation [28] (compound 11, 82%) and subsequent introduction on C-6 of the tosyl function (compound 12, 80%) permitted the insertion of a thioacetate group (compound 13, 93%) [29].Final oxidation by hydrogen peroxide gave the corresponding sulfonic acid (compound 14, 65%), which was fully deacetylated by monohydrate hydrazine to give Sulfavant A (1) (68%), whose NMR data are reported in Table 1.Synthesis of Sulfavant A (1) was characterized by clean and efficient steps, along with a simple and easy workup, which makes the procedure suited for subsequent and crucial scaling-up.In fact, we did not face any problems in the transition from small to larger volumes.Furthermore, the workup did not require any particular experimental change in order to prepare hundreds of milligrams (400 mg) of the product.On the basis of this process, we also expect that the synthetic procedure is able to provide grams of the vaccine adjuvant for in vivo tests and future pharmacological development.
The design of the synthetic strategy also opens possibilities for the preparation of a large range of analogs, including compounds with mixed composition of fatty acids or unsaturated acyl chains on glycerol moiety.An example of this is reported in Scheme 3 for the synthesis of 17, 1-O-stearoyl-2-O-palmitoyl-3-O-β-D-sulfoquinovosylglycerol, which contains two different fatty acids resgiospecifically distributed at the two positions of glycerol.
In this case, use of DCC at 4 • C permitted a good control of the step-wise introduction of two different acyl chains and gave yields of the monoacyl derivative 15 (81%) significantly higher than those obtained by other methods [30][31][32].Acylation of the primary hydroxy residue was largely favored in the first coupling reaction with equimolar amounts of fatty acid, and we did not observe any byproduct at C-2 of 15, as indicated by the NMR spectra, which showed deshielding of H 2 -1 signals at δ H 4.21-4.11(long range correlation of H 2 -1 to a carboxylic ester at δ C 171.1) and δ H value of H-2, linked to non-acylated oxygen, at 3.91 ppm (Figure 2).The design of the synthetic strategy also opens possibilities for the preparation of a large range of analogs, including compounds with mixed composition of fatty acids or unsaturated acyl chains on glycerol moiety.An example of this is reported in Scheme 3 for the synthesis of 17, 1-O-stearoyl-2-O-palmitoyl-3-O-β-D-sulfoquinovosylglycerol, which contains two different fatty acids resgiospecifically distributed at the two positions of glycerol.In this case, use of DCC at 4 °C permitted a good control of the step-wise introduction of two different acyl chains and gave yields of the monoacyl derivative 15 (81%) significantly higher than those obtained by other methods [30][31][32].Acylation of the primary hydroxy residue was largely favored in the first coupling reaction with equimolar amounts of fatty acid, and we did not observe any byproduct at C-2 of 15, as indicated by the NMR spectra, which showed deshielding of H2-1 signals at δH 4.21-4.11(long range correlation of H2-1 to a carboxylic ester at δC 171.1) and δH value of H-2, linked to non-acylated oxygen, at 3.91 ppm (Figure 2).The introduction of acetate as protecting group also offers access to unsaturated analogs, as long as it is possible to control the sensitivity of double bonds to both the iodine and the hydrogen peroxide/acetic acid/acetate systems that are used for removal of the trityl group and sulfur oxidation, respectively.As shown in Scheme 4, this could be achieved by a slight modification of the synthetic procedure, and the use of an acidic resin to deprotect the primary hydroxy group and molybdate oxidation for the final step.Soft acidic conditions by Dowex H + at 50 °C in methanol/water 95/5 allowed a complete removal of the trityl group while preserving the glycosidic bond of 19.The reaction led to formation of a small amount of other hydrolyzed byproducts, which were removed by chromatographic purification, but resulted in a slight decrease in the synthetic yield in comparison with the original step used for Sulfavant A. Molybdate oxidation has been used already for synthesis of α-SQDGs [14], but application in the preparation of the β-SQDG required a careful adjustment of the reaction time.Increase in the temperature from room value to 50 °C, together with the use of twice the amount of ammonium heptamolybdate tetrahydrate and hydrogen peroxide, favoured the course of the reaction and led to the dioleoyl analog 22 with yield close to 60%.The introduction of acetate as protecting group also offers access to unsaturated analogs, as long as it is possible to control the sensitivity of double bonds to both the iodine and the hydrogen peroxide/acetic acid/acetate systems that are used for removal of the trityl group and sulfur oxidation, respectively.As shown in Scheme 4, this could be achieved by a slight modification of the synthetic procedure, and the use of an acidic resin to deprotect the primary hydroxy group and molybdate oxidation for the final step.Soft acidic conditions by Dowex H + at 50 • C in methanol/water 95/5 allowed a complete removal of the trityl group while preserving the glycosidic bond of 19.The reaction led to formation of a small amount of other hydrolyzed byproducts, which were removed by chromatographic purification, but resulted in a slight decrease in the synthetic yield in comparison with the original step used for Sulfavant A. Molybdate oxidation has been used already for synthesis of α-SQDGs [14], but application in the preparation of the β-SQDG required a careful adjustment of the reaction time.Increase in the temperature from room value to 50 • C, together with the use of twice the amount of ammonium heptamolybdate tetrahydrate and hydrogen peroxide, favoured the course of the reaction and led to the dioleoyl analog 22 with yield close to 60%.

Materials and Methods
1D-and 2D-NMR spectra were recorded on a Bruker Avance-400 (400.13MHz) and on a Bruker DRX-600 equipped with a TXI CryoProbe in CDCl3 and CD3OD (δ values are referred to CHCl3 and CH3OH at 7.26 and 3.34 ppm respectively) and 13 C NMR spectra were recorded on a Bruker DPX-300 (75.47 MHz) (δ values are referred to CDCl3 and CD3OD at 77.0 and 49.0 ppm respectively).HRESIMS was conducted on a Micromass Q-TOF micro.TLC plates (Kieselgel 60 F254) and silica gel powder (Kieselgel 60, 0.063-0.200mm) were from Merck.
All the reagents were purchased from Sigma-Aldrich and used without any further purification.

Preparation and Characterization of Synthetic Molecules
All the experimental procedures and spectral data of Sulfavant A (1) and synthetic intermediates 2-14 were reported in reference 4.

Materials and Methods
1D-and 2D-NMR spectra were recorded on a Bruker Avance-400 (400.13MHz) and on a Bruker DRX-600 equipped with a TXI CryoProbe in CDCl 3 and CD 3 OD (δ values are referred to CHCl 3 and CH 3 OH at 7.26 and 3.34 ppm respectively) and 13 C NMR spectra were recorded on a Bruker DPX-300 (75.47 MHz) (δ values are referred to CDCl 3 and CD 3 OD at 77.0 and 49.0 ppm respectively).HRESIMS was conducted on a Micromass Q-TOF micro.TLC plates (Kieselgel 60 F 254 ) and silica gel powder (Kieselgel 60, 0.063-0.200mm) were from Merck.
All the reagents were purchased from Sigma-Aldrich and used without any further purification.

Preparation and Characterization of Synthetic Molecules
All the experimental procedures and spectral data of Sulfavant A (1) and synthetic intermediates 2-14 were reported in reference 4.