Effects of Heparan Sulfate Trisaccharide Containing Oleanolic Acid in Attenuating Hyperphosphorylated Tau-Induced Cell Dysfunction Associated with Alzheimer’s Disease

Abstract

Alzheimer’s disease (AD) is the most common form of dementia, marked by progressive brain degeneration and cognitive decline. A major pathological feature of AD is the accumulation of hyperphosphorylated tau (p-tau) in the form of neurofibrillary tangles (NFTs), which leads to neuronal death and neurodegeneration. P-tau also induces endoplasmic reticulum (ER) stress and activates the unfolded protein response, causing inflammation and apoptosis. Additionally, p-tau spreads in the brain through interactions with heparan sulfate (HS) proteoglycans, promoting aggregation and internalization. Targeting the tau–HS interaction offers a potential therapeutic strategy for AD. We present a novel HS mimetic with a lipophilic oleanolic acid linker and a sulfated trisaccharide, which shows strong cytoprotective effects against p-tau. Moreover, this compound alleviates p-tau-induced ER stress and inflammation. Molecular docking studies indicate that the conjugation of oleanolic acid enhances binding between the ligand and tau protofilament cores, facilitating protective interactions. These findings provide a foundation for the development of novel HS mimetics, enabling further investigation of tau-HS interactions in AD and other tauopathies.

Graphical Abstract

INTRODUCTION

Alzheimer’s disease (AD) is characterized by a progressive decline in memory and cognitive dysfunctions in affected individuals. The key pathological features of AD include the accumulation and deposition of extracellular amyloid β (Aβ) plaques and the aggregation of hyperphosphorylated tau (p-tau) in the brain (Figure 1). Current clinical treatment options for AD include three acetylcholine (ACh) inhibitors (donepezil, rivastigmine, galantamine) and one N-methyl-d-aspartate receptor antagonist (memantine).¹ While these medications can improve cognition and relieve dementia symptoms, they do not stop or reverse disease progression. Initially, the deposition of Aβ-peptide in senile plaques was considered crucial in AD development.² However, the lack of success in clinical trials and the limited efficacy of the two newly FDA-approved antiamyloid antibodies, aducanumab and lecanemab, have recently shifted the focus toward antitau strategies. This shift is supported by growing evidence demonstrating the involvement of tau abnormalities in AD.³

Figure 1.

Healthy human brain and human brain with Alzheimer’s Disease.

Tau is a neuronal microtubule-associated protein (MAP) that aids tubulin assembly and regulates axonal transport (Figure 1).⁴ In the central nervous system, six tau isoforms are regulated by alternative splicing and range from 352 to 441 amino acids. The longest tau isoform has 85 serine, threonine, and tyrosine residues that can be sites for phosphorylation. Our prior findings indicate that the phosphorylated tau protein exhibits phosphorylation levels and site-specific modifications that closely resemble those associated with disease-related p-tau observed in patients with Alzheimer’s disease.⁵ While phosphorylated tau is crucial for cytoskeletal plasticity during embryonic development, its levels are lower in a healthy adult brain.^6,7 In AD, tau becomes hyperphosphorylated,^6,8 disrupting the microtubules and aggregating into filamentous neurofibrillary tangles (NFTs) (Figure 1),^5,9 which are neurotoxic and associated with cognitive decline.^5,10–12 The presence and spread of NFTs closely correlate with the severity of cognitive impairment.¹³ Notably, there is a lack of spatiotemporal association between NFT and cognitive decline in Aβ plaques.

Recent findings show that p-tau spreads transcellularly in the brain in a diffusible prion-like fashion.^14–23 One of the likely mechanisms of p-tau spread between neurons is the interaction between tau and cell surface heparan sulfate proteoglycans (HSPGs, Figure 2A).^24–30 This includes the initial release of intracellular tau protein into the extracellular matrix, the uptake of pathogenic tau seeds by cells, and the self-assembly of tau into larger aggregates. Other potential pathways have also been suggested, including binding and internalizing pathogenic tau through the LRP1 receptor.³⁰ Preventing the prion-like spread of p-tau is a promising approach for attenuating the progression of AD-affected individuals.

Figure 2.

(A) Interactions of HSPG and tau protein. (B) Structure of HS polysaccharide.

Heparan sulfate (HS) chains attached to core proteins and interact with extracellular matrix (ECM) proteins, which are crucial for ECM integrity and cell-ECM communication.³¹ HS regulates the activity of various growth factors and blood coagulation factors at the cell surface and within the ECM.^31,32 It is a linear, highly sulfated, and polydisperse glycosaminoglycan made up of repeating disaccharide units of d-glucosamine (GlcN) and uronic acid [d-glucuronic acid (GlcA) or l-iduronic acid (IdoA)] with varying degrees of sulfation pattern at various sites (Figure 2B).³¹ Structurally, HS is similar to the anticoagulant heparin. However, they differ in two aspects: (1) heparin is found in mast cells within the peripheral system, whereas HS is located on the cell surface; and (2) heparin contains over 70% IdoA, while HS has less than 50% (Figure 2B).³³ The limited ability of unfractionated heparin (UHP) to cross the blood-brain barrier (BBB) has led to the exploration of ultralow molecular weight heparin and heparin-derived oligosaccharides, which can better penetrate the BBB.^34,35 Low molecular weight heparin (LMWH or enoxaparin) has been tested in AD mice for neurological disorders, including stroke and Alzheimer’s disease.³⁶ However, the administration of LMWH is associated with a risk of bleeding.^37–39

The interaction between tau and heparin has been extensively studied.⁴⁰ Ionic interactions primarily mediate the binding of tau to heparin between the negatively charged sulfate and carboxylate groups of heparin and the positively charged amino acid clusters in tau. Recent NMR studies have identified positively charged lysine and histidine residues in tau for binding to heparin.^40–42 Studies have also shown the influence of HS/heparin’s sulfation patterns and chain lengths on their interactions with tau.^43,44 HS containing 6-O-sulfation has been found to regulate tau internalization,²⁷ while HS containing 3-O-sulfation enhances tau interaction and uptake.²⁹ Removal of N- and 6-O-sulfation significantly reduced tau-heparin binding, while the removal of 2-O-sulfation had less impact.²⁷ Additionally, HS containing 2-O-sulfation may promote tau aggregation.^45,46

In eukaryotic cells, the accumulation of unfolded or misfolded proteins causes endoplasmic reticulum (ER) stress, activating the unfolded protein response (UPR). This process is crucial for the maintenance of ER proteostasis and for aiding cells in adapting to and surviving under stress conditions.⁴⁷ However, prolonged or unresolved ER stress can lead to inflammatory responses and programmed cell death, contributing to the development of a variety of complex diseases.^47,48 Upon cellular uptake, p-tau may initiate various types of cellular damage, including apoptosis and mitochondrial dysfunction.^49–53 P-tau neuropathy is linked to cellular defects such as ER stress, unfolded protein responses, and inflammation.^49–53 Therefore, understanding the pathways and cellular functions impaired by p-tau can facilitate the development of effective therapies for AD. Furthermore, investigating the interaction between HS and p-tau using molecular tools, such as HS mimetics, could lead to the development of new compounds that specifically target tau-induced pathology in AD.

Herein, we report the discovery of an HS mimetic, a heparan sulfate trisaccharide conjugated with oleanolic acid (compound 1, Figure 3). Our approach involves installing an oleanolic acid linker to the reducing end of the trisaccharide unit for the following reasons: Oleanolic acid is known for its neuro-protective effects, improving Aβ-induced memory loss in AD rats by maintaining synaptic plasticity.⁵⁴ Second, the lipophilic nature of oleanolic acid could enhance cellular uptake.⁵⁵ Our findings demonstrate that the lipophilic oleanolic acid linker significantly enhances the compound’s ability to protect cells from dysfunction caused by p-tau. Molecular docking studies suggest that the oleanolic acid linker enhances the binding of compound 1 between the two protofilament cores, leading to increased hydrophobic interactions and positioning the sulfated trisaccharide backbone to interact with lysine residues. Mechanistically, this compound reduces p-tau-induced ER stress, UPR activation, and ER stress-associated apoptosis and inflammation in the SH-SY5Y neuroblastoma cell line. We employed the PIMAX (protein interaction modules-assisted function X) p-tau model, which can aggregate without requiring an artificial inducer such as heparin, making it ideal for tau-induced AD research.⁵⁴ This work highlights potential pathways impaired by p-tau as targets for Alzheimer’s treatment. It lays the foundation for developing new HS mimetics to explore tau-HS interactions in Alzheimer’s disease and related tauopathies.

Figure 3.

Heparan sulfate mimetic ligands 1 and 2 consist of two 6-O and 2-N-glucosamine (GlcN) and one glucuronic acid (GlcA). Apomorphine is a cell penetrant p-tau aggregation and cytotoxicity inhibitor.

RESULTS AND DISCUSSION

Our previous report identified HS trisaccharide mimetic 2 (Figure 3) as a potent heparanase inhibitor. This enzyme plays a crucial role in tumor growth, metastasis, and chemoresistance due to its ability to cleave HS chains from proteoglycans.⁵⁶ This compound contains N- and 6-O-sulfation on its sugar backbone, which has been shown to regulate tau internalization and enhance tau binding to heparin.²⁷ Initially, we explored its potential to attenuate p-tau-induced cytotoxicity. However, cell viability results indicate that compound 2 (vide infra) exhibits limited cytoprotectant activity against p-tau toxicity. As a result, we sought to modify this compound further. The interface between tau and HS is primarily driven by electrostatic interactions involving positively charged amino acid residues (Lys, Arg, His) of tau and negatively charged sulfate moieties of HS.⁵⁷ We hypothesized that introducing a lipophilic group at the reducing end of the sulfated trisaccharide unit of compound 2 could increase the cellular uptake⁵⁵ and hydrophobic interactions. This modification may facilitate the positioning of the designed ligands, allowing their sugar region to interact more effectively with lysine and histidine residues. Cryo-electron microscopy maps of paired helical filament (PHF) tau have recently been reported.⁵⁸ To advance our efforts, we replaced the sulfate aminoethyl linker with various lipophilic linkers at the reducing end of ligand 2 (see Figure S2). We found that ligand 1 (Figure 3), a trisaccharide conjugated to oleanolic acid, had the highest docking score with PHF tau. Unlike ligand 2, ligand 1 was anchored in the dimerization region of the two protofilament cores of PHF tau, forming ionic interactions with numerous lysine and histidine residues. Recent studies have shown that oleanolic acid can potentially preserve neuron injury, enhance neuronal arrangements and memory retention, and maintain synaptic integrity to restore synaptic plasticity in rats with AD.⁵⁴ Consequently, ligand 1 was selected for synthesis and subsequent evaluation for its ability to attenuate p-tau-induced toxicity, endoplasmic reticulum (ER) stress, apoptosis, and pro-inflammation in comparison to positive control apomorphine (Figure 3), which is reported as a cell penetrant p-tau aggregation and cytotoxicity inhibitor.⁵⁹

Synthesis of Compound 1.

The synthesis of compound 2 has been previously reported,⁵⁶ prompting our focus on a modular strategy for the synthesis of compound 1 (Scheme 1). We selected a well-protected set of glucosamine (GlcN) and glucuronic acid (GlcA) units with specific criteria: (1) functionalization of the GlcN-1 C2-amino group to form the challenging α-1,2-cis glycosidic linkage with GlcA; (2) C2-hydroxyl protecting group on GlcA to direct β-1,2-trans linkage with GlcN-2; (3) C2-amino protecting group GlcN-2 for β-1,2-trans-glycosidic linkage to oleanolic acid methyl ester; and (4) selective removal of protecting groups for selective sulfation at C6 of GlcN units. Napthylmethyl (Nap) and p-methoxybenzyl (PMB) protecting groups are used for chemoselective sulfation of C6 hydroxyls, which can be removed using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) in the presence of the benzyl (Bn) group. The azido group masks the C2-amino of GlcN-1 for the formation of the α-1,2-cis linkage, while the phthalimide group masks the C2-amino of GlcN-2 for the formation of β-1,2-trans linkage with oleanolic acid. A C2-acetyl group on GlcA directs β-1,2-trans glycoside formation through neighboring group participation strategies.²⁰ Finally, benzyl (Bn) ether groups are permanent protecting groups for C3- and C4-hydroxyls.

Scheme 1. Stereoselective Synthesis of Heparan Sulfate Mimetic Compound 1.

^{a a}Reagents and conditions: (a) 2 mol % TMSOTf, CH₂Cl₂, −78 °C to rt, 8 h, 97%, β only; (b) NaOMe, MeOH (0.2 M), rt, 5 h; then 4-methoxybenzaldehyde dimethyl acetal, 20 mol % CSA, CH₃CN, rt, 5 h, 72%; (c) ethylenediamine, n-butanol; then 1H-imidazole-1-sulfonyl azide·HCl,CuSO₄·5H₂O,K₂CO₃,0 °C to rt, 12 h, 67%; (d) NaH, BnBr, DMF, 0 °C, 2 h, 81%; (e) NaBH₃CN, TFA, DMF, 0 °C to rt, 24 h, 87%; (f) NIS/AgOTf, 4Å MS, CH₂Cl₂/Et₂O (6/1), 80%, β only; (g) AcOH/H₂O, 80 °C, 86%; (h) (i) BAIB, TEMPO, CH₂Cl₂/H₂O (10/1), rt; then (ii) MeI, K₂CO₃, DMF, 82%; (i) TfOH (1 equiv), DMF (9 equiv), 4ÅMS, CH₂Cl₂/Et₂O (6/1), 0 °Ctort, 85%, α/β > 20:1; (j) DDQ, β-pinene, CH₂Cl₂/H₂O (2/1), 5h, 66%; (k) Zn, AcOH, CH₂Cl₂, rt, 95%; (l) SO₃·Et₃N, pyridine/Et3N, microwave, 100 °C, 15 min, 86%; (m) (i) H₂, Pd (OH)₂, H₂O/pH7 buffer; (ii) LiOH, H₂O, 24 h then (iii) Amberlite IR120 Na + resin, 69%.

The synthesis is initiated with the coupling of oleanolic acid methyl ester (OAME) acceptor with thioglycoside donor 3 mediated by TMSOTf (Scheme 1), yielding product 4 predominantly as a single β-isomer (97%) due to the neighboring group participation of the phthalimide (PhthN) group.⁶⁰ The acetyl-protecting groups in 4 were removed, and the resulting 4,6-diol was protected as a benzylidene acetal 5. The C2-PhthN group was transformed into the azido 6, and the C3-hydroxyl was benzylated to give 7. The benzylidene acetal of 7 was cleaved to create the C4-hydroxyl acceptor 8. Subsequent glycosylation of 8 with thioglycoside donor 9 resulted in the oleanolic acid conjugated disaccharide 10 (80% yield). The benzylidene acet al in 10 was removed with acetic acid, followed by the selective oxidation of the C6-hydroxyl to methyl ester 11. Next, the coupling of C4-hydroxyl of acceptor 11 with N-phenyl trifluoroacetimidate 12 mediated by triflic acid (TfOH) in the presence of DMF as an α-modulating additive to direct the stereochemical course of the reaction⁶¹ afford the protected oleanolic acid conjugated trisaccharide 13 in 85% with α/β > 20:1. The removal of Nap and PMB groups using DDQ and β-pinene as an acid scavenger of the acidic 2,3-dichloro-5,6-dicyanohydroquinone byproduct delivered diol 14.⁶² The azido moiety of 14 was reduced with Zn in acetic acid to yield the corresponding amine, which was sulfated to afford 15. Finally, benzyl ethers were removed by hydro-genolysis followed by hydrolysis of the acetyl group to deliver the sodium sulfated ligand 1 in 69% yield.

Evaluation of Compound 1 as a Cytoprotectant against Hyperphosphorylated tau.

After obtaining oleanolic acid methyl ester conjugated trisaccharide 1, we investigated its potential to mitigate the cytotoxic effects of p-tau in the SH-SY5Y neuroblastoma cell line (Figure 4). We assessed the viability of p-tau-treated SH-SY5Y cells in the presence of compound 1, trisaccharide 2, oleanolic acid methyl ester (OAME), and apomorphine (ApoM) as a positive control for p-tau cytotoxicity neutralization.⁵⁹ We utilized relative viability to express the ligand’s cytoprotectant activity against p-tau with p-tau treatment as the reference to account for baseline differences between experiments or conditions. As illustrated in Figure 4a, 0.5 μM of p-tau resulted in a significant reduction of cell viability after overnight treatment. In the presence of 8 μM of ligand 1, the relative viability was maintained at around 0.6. In comparison, apomorphine (8 μM), a brain permeant compound found previously to afford potent protection from p-tau attack, restored the relative viability to 0.7. In contrast, oleanolic acid (OAME) had no effect. Intriguingly, ligand 2 (8 μM) had a moderate effect (0.3 relative viability), likely due to a low-affinity binding without the OAME linker.^17,24 This suggests that the trisaccharide moiety of compound 1 enhances binding to p-tau, while the OAME component plays a crucial role in counteracting p-tau’s cytotoxicity. Furthermore, Figure 4b demonstrated that p-tau caused the death of almost 85% of cells, while compounds 1 and 2 and OAME did not exhibit cytotoxicity at a concentration of 8 μM. The cytoprotective activity of ligand 1 was further quantified (Figure 4c). The concentration of ligand 1 to maintain 50% cell viability (EC₅₀) in the presence of 0.5 μM of p-tau was found to be 2.69 μM, comparable to the positive control apomorphine (EC₅₀ = 2.90 μM).⁵⁹ These findings indicate that ligand 1 exhibits a noticeable protective function in enabling cells to withstand the negative impact of p-tau.

Figure 4.

(A) SH-SY5Y cells were treated with vehicle as a control, 0.5 μM p-tau alone, p-tau (0.5 μM) plus ligand 2 (8 μM), p-tau (0.5 μM) plus OAME (8 μM), p-tau (0.5 μM) plus ligand 1 (8 μM), p-tau (0.5 μM) plus apomorphine (8 μM). Ligand 1 has potent cytoprotectant activity against hyperphosphorylated tau, while ligand 2 exhibits a moderate effect. Oleanolic acid methyl ester (OAME) has no effect. Apomorphine (ApoM) was used as a positive control. (B) Investigating the cytotoxicity of the ligands. P-tau killed almost 85% of the cells, while the compounds did not exhibit cytotoxicity at 8 μM concentration. (C) Measuring EC₅₀ value of ligand 1. (D) Morphology and SH-SY5Y cells density after treatments with vehicle control (PBS), p-tau only (0.5 μM), p-tau (0.5 μM) plus ligand 1 (5 μM) for 24 h. The experiment was repeated three times, and the representative images are shown.

To evaluate changes in cell morphology following treatments, the human neuroblastoma cell line SH-SY5Y was treated with a control vehicle, p-tau (0.5 μM) alone, and p-tau (0.5 μM) plus ligand 1 (5 μM) for 24 h (Figure 4d). Cell morphology was subsequently documented using contrast microscopy to capture cell shape and arrangement nuances. The results demonstrated a significant perturbation in cell viability and morphological characteristics following p-tau administration, revealing a marked reduction in the total number of viable cells (Figure 4d). The p-tau-treated cells appeared more irregular and fragmented compared to the cohesive structure of the vehicle control group. In contrast, SH-SY5Y cells treated with p-tau plus ligand 1 remarkably recovered cell density and morphology. These cells retained a structure similar to those observed in the vehicle control group, characterized by the typical rod-like colony formation observed in the control that indicates healthy cell growth. In contrast, the cells treated solely with p-tau displayed an aberrant morphology, with much less uniformity and more pronounced irregular shapes, highlighting the detrimental effects of p-tau compared to the protective influence of ligand 1.

Evaluation of Compound 1 to Interact with p-tau and Platelet Factor 4 Protein.

Next, we evaluated the efficacy of compound 1 in inhibiting p-tau activity using a solution-based biolayer interferometry (BLI) competition assay. In this assay, we assessed the ability of compound 1 to compete with heparin bound to the BLI biosensor for binding to p-tau. Consequently, we measured the response from the remaining p-tau interacting with the immobilized heparin. This allowed us to determine the percentage activity of the residual p-tau, and we utilized GraphPad Prism to calculate the IC₅₀ values. Compound 1 demonstrated a low IC₅₀ value in micromolar concentration (28.8 ± 1.03 μM, Figure 5a), indicating a strong binding affinity to p-tau. This result is comparable to the data obtained for a small molecule CLR01.⁶³ We also investigated whether apomorphine could bind to p-tau. We observe nonspecific binding of heparin biotin with apomorphine (see Figure S5); therefore, this BLI assay cannot establish whether apomorphine can compete with heparin for binding to p-tau.

Figure 5.

Binding affinity interactions of compound 1 with p-tau (a) and PF4 (b) using solution-based biolayer interferometry (BLI) competition assay.

Both unfractionated heparin and LMWH heparin are associated with the risk of bleeding. Yet, heparin-induced thrombocytopenia (HIT) presents a severe adverse effect linked with these treatments, with LMWH exhibiting a lower risk for HIT.^37–39 This complication is challenging to mitigate, as the process of producing both heparin forms is less refined than those employed for many other pharmaceuticals. The binding of heparin to platelet factor 4 (PF4), a natural chemokine,⁶⁴ triggers an autoimmune response that contributes to the onset of HIT.^65,66 HIT is characterized by a decrease in platelet count and a hypercoagulable state, which results in bleeding complications, embolisms, and thrombosis.⁶⁷ To investigate this further, we evaluated the ability of compound 1 to form a complex with PF4 using BLI competition assay (Figure 5b). Similar to the previous BLI assay examining the interaction of p-tau with compound 1, we assessed the percent activity of the remaining PF4. Heparin exhibited a markedly low IC₅₀ value in nanomolar concentration (1.46 ± 0.09 nM, see Figure S6), suggesting excellent binding strength to PF4. In contrast, compound 1 binding to PF4 showed markedly reduced affinity (IC₅₀ = 21.8 ± 3.33 μM). The significantly reduced binding affinity of compound 1 compared to heparin can be partly attributed to heparin’s multiple iduronic acid (IdoA) residues, which are absent in compound 1. Studies have shown that the highly sulfated IdoA-GlcN region of heparin is responsible for binding to PF4.⁶⁸

Compound 1 Protects Cells by Blocking p-tau-Induced ER Stress and Unfolded Protein Response.

AD is a protein-misfolding disease.⁶⁹ The first line of defense against misfolded proteins is the ER, which initiates the UPR signaling pathways to protect cells from stress challenges.⁴⁷ Since the ER is an organelle responsible for protein folding and assembly, lipid and sterol biosynthesis, and free calcium storage,⁷⁰ it is highly sensitive to alterations in ER protein homeostasis. Various biochemical stimuli that cause ER calcium depletion, altered glycosylation, nutrient deprivation, or oxidative stress can disrupt the protein folding process, leading to the accumulation of unfolded or misfolded proteins in the ER—a condition referred to as “ER stress”.⁴⁸ If left untreated, prolonged ER stress may lead to apoptosis.⁷¹ Under prolonged or severe ER stress conditions, the UPR activates ER stress-associated pro-apoptotic and/or inflammatory pathways to kill the stressed cells.⁷¹ The ER stress-associated apoptotic pathway is mediated by multiple pro-apoptotic factors, including ATF4, CHOP, GADD34, and TRB3 (Figure 6), downstream of the PERK/elF2α UPR pathway.⁷² PERK phosphorylates translation initiation factor elF2α, which, in turn, induces the translation of mRNA encoding ATF4 to trigger the production of CHOP and several other pro-apoptotic factors to mediate apoptosis. Accordingly, we studied whether compound 1 can prevent p-tau-induced ER-stress-associated apoptosis. SH-SY5Y cells were exposed to vehicle control, p-tau (0.5 μM), p-tau plus ligand 1 (5 μM) or p-tau plus ligand 2 (5 μM) for 24 h. Total RNA was subsequently purified and examined by quantitative real-time PCR (qRT-PCR) to quantify the expression of the genes encoding ER stress-associated pro-apoptotic factors, including ATF4, CHOP, GADD34, and TRB3.^73–75 Our findings in Figure 6 revealed that p-tau treatment activated UPR signaling pathways, leading to elevated ER stress-associated pro-apoptotic factors. In stark contrast, ligand 1 significantly lowered the expression of ATF4, CHOP, GADD34, and TRB3, indicating a protective effect against p-tau-induced ER stress-associated apoptosis. To compare, ligand 2 did not exhibit any protective effect and did not reduce the expression of these pro-apoptotic factors in SH-SY5Y cells.

Figure 6.

Ligand 1 protects neuronal cells from p-tau-induced ER stress-associated apoptosis. Expression levels of the pro-apoptotic factors, including ATF4, CHOP, GADD34, and TRB3, were determined by quantitative real-time PCR (qRT-PCR) analyses. Data are shown as mean ± SEM (n = 3). * **p < 0.001 for ligand 1 or ligand 2 plus p-tau vs p-tau and p-tau vs vehicle control (Ctl).

Under ER stress, the primary UPR transducer IRE1α functions as an RNase to splice Xbp1 mRNA, resulting in transcriptional reprogramming.⁴⁷ We conducted an experiment to determine if ligand 1 could prevent p-tau-induced ER stress and UPR in the neuroblastoma SH-SY5Y cells (Figure 7A). Accordingly, human neuronal SH-SY5Y cells were treated with the vehicle, p-tau (0.5 μM), p-tau plus ligand 2 (5 μM), or p-tau plus ligand 1 (5 μM) for 24 h before being harvested for total RNA and cell protein lysate preparation. RNA and protein samples were then used to profile ER stress markers and UPR mediators using qRT-PCR analyses established in our laboratory.^73–75 The results indicated that ligand 1 exhibited strong protective effects on p-tau-induced neuronal ER stress response or UPR as evidenced by the reduced expression of spliced Xbp1 mRNA, BiP, and ERdj4 (Figure 7A).

Figure 7.

Compound 1 protects neuronal SH-SY5Y cells from p-tau-induced ER stress response (A) and ER-stress-associated inflammatory responses (B). Expression levels of the transcripts encoding ER chaperones or UPR mediators, including BiP, ERdj4, and spliced Xbp1, and qRT-PCR determined the transcripts encoding pro-inflammatory factors, including TNFα, IL1β, and IRF3. Data are shown as mean ± SEM (n = 3). *p < 0.01 and **p < 0.001 for ligand 1 or 2 plus p-tau vs p-tau and p-tau vs vehicle control (Ctl).

As ER stress-associated inflammatory response is one of the major causes of neurotoxicity, we tested if ligand 1 could effectively mitigate the pro-inflammatory response caused by p-tau. We previously demonstrated that IRE1α interacts with TRAF6, a key mediator of IKK/NF-kB signaling pathway, to activate XBP1, which triggers a potent inflammatory response in macrophages under ER stress conditions.^71,76 We examined both ER stress- and general disease-associated inflammatory cytokines, as well as other immune markers, in SH-SY5Y cells challenged with p-tau. As illustrated in Figure 7B, ligand 1 was able to protect cells against p-tau-induced inflammatory cytokines IL1β and TNFα as well as reduced expression levels of IRF3 (IFN regulatory factor-3). In comparison, ligand 2 did not offer protection to the neuronal cell line SH-SY5Y from p-tau-induced ER stress and inflammatory responses.

To investigate whether ligand 1 acted as a general suppressor of the ER stress response, we exposed SH-SY5Y cells to tunicamycin (Tm) or thapsigargin (Tg), two well-known ER stress-inducing agents that lead to the accumulation of unfolded or misfolded proteins in the ER (Figure 8A).^48,74 While Tm or Tg significantly induced ER stress response, as reflected by upregulation of the genes encoding the UPR mediators Xbp1s, P58ipk, and ATF4, cotreatment of ligand 1 failed to attenuate Tm- or Tg-induced ER stress response (Figure 8A). Interestingly, the presence of ligand 1 enhanced the effect of Tg. These results suggest that ligand 1 is a specific inhibitor for ER stress response induced by p-tau but not by Tm or Tg.

Figure 8.

(A) Ligand 1 does not attenuate ER stress response induced by both tunicamycin (Tm) and thapsigargin (Tg). The qRT-PCR analysis of the transcripts encoding representative mediators of ER stress response, including Xbp1s, P58ipk, and ATF4 in SH-SY5Y cells treated with vehicle, Tm (5 μg/mL), Tg (0.5 μM). Tm (5 μg/mL) plus ligand 1 (5 μM), or Tg (0.5 μM) plus ligand 1 (5 μM) for 24 h. Mean ± SEM (n = 3 biological repeats). *p ≤ 0.05; **p ≤ 0.01. (B) Transmission electron microscopy reveals structural impacts on p-tau by apomorphine and ligand 1. Shown are 3,000× magnification; scale bar = 200 nm. Arrows indicate more organized structures seen in the presence of either compound.

The findings indicate that ligand 1 selectively suppressed the p-tau-induced ER stress response (Figure 7A), while showing no effect on the ER stress responses induced by the tunicamycin or thapsigargin (Figure 8A). This suggests a direct influence of ligand 1 on the conformation of p-tau. To investigate this possibility, transmission electron microscopy (TEM) was used to analyze the morphology p-tau after a 24 h incubation in the absence or presence of apomorphine (apo) or ligand 1 (Figure 8B). Apomorphine was utilized as the positive control.⁵⁹ The results illustrated that both compounds, consistent with their cytoprotective activity against p-tau, induced a reorganization of p-tau from an amorphous structure into more ordered clusters of smaller, 30 – 50 nm units. Notably, the final shapes of p-tau structures seem to be different between the apomorphine and ligand 1 incubation products, raising the potential involvement of multiple distinct mechanisms in regulating the cytotoxicity of p-tau.

Global Molecular Docking.

Currently, there is no high-resolution structures of tau filaments are available. However, cryo-electron microscopy (cryo-EM) maps of paired helical filament (PHF) tau at 3.4 Å resolution have recently been reported (Figure 9).⁵⁸ Since PHF tau is the most predominant neurofibrillary tangle (NFT),⁷⁷ we conducted an in silico global docking of ligands 1 and 2 into PHF tau to determine their binding mode and interaction (Figure 10), comparing them to apomorphine. Neurofibrillary tangles (NFTs) are comprised of 95% paired helical filaments (PHFs) and 5% straight filaments (SFs).⁷⁷ The cryo-EM has provided insights into the structures of both PHFs and SFs (Figure 9). Microscopy studies show that these higher-order structures of p-tau are made up of double helical stacks of C-shaped subunits (Figure 9A), with cores featuring eight β-sheets aligned along the length of the protofilament (Figure 9B).⁵⁸ Dimerization of these protofilaments is a critical step in fibril-formation, with the hexapeptide ³⁰⁶VQIVYK³¹¹ playing a crucial role in tau self-aggregation. It is hypothesized that the interaction between tau and heparan sulfate (HS) is predominantly driven by electrostatic forces involving the positively charged amino acid residues (Lys, Arg, His) of tau and the negatively charged sulfate groups of HS.⁵⁷

Figure 9.

Paired helical filament (PHF) tau atomic model.⁵⁸ (A) Cryo-EM structure of PHF tau (B) amino acid sequence comprising the protofilament core of PHF.

Figure 10.

Molecular docking of compound 2 (A), compound 1 (B), and apomorphine (C) into the cryo-EM structure of PHF tau (PDB: 503L). Yellow surface = lysine residues, solid green line = hydrophobic interaction, solid blue line = cation-pi, solid magenta line = ionic, dashed yellow line = hydrogen bond.

Molecular docking analysis indicates that ligand 1 forms more hydrogen bonding, hydrophobic, and ionic interactions with lysine and histidine residues within the PHF protofilament core compared to ligand 2 (see Table S1). The optimal binding pose for ligand 2 reveals its binding to the residues in the inner loop of the C-shaped protofilament of PHF tau (Figure 10A) with fewer interactions with lysine and histidine residues than that of ligand 1. The addition of oleanolic acid methyl ester drove the binding of ligand 1 to the interface between the two protofilament cores of PHF tau (Figure 10B). Moreover, the lipophilic oleanolic acid linker facilitates more hydrophobic interactions and positions ligand 1 such that its trisaccharide region interacts with a greater number of lysine and histidine residues (Figure 10B). The binding of ligand 1 to the dimerization region of the protofilament cores may disrupt fibrillization or the formation of tangles, as dimerization is a critical step in PHF formation. To further investigate this, we conducted molecular docking studies of apomorphine.⁵⁹ Docking analysis shows that apomorphine binds to the dimerization point (Figure 10C), but its binding site is opposite to that of ligand 1. These findings are consistent with the TEM and BLI assay results and may provide evidence for mechanisms that underline the cytoprotective activities of ligand 1 and apomorphine against p-tau.

CONCLUSIONS

In summary, we have discovered a heparan sulfate mimetic, consisting of a sulfated trisaccharide and oleanolic acid methyl ester, which effectively protects neuron cells against p-tau-caused cytotoxicity. This compound demonstrates significant efficacy in mitigating p-tau-induced ER stress, UPR activation, and the associated apoptotic and neuroinflammatory pathways. Notably, the sulfated trisaccharide backbone alone exhibits moderate cell viability, and oleanolic acid alone has no effect. Furthermore, the trisaccharide alone does not protect neuronal cells from p-tau-induced ER stress-associated apoptotic and inflammatory responses. Importantly, this hybrid ligand specifically inhibits ER stress response induced by p-tau, but not by tunicamycin or thapsigargin, two known ER stress-inducing agents. TEM analysis reveals that this hybrid ligand reorganizes p-tau from an amorphous structure into more ordered clusters of smaller units. This study serves as a promising starting point for the development of novel heparan sulfate mimetics to address tau-associated Alzheimer’s disease and other tauopathies.

EXPERIMENTAL SECTION

General Synthesis.

All reactions were performed in oven-dried flasks fitted with septa under a positive pressure of nitrogen atmosphere. Organic solutions were concentrated using a Buchi rotary evaporator below 40 °C at 25 Torr. Analytical thin-layer chromatography was routinely utilized to monitor the progress of the reactions. It was performed using precoated glass plates with 230–400 mesh silica gel impregnated with a fluorescent indicator (250 nm). Visualization was then achieved using UV light, iodine, or ceric ammonium molybdate. Flash column chromatography was performed using 40–63 μm silica gel (SiliaFlash F60 from Silicycle). Dry solvents were obtained from an SG Waters solvent system utilizing activated alumina columns under an argon pressure. All other commercial reagents were used as received from Sigma-Aldrich, Alfa Aesar, Acros Organics, TCI, and Combi-Blocks, unless otherwise noted. All new compounds were characterized by Nuclear Magnetic Resonance (NMR) spectroscopy and High-Resolution Mass spectrometry (HRMS). All ¹H NMR spectra were recorded on Agilent 400 or Bruker 500 MHz spectrometers. All ¹³C NMR spectra were recorded on Agilent 100 or Bruker 125 MHz spectrometer. Chemical shifts are expressed in parts per million (δ scale) referenced to the residual proton in the NMR solvent (CDCl₃: δ 7.26 ppm, δ 77.16 ppm, CD₃OD: 3.31). Data are presented: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, and bs = broad singlet), integration, and coupling constant in hertz (Hz). High-resolution mass spectra (HRMS) were recorded using a Micromass LCT Premier XE instrument (Waters) and were determined by electrospray ionization (ESI). NMR and mass spectrometry analysis confirmed the purity of all synthetic compounds 4–15 and 1 (>95%).

3β-(2-Pthalimido-2-deoxy-3,4,6-tri-O-acetyl-β-d-glucopyranosyloxy)-methyloleanolate (4).

A solution of trichloroacetimidate 3⁷⁸ (835 mg, 1.44 mmol) and oleanolic acid methyl ester (521.5 mg, 1.11 mmol) in dry CH₂Cl₂ (10 mL) was placed at −78 °C and stirred for 10 min under N2. TMSOTf (5.2 μL, 28.80 μmol) were added to the solution. After 12 h of stirring from −78 °C to rt, the reaction mixture was quenched with Et₃N and concentrated in vacuo. The resulting residue was purified by silica gel column chromatography (1:3, EtOAc/hexanes) to give 4 as a white foam (951 mg, 97%). ¹H NMR (500 MHz, CDCl₃) δ: 7.84 (dd, J = 5.4, 3.1 Hz, 2H), 7.72 (dd, J = 5.5, 3.0 Hz, 2H), 5.82 (dd, J = 10.8, 9.0 Hz, 1H), 5.37 (d, J = 8.4 Hz, 1H), 5.26 (t, J = 3.9 Hz, 1H), 5.12 (dd, J = 10.2, 8.9 Hz, 1H), 4.39–4.28 (m, 2H), 4.17–4.12 (m, 1H), 3.87 (ddd, J = 10.1, 5.5, 2.6 Hz, 1H), 3.60 (s, 3H), 3.05 (dd, J = 11.9, 4.6 Hz, 1H), 2.84 (dd, J = 13.7, 4.2 Hz, 1H), 2.09 (s, 3H), 2.03 (s, 3H), 1.98–1.89 (m, 1H), 1.86 (s, 3H), 1.82 (d, J = 13.1 Hz, 2H), 1.71–1.63 (m, 2H), 1.62–1.55 (m, 3H), 1.53–1.43 (m, 3H), 1.37–1.27 (m, 3H), 1.23 –1.10 (m, 4H), 1.07 (s, 3H), 1.03–0.94 (m, 2H), 0.92 (s, 3H), 0.89 (s, 3H), 0.84 (s, 3H), 0.66 (s, 3H), 0.58 (s, 3H), 0.42 (s, 3H). ¹³C NMR (126 MHz, ₃) δ: 178.25, 170.69, 170.24, 169.53, 143.78, 134.27, 123.52, 122.33, 99.97, 90.73, 71.61, 70.80, 69.39, 62.45, 55.17, 54.88, 51.51, 47.57, 46.71, 45.89, 41.60, 41.29, 39.25, 38.39, 38.24, 36.71, 33.86, 33.11, 32.56, 32.37, 30.93, 30.69, 27.64, 27.47, 25.88, 25.45, 23.64, 23.40, 23.05, 20.80, 20.67, 20.51, 18.07, 16.78, 16.36, 15.21. HRMS: m/z calcd for C₅₁H₇₃N₂O₁₂ [M + NH₄]⁺, 905.5158; found, 905.5158.

3β-(2-Pthalimido-2-deoxy-4,6-O-(p-methoxybenzylidene)-β-d-glucopyranosyloxy)-methyloleanolate (5).

To a solution of 4 (930 mg, 1.05 mmol) in MeOH (10 mL), NaOMe was added (108.04 mg, 2 mmol). After 3 h of stirring at rt, the reaction mixture was quenched by Amberlyst 15 hydrogen resin and filtered. The filtrate was concentrated in vacuo, and the resulting residue was used for the next step without further purification. The residue was dissolved in dry CH₃CN, CSA (48.65 mg, 0.2 equiv., 0.21 mmol), and p-methoxy benzaldehyde dimethyl acetal (0.36 mL, 2 equiv, 2.09 mmol) was added. After 5 h of stirring at rt, the reaction mixture was quenched with Et₃N and concentrated in vacuo. The crude product was purified by silica gel column chromatography (1:3, EtOAc/hexanes) as the eluent to give 5 (664.1 mg, 72%) as a white foam. ¹H NMR (500 MHz, CDCl₃) δ: 7.87–7.82 (m, 2H), 7.74–7.70 (m, 2H), 7.42 (d, J = 8.8 Hz, 2H), 6.90 (d, J = 8.9 Hz, 2H), 5.52 (s, 1H), 5.32–5.28 (m, 1H), 5.26 (t, J = 3.8 Hz, 1H), 4.64 (dd, J = 10.6, 8.4 Hz, 1H), 4.34 (dd, J = 10.6, 4.6 Hz, 1H), 4.31–4.25 (m, 1H), 3.84 (d, J = 9.9 Hz, 1H), 3.80 (s, 3H), 3.66–3.60 (m, 2H), 3.59 (s, 3H), 3.04 (dd, J = 11.8, 4.6 Hz, 1H), 2.84 (dd, J = 13.7, 4.8 Hz, 1H), 1.93 (td, J = 14.4, 13.9, 4.1 Hz, 1H), 1.82 (ddt, J = 17.9, 13.9, 4.0 Hz, 3H), 1.71–1.62 (m, 3H), 1.62–1.54 (m, 4H), 1.48 (ddt, J = 18.0, 10.6, 5.5 Hz, 3H), 1.36–1.24 (m, 4H), 1.23–1.10 (m, 4H), 1.07 (s, 3H), 1.02–0.96 (m, 1H), 0.91 (s, 3H), 0.89 (s, 3H), 0.84 (s, 3H), 0.66 (s, 3H), 0.59 (s, 3H), 0.44 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.27, 160.33, 143.76, 134.12, 132.01, 131.70, 129.55, 127.66, 122.35, 114.33, 113.78, 101.88, 100.80, 90.48, 77.28, 77.03, 76.77, 68.81, 68.71, 66.02, 56.85, 55.60, 55.34, 55.20, 51.51, 47.56, 46.71, 45.88, 41.58, 41.28, 39.24, 38.43, 38.38, 38.24, 36.70, 33.86, 33.11, 32.56, 32.38, 30.92, 30.69, 27.65, 27.48, 25.88, 25.68, 23.64, 23.39, 23.05, 18.07, 16.78, 16.37, 16.33, 15.21, 14.21. HRMS: m/z calcd for C₅₃H₇₀NO₁₀ [M + H]⁺, 880.4994; found, 880.4979.

3β-(2-Azido-2-deoxy-4,6-O-(p-methoxybenzylidene)-β-d-glucopyranosyloxy)-methyloleanolate (6).

Ethylenediamine (0.5 mL, 7.55 mmol) was added to a suspension of compound 5 (664 mg, 0.75 mmol) in n-BuOH (10 mL). The reaction mixture was stirred for 2 h at 85 °C. The solvent was removed under reduced pressure and repeatedly co-evaporated with toluene (3 × 10 mL), and dried in vacuo. The crude product was used for the next step without further purification. The residue was dissolved in H₂O: MeOH (2:1, 20 mL), followed by the addition of pyridine (5 mL) and stirred for 10 min. Sticks reagent (474.5 mg, 2.26 mmol) and CuSO₄·5H₂O (37.7 mg, 0.15 mmol) were added at 0 °C. After overnight stirring at rt and completion of the reaction, the workup was done with ethyl acetate and water. The organic layer was separated, dried over anhydrous Na₂SO₄, and evaporated under reduced pressure. The residue was purified by silica column chromatography (1:4, EtOAc/hexanes) as the eluent to give 6 (374.1 mg, 67%) as a white foam. ¹H NMR (500 MHz, CD₃OD) δ: 7.43 (d, J = 8.9 Hz, 2H), 6.92 (d, J = 8.9 Hz, 2H), 5.56 (s, 1H), 5.29 (t, J = 3.8 Hz, 1H), 4.52 (d, J = 8.0 Hz, 1H), 4.27 (dd, J = 10.3, 5.0 Hz, 1H), 3.82 (s, 3H), 3.78 (t, J = 10.1 Hz, 1H), 3.66 (s, 3H), 3.61 (t, J = 9.3 Hz, 1H), 3.52 (d, J = 9.1 Hz, 1H), 3.43 (td, J = 9.8, 4.9 Hz, 1H), 3.29–3.23 (m, 2H), 2.95–2.86 (m, 1H), 2.13–2.06 (m, 1H), 1.94 (dq, J = 8.5, 5.0, 4.0 Hz, 2H), 1.85 (dt, J = 11.6, 4.0 Hz, 1H), 1.78–1.59 (m, 8H), 1.59–1.51 (m, 2H), 1.50–1.39 (m, 2H), 1.37–1.31 (m, 1H), 1.28–1.22 (m, 2H), 1.20 (s, 3H), 1.17–1.12 (m, 1H), 1.11 (s, 3H), 1.07–1.02 (m, 1H), 1.00 (s, 3H), 0.97 (s, 3H), 0.95 (s, 3H), 0.91 (s, 3H), 0.88–0.82 (m, 1H), 0.79 (s, 3H). ¹³C NMR (126 MHz, CD₃OD) δ: 180.06, 161.64, 145.02, 128.81, 114.35, 105.43, 102.98, 90.89, 82.23, 73.25, 70.29, 69.62, 67.47, 61.54, 56.97, 55.70, 52.19, 48.13, 47.10, 42.83, 42.79, 40.61, 39.98, 39.71, 37.90, 34.80, 33.91, 33.52, 31.59, 28.80, 28.36, 27.16, 26.45, 24.53, 24.09, 23.96, 19.34, 17.61, 16.93, 15.92, 14.47.

3β-(2-Azido-2-deoxy-3-O-benzyl-4,6-O-(p-methoxybenzylidene)-β-d-glucopyranosyloxy)-methyl-oleanolate (7).

To a solution of 6 (779.7 mg, 0.98 mmol) in dry DMF (8 mL) under N₂at 0 °C, NaH was added (70.9 mg, 2.95 mmol). The slurry solution was stirred for 20 min, and BnBr (0.35 mL, 2.95 mmol) was added. After 2 h of stirring, the reaction mixture was quenched with MeOH, diluted with ethyl acetate, and washed with brine solution (2×). The organic layer was separated, dried over Na₂SO₄, and evaporated under reduced pressure. The residue was purified by silica gel column chromatography (1:4, EtOAc/hexanes) as the eluent to give 7 as a viscous oil (702.6 mg, 81%). ¹H NMR (400 MHz, CDCl₃) δ: 7.43–7.27 (m, 7H), 6.90 (d, J = 8.4 Hz, 2H), 5.52 (s, 1H), 5.28 (s, 1H), 4.94–4.73 (m, 2H), 4.35 (d, J = 8.1 Hz, 1H), 4.28 (dd, J = 10.6, 5.2 Hz, 1H), 3.82 (s, 3H), 3.77 (t, J = 10.3 Hz, 1H), 3.68 (t, J = 9.2 Hz, 1H), 3.62 (s, 3H), 3.53 (t, J = 9.3 Hz, 1H), 3.42 (t, J = 8.7 Hz, 1H), 3.38–3.29 (m, 1H), 3.17 (dd, J = 11.3, 5.3 Hz, 1H), 2.91–2.81 (m, 1H), 2.02–1.91 (m, 1H), 1.88 (d, J = 14.1 Hz, 2H), 1.79–1.67 (m, 3H), 1.62 (d, J = 10.2 Hz, 5H), 1.55 (d, J = 4.4 Hz, 1H), 1.50 (s, 1H), 1.46–1.24 (m, 6H), 1.19 (d, J = 13.0 Hz, 2H), 1.12 (s, 3H), 1.05 (s, 3H), 0.97 (d, J = 6.7 Hz, 1H), 0.92 (s, 6H), 0.90 (s, 3H), 0.86 (s, 3H), 0.72 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.43, 160.25, 143.99, 138.13, 129.84, 128.51, 128.28, 127.96, 127.47, 122.50, 113.78, 104.47, 101.41, 90.52, 81.62, 79.44, 75.01, 68.81, 67.57, 66.25, 60.54, 55.82, 55.47, 51.69, 47.81, 46.88, 46.05, 41.79, 41.46, 39.47, 39.10, 38.68, 36.90, 34.02, 33.27, 32.83, 32.54, 30.85, 27.97, 27.84, 26.11, 26.08, 23.80, 23.58, 23.22, 18.36, 16.99, 16.51, 15.45, 14.35. HRMS: m/z calcd for C₅₂H₇₂N₃O₈ [M + H]⁺, 866.5314; found, 866.5307.

3β-(2-Azido-2-deoxy-3-O-benzyl-6-O-(p-methoxybenzyl)-β-d-glucopyranosyloxy)-methyloleanolate (8).

To a solution of 7 (180 mg, 0.2 mmol) in dry DMF (2 mL) under N₂ at 0 °C, NaBH₃CN was added (128.2 mg, 2.04 mmol) followed by the addition of trifluoroacetic acid (156 μL, 2.04 mmol). After 30 h of stirring, the reaction mixture was quenched with Et₃N, diluted with ethyl acetate, and washed with brine solution (2×). The organic layer was separated, dried over anhydrous Na₂SO₄, and evaporated under reduced pressure. The residue was purified by silica gel column chromatography (1:4, EtOAc/hexanes) to give 8 as a viscous oil (157 mg, 87%). ¹H NMR (500 MHz, CDCl₃) δ: 7.41–7.34 (m, 4H), 7.31 (ddt, J = 8.6, 6.0, 1.9 Hz, 1H), 7.24 (d, J = 8.6 Hz, 2H), 6.86 (d, J = 8.8 Hz, 2H), 5.28 (t, J = 3.7 Hz, 1H), 4.91 (d, J = 11.4 Hz, 1H), 4.77 (d, J = 11.4 Hz, 1H), 4.49 (q, J = 11.6 Hz, 2H), 4.29 (d, J = 8.0 Hz, 1H), 3.80 (s, 3H), 3.71 (dd, J = 10.1, 4.5 Hz, 1H), 3.66 (dd, J = 10.2, 5.6 Hz, 1H), 3.63 (s, 3H), 3.58 (t, J = 9.1 Hz, 1H), 3.41–3.34 (m, 2H), 3.23 (dd, J = 9.9, 8.6 Hz, 1H), 3.15 (dd, J = 11.8, 4.5 Hz, 1H), 2.86 (dd, J = 13.9, 4.8 Hz, 1H), 2.63 (s, 1H), 2.01–1.93 (m, 1H), 1.87 (ddq, J = 15.3, 7.6, 3.4 Hz, 3H), 1.78–1.67 (m, 2H), 1.67–1.49 (m, 9H), 1.48–1.29 (m, 5H), 1.22–1.15 (m, 2H), 1.12 (s, 3H), 1.09–1.04 (m, 1H), 1.03 (s, 3H), 0.93 (d, J = 3.3 Hz, 6H), 0.90 (s, 3H), 0.87 (s, 3H), 0.72 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.41, 159.48, 144.02, 138.36, 130.02, 129.44, 128.76, 128.22, 128.15, 122.48, 114.01, 104.20, 90.24, 82.98, 77.16, 77.05, 76.91, 75.27, 73.82, 73.46, 72.45, 70.45, 67.07, 55.83, 55.44, 51.69, 47.83, 46.87, 46.07, 41.79, 41.46, 39.47, 39.08, 38.79, 36.89, 34.01, 33.28, 32.84, 32.54, 30.85, 27.92, 27.83, 26.17, 26.08, 23.81, 23.61, 23.22, 21.19, 18.36, 16.98, 16.47, 15.47, 14.35. HRMS: m/z calcd for C₅₂H₇₄N₃O₈ [M + H]⁺, 868.5470; found, 868.5405.

3β-(2-O-Acetyl-3-O-benzyl-4,6-O-benzylidene-β-d-glucopyranosyl-(1 → 4)-2-Azido-2-deoxy-3-O-benzyl-6-O-(p-methoxybenzyl)-β-d-glucopyranosyloxy)-methyloleanolate (10).

A solution of thioglycoside donor 9⁵⁶ (847 mg, 1.91 mmol), nucleophilic acceptor 8 (827 mg, 0.95 mmol), and freshly activated MS 4 Å in anhydrous CH₂Cl₂ (20 mL) was stirred at rt for 1 h and then cooled to 0 °C. NIS (857.4 mg, 3.81 mmol) and AgOTf (48.95 mg, 0.19 mmol) were added to the solution. After 20 min of stirring, the reaction mixture was quenched with Et₃N, diluted with CH₂CI₂, and filtered. The filtrate was concentrated in vacuo, and the residue was purified by silica gel column chromatography with EtOAc and hexanes (1:4) to give 10 (1.09 g, 92%) as a foam. ¹H NMR (500 MHz, CDCl₃) δ: 7.46 (dd, J = 7.6, 2.1 Hz, 2H), 7.40–7.31 (m, 8H), 7.31–7.27 (m, 3H), 7.25–7.18 (m, 4H), 6.85 (d, J = 8.6 Hz, 2H), 5.47 (s, 1H), 5.28 (t, J = 3.8 Hz, 1H), 4.94 (t, J = 8.6 Hz, 1H), 4.86 (t, J = 11.2 Hz, 2H), 4.73 (d, J = 10.8 Hz, 1H), 4.61 (dd, J = 11.8, 6.9 Hz, 2H), 4.50 (d, J = 8.0 Hz, 1H), 4.37 (d, J = 11.8 Hz, 1H), 4.20 (d, J = 7.8 Hz, 1H), 4.12 (dd, J = 10.7, 4.8 Hz, 1H), 3.84 (t, J = 9.1 Hz, 1H), 3.72 (s, 3H), 3.69–3.63 (m, 3H), 3.61 (s, 3H), 3.52 (t, J = 9.2 Hz, 1H), 3.45 (t, J = 10.3 Hz, 1H), 3.38–3.29 (m, 2H), 3.26 (dt, J = 9.9, 3.1 Hz, 1H), 3.19–3.07 (m, 2H), 2.86 (dd, J = 13.9, 4.9 Hz, 1H), 1.95 (s, 3H), 1.86 (tdd, J = 18.8, 8.3, 3.5 Hz, 3H), 1.78–1.67 (m, 2H), 1.67–1.56 (m, 5H), 1.51 (dt, J = 15.3, 5.1 Hz, 3H), 1.46–1.23 (m, 5H), 1.22–1.13 (m, 2H), 1.11 (s, 3H), 1.09–1.02 (m, 2H), 1.01 (s, 3H), 0.92 (d, J = 2.6 Hz, 6H), 0.89 (s, 3H), 0.85 (s, 3H), 0.72 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.34, 169.19, 159.46, 144.02, 138.56, 138.42, 137.31, 130.16, 129.66, 129.11, 128.42, 128.35, 128.34, 127.86, 127.76, 127.73, 127.71, 126.12, 122.42, 113.99, 104.04, 101.26, 100.97, 90.16, 81.71, 81.39, 78.87, 76.96, 75.37, 74.93, 74.32, 73.48, 73.29, 68.62, 67.62, 67.09, 66.17, 55.80, 55.34, 51.64, 47.80, 46.82, 46.05, 41.75, 41.43, 39.43, 39.06, 38.76, 36.86, 33.96, 33.25, 32.80, 32.49, 30.81, 27.92, 27.79, 26.10, 26.04, 23.78, 23.57, 23.18, 20.95, 18.31, 16.94, 16.47, 15.43. HRMS: m/z calcd for C₇₄H₉₉N₄O₁₄ [M + NH₄]⁺, 1267.7152; found, 1267.7149.

3β-(Methyl-2-O-acetyl-3-O-benzyl-β-d-glucopyranosyluronate-(1 → 4)-2-azido-2-deoxy-3-O-benzyl-6-O-(p-methoxybenzyl)-β-d-glucopyranosyloxy)-methyloleanolate (11).

Step-1: Disaccharide 10 (450 mg, 0.36 mmol) was dissolved in a mixture of acetic acid/H₂O (4:1, 8 mL). After the reaction mixture had been stirring at 80 °C for 6 h, it was concentrated. The residue was subjected to silica gel column chromatography (1:1, EtOAc/hexanes) to give diol (361 mg, 86%) as a syrup. Step-2: To a solution of resulting diol in CH₂Cl2/H₂O (10:1, 6 mL) were added BAIB (250 mg, 0.77 mmol) and TEMPO (10 mg, 63 μmol) at 0 °C. After the reaction mixture had been stirring at room temperature for 4 h, it was diluted with CH₂Cl₂, washed with aq. Na₂S₂O₃, dried over anhydrous Na₂SO₄, and concentrated. The resulting residue was then dissolved in DMF (8 mL), and KHCO₃ (124 mg, 1.2 mmol) and MeI (116 μL, 1.9 mmol) were added. After the reaction mixture had been stirring at room temperature for 8 h, it was diluted with EtOAc, washed with water and brine, dried over anhydrous Na₂SO₄, and concentrated. The residue was subjected to silica gel column chromatography (1:2, EtOAc/hexanes) to give 11 (302 mg, 82%, over 2 steps) as a syrup. ¹H NMR (500 MHz, CDCl₃) δ: 7.32–7.27 (m, 4H), 7.25–7.17 (m, 8H), 6.82 (d, J = 8.6 Hz, 2H), 5.22 (q, J = 3.3 Hz, 1H), 4.91 (d, J = 11.3 Hz, 1H), 4.85 (dd, J = 9.6, 8.0 Hz, 1H), 4.76 (d, J = 11.8 Hz, 1H), 4.66 (d, J = 11.3 Hz, 1H), 4.62–4.55 (m, 2H), 4.49 (d, J = 8.1 Hz, 1H), 4.33 (d, J = 11.8 Hz, 1H), 4.15 (d, J = 7.6 Hz, 1H), 3.85 (d, J = 2.1 Hz, 2H), 3.74 (s, 3H), 3.71 (s, 1H), 3.64–3.58 (m, 2H), 3.56 (d, J = 2.5 Hz, 3H), 3.53 (d, J = 6.1 Hz, 3H), 3.50 (d, J = 9.9 Hz, 1H), 3.35–3.25 (m, 3H), 3.21 (ddd, J = 9.9, 4.0, 2.1 Hz, 1H), 3.06 (dd, J = 11.8, 4.5 Hz, 1H), 2.99–2.91 (m, 1H), 2.81 (dd, J = 13.9, 4.6 Hz, 1H), 1.87 (s, 3H), 1.85–1.74 (m, 3H), 1.70–1.59 (m, 3H), 1.56 (dd, J = 13.9, 3.6 Hz, 5H), 1.49–1.43 (m, 3H), 1.38–1.24 (m, 4H), 1.17–1.08 (m, 2H), 1.06 (s, 3H), 0.99 (dd, J = 13.9, 3.2 Hz, 1H), 0.95 (d, J = 2.4 Hz, 3H), 0.87 (d, J = 2.8 Hz, 6H), 0.84 (s, 3H), 0.79 (s, 3H), 0.66 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.26, 169.58, 169.11, 159.34, 143.92, 138.76, 138.31, 130.11, 129.60, 128.44, 128.27, 128.20, 127.81, 127.74, 127.64, 127.56, 127.40, 127.35, 122.32, 113.92, 103.99, 100.58, 90.00, 81.45, 81.37, 74.94, 74.66, 74.53, 73.93, 73.22, 72.60, 72.19, 67.55, 67.11, 60.40, 55.69, 55.27, 52.65, 51.55, 47.70, 46.73, 45.95, 41.66, 41.33, 39.33, 38.94, 36.77, 33.87, 33.15, 32.70, 32.39, 30.71, 27.80, 25.94, 23.67, 20.87, 18.21, 16.84, 16.35, 15.34, 14.21. HRMS: m/z calcd for C₆₈H₉₁N₃O₁₅Na [M + Na]⁺, 1212.6342; found, 1212.6331.

3β-(2-Azido-2-deoxy-3,4-di-O-benzyl-6-O-(2-naphthylmethyl)-α-d-glucopyranosyl-(1 → 4)-methyl-2-O-acetyl-3-O-benzyl-β-d-glucopyranosyluronate-(1 → 4)-2-Azido-2-deoxy-3-O-benzyl-6-O-(p-methoxy-benzyl)-β-d-glucopyranosyloxy)-methyloleanolate (13).

A mixture of donor 12⁵⁶ (1.05 g, 1.51 mmol), acceptor 11 (600 mg, 0.50 mmol), DMF (0.35 mL, 4.5 mmol), and freshly activated MS 4 Å in anhydrous CH₂Cl₂/Et₂O (15 mL, 6/1) was stirred at room temperature for 1 h. TfOH (88 μL, 1.01 mmol) was then added. After 8 h of stirring at rt, the reaction mixture was quenched with Et₃N, diluted with CH₂CI₂, and filtered. The filtrate was concentrated in vacuo, and the resulting residue was purified by silica column chromatography (1:2, EtOAc/hexanes) to give 13 (725 mg, 85%) as a white foam. ¹H NMR (500 MHz, CDCl₃) δ: 7.84–7.76 (m, 5H), 7.52–7.44 (m, 4H), 7.43–7.40 (m, 2H), 7.37–7.33 (m, 8H), 7.30 (dd, J = 5.4, 3.2 Hz, 4H), 7.27–7.23 (m, 2H), 7.20 (dd, J = 7.9, 6.2 Hz, 2H), 7.07–7.02 (m, 2H), 6.94 (d, J = 8.8 Hz, 2H), 5.48 (d, J = 3.8 Hz, 1H), 5.32 (t, J = 3.8 Hz, 1H), 5.07 (dd, J = 9.3, 8.0 Hz, 1H), 4.98 (d, J = 11.1 Hz, 1H), 4.91–4.85 (m, 3H), 4.79 (t, J = 11.3 Hz, 2H), 4.70–4.64 (m, 4H), 4.61 (d, J = 12.3 Hz, 1H), 4.49 (dd, J = 22.8, 11.4 Hz, 2H), 4.26–4.20 (m, 2H), 3.89–3.83 (m, 3H), 3.82 (s, 3H), 3.78 (td, J = 11.6, 11.1, 2.9 Hz, 3H), 3.68 (p, J = 5.1, 4.4 Hz, 3H), 3.65 (s, 3H), 3.52 (d, J = 1.6 Hz, 3H), 3.50 (dd, J = 5.9, 3.8 Hz, 1H), 3.40–3.34 (m, 3H), 3.28 (ddd, J = 9.9, 3.9, 2.3 Hz, 1H), 3.14 (dd, J = 11.8, 4.5 Hz, 1H), 2.90 (dd, J = 13.7, 4.7 Hz, 1H), 1.95 (s, 3H), 1.93–1.83 (m, 3H), 1.79–1.68 (m, 3H), 1.65 (dd, J = 14.2, 3.7 Hz, 4H), 1.55 (qd, J = 7.6, 6.9, 3.4 Hz, 4H), 1.48–1.31 (m, 5H), 1.25–1.17 (m, 2H), 1.15 (s, 3H), 1.10–1.06 (m, 1H), 1.03 (s, 3H), 0.96 (d, J = 3.3 Hz, 6H), 0.93 (s, 3H), 0.88 (s, 3H), 0.75 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.26, 171.14, 168.04, 159.39, 143.93, 138.49, 138.03, 137.83, 137.66, 135.23, 133.22, 133.05, 130.02, 129.59, 128.49, 128.47, 128.29, 128.26, 128.11, 128.05, 128.04, 127.97, 127.91, 127.78, 127.70, 127.62, 127.50, 127.42, 127.40, 127.38, 126.85, 126.16, 126.01, 125.95, 122.33, 113.96, 103.97, 100.57, 97.74, 90.00, 82.50, 81.20, 79.97, 77.87, 75.39, 75.29, 74.99, 74.80, 74.74, 74.59, 74.57, 73.77, 73.29, 71.38, 67.63, 67.10, 63.36, 60.40, 55.70, 55.28, 52.51, 51.55, 47.70, 46.73, 45.95, 41.66, 41.34, 39.33, 39.32, 38.94, 38.67, 36.77, 33.87, 33.15, 32.70, 32.40, 31.60, 30.72, 27.81, 27.70, 26.02, 25.95, 23.68, 23.48, 23.09, 22.67, 21.06, 20.84, 18.21, 16.84, 16.35, 15.34, 14.22, 14.13.

3β-(2-Azido-2-deoxy-3,4-di-O-benzyl-α-d-glucopyranosyl-(1 → 4)-methyl-2-O-acetyl-3-O-benzyl-β-d-glucopyranosyluronate-(1 → 4)-2-Azido-2-deoxy-3-O-benzyl-β-d-glucopyranosyloxy)-methyloleanolate (14).

To a solution of trisaccharide 13 (170 mg, 0.10 mmol) in CH₂Cl₂/H₂O (10:1, 2 mL), DDQ (81.9 mg, 0.60 mmol) and β-pinene (94 μL, 0.60 mmol) were added. After 8 h of stirring, the reaction mixture was diluted with CH₂Cl₂, washed with saturated aqueous NaHCO₃ and then brine, dried over anhydrous Na₂SO₄, and concentrated. The resulting residue was purified by silica gel column chromatography (1:2, EtOAc/hexanes) to give the corresponding product 14 (95.1 mg, 66%) as a white foam. ¹H NMR (500 MHz, CDCl₃) δ: 7.40–7.31 (m, 13H), 7.30–7.26 (m, 7H), 5.38 (d, J = 3.8 Hz, 1H), 5.28 (t, J = 3.8 Hz, 1H), 5.10 (t, J = 8.4 Hz, 1H), 4.97 (d, J = 11.0 Hz, 1H), 4.90–4.81 (m, 4H), 4.75 (d, J = 7.9 Hz, 1H), 4.72–4.63 (m, 3H), 4.28 (d, J = 8.0 Hz, 1H), 4.22 (t, J = 8.9 Hz, 1H), 3.95 (d, J = 9.1 Hz, 1H), 3.89–3.82 (m, 3H), 3.80 (d, J = 9.0 Hz, 1H), 3.75 (dd, J = 12.6, 8.0 Hz, 2H), 3.69–3.64 (m, 1H), 3.62 (s, 3H), 3.58 (s, 3H), 3.53 (t, J = 9.4 Hz, 1H), 3.42 (dt, J = 10.0, 3.3 Hz, 1H), 3.39–3.35 (m, 1H), 3.31 (dd, J = 9.9, 7.9 Hz, 1H), 3.27 (dd, J = 10.4, 3.8 Hz, 1H), 3.21 (dt, J = 9.6, 2.9 Hz, 1H), 3.13 (dd, J = 9.6, 6.8 Hz, 1H), 2.86 (dd, J = 13.8, 4.9 Hz, 1H), 2.00 (s, 3H), 1.88 (ddq, J = 11.4, 6.9, 4.2 Hz, 3H), 1.77–1.70 (m, 4H), 1.64–1.58 (m, 5H), 1.55–1.50 (m, 3H), 1.44–1.37 (m, 2H), 1.32 (ddt, J = 16.3, 6.9, 3.3 Hz, 3H), 1.22–1.15 (m, 2H), 1.12 (s, 3H), 1.07–1.03 (m, 1H), 1.02 (s, 3H), 0.93 (s, 3H), 0.92 (s, 3H), 0.90 (s, 3H), 0.85 (s, 3H), 0.72 (s, 3H). ¹³C NMR (126 MHz, CDCl₃) δ: 178.27, 169.21, 168.22, 143.88, 138.39, 137.81, 137.69, 137.61, 128.53, 128.52, 128.50, 128.32, 128.07, 127.98, 127.94, 127.84, 127.80, 127.51, 127.45, 122.31, 103.90, 100.91, 97.71, 90.37, 82.34, 81.10, 79.78, 77.73, 76.92, 75.46, 75.31, 75.25, 75.00, 74.86, 74.72, 74.63, 73.30, 72.14, 67.24, 65.89, 63.38, 61.23, 60.86, 55.68, 52.72, 51.55, 47.68, 46.73, 45.92, 41.65, 41.32, 39.32, 38.93, 38.57, 36.77, 33.87, 33.13, 32.67, 32.39, 30.71, 27.81, 27.70, 26.19, 25.93, 25.64, 23.66, 23.44, 23.08, 20.83, 18.21, 16.84, 16.35, 15.31, 14.13. HRMS: m/z calcd for C₈₀H₁₀₄N₆O₁₈Na [M + Na]⁺, 1459.7299; found, 1459.7287.

3β-(2-N-Sulfo-2-deoxy-3,5-di-O-benzyl-6-O-sulfo-α-d-glucopyranosyl-(1 → 4)-methyl-2-O-acetyl-3-O-benzyl-β-d-glucopyranosyluronate-(1 → 4)-2-N-sulfo-2-deoxy-3-O-benzyl-6-O-sulfo-β-d-glucopyrano-syloxy)-methyloleanolate (15).

Step1: To a solution of trisaccharide 14 (60 mg, 42 μmol) in AcOH (3 mL), zinc (28 mg, 0.42 mmol) was added. After 12 h of stirring, the reaction mixture was filtered and concentrated in vacuo. The resulting residue was purified by silica gel column chromatography (1:20, MeOH/CH₂Cl₂) to give the corresponding amine (55 mg, 95%). Step 2: The amine intermediate (55 mg, 0.04 mmol) was dissolved in pyridine (2 mL), and SO₃·Et₃N (110.5 mg, 0.8 mmol) and Et₃N (0.15 mL, 0.8 mmol) were added. The reaction mixture was stirred in a microwave synthesizer (Manufacturer: Biotage, model: Initiator+, Serial #: 013596–27J) at 100 °C and 1 atm for 30 min. The reaction progress was monitored by ESI mass spectrometry and directly loaded onto a Sephadex LH-20 column for purification. The fractions containing the product were monitored using ESI mass spectrometry. The combined fractions were lyophilized to give sulfated trisaccharide 15 (58 mg, 86%). ¹H NMR (600 MHz, CD₃OD) δ: 7.44–7.40 (m, 2H), 7.38 (ddd, J = 6.9, 5.2, 1.6 Hz, 4H), 7.32–7.25 (m, 8H), 7.24–7.17 (m, 5H), 7.17–7.12 (m, 1H), 5.49 (d, J = 3.3 Hz, 1H), 5.23 (t, J = 3.9 Hz, 1H), 4.98 (t, J = 11.4 Hz, 2H), 4.90 (t, J = 8.5 Hz, 1H), 4.87–4.85 (m, 2H), 4.81 (d, J = 10.6 Hz, 1H), 4.77 (q, J = 4.4, 3.8 Hz, 3H), 4.70 (d, J = 11.3 Hz, 1H), 4.63 (d, J = 11.7 Hz, 1H), 4.29 (td, J = 11.0, 3.8 Hz, 2H), 4.24 (dd, J = 10.6, 4.6 Hz, 1H), 4.15 (dd, J = 10.6, 2.1 Hz, 1H), 4.12–4.09 (m, 2H), 4.04–3.98 (m, 2H), 3.94 (t, J = 6.3 Hz, 1H), 3.73 (dt, J = 7.8, 5.4 Hz, 2H), 3.69 (s, 3H), 3.60 (s, 3H), 3.59–3.56 (m, 2H), 3.44 (q, J = 7.2 Hz, 3H), 3.41–3.37 (m, 1H), 2.87–2.81 (m, 1H), 2.06–1.99 (m, 1H), 1.96 (s, 3H), 1.93–1.83 (m, 3H), 1.72–1.63 (m, 3H), 1.63–1.56 (m, 3H), 1.56–1.44 (m, 4H), 1.44–1.34 (m, 3H), 1.22–1.15 (m, 2H), 1.13 (s, 3H), 1.12–1.09 (m, 1H), 1.07 (t, J = 6.1 Hz, 1H), 1.03 (s, 3H), 0.99 (d, J = 8.4 Hz, 1H), 0.91 (d, J = 3.7 Hz, 6H), 0.89 (s, 3H), 0.78 (s, 3H), 0.71 (s, 3H). ¹³C NMR (151 MHz, cd₃od) δ: 178.63, 170.20, 168.71, 143.51, 138.96, 138.94, 138.62, 138.26, 128.04, 127.99, 127.98, 127.95, 127.77, 127.75, 127.73, 127.70, 127.19, 127.07, 126.96, 126.82, 122.48, 102.37, 100.00, 98.59, 89.11, 80.71, 79.55, 78.62, 77.22, 77.17, 75.98, 74.89, 74.54, 74.45, 73.20, 72.88, 72.20, 70.36, 66.80, 65.52, 58.55, 58.46, 57.28, 55.53, 53.66, 52.04, 50.73, 46.51, 45.61, 42.13, 41.36, 41.34, 39.15, 38.43, 38.24, 36.48, 33.36, 32.49, 32.19, 32.07, 30.14, 27.60, 27.34, 25.11, 24.99, 23.10, 22.64, 22.51, 22.30, 19.73, 17.93, 16.17, 15.86, 7.88, 6.83. ESI-TOF MS (m/z): calcd for C₈₀H₁₀₆N₂O₃₀S₄ [M – 2H]²⁻, 851.28; found, 851.70.

3β-(2-N-Sulfo-2-deoxy-6-O-sulfo-α-d-glucopyranosyl-(1 → 4)-β-d-glucopyranosyluronate-(1 → 4)-2-N-sulfo-2-deoxy-6-O-sulfo-β-d-glucopyranosyloxy)-methyloleanolate (1).

Step 1: To a solution of sulfated trisaccharide substrate 15 (58 mg) in CH₃OH (5 mL), Amberlite IR120 Na + resin was added. After stirring for 24 h at rt, the mixture was filtered and concentrated to afford the trisaccharide sodium salt quantitatively. The resulting residue was then dissolved in a mixture of CH₃OH/pH 7 buffer (1:1, 3 mL), and Pd(OH)₂/C (20%, 174 mg) (3 × weight of starting material) was added. After the reaction mixture had been stirred under 100 psi H₂ at rt for 24 h, it was filtered and concentrated in vacuo. The resulting residue was subjected to Sephadex LH-20 column chromatography (H₂O). The fractions containing the corresponding trisaccharide intermediate were identified using ESI Mass spectrometry. The combined trisaccharide intermediate fractions were combined and lyophilized. Step 2: To a solution of the trisaccharide intermediate generated in H₂O (9 mL), 1 M LiOH (1 mL) was added at 0 °C. The reaction mixture was gradually warmed to rt and stirred for 24 h. The mixture was lyophilized, loaded to the Sephadex LH-20 column, and eluted with H₂O. The fractions containing the product were identified using ESI Mass spectrometry. The combined fractions were charged with Na⁺ exchange resin for 24 h, followed by filtration and lyophilization to deliver the final product 1 (30 mg, 69%). ¹H NMR (500 MHz, D₂O) δ: 5.59 (d, J = 3.8 Hz, 1H), 5.26 (s, 1H), 4.55 (dd, J = 13.1, 8.2 Hz, 2H), 4.35–4.23 (m, 3H), 4.11 (d, J = 10.9 Hz, 1H), 3.86–3.80 (m, 2H), 3.80–3.73 (m, 2H), 3.72–3.64 (m, 3H), 3.62 (d, J = 3.3 Hz, 3H), 3.59–3.49 (m, 2H), 3.35–3.29 (m, 1H), 3.20 (dt, J = 10.4, 5.1 Hz, 2H), 3.01 (t, J = 9.0 Hz, 1H), 2.75 (d, J = 11.9 Hz, 1H), 2.11–1.97 (m, 1H), 1.86 (d, J = 1.1 Hz, 3H), 1.71–1.58 (m, 4H), 1.57–1.44 (m, 5H), 1.43–1.29 (m, 3H), 1.21 (dd, J = 29.0, 12.0 Hz, 3H), 1.12 (s, 3H), 1.02 (s, 3H), 0.99–0.92 (m, 2H), 0.87 (d, J = 11.1 Hz, 9H), 0.79 (s, 3H), 0.73 (d, J = 11.6 Hz, 1H), 0.66 (s, 3H). ¹³C NMR (126 MHz, D₂O) δ: 181.63, 175.01, 144.24, 129.16, 125.85, 122.66, 101.84, 97.41, 91.18, 77.75, 76.47, 76.33, 76.22, 73.40, 73.00, 72.23, 71.10, 69.77, 69.00, 66.31, 60.00, 57.93, 55.25, 52.27, 48.88, 47.12, 41.48, 38.97, 38.72, 36.31, 32.45, 32.07, 30.25, 30.05, 27.50, 25.85, 25.46, 23.27, 22.96, 16.40, 15.88, 14.74. HRMS: m/z calcd for C₄₉H₇₆N₂Na₂O₂₉S₄ [M-3Na + H]²⁻, 665.1611; found, 665.1599.

Biological Assays.

150-Mesh copper square grids (100489–712) were from VWR. DA9 tau antibody was a generous gift from Dr. Peter Davies. Chemicals were purchased from Sigma-Aldrich unless indicated otherwise. SH-SY5Y cells were authenticated and purchased from the American Type Culture Collection.

Expression and Purification of Hyperphosphorylated tau.

The procedure of the PIMAX p-tau preparation is based on the methods established in our lab. 3 In brief, the construct pMK1013-GSK-tau bearing the 1N4R isoform of tau and GSK-3β kinase, each fused to the leucine zipper domain of Jun and Fos, was transformed into BL21 codon plus cells. BL21 overnight culture was grown in 50 mL LB medium with 100 μg/mL ampicillin and 10 μg/mL chloramphenicol overnight at 37 °C, 250 rpm shaking. The overnight culture was diluted in LB medium with 100 μg/mL ampicillin and 2 mM MgSO₄ for further growing. Induction was done with 0.2 mM IPTG for 2 h when OD₆₀₀ reached between 0.4 and 0.5. To purify p-tau, the cell pellet was suspended with lysis buffer (20 mM Tris, pH 5.8, 100 mM NaCl) and incubated in the presence of lysozyme, PMSF, orthovanadate, and protease inhibitor. The suspension was sonicated and centrifuged. The sonication supernatant was boiled for 30 min and then centrifuged again to remove precipitated proteins. The supernatant was treated with TEV protease in the presence of DTT and EDTA overnight at 4 °C. The solution was centrifuged to remove TEV protease precipitation and protein aggregates. The supernatant was concentrated with Amicon spin column. The concentrated sample in storage buffer (lysis buffer) (20 mM Tris, pH 7.4, 100 mM NaCl) was supplemented with 10% glycerol (v/v) and stored at −80 °C for further usage.

P-tau produced for this work was free of lipopolysaccharides (endotoxins) and was estimated to have at least 90% purity.^79,80

SH-SY5Y Cell Culture.

SH-SY5Y (ATCC cat #: CRL-2266) cells were defrosted and grown in accordance with cell bank protocols. Cells were grown to confluence in tissue culture plates in a 1:1 solution of Eagle’s minimum essential medium (EMEM, ATCC cat #: 30-2003) and Ham’s F-12 Nutrient Mix (Gibco cat #: 11765054). The media is supplemented with 10% Fetal Bovine Serum (FBS, ATCC cat #: 30-2020) and treated with 1% Penn Strep (ATCC cat #: 30-2300) solution. Cells were trypsinized in 0.25% Trypsin/0.53 mM EDTA (ATCC cat #: 30-2101) and incubated for 2–3 min until all adherent cells detached from the plate. The cells were then pelletized through centrifugation at 1000 rpm for 5 min. Then, cells were resuspended and plated on either 96-well (10,000 cells/well; Corning), 24-well (40,000 cells/well; Falcon), or 12-well (80,000 cells/well; Falcon) in culture media. All cells used in these experiments were of passage number 8–19.

Cell Viability Assay.

To measure the cytoprotective activity of the compounds in the presence of p-tau, cell counting kit-8 (CCK-8, GLPBIO cat #: GK10001) assay was used. SH-SY5Y cells were seeded in a 24-well plate with a density of 40,000/500 μL of EMEM/F12 media and grown overnight. Then, cells were pretreated with different concentrations of compound for 12 h. After pretreatment, 50 μL of p-tau was added to each well and incubated overnight. The cell viability was assessed by adding 55 μL of CCK8 solution to each well. Hereafter, the cells were incubated for 2 h at 37 °C before measurement of the OD values at 450 nm using a SpectraMax iD5 Microplate Reader (Molecular Devices). The mean value of the no-treatment group was used to normalize the data. The toxicity of p-tau to cells was expressed as LD₅₀, the concentration of p-tau in μM that caused 50% of cell death. The absolute LD₅₀ value of p-tau to SH-SY5Y cells, and hence the EC₅₀ for a given compound’s potency, may vary depending on the batch of the cells. We typically used 0.8 μM of p-tau for these experiments. When the batch of cells was more sensitive to p-tau, the dose of the protein was reduced to 0.5 μM for EC₅₀ quantification.

To obtain the EC₅₀ curve, cells were seeded in a 96-well plate, pretreated with different concentrations of compounds, and treated with p-tau. Relative viability was calculated using the equation

$$\text{Relative viability}=\frac{{\text{Abs}}_{\text{samp}}-{\text{Abs}}_{\text{ptau}}}{{\text{Abs}}_{\text{ctrl}}-{\text{Abs}}_{\text{ptau}}}$$

To further quantify cell viability, we also utilized PI and FDA staining in which 2000 SH-SY5Y cells/well were seeded and grown overnight in 96-well plates. After seeding, cells were with 0.5 μM p-tau for 16–20 h before propidium iodide (PI) and fluorescein diacetate (FDA) staining for viability assessment. Relative cell viability was calculated by FDA positive cells/(FDA positive cells + PI positive cells).

Cell Morphology.

2000–4000 SH-SY5Y cells per well were seeded and grown overnight in 96-well plates. They were then treated with 0.8 μM p-tau, with or without the test compounds, for 24 h. The cells were visualized by bright field and phase contrast microscopy at 20× magnification.

Transmission Electron Microscopy.

P-Tau (10 μM) was incubated with or without compounds in 20 mM Tris buffer (pH 7.4) at 37 °C for 24 h. The protein solution was 10-fold diluted and incubated with 2.5% glutaraldehyde for 5 min. Then, 20 μL of the sample was fixed on a 150-mesh copper square grid. The grid was negatively stained by 1% uranyl acetate for 10 s and then carefully dried with filter paper. The sample was imaged by JEOL 1400 Flash TEM at 3000 to 10000 magnifications.

Quantitative Real-Time PCR.

Total RNAs from SH-SY5Y cells were extracted using QIAwave RNA Mini (QIAGEN 74534). RNA quality was determined by measuring the OD 260/280 values and complementary DNA (cDNA) was synthesized using superscript III reverse transcriptase (Invitrogen 18080085). Quantitative PCR (qPCR) analysis was performed using 2× Universal SYBR Green Fast qPCR Mix (Abclonal RK21203). The sequences of the primers (forward and reverse): for BiP: 5′-CCTGGGTGGCGGAACCTTCGATGTG-3′ and 5′-CTGGACGGGCTTCATAGTAGACCGG-3′; for XBP1s: 5′-CCGCAGCAGGTGCAGG-3′ and 5′-GAGTCAATACCGCCAGAATCCA-3′; for ERdj4: 5′-TCTTAGGTGTGCCAAAATCGG-3′ and 5′-TGTCAGGGTGGTACTTCATGG-3′; for IL1β: 5′-GCTCGCCAGTGAAATGATGG-3′ and 5′-GTCCTGGAAGGAGCACTTCAT-3′; for P58ipk: 5′-GGTGCTGAATGTGGAGTAAATGC-3′ and 5′-AGTAGCCCTCCGATAATAAGCAA-3′; for ATF4: 5′-TCAAACCTCATGGGTTCTCCA-3′; for IRF3: 5′-CCTGCACATTTCCAACAGCC-3′ and 5′-TGGAAATCCATGCCCTCCAC-3′; for TRB3: 5′-TACCTGCAAGGTGTACCCC-3′; for CHOP: 5′-GCCTTTCTCCTTTGGGACACTGTCCAGC-3′ and 5′-CTCGGCGAGTCGCCTCTACTTCCC-3′; for TNFa: 5′-CCCATCTATCTGGGAGGGGT-3′; for GADD34: 5′-ATGATGGCATGTATGGTGAGC-3′ and 5′-AACCTTGCAGTGTCCTTATCAG-3′; and for GAPDH: 5′-TGAAGGTCGGAGTCAACGG-3′ and 5′-AGAGTTAAAAGCAGCCCTGGTG-3′.

P-Tau and Platelet Factor 4 Binding Assays.

Biolayer interferometry (BLI) assays were conducted on an Octet Red instrument (fortéBIO) at 25 °C. A solution-phase affinity assay, adapted from SPR analysis, was used to determine compound 1 binding affinities by BLI affinities.⁶⁴ PIMAX p-tau protein or human recombinant CXCL4/platelet factor 4 (PF4) protein (R&D systems, cat. 795-P4) was incubated with varying concentrations of compound 1. The remaining free protein in this equilibrium mixture was tested for binding to immobilized heparin. Heparin–biotin (Creative PEGworks, 18 kDa, with one biotin per polymer) at 5 μg/mL was immobilized onto Octet Streptavidin (SA) biosensors (Sartorius, cat. 18-5019) for 5 min. The immobilization and binding studies were performed at a shaking speed of 1000 rpm in 1× HBS-EP buffer (Cytiva Life Sciences, cat. BR100669). Binding events were analyzed using GraphPad Prism, and the percentage of p-tau/PF4 binding was calculated based on the signal response relative to control conditions.

Computational Studies.

To investigate the binding mode and interaction of our compounds with tau fibrils, we performed in silico docking studies of ligand 1, ligand 2, and apomorphine into the cryo-EM structure of paired-helical filament (PHF) tau (PDB ID: 503L).⁵⁸ Missing hydrogens from the protein. pdb file were added using Molprobity,⁸¹ and water molecules were removed in YASARA. The sugar ligands were built using ChemDraw and transported to Discovery Studio in pdb format. Connectivities were further modified using the Avogadro software. The modified ligand and protein. pdb files were subjected to in vacuo energy minimization in YASARA and saved in sdf and pdb format, respectively. Each ligand was docked separately into PHF tau using the AutoDock Vina⁸² suite default parameters in YASARA. The ligand and protein were subjected to 25 global docking runs, where the AMBER14⁸³ force field was applied to the protein, and the GLYCAM06⁸⁴ and GAFF/Am1BCC force field was applied to each ligand. The docking pose with the highest docking score was then subjected to analysis.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.4c02563.

NMR spectra for the key intermediates and HS trisaccharide-oleanolic acid conjugate 1 and molecular docking parameters for the top three populated clusters (PDF) (PDF)

pdb file were added using Molprobity (PDB)

pdb files were subjected to in vacuo energy (PDB)

In vacuo energy minimization in YASARA (PDB)

Molecular formula strings (CSV) (CSV)

ABBREVIATIONS

AD

Alzheimer’s disease

ApoM

apomorphine

ATF4

activating transcription factor 4

BiP

78 kDa glucose-regulated protein

BBB

blood-brain barrier

CHOP

C/EBP homologous protein

ECM

extracellular matrix

ER

endoplasmic reticulum

FDA

fluoresceine diacetate

GADD34

growth arrest and DNA damage-inducible protein 34

GlcA

d-glucuronic acid

GlcN

d-glucosamine

HS

heparan sulfate

HSPG

heparan sulfate proteoglycans

IdoA

l-iduronic acid

MAP

microtubule-associated protein

NFT

neurofibrillary tangle

NMR

nuclear magnetic resonance

OAME

oleanolic acid methyl ester

PERK

PRKR-like endoplasmic reticulum kinase

P58ipk

58 kDa inhibitor protein kinase

PHF

paired helical filament

P-tau

hyperphosphorylated tau

PI

propidium iodide

qPCR

quantitative PCR

TEM

transmission electron microscopy

Tg

thapsigargin

Tm

tunicamycin

TRB3

Tribbles 3

UPR

unfolded protein response

XBP1

X-box binding protein 1