Design and synthesis of 4-azido-phosphatidylinositol as a potential probe for metabolic engineering of glycosylphosphatidylinositol on cells

Abstract

A diacyl phosphatidylinositol (PI) derivative with an azide linked to its inositol C4-position was effectively synthesized in 19 steps for the longest linear sequence and in a ca. 1% overall yield from 1,2-distearoyl-sn-glycerol and D-glucose. This compound was designed as a biosynthetic precursor of glycosylphosphatidylinositol (GPI) anchors. Its azide would enable further modification to introduce other molecular tags by biocompatible click reaction. Therefore, it can be a useful probe for metabolic engineering of cell surface GPI anchors and GPI-anchored proteins.

Graphical Abstract

Introduction

Glycosylphosphatidylinositol (GPI) anchors are a class of complex glycolipids, whose attachment to the protein C-terminus is one of the most common and critical post-translational modifications in eukaryotic cells.^1–3 A key function of GPIs is to anchor proteins onto the cell surface through embedding the GPI lipophilic tails in the lipid bilayer of cell membrane (Figure 1). To date, about 300 GPI-anchored proteins (GPI-APs) have been discovered and demonstrated to play a pivotal role in many biological processes, such as cell signaling, immune responses, etc.^4–6 It has also been shown that GPI-APs are involved in various human diseases. For example, there is an increase in GPI-APs on cancer cell surfaces, suggesting the potentials of GPI anchors and GPI-APs as targets for cancer diagnosis and therapy.^{7, 8}

Figure 1.

The conserved core structure of GPI anchors, as well as GPI insertion into the cell membrane

To investigate the properties and functions of GPI anchors and GPI-APs on the cell surface, it is necessary to have proper approaches and molecular tools for their selective labeling and analysis on live cells. To this end, we have investigated a metabolic engineering-based strategy for the labeling of cell surface GPI anchors and GPI-APs with azide-modified inositol derivatives as the GPI biosynthetic precursors.^{9, 10} Although this strategy was proved to be efficient, the inositol derivatives needed to be O-acetylated to achieve the desired cell membrane permeability and metabolic engineering efficiency,¹⁰ or needed to be modified with a peptide that is processed only in cancer cells to accomplish cancer-selective metabolic engineering.⁹ Furthermore, these inositol derivatives can also be used by other biosynthetic pathways involving inositol,¹¹ thus they are not very selective to GPI anchors and GPI-APs. The present work intended to address these issues by developing new probes for more efficient and selective metabolic engineering of cell surface GPI anchors and GPI-APs.

Results and Discussion

Natural GPI anchors are structurally diverse, but they share the highly conserved core structure as depicted in Figure 1, suggesting the conserved biosynthetic pathways for eukaryotes from yeasts to humans. This GPI core consists of a tetrasaccharide chain and a phosphatidylinositol (PI) motif, with the former moiety attached to the inositol 6-O-position of the latter.¹² The GPI biosynthetic pathway, as outlined Figure 2, involves the PI as a key synthetic intermediate. Starting from the PI, the GPI biosynthesis proceeds through two distinctive and well-regulated stages. During the first stage, an N-acetylglucosamine (GlcNAc) residue is enzymatically linked to the inositol 6-O-position of the PI, which is followed by enzymatic de-N-acetylation. These transformations occur on the cytoplasmic membrane surface of the endoplasmic reticulum (ER). During the second stage, the glycolipid is translocated onto the lumen side of the ER membrane, where other modifications are performed to generate GPI anchors that are eventually attached to the target proteins.^{13, 14} As the first biosynthetic stage for GPI anchors occurs on the cytoplasmic site of ER membrane, the PI can be an excellent target or point of intervention for the metabolic engineering of GPI anchor biosynthesis. This research will develop new molecular tools for the metabolic engineering of GPI anchors and GPI-APs based on PI.

Figure 2.

Schematic representation of the first few steps in the biosynthesis of GPI anchors and GPI-APs

Figure 3 has illustrated our newly designed azido derivative 1 of diacyl PI as the biosynthetic precursor of GPI anchors and GPI-APs. A key feature of this metabolic engineering probe is the location of its azido group. The inositol 2-O-position is directly involved in the GPI biosynthesis while its 6-O-position is reserved for glycan installation,^{12, 15} hence we chose to use an azido group to substitute the inositol 4-hydroxyl group, which is not directly involved in GPI-AP biosynthesis. In the meantime, the inositol C4-position also has the furthest distance from 2-O-and 6-O-positions. Moreover, the azide group is small and, indeed, not much bigger than a hydroxyl group. Thus, we anticipated that an azido group at the C4-position would not significantly affect the incorporation of 1 into the biosynthetic pathways of GPI anchors and GPI-APs. Finally, replacing the inositol 4-OH group with an azido group will prohibit the probe from participating in 4-O-phophorylation and subsequently 5-O- and 3-O-phophorylations involved in phosphoinositide biosynthesis,¹⁶ thereby to inhibit the formation of azide-labelled phosphoinositides, such as PI-(4), (4,5) and (3,4,5)-phosphates, on plasma membranes. As a result, probe 1 would be selective for the metabolic engineering of GPI anchors and GPI-APs.

Figure 3.

Designed PI analog 1 as a GPI probe

The new PI analog 1 as a probe for GPI anchor metabolic engineering will also have several other advantages over our previous probes.^{9, 10} First, intervening the GPI biosynthetic pathway at the PI stage will skip several necessary steps for GPI anchor biosynthesis, which are some of the limiting factors for the metabolic engineering of GPI anchors using inositol derivatives. Second, many studies have demonstrated that cells can readily incorporate glycolipids,^{17, 18} thus it would be feasible to directly apply 1 to cell metabolic engineering. Third, 1 will participate in the GPI biosynthetic pathway without the inositol deacetylation step, which was required for the previous probes. Finally, 1 will target the cytoplasm side of ER membrane, which should be more easily achievable than targeting the ER lumen. All these factors will help the cell incorporation of 1 and its participation in GPI anchor and GPI-AP biosynthesis. Thus, we anticipate 1 to have increased metabolic engineering efficiency as compared to previous probes. Once 1 is incorporated in GPI anchors and GPI-APs by cells, the azide group in its structure can be used as a molecular handle to introduce other functionalities, such as affinity and fluorescent tags, through a biocompatible click reaction to facilitate various biological studies and other applications.^19–21

For the synthesis of 1, a challenge is that the final product contains an azido group and two ester linkages. As a result, the commonly used protecting groups, such as the acyl and benzyl groups, to mask hydroxyl groups were not applicable, since the acyl and azido groups in the final product would be sensitive to the basic conditions involved in deacetylation and catalytic hydrogenolysis reaction employed for debenzylation, respectively. Thus, we decided to employ the acid-labile protecting groups, such as the para-methoxybenzyl (PMB) and methoxymethyl (MOM) groups, for permanent protection of the hydroxyl groups. In accordance with this protection tactic, we designed a synthesis for 1, as outlined in Scheme 1, with D-glucose 2 as a starting material. First, 2 was converted into 3 in 13 steps, following a method reported by the Swarts group.²² Then, the keto group was reduced with NaB(OAc)₃H under mildly acidic conditions, which was show to be stereoselective to afford the desired PMB-protected azido inositol derivative 4.²² This was followed by the protection of the two free hydroxyl groups in 4 with MOM groups to obtain fully protected 5. Here, we used MOM as it could be introduced under mildly basic conditions that would not affect the O-acetyl group. Subsequently, the acetyl group in 5 was chemoselectively removed with NaOMe to give the key intermediate 6, which was ready for the installation of the phospholipid moiety. All the reactions went smoothly to afford good to very good yields.

Scheme 1.

Synthesis of azido PI 1

In the meanwhile, the phosphorylating reagent—phosphorimidate 8—was prepared from 7 following a literature procedure.²³ Phospholipidation of 6 was accomplished via a one-pot two-step method, involving the reaction of 6 and 8 in the presence of 1H-tetrazole and then oxidation of the resultant phosphite ester with t-BuO₂H, to provide 9 (79%) as a diastereomeric mixture, caused by the stereogenic phosphorous atom. It should be noted that since phosphorimidate 8 was quite sensitive to moisture, it was used in large excess (> 5 equivalents) in this reaction. Finally, global deprotection of 9 was achieved in two steps, including removal of the cyanoethyl group utilizing 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) and then deprotection of the PMB and MOM ethers employing 10% trifluoroacetic acid (TFA),²⁴ to provide the synthetic target 1 in a 77% yield. The final product, as well as all the synthetic intermediates, was also verified with NMR and HR-MS data (Supporting Information).

Conclusion

Currently, the investigation of cell surface GPI anchors and GPI-APs represents a significant challenge, whereas an effective probe for GPI anchor labeling by metabolic engineering is highly attractive. In this study, we have designed an azide-modified PI derivative 1 as a novel probe for metabolic engineering of GPI anchors and GPI-APs on cells, and developed a facile and effective method for the chemical synthesis of this molecular probe. As mentioned above, compared to the probes reported previously for GPI anchor metabolic engineering^{9, 10} or for phosphoinositide labeling,²⁵ the new probe 1 can have several advantages. For example, it targets the cytoplasm side, instead of the lumen side, of the ER membrane and will be incorporated into the GPI biosynthesis pathway at a later stage and be more selective for GPI anchors. Moreover, cells may use 1 directly without the need to process it before incorporation in GPI biosynthesis. Therefore, 1 should be more effective for the metabolic engineering of cell surface GPIs and GPI-APs than our previous probes. Presently, biological studies on the new molecular probe are underway in our lab, and the results will be reported in due course.

Experimental

General protocols

Commercial chemicals and materials were used without further purification. Flash chromatography was performed using silica gel 60 (230–400 mesh). Thin layer chromatography (TLC) was performed on silica gel 60G F₂₅₄ 25 glass plates detected with UV, charring with 10% (v/v) H₂SO₄ in ethanol, and p-anisaldehyde staining. ¹H, ¹³C, COSY, and HSQC NMR spectra were obtained with 400 and 600 MHz NMR spectrometers. Coupling constants (J) were reported in hertz (Hz). Chemical shifts (δ) were reported in ppm referenced to CDCl₃ (¹H NMR: δ 7.26 ppm; ¹³C NMR: δ 77.16 ppm) or MeOD (¹H NMR: δ 3.31 ppm; ¹³C NMR: δ 49.0 ppm). Abbreviations used to describe peak splitting patterns are: s, singlet; d, doublet; t, triplet; dd, double doublets; ddd, double of double doublets; m, multiplet. High Resolution Electrospray Ionization (HR-ESI) MS spectra were obtained with an Agilent 6230 Time-of-Fly (TOF) machine. Aluminum heating blocks were used for heating reaction mixtures.

Synthesis of 1-O-Acetyl-4-azido-4-deoxy-3,5-di-O-(p-methoxybenzyl)-myo-inositol (4)

Compound 3 (0.187 g, 0.385 mmol), which was prepared from D-glucose 2 according to a reported method,²² was dissolved in CH₃CN and cooled to 0 °C. Then, NaB(OAc)₃H (0.644 g, 2.90 mmol) was added, followed by AcOH (1 mL, 17.47 mmol). After being stirred at this temperature for 2 h, the reaction mixture was diluted with water and extracted with CH₂Cl₂. The organic layers were combined and dried with Na₂SO₄, filtered, and condensed under vacuum. The crude product was purified via silica gel column chromatography to give 4 as a white solid (120 mg, 69%). ¹H NMR (400 MHz, CDCl₃) δ 7.34 − 7.28 (m, 4H), 6.93 − 6.87 (m, 4H), 4.87 (d, J = 10.7 Hz, 1H), 4.71 − 4.63 (m, 2H), 4.61 (d, J = 2.7 Hz, 2H), 4.25 (t, J = 2.7 Hz, 1H), 4.06 (t, J = 9.8 Hz, 1H), 3.84−3.80 (m, 7H), 3.34 (dd, J = 10.0, 2.7 Hz, 1H), 3.13 (t, J = 9.6, 1H), 2.36 (s, 1H), 2.20 (s, 1H), 2.16 (s, 3H). ¹H NMR data matched those in the literature.²²

Synthesis of 1-O-Acetyl-4-azido-4-deoxy-3,5-di-O-(p-methoxybenzyl)-1,2-di-O-(methoxymethyl)-myo-inositol (5).

Compound 4 (49 mg, 0.085 mmol) dissolved in anhydrous DMF (0.2 mL) was cooled to 0 °C. DIPEA (210 μL, 1.2 mmol) was added to this solution, which was followed by MOMCl (60 μL, 0.79 mmol). The reaction mixture was heated to 45 °C and stirred at this temperature for 16 h. The reaction mixture was diluted with water and then extracted with CH₂Cl₂ (15 mL × 3). The organic layers were combined, dried with Na₂SO₄, filtered, and condensed under reduced pressure. The crude product was purified with silica gel column chromatography to provide 5 as a colorless syrup (39 mg, 80%). R_f = 0.45 (EtOAc:Hex 4:6). ¹H NMR (600 MHz, CDCl₃): δ 7.30 (m, 4H), 6.88, (m, 4H), 4.82 (d, J = 6.5, 1H), 4.80 (d, J = 10.2, 1H), 4.73 − 4.60 (m, 6H), 4.53 (d, J = 11.0, 1H), 4.24 (t, J = 2.5, 1H), 4.03 (t, J = 9.4, 1H), 3.87 (t, J = 10.1, 1H), 3.81 (s, 3H), 3.80 (s, 3H), 3.36 (s, 3H), 3.35 (s, 3H), 3.27 (dd, J = 10.3, 2.3, 1H), 3.20 (t, J = 9.5, 1H), 2.12 (s, 3H). ¹³C NMR (151 MHz, CDCl₃): δ 170.5, 159.6, 159.5, 130.2, 129.9, 129.7, 129.4, 114.1, 114.0, 98.4, 97.5, 81.4, 77.8, 77.3, 75.6, 72.5, 72.2, 65.2, 56.2, 55.9, 55.4 (2C), 21.2. HR-MS (ESI-TOF) m/z: Calcd for C₂₈H₃₇N₃O₁₀Na [M + Na]⁺ 598.2371; Found 598.2378.

Synthesis of 4-Azido-4-deoxy-3,5-di-O-(p-methoxybenzyl)-1,2-di-O-(methoxymethyl)-myo-inositol (6)

Compound 5 (50 mg, 0.094 mmol) was dissolved in MeOH (0.86 mL). To this solution was added 0.5 M NaOMe (0.2 mL) at rt. The reaction mixture was stirred at rt for 2 h before neutralization with Amberlyst H⁺ resin followed by filtration and condensation. The product was purified via silica gel column chromatography to afford 6 (42 mg, 84%) as a white solid. R_f = 0.21 (EtOAc:Hex 4:6). ¹H NMR (600 MHz, CDCl₃): δ 7.30 (m, 4H), 6.87 (m, 4H), 4.84 (d, J = 6.7, 1H), 4.76 (m, 3H), 4.70−4.67 (m, 2H), 4.65 (d, J = 11.4, 1H), 4.57 (d, J = 11.4, 1 H), 4.12 (t, J = 2.5, 1H), 3.86 (t, J = 10.1, 1H), 3.81 (s, 3H), 3.80, (s, 3H), 3.72 (t, J = 9.4, 1H), 3.43 (s, 3H), 3.39 (s, 3H), 3.32 (m, 1H), 3.20 (dd, J = 10.4, 2.4, 1H), 3.17 (t, J = 9.5, 1H). ¹³C NMR (151 MHz, CDCl₃): δ 159.4, 159.3, 129.9, 129.7, 114.0 (2C), 98.7, 96.0, 83.3, 82.8, 81.1, 77.4, 75.5, 75.3, 71.8, 71.0, 65.3, 56.1, 56.0, 55.4 (2C). HR-MS (ESI-TOF) m/z: Calcd for C₂₆H₃₅N₃O₉Na [M + Na]⁺ 556.226; Found 556.2275.

Synthesis of 4-Azido-1-O-[(2-cyanoethoxyl)-(2,3-di-O-stearoly-sn-glycerol)-phosphono]-4-deoxy-3,5-di-O-(p-methoxybenzyl)-1,2-di-O-(methoxymethyl)-myo-inositol (9).

Compound 6 (30 mg, 0.056 mmol) was dissolved in CH₂Cl₂ and CH₃CN (3:1, 10.8 mL). Activated 4 Å molecular sieves and tetrazole (0.45 M in hexane, 1.25 mL, 0.313 mmol) were added at rt, followed by dropwise addition of 8 (240 mg, 0.291 mmol) that was dissolved in CH₂Cl₂. After 30 min of stirring, the mixture was cooled to 0 °C, and t-BuOOH (0.2 mL, 2.07 mmol) was added. After 1 h of stirring, the mixture was diluted with CH₂Cl₂ and washed with aqueous NaHCO₃. The aqueous layer was extracted with CH₂Cl₂. The organic layers were combined, washed with brine, dried with Na₂SO₄, filtered, and condensed. The product was purified via silica gel column chromatography to give 9 (56 mg, 79%) as a white solid. R_f = 0.18 (EtOAc:Hex 4:6). ¹H NMR (600 MHz, CDCl₃): δ 7.34 − 7.28 (m, 4H), 6.92 − 6.85 (m, 4H), 5.32 – 5.37 (m, 1H), 4.85 − 4.82 (m, 1H), 4.81 (d, J = 10.2 Hz, 1H), 4.78 − 4.75 (m, 2H), 4.74 − 4.70 (m, 1H), 4.70 − 4.61 (m, 2H), 4.55 (d, J = 11.0 Hz, 1H), 4.38 − 4.27 (m, 5H), 4.27 − 4.21 (m, 2H), 4.20 − 4.14 (m, 2H), 4.14 − 4.08 (m, 1H), 4.03 − 3.98 (m, 1H), 3.87 − 3.83 (m, 2H), 3.82 − 3.80 (m, 6H), 3.40 − 3.38 (m, 3H), 3.35 − 3.33 (m, 3H), 3.27 (td, J = 10.0, 2.4 Hz, 1H), 3.19 − 3.13 (m, 1H), 2.81 − 2.75 (m, 2H), 2.37 − 2.29 (m, 4H), 1.67 − 1.50 (m, 8H), 1.34 − 1.21 (m, 67H), 0.88 (t, J = 7.0 Hz, 6H). ¹³C NMR (151 MHz, CDCl₃): δ 173.4, 173.0, 159.6, 159.5, 130.1, 129.9 (2C), 129.8 (3C), 129.6, 129.5, 116.6 (2C), 114.1, 114.0, 98.6, 97.7 (2C), 81.0 (2C), 77.7, 77.6, 75.5 (2C), 73.3, 73.2, 72.4, 72.3, 71.9, 71.0, 69.5 (2C), 66.2 (2C), 66.0 (2C) 65.2, 62.4, 62.2, 61.7, 60.6, 56.8, 56.7, 56.1, 56.0, 55.4 (2C), 46.5 (2C), 34.3, 34.2 (2C), 32.1, 31.6, 29.9, 29.8, 29.7 (3C), 29.5 (4C), 29.3 (2C), 25.0 (2C), 22.8, 22.6, 21.2, 19.9, 19.8 (3C), 14.4, 14.3. ³¹P NMR (243 MHz, CDCl₃): δ −1.98, −2.04. HR-MS (ESI-TOF) m/z: Calcd for C₆₈H₁₁₃N₄O₁₆PNa [M + Na]⁺ 1295.7781; Found 1295.7791.

Synthesis of 4-Azido-4-deoxy-1-O-[(2,3-di-O-stearoyl-sn-glycerol)-phosphono]-myo-inositol (1).

To a solution of 9 (20 mg, 0.022 mmol) in anhydrous CH₂Cl₂ (7.8 mL) was added DBU (22 uL, 0.15 mmol) at rt. The solution was stirred for 1 h and then cooled to 0 °C. A 20% TFA solution in CH₂Cl₂ (7.8 mL) was added to the reaction mixture to yield an overall concentration of 10% TFA. The solution was warmed to room temperature and stirred for 3 h, and then diluted with toluene and condensed under reduced pressure. The remaining was co-evaporated with toluene 3 times, and the product was purified via silica gel column chromatography to yield 1 as a white solid (15 mg, 77%). R_f = 0.15 (MeOH:CH₂Cl₂ 15:85) ¹H NMR (600 MHz, MeOD:CDCl₃ 3:2): δ 5.22 (bs, 1H), 4.40 (dd, J = 12.0, 3.1 Hz, 1H), 4.17 − 4.04 (m, 2H), 4.04 − 3.94 (m, 2H), 3.81 − 3.78 (m, 2H), 3.59 − 3.55 (m, 1H), 3.39 (dd, J = 10.2, 2.9 Hz, 1H), 3.16 (t, J = 9.3 Hz, 1H), 2.33 − 2.23 (m, 4H), 1.61 − 1.53 (m, 4H), 1.34 − 1.15 (m, 56H), 0.85 (t, J = 6.9 Hz, 6H). ¹³C NMR: (151 MHz, MeOD:CDCl₃ 3:2): δ 174.6, 174.2, 76.9 (2C), 73.9, 72.4, 72.2, 71.1, 71.0, 66.4, 64.2, 63.2, 34.7, 34.6, 34.5, 32.4, 31.2, 30.1 (3C) 30.0 (3C), 29.9, 29.8 (2C), 29.7, 29.6 (2C), 25.4 (2C), 25.3, 24.7, 23.1, 14.3. ³¹P NMR (243 MHz, MeOD:CDCl₃ 2:3): δ 0.52. HR-MS (ESI-TOF) m/z: Calcd for C₄₅H₈₅N₃O₁₂P [M - H]⁻ 890.5876; Found 890.5855.