Ligand Valency and Linker Design Dictate the Efficacy of CI-M6PR-Mediated Targeted Delivery of M6P–siRNA Conjugates

Abstract

Targeted delivery of small interfering RNAs (siRNAs) to nonhepatic tissues remains a major challenge in RNAi therapeutics owing to inefficient cellular uptake and off-target effects. In this study, mannose-6-phosphate (M6P)-conjugated siRNAs are developed to target the cation-independent M6P receptor (CI-M6PR/IGF2R), which is overexpressed in various cancers, including chronic myeloid leukemia (CML). Several mono-, di-, and tetravalent M6P–siRNA conjugates are synthesized via solid-phase methods using either flexible hexanediol or rigid proline-based linkers. Thermal melting analysis indicated that M6P conjugation modestly reduced duplex stability, with high-valence constructs and those containing more flexible linkers being particularly affected. By contrast, rigid proline linkers mitigated destabilization. Circular dichroism spectroscopy confirmed that the native A-form RNA structure was maintained after M6P conjugation, while in vitro gene silencing studies in CI-M6PR-positive K562 cells targeting KNTC2 mRNA demonstrated that ligand valency and linker rigidity significantly influenced the activity under transfection-free conditions. Tetravalent siRNA 7 (partially proline-linked) achieved the most efficient knockdown (∼36% reduction in KNTC2 mRNA), whereas the reduction in the activity of tetravalent siRNA 5 (fully hexanediol-linked) was similar to that of the divalent constructs, likely owing to impaired receptor engagement. Flow cytometry analysis further established that this superior activity was attributable to the markedly higher cellular uptake of siRNA 7 relative to the other conjugates. Electroporation experiments confirmed that all siRNA variants retained RNAi activity, indicating that delivery efficiency, rather than functional impairment, is responsible for the observed activity differences. These findings illustrate the pivotal role of ligand valency and linker architecture in optimizing CI-M6PR-targeted siRNA delivery and provide a framework for designing ligand-conjugated siRNA therapeutics against CI-M6PR-expressing tumors.

1. Introduction

Small interfering RNA (siRNA)-based therapeutics offer a precise and programmable approach for post-transcriptional gene silencing and are a rapidly advancing Frontier in nucleic acid medicine. , siRNAs harness the endogenous RNA interference (RNAi) pathway to degrade specific sequences of the target mRNA, enabling virtually any gene of interest to be silenced. This capability holds considerable promise for the treatment of a broad spectrum of diseases including cancers, viral infections, metabolic syndromes, and rare genetic disorders. Notably, siRNAs offer the distinct advantage of targeting genes that are often deemed “undruggable” by conventional small-molecule or antibody-based therapeutics, thereby significantly expanding the druggable genome.

Despite their conceptual appeal and mechanistic precision, the clinical translation of siRNA therapies is impeded by challenges related to delivery, stability, and off-target effects. , Naked siRNAs are large, negatively charged macromolecules that suffer from poor membrane permeability, rapid nuclease-mediated degradation, and unfavorable pharmacokinetics characterized by short circulation half-lives and rapid renal clearance. To address these challenges, receptor-mediated endocytosis offers a highly selective and efficient route for cell-specific delivery with the potential to enhance intracellular uptake, reduce systemic toxicity, and minimize off-target effects. − Among these, N-acetylgalactosamine (GalNAc) conjugation has demonstrated the most notable clinical success. The approval of Givosiran (a GalNAc–siRNA for the treatment of acute hepatic porphyria) by the U.S. Food and Drug Administration is a milestone in RNAi-based drug development. The subsequent approval of the GalNAc–siRNA therapeutics Lumasiran, Inclisiran, and Vutrisiran, − which target hepatocyte-specific genes, further emphasized the therapeutic potential of receptor-targeted siRNA delivery.

GalNAc conjugation leverages the high binding affinity of GalNAc for the asialoglycoprotein receptor, which is highly expressed in hepatocytes. , This interaction enables efficient receptor-mediated endocytosis following subcutaneous administration, enabling potent gene silencing with a favorable safety profile however, the success of GalNAc–siRNA delivery is largely confined to liver-specific diseases, and expanding this platform to extrahepatic tissues remains a significant challenge. Consequently, the development of alternative receptor–ligand systems is essential to enable targeted siRNA delivery beyond the liver.

The cation-independent mannose-6-phosphate receptor (CI-M6PR), also known as the insulin-like growth factor 2 receptor (IGF2R), has emerged as a promising target for receptor-mediated endocytosis. CI-M6PR is a type I transmembrane glycoprotein that is widely expressed in various cell types, including hepatic stellate cells, tumor-associated fibroblasts, and several types of cancer cells (e.g., prostate, breast, pancreatic, and leukemic cells). − The physiological role of CI-M6PR involves trafficking lysosomal enzymes and insulin-like growth factor 2, primarily through interactions with ligands tagged with mannose-6-phosphate (M6P). CI-M6PR is particularly attractive for drug delivery owing to its elevated expression in pathological tissues and ability to undergo clathrin-mediated endocytosis and lysosomal trafficking. , Several studies have demonstrated its utility in targeted cancer therapies. M6P-decorated carriers, including DOX-HSA-M6P, M6P-liposomes, and M6P-functionalized nanocarriers, have shown enhanced cellular uptake, lysosomal localization, and tumor-selective cytotoxicity in preclinical models. Moreover, the use of M6P ligands in lysosome-targeting chimeras (LYTACs) degrades extracellular and membrane proteins via the CI-M6PR trafficking pathways. , Collectively, these studies emphasize the versatility and translational potential of this receptor for targeted drug delivery.

Beyond ligand choice, linker chemistry and ligand architecture are critical for determining siRNA conjugate performance. In GalNAc–siRNAs, triantennary architectures consistently outperform mono- and divalent formats due to higher avidity for ASGPR, while linkers positioned at the 3′ end of the sense strand maintain RNAi activity by preserving RISC accessibility. , Similarly, RGD–siRNA conjugates show valency-dependent silencing efficiency, with trivalent designs yielding the strongest effects before steric hindrance reduces activity. In folate–siRNA systems, cleavable disulfide linkers achieve superior knockdown compared to stable noncleavable tethers, emphasizing the need for intracellular release. , Cholesterol–siRNA conjugates illustrate another key principle: linker stability directly affects the balance between systemic circulation and cytoplasmic release. , Finally, aptamer–siRNA chimeras rely on flexible linkers and careful positioning to avoid interference with Dicer processing or RISC incorporation. , Taken together, these studies establish that both ligand valency and linker design are indispensable for efficient receptor engagement, internalization, and RNAi activity.

Despite extensive research into M6P-conjugated drug carriers, the application of M6P–siRNA conjugates remains underexplored. Zhu and Mahato (2010) synthesized M6P–PEG–siRNA conjugates and demonstrated ∼40% luciferase knockdown in hepatic and hepatic stellate cells in the absence of a carrier: however, this study was limited to monovalent conjugates and did not evaluate the influence of ligand valency, spatial configuration, or linker chemistry, which critically influence receptor binding and internalization efficiency. Furthermore, CI-M6PR–expressing cancer cell lines were not explored beyond hepatic models: thus, the performance of M6P–siRNA conjugates in a broader oncological context remains unknown.

The present study seeks to addresses these knowledge gaps by designing and evaluating M6P–siRNA conjugates with systematic variations in both ligand valency and linker chemistry (Figure ). We targeted the KNTC2 gene, a mitotic kinesin critical for spindle checkpoint regulation and frequently dysregulated in cancer, using CI-M6PR–positive K562 chronic myeloid leukemia cells as a model. Aberrant KNTC2 expression is linked to uncontrolled proliferation and genomic instability in various cancers: thus, silencing this gene in K562 cells disrupts mitotic progression and triggers apoptosis, highlighting KNTC2 as a compelling therapeutic target for the treatment of CI-M6PR–positive leukemias. Mono-, di-, and tetravalent M6P conjugates were synthesized to investigate the influence of multivalency on receptor engagement and silencing activity. In parallel, the effects of linker chemistry on siRNA activity and receptor engagement were systematically explored. The synthesis of branched or dendritic linkers is complex: thus, we employed simple, synthetically accessible linkers such as 1,6-hexanediol and proline derivatives, which are readily incorporated using standard automated protocols for the synthesis of oligonucleotides. These monomeric linkers offer structural diversity while minimizing the synthetic burden, enabling the efficient exploration of linker length, flexibility, and spatial orientation.

1.

Strategy to optimize the ligand valency and the linker architecture of M6P-conjugated siRNAs.

By integrating multivalent ligand design with rational linker engineering, this work aims to establish a structure–function framework for M6P–siRNA conjugates, enabling more effective CI-M6PR-mediated delivery. Our findings provide valuable insights into the structure–activity relationship of M6P–siRNA conjugates and will facilitate the development of versatile receptor-targeted platforms for RNAi-based cancer therapy.

2. Results and Discussion

2.1. Synthesis of M6P Ligand

A mannose-6-phosphate (M6P) ligand was synthesized to enable the site-specific conjugation of oligonucleotides for targeted siRNA delivery. The ligand incorporates an M6P moiety tethered via a five-carbon alkyl spacer to a triazole ring, formed through copper­(I)-catalyzed azide–alkyne cycloaddition (CuAAC), and further tethered to a prolinol residue via a succinimide linker via amide bond formation. This molecular structure balances hydrophilic and hydrophobic characteristics, thereby improving the pharmacokinetic profile of the resulting siRNA–ligand conjugates. Additionally, the modularity and synthetic efficiency of CuAAC chemistry enabled a streamlined synthesis. The phosphoramidite product (compound 12) was directly incorporated into oligonucleotides via automated solid-phase synthesis (Scheme ).

1. Synthesis of Compound 12 .

^a Reagents and conditions: (i) TBDPSCl, pyridine, rt, 3.5 h; (ii) Ac₂O, pyridine, rt, 4.5 h, 76% (2 steps); (iii) BnNH₂, DMF, 60 °C, 3.5 h, 79%; (iv) Trichloroacetonitrile, DBU, CH₂Cl₂, rt, 1.5 h; (v) 5-bromo-1-pentanol, BF₃·OEt₂, CH₂Cl₂, −20 °C, 2.3 h, 51% (2 steps); (vi) NaN₃, DMF, 60 °C, 6 h, 96%; (vii) TBAF/THF, AcOH, THF, r.t., 3 h, 73%; (viii) 1H-tetrazole, bis­(2-cyanoethyl)-N,N-diisopropylphosphoramidite, THF, −40 °C, 1.5 h, then 30% tert-butylperoxide, 0 °C, 1.5 h, 92%; (ix) Compound 13, CuSO₄, sodium ascorbate, sodium carbonate, t-BuOH:H₂O (1:3, v/v), rt, 1.3 h, 67%; (x) Compound 18, DIPEA, HBTU, DMF, rt, 50 min, 60%; (xi) 2-cyanoethyl-N,N,-diisopropylchlorophosphoroamidite, DIPEA, CH₂Cl₂, r.t., 1.5 h, 64%.

The synthetic route commenced with the transformation of α-d-mannopyranose (compound 1) into the trichloroacetimidate-functionalized intermediate (compound 5) through a sequence of four steps, following previously reported methodologies. Initially, selective silylation of the 6′-hydroxy group of compound 1 was achieved using tert-butyldiphenylsilyl chloride (TBDPSCl) in pyridine. Subsequent acetylation with acetic anhydride furnished the fully acetylated intermediate 3 in an overall yield of 76% over the two steps. Selective deacetylation of the anomeric (1′-) acetyl group using benzylamine in DMF yielded intermediate 4 in 79% yield.

To introduce a five-carbon spacer at the 1′-hydroxy position of compound 4, a two-step Schmidt glycosylation was employed. In the first step, activation of the anomeric hydroxyl group with trichloroacetonitrile in the presence of DBU yielded the corresponding trichloroacetimidate (compound 5). This activated intermediate subsequently underwent glycosylation with 5-bromo-1-pentanol, catalyzed by boron trifluoride diethyl etherate (BF₃·Et₂O), to afford the M6P-alkyl bromide derivative 6. The combined yield for these two steps was 51%.

Conversion of the terminal bromide in compound 6 to an azide was achieved using sodium azide, affording compound 7 in 96% yield. Compound 8 was obtained in 73% yield via subsequent desilylation of the 6′-O-TBDPS group with 1 M tetrabutylammonium fluoride (TBAF) in THF containing acetic acid, which suppressed undesired rearrangements.

Phosphorylation at the 6′-hydroxy group of compound 8 was performed by treatment with bis­(2-cyanoethyl)-N,N-diisopropylphosphoramidite and 1H-tetrazole to generate a trivalent phosphite, which was oxidized in situ with tert-butyl hydroperoxide (t-BuOOH) to afford the corresponding phosphate ester (compound 9) in 92% yield.

The triazole linkage was constructed via the CuAAC reaction of compound 9 (bearing an azide group) with alkyne-bearing compound 13 to yield triazole-linked intermediate 10 in 67% yield. The latter was conjugated with a proline-derived secondary amine (compound 18) via HBTU/DIPEA-mediated amide bond formation, affording compound 11 in 60% yield. Phosphitylation of compound 11 using 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (CEPCl) and DIPEA yielded the desired phosphoramidite monomer (12) in 64% yield, which was suitable for automated oligonucleotide synthesis.

Compound 13 was synthesized in 61% yield by the reaction of propargylamine with succinic anhydride and used in the CuAAC reaction.

To incorporate the prolinol moiety into the M6P-ligand structure, a key intermediate (compound 18) was synthesized using an established four-step protocol (Scheme ). The secondary amine of commercially available prolinol (compound 14) was protected with an Fmoc group (compound 15) before the carboxylic acid group was reduced using NaBH₄/BF₃·OEt₂ to yield the primary alcohol (compound 16). Selective protection of the primary alcohol with DMTr furnished compound 17, and the Fmoc group was removed using triethylamine to provide secondary amine intermediate 18 in 75% overall yield.

2. Synthesis of Compounds 18 and 19 .

^a Reagents and conditions: (i) FmocCl, toluene, NaHCO₃,H₂O/Dioxane (1:1, v/v), rt, 18 h; (ii) NaBH₄, THF, then BF₃·OEt₂, rt, 6 h; (iii) DMTrCl, pyridine, rt, 18 h, 82% (3 steps); (iv) Triethylamine, DMF, 80 °C, 1.5 h, 75%, (v) 2-cyanoethyl-N,N,-diisopropylchlorophosphoramidite, DIPEA, THF, rt, 1.5 h, 80%.

Both proline- and hexanediol-based phosphoramidite linkers were synthesized to enable the modular modification of the linker structure between the M6P monomer units in the M6P-siRNA conjugates. The proline-based linker (compound 19) was synthesized via phosphitylation of compound 17 with 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (CEPCl) in the presence of DIPEA, affording the corresponding phosphoramidite in 80% yield (Scheme ).

Simultaneously, a hexanediol-based linker was synthesized via a two-step process. The selective protection of one hydroxyl group of 1,6-hexanediol with DMTrCl in an ice bath afforded compound 21 in 76% yield. Subsequent phosphitylation produced the corresponding phosphoramidite (compound 22) in 70% yield. Phosphoramidites with proline- and hexanediol-based linkers were both compatible with automated oligonucleotide synthesis, enabling the synthesis of structurally diverse M6P-siRNA conjugates. The physicochemical properties and gene-silencing efficiency of these siRNA conjugates were systematically investigated to determine the influence of the linker.

2.2. Design of M6P Ligand-Conjugated siRNAs

To systematically evaluate the influence of ligand valence and linker architecture on the physicochemical and biological performance of siRNA conjugates, several mannose-6-phosphate (M6P) ligand-conjugated siRNAs were synthesized (Figure ), including monovalent, divalent, and tetravalent constructs incorporating various linkers between the M6P moieties and siRNA strand. All constructs were synthesized using an automated DNA/RNA synthesizer to facilitate precise structural control and reproducibility. To enhance their stability and nuclease resistance, all siRNAs were chemically modified with 2′-O-methyl and 2′-fluoro ribonucleotides, and phosphorothioate linkages were partially incorporated into the siRNA structure. The constructs were synthesized using an Advanced Enhanced Stabilization Chemistry (Adv. ESC)-like modification pattern, combining 2′-OMe/2′-F substitutions with terminal P = S linkages to ensure consistent backbone stability across the study.

2.

Chemical structures of mono-, di-, and tetravalent M6P ligands.

The mannose-6-phosphate receptor (M6PR) mediates the uptake and intracellular trafficking of M6P-tagged ligands, primarily to lysosomes. This receptor–ligand interaction is a particularly attractive target for siRNA delivery strategies that aim to bypass the need for exogenous transfection reagents. Although multivalency enhances the binding affinity between M6P and M6PR, − the precise structural requirements for optimal interactions, particularly in the context of siRNA conjugates, remain poorly defined. M6P binding sites have been mapped to domains 3 and 9 of M6PR, which exhibit flexible side-chain orientations and may therefore adopt various conformations. Structural studies have suggested that M6PR functions as a dimer, increasing the probability of cooperative ligand engagement across multiple binding domains.

Based on these considerations, several structurally distinct M6P-siRNA conjugates were synthesized: siRNA 2, in which a single M6P ligand is directly conjugated to the siRNA strand without a spacer, served as the foundational monovalent construct. This design was inspired by a monovalent M6P-siRNA conjugate developed by Zhu and Mahato and was used as a baseline for comparison. siRNA 4 was also monovalent, but incorporated one proline and two 1,6-hexanediol linkers between the RNA strand and the M6P ligand, enabling the effects of the linker flexibility/rigidity on siRNA performance to be determined. siRNA 3 and siRNA 6 are divalent conjugates, each bearing two M6P ligands. In siRNA 3, the M6P units were separated by a single hexanediol linker, whereas five hexanediol linkers were used in siRNA 6 to span a larger distance between M6PR domains 3 and 9. These constructs were designed to evaluate the roles of multivalency and interligand spacing in M6PR engagement and subsequent cellular uptake. siRNA 5 and siRNA 7 were tetravalent conjugates that incorporate four M6P ligands and target all four putative M6P-binding domains of the M6PR dimer. The M6P moieties in siRNA 5 were exclusively connected to the siRNA strand via flexible hexanediol linkers. By contrast, siRNA 7 incorporated fewer hexanediol linkers and more rigid proline linkers. This design aimed to assess the influence of conformational entropy on the receptor binding strength by altering the spatial alignment and entropic penalties.

The choice between the hexanediol and proline linkers was based on their distinct physicochemical properties. The flexibility and hydrophobicity of hexanediol promote membrane interactions and may facilitate endosomal escape. By contrast, the rigid, hydrophilic proline residue offers a defined spatial orientation, enhanced solubility, and improved stability in aqueous environments. The complementary properties of these linkers enable the properties of the siRNA conjugates to be tailored to balance cellular uptake, stability, and receptor-targeting efficiency. All conjugate variations were efficiently synthesized by modifying an automated oligonucleotide synthesizer program, enabling rapid iteration and precise structural modulation. This modular approach facilitates the systematic optimization of siRNA delivery vehicles through rational design.

2.3. Oligonucleotide Synthesis and Purification of M6P Ligand-Conjugated siRNAs

M6P ligand-conjugated siRNAs were synthesized using standard solid-phase oligonucleotide synthesis protocols with an automated DNA/RNA synthesizer. Site-specific incorporation of M6P ligands was achieved using an M6P phosphoramidite (compound 12), whereas hexanediol and proline linkers were introduced using the corresponding phosphoramidite derivatives, compounds 19 and 22, respectively, to adjust the spatial orientation, rigidity, and flexibility of the ligand–siRNA interface. Following the chain assembly, cleavage and deprotection steps were performed. The controlled pore glass support was treated with 10% diethylamine in CH₃CN at 25 °C for 5 min to remove the cyanoethyl protecting groups. Subsequent treatment with a 1:1 (v/v) mixture of concentrated aqueous ammonia and 40% methylamine at 65 °C for 10 min completely removed the base-protecting groups and cleaved the oligonucleotides from the solid support.

The crude products were purified using 20% denaturing polyacrylamide gel electrophoresis to ensure high resolution of the full-length RNAs. Further purification was performed using Sep-Pak C18 cartridge chromatography to remove small-molecule impurities and ensure sample desalting. The structural integrity and molecular identity of the purified RNAs were determined using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Figure S1, Supporting Information). All measured molecular weights were consistent with the predicted masses of the designed conjugates, indicating that the M6P-RNA conjugates exhibited high stability under MALDI-TOF MS conditions, in contrast to the M6P-conjugates reported by Mahato et al., which degraded during similar mass spectrometric analysis. The purity of the synthesized RNAs was confirmed using reverse-phase high-performance liquid chromatography (RP-HPLC) (Figure S2, Supporting Information). The high purity of the synthesized M6P ligand-conjugated oligonucleotides confirmed the robustness, efficiency, and reproducibility of the overall synthetic strategy. The sequences and structural features of all synthesized siRNAs are summarized in Table .

1. Sequences of ssRNAs, siRNAs, and T _m Values of siRNAs.

^a Y (black), L1 (blue), L2 (red), small letter, capital letter, and bold dot (•) denote an M6P ligand-bearing unit, proline-based linker, hexanediol-based linker, 2′-OMe, 2′-F and phosphorothioate interlinkage, respectively.

^bThe T _m values were measured in buffer containing 10 mM sodium phosphate (pH 7.0) and 100 mM NaCl. The concentrations of the duplexes were 3 μM. All experiments were performed in triplicate, and data are represented as the mean ± SD.

^cΔT _m is given by [T _m (siRNA 2–7) – T _m (siRNA 1)].

2.4. Thermal Stability of M6P Ligand-Conjugated siRNAs

The duplex stability of M6P ligand-conjugated siRNAs is a critical determinant of their biological performance: thus, the thermal stability of the siRNAs was evaluated using temperature-dependent UV melting analysis to assess the impact of chemical modification on their duplex stability. Experiments were conducted in 10 mM sodium phosphate buffer (pH 7.0) containing 100 mM NaCl, and the melting temperatures (T _m) of all siRNA variants were determined at a duplex concentration of 3 μM (Figure S3, Supporting Information).

The unmodified siRNA 1, which served as the reference standard, exhibited a T _m of 76.4 °C. Incorporation of a single M6P ligand (siRNA 2) reduced the T _m only slightly (75.8 °C; ΔT _m = −0.6 °C), indicating that monovalent M6P conjugation exerts a minimal destabilizing effect. Comparable observations have been reported for GalNAc–oligonucleotide and glucosamine-modified systems, where the introduction of a single sugar unit or monovalent ligand typically produces negligible or marginal reductions in duplex stability. , Increasing ligand valency, however, was associated with a progressive decrease in duplex stability. For instance, siRNA 3, containing two M6P moieties conjugated via flexible hexanediol (L2) linkers, exhibited a T _m of 75.4 °C (ΔT _m = −1.1 °C). This observation mirrors findings from GalNAc–oligonucleotide architectures, which demonstrate incremental destabilization proportional to ligand number. The reduction in T _m is plausibly explained by steric hindrance and entropic penalties introduced by multiple flexible linkers.

Interestingly, siRNA 4, which contained only a single M6P ligand, displayed a markedly lower T _m (74.1 °C; ΔT _m = −2.3 °C). This disproportionate reduction in stability may arise from the combined use of flexible hexanediol linkers with a rigid proline-based linker. While proline introduces conformational rigidity, its juxtaposition with flexible segments can generate local conformational strain, ultimately compromising duplex integrity. Previous reports have shown that rigid, proline-rich or helical linkers enhance structural stability, , whereas excessive flexibility or mismatched linker architectures can impair base stacking and duplex rigidity, leading to reduced stability or binding affinity. ,

Consistent with these findings, the tetravalent construct (siRNA 5), bearing multiple M6P ligands predominantly linked via hexanediol moieties, exhibited a further reduction in duplex stability (74.0 °C; ΔT _m = −2.4 °C). These results highlight the cumulative destabilizing effects of high ligand valency and flexible linker architectures. Similar conjugation-induced destabilization trends have been reported for cholesterol-modified siRNAs, where lipid attachment can influence duplex thermodynamics and activity, and for GalNAc-siRNAs, where thermal destabilization strategies have been deliberately employed to tune potency and specificity. ,

The destabilizing role of linker flexibility was further reinforced by siRNA 6, which contained the same number of ligands as siRNA 3 but with additional hexanediol linkers, and displayed a T _m of 74.0 °C (ΔT _m = −2.4 °C), comparable to the tetravalent construct. In contrast, siRNA 7, also tetravalent but incorporating several rigid proline linkers in place of flexible hexanediol segments, exhibited improved thermal stability (74.7 °C; ΔT _m = −1.8 °C). This suggests that conformational rigidity imparted by proline reduces entropy-driven destabilization, thereby partially preserving duplex integrity. ,

Taken together, these findings feature a clear interplay between ligand valency, linker architecture, and duplex thermal stability. While increasing ligand valency and the use of long, flexible linkers consistently compromise duplex integrity, the incorporation of rigid elements such as proline can partially offset these destabilizing effects. Achieving an optimal balance between flexible and rigid linkers is therefore essential, not only to preserve duplex stability, but also to ensure appropriate spatial orientation of ligands and maximize silencing efficiency. Accordingly, the rational design of M6P–siRNA conjugates should carefully integrate both valency and linker composition to yield constructs that are thermodynamically stable while maintaining robust biological activity.

2.5. Circular Dichroism Spectroscopy of Ligand-Conjugated siRNAs

To evaluate the impact of M6P ligand conjugation on the overall tertiary structure of siRNAs, both unmodified and M6P-conjugated siRNA duplexes were analyzed using circular dichroism (CD) spectroscopy in a 10 mM buffer solution over a spectral range of 200–350 nm (Figure ). The CD spectrum of the unmodified siRNA (siRNA 1) displayed the characteristic features of an A-form RNA helix, with a positive peak at approximately 260 nm and a negative peak near 210 nm. Importantly, the CD spectra of all M6P-conjugated siRNA variants closely overlapped with those of the unmodified control, preserving the canonical A-form profile. These results indicate that conjugation of M6P ligands does not disrupt the global helical conformation of the siRNA duplex, thereby preserving its structural integrity. This observation is consistent with previous reports showing that ligand or linker modifications, when positioned at the termini or noncritical regions of siRNAs, generally maintain the A-form structure required for efficient RISC loading and RNAi activity. , By contrast, modifications that introduce steric hindrance or alter backbone geometry can distort the helical conformation and compromise silencing efficiency. , The fact that our M6P-conjugated siRNAs retain canonical CD profiles therefore highlights the structural compatibility of this modification, supporting its application in ligand-mediated delivery strategies without compromising conformational stability.

3.

CD spectra of M6P ligand–conjugated siRNAs (4 μM) recorded in a buffer containing 10 mM sodium phosphate (pH 7.0) and 100 mM NaCl at 20 °C.

2.6. Evaluation of M6P Ligand-Conjugated siRNA Activity In Vitro

The gene-silencing efficacy of the M6P-conjugated siRNA constructs targeting the KNTC2 gene was assessed using K562 cells, a chronic myeloid leukemia cell line. K562 cells exhibit approximately 4-fold higher expression levels of the cation-independent mannose 6-phosphate receptor (CI-M6PR) than HeLa cells, thereby providing a more appropriate in vitro model of receptor-mediated uptake. All experiments were performed under transfection-free conditions to evaluate the intrinsic ability of the constructs to enter cells via endogenous CI-M6PR-mediated internalization. The gene knockdown efficiency was determined by quantitative PCR (qPCR), while KNTC2 silencing was evaluated initially across siRNA constructs 1–5 at a final concentration of 3 μM in a carrier-free format (data not shown).

None of the tested M6P-conjugated siRNA constructs exhibited significant KNTC2 gene silencing activity. Previous reports have linked increased ligand valency to enhanced receptor-mediated uptake: thus, we hypothesized that siRNA 5, a tetravalent construct bearing four M6P ligands connected via extended hexanediol linkers, would demonstrate superior gene knockdown efficacy. Contrary to expectations, at equivalent concentrations (3 μM) siRNA 5 exhibited only minimal activity, similar to that of the divalent siRNA 3. The architecture of siRNA 5 was designed to enhance receptor binding and internalization: however, no measurable improvement in functional gene silencing was observed.

The limited efficacy of siRNA 5 prompted us to hypothesize that the extended and flexible hexanediol linker compromised receptor engagement. Excessive linker length and flexibility may hinder effective receptor interactions and disrupting optimal endocytic processing by shielding M6P ligands. Such structural features can reduce both specific (CI-M6PR-mediated) and nonspecific uptake, thereby limiting intracellular delivery and silencing efficacy.

To test this hypothesis, two siRNA constructs with systematically modified linker structures were synthesized. siRNA 6, a divalent construct similar to siRNA 3, featured an elongated hexanediol linker to assess the effect of increased flexibility. By contrast, siRNA 7 incorporated a proline-based linker and lower hexanediol content but retained the tetravalent configuration of siRNA 5. This rigid cyclic structure was intended to limit conformational flexibility and potentially improve both the stability of receptor binding and endosomal escape owing to the susceptibility of proline to enzymatic cleavage in the lysosomal environment.

Cellular uptake efficiency was evaluated by flow cytometry (Figure S4, Supporting Information), which revealed distinct differences in the internalization of M6P–siRNA conjugates. Compared with the mock control, all tested constructs (siRNAs 1, 3, 6, and 7) exhibited a rightward fluorescence shift, confirming effective uptake by K562 cells. Among these, siRNA 7 showed the most pronounced shift, indicating superior uptake efficiency, while the divalent constructs (siRNAs 3 and 6) displayed intermediate levels. These findings emphasize the critical role of ligand valency and linker architecture in facilitating receptor-mediated uptake.

In parallel, the gene silencing activity of siRNAs 1, 3, 6, and 7 was quantitatively assessed using K562 cells at 3 μM using qPCR (Figure ). The tetravalent constructs outperformed their divalent counterparts. In particular, siRNA 7 achieved the highest knockdown efficiency (∼36% reduction in KNTC2 mRNA relative to siRNA 1), significantly surpassing that of siRNA 5 despite exhibiting the same ligand valency. This superior activity was attributed to the optimized linker rigidity and exposure of the M6P moieties, which facilitated receptor-mediated uptake and intracellular trafficking. In comparison, siRNA 3 (divalent, short hexanediol linker) and siRNA 6 (divalent, extended hexanediol linker) showed activities of 24% and ∼20%, respectively, further demonstrating that the length and flexibility of the linker can inhibit functional activity despite a constant ligand valency.

4.

Relative KNTC2 mRNA levels in myeloid leukemia cell line K562 transfected with siRNAs (3 μM) without transfection reagent and incubated for 48 h, followed by an additional 24 h incubation after medium exchange. The Mock group was used as a control group. Total intracellular mRNA was extracted and subjected to reverse transcription. Quantitative real-time polymerase chain reaction (qRT-PCR) was then performed, and the relative KNTC2 mRNA levels were calculated using the ΔΔCT method. Data are presented as the mean ± standard deviation (SD) of three independent experiments performed in duplicate (Figure S5 and Table S1, Supporting Information). Statistical significance was indicated by p < 0.05 (*), and p < 0.01 (**).

The superior performance of siRNA 7 highlights the importance of incorporating rational linker designs that balance rigidity and degradability to optimize receptor engagement, cellular uptake, and intracellular trafficking.

To validate the factors responsible for the observed differences in gene silencing, all constructs were evaluated via direct cytoplasmic delivery using electroporation. Under these conditions, which bypass membrane-associated barriers, all siRNAs (1, 3, 6, and 7) demonstrated robust gene silencing efficiency (∼80% knockdown), confirming that the chemical modifications did not inhibit RNAi functionality (Figure ).

5.

Relative KNTC2 mRNA levels in K562 cells following transfection with M6P–siRNA conjugates (1, 3, 6, and 7). Cells were electroporated with 100 nM siRNAs using the Neon Transfection System (1000 V, single 50 ms pulse) and incubated for 48 h. Total RNA was isolated, reverse-transcribed into cDNA, and relative KNTC2 mRNA expression was quantified by qRT-PCR using the ΔΔCT method. Data are presented as mean ± SD of three independent experiments (Figure S6, Supporting Information) and normalized to siLuc control (siLuc = 1).

The observed variability under carrier-free conditions is therefore primarily attributed to differences in pharmacokinetics and intracellular trafficking rather than to impairment of the RNAi machinery. Thus, ligand valency and linker structure primarily modulate delivery efficiency rather than the intrinsic silencing capability. Collectively, these results demonstrate that CI-M6PR-targeted delivery of siRNA is strongly influenced by both the ligand valency and physicochemical properties of the linker. Although higher-valency constructs generally promote uptake, this alone is not sufficient to ensure enhanced gene silencing. Rather, the receptor interactions, endosomal trafficking, and cytoplasmic release are influenced by the spatial presentation of the M6P ligands, linker rigidity, and degradability. Among the tested designs, siRNA 7 emerged as the most effective construct owing to its proline-based linker and tetravalent configuration, combining high receptor engagement with efficient intracellular delivery and gene silencing. Our findings demonstrate the importance of rational linker engineering in the development of receptor-targeted siRNA therapeutics and provide a framework for the design of future ligand-conjugated oligonucleotide delivery systems targeting CI-M6PR-expressing tumors.

3. Conclusions

This study presents the design, synthesis, and systemic evaluation of several M6P–siRNA conjugates targeting the cation-independent mannose-6-phosphate receptor (CI-M6PR) to enhance receptor-mediated siRNA delivery by precisely modulating ligand valency and linker structure. Both parameters critically influence the physicochemical properties, structural integrity, and gene silencing efficacy of the conjugates. Although increased ligand valency enhances the potential for receptor binding and cellular uptake, it does not inherently necessarily improve gene silencing. Instead, the spatial configuration and rigidity of the linker architecture are the key determinants of functional delivery. In particular, rigid proline-based linkers conferred superior gene knockdown efficacy than their flexible hexanediol counterparts, likely owing to improved ligand exposure, receptor engagement, and intracellular trafficking. Consistently, flow cytometry analysis confirmed that the superior activity of siRNA 7 was attributable to its markedly higher cellular uptake relative to the other conjugates.

Thermal melting analysis revealed an inverse correlation between the linker flexibility and duplex stability, whereas circular dichroism spectroscopy confirmed that all M6P conjugates retained their native A-form RNA structure. In vitro studies using K562 cells, which overexpress CI-M6PR, identified siRNA 7, a tetravalent construct with proline-based linkers, as the most effective, achieving an approximately 36% knockdown of KNTC2 mRNA in the absence of a transfection carrier. Electroporation studies further confirmed that the RNAi functionality remained intact across all constructs, demonstrating that differences in gene silencing were attributable to delivery efficiency rather than to the loss of RNAi activity. These findings highlight the importance of the rational design of ligand-conjugated siRNA platforms. This study provides a mechanistic framework for optimizing multivalent receptor-targeted oligonucleotide therapeutics and facilitates the broader applicability of CI-M6PR-mediated delivery strategies in cancer and lysosome-targeting applications beyond hepatic systems.

4. Experimental Section

4.1. General Remark

All chemicals and dry solvents (THF, DMF, CH₂Cl₂, MeOH, and pyridine) were obtained from commercial sources and used without any further purification. Thin layer chromatography (TLC) was performed on silica gel plates precoated with fluorescent indicator with visualization by UV light or by dipping into a solution of 5% (v/v) concentrated H₂SO₄ in mixture of p-anisaldehyde and methanol and then heating. Silica gel (63–210 mesh) was used for column chromatography. ¹H NMR (400 or 600 MHz), ¹³C {¹H} NMR (151 MHz), ³¹P NMR (162 MHz) were recorded on 400 or 600 MHz NMR equipment. CDCl₃ was used as a solvent for obtaining NMR spectra. Chemical shifts (δ) are reported in parts per million (ppm) relative to CDCl₃ (7.26 ppm) for ¹H NMR spectra, CDCl₃ (77.16 ppm) for ¹³C­{¹H} NMR spectra, and 80% H₃PO₄ (0.0 ppm) for ³¹P NMR. The abbreviations s, d, t, q, and m signify singlet, doublet, triplet, quadruplet, and multiplet, respectively. High resolution mass spectra (HRMS) were obtained in positive ion electrospray ionization (ESI-TOF) mode.

4.1.1. (2S,3S,4S,5R,6R)-2-((5-bromopentyl)­oxy)-6-(((tert-butyldiphenylsilyl)­oxy)­methyl)­tetrahydro-2H-pyran-3,4,5-triyl triacetate (6)

To a solution of compound 4 (3.66 g, 6.72 mmol) in anhydrous CH₂Cl₂ (26.8 mL), trichloroacetonitrile (8.82 mL, 87.31 mmol) and 1,8-diazabicyclo [5.4.0]­undec-7-ene (DBU, 0.20 mL, 1.34 mmol) were added sequentially under an ice bath (0 °C) with stirring under an inert atmosphere. The reaction mixture was stirred at 0 °C for 15 min, then allowed to warm to room temperature and stirred for an additional 1.5 h. Upon completion of the reaction, as monitored by TLC, the mixture was quenched with cold saturated aqueous NaHCO₃ solution. The aqueous phase was extracted with CHCl₃, and the combined organic extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was purified by column chromatography on silica gel using 17% EtOAc in hexane and 1% Et₃N to furnish the product 5 as a yellow solid. Intermediate 5 was dissolved in anhydrous CH₂Cl₂ (44.8 mL) and cooled to −20 °C under an argon atmosphere. To this solution, 5-bromo-1-pentanol (1.18 mL, 9.75 mmol) was added, and the mixture was stirred for 10 min. Subsequently, boron trifluoride diethyl etherate (BF₃·Et₂O, 1.63 mL, 13.00 mmol) was added dropwise while maintaining the temperature at −20 °C. The reaction mixture was stirred at −20 °C for 2.3 h. The reaction was quenched by the addition of saturated aqueous NaHCO₃ (10.0 mL), and the mixture was extracted with CHCl₃. The combined organic layers were washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The residue was purified by column chromatography on silica gel using 14% EtOAc in hexane as the eluent to afford compound 6 as an oily product (2.38 g, 6.35 mmol, 51% yield over two steps). ¹H NMR (600 MHz, CDCl₃) δ: 7.67 (ddd, J = 13.4, 8.0, 1.3 Hz, 4H), 7.44–7.36 (m, 6H), 5.35–5.30 (m, 2H), 5.20 (dd, J = 2.9, 1.9 Hz, 1H), 4.82 (d, J = 1.6 Hz, 1H), 3.84–3.81 (m, 1H), 3.78–3.72 (m, 2H), 3.68 (dd, J = 11.3, 1.9 Hz, 1H), 3.46–3.43 (m, 1H), 3.42–3.39 (m, 2H), 2.13 (s, 3H), 1.99 (s, 3H), 1.91–1.85 (m, 5H), 1.65–1.60 (m, 2H), 1.54–1.47 (m, 2H), 1.06 (s, 9H); ¹³C NMR (151 MHz, CDCl₃) δ: 170.3, 170.2, 169.7, 135.8, 135.7, 133.4, 133.3, 129.8, 128.0, 127.9, 127.8, 127.7, 97.2, 71.5, 70.0, 69.6, 67.7, 66.5, 63.0, 33.6, 32.6, 29.8, 28.5, 26.8, 24.9, 21.0, 20.9, 20.7, 19.3; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₃₃H₄₅BrNNaO₉Si, 715.1914; found, 715.1928.

4.1.2. (2S,3S,4S,5R,6R)-2-((5-azidopentyl)­oxy)-6-(((tert-butyldiphenylsilyl)­oxy)­methyl)­tetrahydro-2H-pyran-3,4,5-triyl triacetate (7)

To a solution of compound 6 (1.81 g, 2.61 mmol) in anhydrous DMF (18.1 mL), sodium azide (NaN₃, 0.85 g, 13.04 mmol) was added under an argon atmosphere. The reaction mixture was stirred at 60 °C for 6 h under inert conditions. Upon completion of the reaction, as monitored by TLC, the mixture was cooled to room temperature and diluted with EtOAc, followed by extraction. The organic phase was washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was purified by column chromatography on silica gel using 20% EtOAc in hexane as the eluent to afford compound 7 as an oily material (1.65 g, 2.51 mmol, 96%). ¹ H NMR (600 MHz, CDCl₃) δ: 7.69–7.65 (m, 4H), 7.44–7.41 (m, 2H), 7.39–7.36 (m, 4H), 5.35–5.30 (m, 2H), 5.20 (dd, J = 2.8, 1.9 Hz, 1H), 4.81 (d, J = 1.6 Hz, 1H), 3.83–3.81 (m, 1H), 3.77 (dd, J = 11.3, 5.5 Hz, 1H), 3.75–3.72 (m, 1H), 3.68 (dd, J = 11.3, 1.9 Hz, 1H), 3.45–3.41 (m, 1H), 3.27 (t, J = 6.9 Hz, 2H), 2.13 (s, 3H), 1.99 (s, 3H), 1.88 (s, 3H), 1.65–1.59 (m, 4H), 1.46–1.40 (m, 2H), 1.06 (s, 9H); ¹³C NMR (151 MHz, CDCl₃) δ: 170.3, 170.2, 169.7, 135.8, 135.7, 133.4, 133.3, 129.8, 127.8, 127.7, 97.2, 71.6, 70.1, 69.6, 67.7, 66.5, 63.0, 51.4, 29.0, 28.8, 26.8, 23.5, 21.0, 20.9, 20.8, 19.4; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₃₃H₄₅N₃NaO₉Si, 678.2823; found, 678.2837.

4.1.3. (2S,3S,4S,5R,6R)-2-((5-azidopentyl)­oxy)-6-(hydroxymethyl)­tetrahydro-2H-pyran-3,4,5-triyl triacetate (8)

To a stirred solution of compound 7 (4.32 g, 6.59 mmol) in anhydrous THF (43.2 mL), n-tetrabutylammonium fluoride (TBAF, 11.2 mL of a 1 M solution in THF, 11.2 mmol) and acetic acid (AcOH, 1.88 mL, 32.94 mmol) were added sequentially at room temperature under an argon atmosphere. The reaction mixture was stirred for 3 h at room temperature. Upon completion of the reaction, as confirmed by TLC, the mixture was diluted with EtOAc and extracted. The organic layer was washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude product was purified by column chromatography on silica gel using 50% EtOAc in hexane as the eluent to afford compound 8 as an oily substance (2.01 g, 4.82 mmol, 73%). ¹ H NMR (600 MHz, CDCl₃) δ: 5.40 (dd, J = 10.2, 3.5 Hz, 1H), 5.25–5.21 (m, 2H), 4.81 (d, J = 1.5 Hz, 1H), 3.76 (dd, J = 10.0, 4.1, 2.3 Hz, 1H), 3.72–3.68 (m, 2H), 3.64–3.61 (m, 1H), 3.46–3.42 (m, 1H), 3.31–3.28 (m, 2H), 2.38 (dd, J = 8.7, 5.6 Hz, 1H), 2.15 (s, 3H), 2.08 (s, 3H), 2.01 (s, 3H), 1.67–1.61 (m, 4H), 1.48–1.43 (m, 2H); ¹³ C NMR (151 MHz, CDCl₃) δ: 170.9, 170.2, 170.0, 97.6, 70.6, 69.7, 68.9, 68.0, 66.6, 61.3, 51.3, 28.9, 28.7, 23.4, 21.0, 20.8; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₁₇H₂₇N₃NaO₉, 440.16465; found, 440.1646.

4.1.4. (2S,3S,4S,5R,6R)-2-((5-azidopentyl)­oxy)-6-(((bis­(2-cyanoethoxy)­phosphoryl)­oxy)­methyl)­tetrahydro-2H-pyran-3,4,5-triyl Triacetate (9)

To a solution of compound 8 (2.08 g, 4.98 mmol) in anhydrous THF (7.00 mL), 1H-tetrazole (1.04 g, 14.96 mmol) was added at – 40 °C under an inert atmosphere. A solution of bis­(2-cyanoethyl) N,N-diisopropylphosphoramidite (3.25 mL, 12.45 mmol) in THF (5.40 mL) was then added dropwise to the reaction mixture while maintaining the temperature at – 40 °C. The reaction was gradually warmed to room temperature and stirred for 1.5 h. Subsequently, the reaction mixture was cooled to 0 °C using an ice bath, and 30% tert-butyl hydroperoxide (2.99 mL, 9.96 mmol) was added dropwise. The mixture was stirred for an additional 1.5 h at 0 °C. The reaction was then quenched with an aqueous solution of sodium thiosulfate (10.3 mL) and extracted with EtOAc and water. The organic layer was washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude residue was purified by column chromatography on silica gel using 66% EtOAc in hexane as the eluent to yield compound 9 as an oily material (2.76 g, 4.57 mmol, 92%). ¹ H NMR (600 MHz, CDCl₃) δ: 5.34 (dd, J = 10.0, 3.4 Hz, 1H), 5.28 (t, J = 10.0 Hz, 1H), 5.24 (q, J = 1.7 Hz, 1H), 4.80 (d, J = 1.5 Hz, 1H), 4.34–4.27 (m, 4H), 4.24–4.18 (m, 2H), 3.98–3.96 (m, 1H), 3.73–3.68 (m, 1H), 3.48–3.44 (m, 1H), 3.29 (t, J = 6.8 Hz, 2H), 2.84–2.75 (m, 4H), 2.15 (s, 3H), 2.06 (s, 3H), 1.99 (s, 3H), 1.66–1.59 (m, 4H), 1.48–1.42 (m, 2H); ¹³ C NMR (151 MHz, CDCl₃) δ: 170.0, 169.9, 116.5, 116.4, 97.6, 69.5, 69.0, 68.8, 68.7, 68.3, 66.3, 65.5, 62.5, 62.4, 51.2, 28.8, 28.6, 23.3, 20.9, 20.8, 20.7, 20.6, 19.7, 19.6, 19.5; ³¹ P NMR (243 MHz, CDCl₃) δ: −1.8; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₂₃H₃₄N₅NaO₁₂P, 626.1839; found, 626.1834.

4.1.5. 4-oxo-4-(((1-(5-(((2S,3S,4S,5R,6R)-3,4,5-triacetoxy-6-(((bis­(2-cyanoethoxy)­phosphoryl)­oxy)­methyl)­tetrahydro-2H-pyran-2-yl)­oxy)­pentyl)-1H-1,2,3-triazol-4-yl)­methyl)­amino)­butanoic acid (10)

To a solution of compound 13 (0.97 g, 6.26 mmol) in water (16.9 mL), copper­(II) sulfate (CuSO₄, 0.17 g, 1.04 mmol) was added under ambient conditions. In a separate vessel, a solution of compound 9 (2.52 g, 4.18 mmol) was prepared in a mixture of t-BuOH/H₂O (2.52 mL/2.00 mL), followed by the addition of sodium ascorbate (0.83 g, 4.18 mmol) and sodium carbonate (0.089 g, 0.84 mmol). This mixture was then added to the aqueous solution of compound 13 containing CuSO₄. The resulting reaction mixture was stirred at room temperature for 1.3 h. Upon completion of the reaction, as monitored by TLC, the mixture was extracted with CHCl₃ and water. The organic phase was separated, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude residue was purified by column chromatography on silica gel using a mixture of CHCl₃ and MeOH (10:1, v/v) as the eluent to afford compound 10 as a white solid (2.11 g, 2.79 mmol, 67%). ¹ H NMR (600 MHz, CDCl₃) δ: 7.59 (s, 1H), 5.30 (dd, J = 10.1, 3.4 Hz, 1H), 5.25 (t, J = 10.0 Hz, 1H), 5.22 (q, J = 1.7 Hz, 1H), 4.79 (d, J = 1.5 Hz, 1H), 4.49 (d, J = 5.8 Hz, 2H), 4.38–4.26 (m, 6H), 4.25–4.18 (m, 2H), 3.96–3.94 (m, 1H), 3.71–3.66 (m, 1H), 3.48–3.44 (m, 1H), 2.86–2.77 (m, 4H), 2.67–2.65 (m, 2H), 2.53–2.51 (m, 2H), 2.15 (s, 3H), 2.07 (s, 3H), 2.00 (s, 3H), 1.96–1.89 (m, 2H), 1.68–1.60 (m, 2H), 1.40–1.34 (m, 2H); ¹³ C NMR (151 MHz, CDCl₃) δ: 172.7, 170.2, 170.1, 170.0, 144.7, 122.6, 116.7, 116.6, 97.7, 69.5, 69.1, 69.0, 68.9, 68.2, 66.7, 66.6, 65.7, 62.7, 62.6, 50.3, 34.8, 30.9, 29.8, 28.6, 23.1, 21.0, 20.8, 19.8, 19.7; ³¹ P NMR (243 MHz, CDCl₃) δ: −2.1.; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₃₀H₄₃N₆NaO₁₅P, 781.2422; found, 781.2399.

4.1.6. (2S,3S,4R,5R,6R)-2-(((bis­(2-cyanoethoxy)­phosphoryl)­oxy)­methyl)-6-((5-(4-((4-((2S,4R)-2-((bis­(4-methoxyphenyl)­(phenyl)­methoxy)­methyl)-4-hydroxypyrrolidin-1-yl)-4-oxobutanamido)­methyl)-1H-1,2,3-triazol-1-yl)­pentyl)­oxy)­tetrahydro-2H-pyran-3,4,5-triyl triacetate (11)

To a stirred solution of compound 10 (1.69 g, 2.23 mmol) in anhydrous DMF (16.9 mL), N,N-diisopropylethylamine (DIPEA, 1.14 mL, 6.68 mmol) and O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU, 0.93 g, 2.45 mmol) were added sequentially under an argon atmosphere. The mixture was stirred at room temperature for 10 min. A solution of compound 18 (1.21 g, 2.90 mmol) in DMF (4.00 mL) was then added dropwise to the reaction mixture. The resulting mixture was stirred for 50 min at room temperature. Upon completion, the reaction was quenched by the addition of cold water under ice bath conditions and extracted with EtOAc and water. The organic phase was washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude residue was purified by column chromatography on silica gel using CHCl₃/MeOH (15:1, v/v) as the eluent to yield compound 11 as a white solid (1.54 g, 1.33 mmol, 60%). ¹ H NMR (600 MHz, CDCl₃) δ: 7.54–7.51 (m, 1H), 7.36–7.33 (m, 2H), 7.27–7.16 (m, 7H), 7.05 (t, J = 5.6 Hz, 1H), 6.81–6.78 (m, 4H), 5.30 (dq, J = 10.0, 1.6 Hz, 1H), 5.25 (t, J = 9.9 Hz, 1H), 5.21 (dd, J = 4.7, 3.0 Hz, 1H), 4.78 (dd, J = 5.3, 1.5 Hz, 1H), 4.56 (s, 0.6H), 4.51–4.45 (m, 1H), 4.41–4.33 (m, 2H), 4.32–4.23 (m, 6.4H), 4.20–4.18 (m, 2H), 3.97–3.93 (m, 1.4H), 3.77–3.75 (m, 6H), 3.69–3.60 (m, 2.4H), 3.45–3.40 (m, 1.6H), 3.32–3.29 (m, 1H), 3.17 (dd, J = 9.8, 5.0 Hz, 0.4H), 3.09 (dd, J = 9.3, 3.1 Hz, 1H), 2.83–2.60 (m, 6H), 2.47–2.41 (m, 1.2H), 2.35–2.30 (m, 0.4H), 2.27–2.19 (m, 1H), 2.17–2.09 (m, 3.8H), 2.06 (s, 3H), 2.01–1.99 (m, 3.6H), 1.92–1.85 (m, 2H), 1.65–1.57 (m, 2H), 1.38–1.32 (m, 2H); ¹³ C NMR (151 MHz, CDCl₃) δ: 172.7, 172.2, 171.0, 170.2, 170.1, 170.0, 158.7, 158.5, 145.2, 136.3, 136.2, 135.9, 130.1, 128.2, 128.0, 127.9, 127.0, 126.8, 122.4, 116.6, 113.3, 113.2, 97.7, 86.6, 85.9, 70.5, 69.6, 69.4, 69.2, 69.0, 68.2, 66.6, 65.7, 65.4, 63.5, 62.6, 56.4, 56.0, 55.9, 55.3, 50.1, 36.7, 35.2, 35.1, 31.7, 31.3, 30.2, 29.9, 28.6, 23.1, 21.0, 20.9, 20.8, 19.8, 19.7; ³¹ P NMR (243 MHz, CDCl₃) δ: −2.0, −2.0; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₅₆H₇₀N₇NaO₁₈₈P, 1182.4413; found, 1182.4387.

4.1.7. (2R,3R,4S,5S,6S)-2-(((bis­(2-cyanoethoxy)­phosphoryl)­oxy)­methyl)-6-((5-(4-((4-((2S,4R)-2-((bis­(4-methoxyphenyl)­(phenyl)­methoxy)­methyl)-4-(((2-cyanoethoxy)­(diisopropylamino)­phosphaneyl)­oxy)­pyrrolidin-1-yl)-4-oxobutanamido)­methyl)-1H-1,2,3-triazol-1-yl)­pentyl)­oxy)­tetrahydro-2H-pyran-3,4,5-triyl triacetate (12)

To a stirred solution of compound 11 (0.69 g, 0.60 mmol) in anhydrous CH₂Cl₂ (6.94 mL), DIPEA (0.41 mL, 2.39 mmol) and 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (0.27 mL, 1.20 mmol) were added sequentially under an argon atmosphere at room temperature. The reaction mixture was stirred for 1.5 h. Upon completion of the reaction, the mixture was diluted with CHCl₃ and washed successively with saturated aqueous NaHCO₃ and brine. The organic phase was dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude product was purified by column chromatography on silica gel using a mixture of CHCl₃ and MeOH (30:1, v/v) as the eluent to afford the desired phosphoramidite compound 12 as a white solid (0.52 g, 0.39 mmol, 64%). ¹ H NMR (600 MHz, CDCl₃) δ: 7.54–7.50 (m, 1H), 7.37–7.34 (m, 2H), 7.29–7.27 (m, 1.5H), 7.25–7.17 (m, 5.5H), 6.83–6.77 (m, 5H), 5.31 (dd, J = 10.0, 3.3 Hz, 1H), 5.26 (t, J = 9.9 Hz, 1H), 5.22 (q, J = 1.5 Hz, 1H), 4.78 (d, J = 1.5 Hz, 1H), 4.48 (d, J = 5.7 Hz, 2H), 4.34–4.25 (m, 7H), 4.21–4.19 (m, 2H), 3.94 (d, J = 9.5 Hz, 1H), 3.89 (q, J = 6.0 Hz, 1H), 3.83–3.65 (m, 10H), 3.60–3.40 (m, 4.7H), 3.21–3.06 (m, 1.3H), 2.83–2.75 (m, 4H), 2.72–2.47 (m, 5.5H), 2.39–2.34 (m, 0.5H), 2.26–2.21 (m, 0.7H), 2.15–2.12 (m, 3.7H), 2.11–2.06 (m, 3.6H), 2.00 (s, 3H), 1.93–1.87 (m, 2H), 1.66–1.63 (m, 2H), 1.41–1.36 (m, 2H), 1.18–1.13 (m, 12H); ³¹ P NMR (243 MHz, CDCl₃) δ: 148.2, 147.8, 147.7, −2.2; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₆₅H₈₇N₉NaO₁₉P₂, 1382.5491; found, 1382.5519.

4.1.8. 4-Oxo-4-(2-propyn-1-ylamino) butanoic acid (13)

This compound was prepared according to the procedure reported in literature.

4.1.9. (2S,4R)-1-(((9H-fluoren-9-yl)­methoxy)­carbonyl)-4-hydroxypyrrolidine-2-carboxylic acid (15)

This compound was prepared according to the procedure reported in literature.

4.1.10. (9H-fluoren-9-yl)­methyl (2S,4R)-4-hydroxy-2-(hydroxymethyl)­pyrrolidine-1-carboxylate (16)

Using the crude product of intermediate 15, NaBH₄ (0.92 g, 24.40 mmol) was added a solution of 15 in THF (53.1 mL) under an ice bath. After stirring under an ice bath for 1 h, the mixture was added dropwise with BF₃·Et₂O (3.83 mL, 30.50 mmol) and stirring under an ice bath for 10 min. After stirring at room temperature for 5 h, the mixture was quenched by H₂O. The mixture was extracted with CHCl₃ and water. The organic layer was washed with brine, dried over Na₂SO₄, filtered, and concentrated to furnish the crude product of 16 which was used in next step without further purification.

4.1.11. (9H-fluoren-9-yl)­methyl (2S,4R)-2-((bis­(4-methoxyphenyl)­(phenyl)­methoxy)­methyl)-4-hydroxypyrrolidine-1-carboxylate (17)

This compound was prepared according to the procedure reported in literature using the crude product 16.

4.1.12. (3R,5S)-5-[[Bis­(4-methoxyphenyl) phenylmethoxy] methyl]-3-pyrrolidinol (18)

To a solution of 17 (1.14 g, 1.78 mmol) in DMF (11.4 mL), Triethylamine (11.4 mL) was added under argon atmosphere. After stirring under 80 °C for 1.5 h, the mixture was extracted with CHCl₃ after quenching with saturated aqueous NaHCO₃. The organic layer was washed with brine, dried over Na₂SO₄, filtered, and concentrated. The crude residue was purified by column chromatography using CHCl₃: MeOH (v/v = 20:1) as the eluent to afford the desired product 18 as a yellow solid (0.56 g, 1.31 mmol, 75%). ¹ H NMR (600 MHz, CDCl₃) δ: 7.43–7.41 (m, 2H), 7.32–7.26 (m, 6H), 7.21–7.19 (m, 1H), 6.83–6.80 (m, 4H), 4.40–4.37 (m, 1H), 3.78 (s, 6H), 3.66–3.62 (m, 1H), 3.11–3.03 (m, 3H), 2.92–2.88 (m, 1H), 2.49 (s, 2H), 1.90–1.86 (m, 1H), 1.69–1.64 (m, 1H).

4.1.13. (9H-fluoren-9-yl)­methyl (2S,4R)-2-((bis­(4-methoxyphenyl)­(phenyl)­methoxy)­methyl)-4-(((2-cyanoethoxy)­(diisopropylamino)­phosphaneyl)­oxy)­pyrrolidine-1-carboxylate (19)

To a solution of compound 17 (1.29 g, 2.01 mmol) in anhydrous THF (20.0 mL) under an argon atmosphere, N,N-diisopropylethylamine (DIPEA, 1.03 mL, 6.06 mmol) and 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (0.54 mL, 2.41 mmol) were added sequentially at room temperature. The reaction mixture was stirred for 1.5 h and then quenched with saturated aqueous NaHCO₃. The resulting mixture was extracted with EtOAc, and the organic layer was washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography using 50% EtOAc in hexane as the eluent to afford compound 19 as a white solid (1.36 g, 1.62 mmol, 80% yield). ¹ H NMR (600 MHz, CDCl₃) δ: 7.77–7.71 (m, 2H), 7.66–7.61 (m, 1H), 7.47 (d, J = 7.3 Hz, 0.5H), 7.40–7.35 (m, 4H), 7.31–7.22 (m, 8H), 7.17 (t, J = 7.0 Hz, 1.5H), 6.81–6.75 (m, 4H), 4.73–4.61 (m, 1H), 4.36–4.30 (m, 1.5H), 4.28–4.18 (m, 1.5H), 4.03 (s, 0.5H), 3.96 (t, J = 6.4 Hz, 0.5H), 3.84–3.66 (m, 10H), 3.65–3.54 (m, 2H), 3.49–3.42 (m, 0.5H), 3.29–3.24 (m, 0.5H), 3.18–3.14 (m, 0.5H), 3.05 (t, J = 10.4 Hz, 0.5H), 2.64–2.54 (m, 2H), 2.33–2.16 (m, 2H), 1.20–1.13 (m, 12H); ³¹ P NMR (243 MHz, CDCl₃) δ: 148.6, 148.4, 148.3, 148.2; HRMS (ESI-TOF) m/z: calcd for [M + Na] ⁺ C₅₀H₅₆N₃NaO₇P, 864.37536; found, 864.37712.

4.1.14. 6-[Bis­(4-methoxyphenyl) phenylmethoxy]-1-hexanol (21)

Compound 21 was prepared according to a previously reported method.

4.1.15. 6-(Bis­(4-methoxyphenyl)­(phenyl)­methoxy)­hexyl (2-cyanoethyl) diisopropylphosphoramidite (22)

To a stirred solution of compound 21 (2.64 g, 6.29 mmol) in CH₂Cl₂ (31.4 mL) under an argon atmosphere, N,N-diisopropylethylamine (DIPEA, 2.14 mL, 12.57 mmol) and 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (1.68 mL, 7.55 mmol) were added sequentially at room temperature. The reaction mixture was stirred for 1 h at ambient temperature, after which it was quenched with saturated aqueous NaHCO₃ and extracted with CHCl₃. The combined organic layers were washed with brine, dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The crude product was purified by column chromatography using 33% EtOAc in hexane containing 0.5% TEA as the eluent to afford compound 22 as a pale-yellow oil (2.75 g, 4.43 mmol, 70% yield). ¹ H NMR (400 MHz, CDCl₃) δ: 7.44–7.42 (m, 2H), 7.34–7.26 (m, 6H), 7.22–7.18 (m, 1H), 6.82 (dt, J = 9.6, 2.5 Hz, 4H), 3.88–3.75 (m, 8H), 3.68–3.52 (m, 4H), 3.03 (t, J = 6.6 Hz, 2H), 2.62 (t, J = 6.6 Hz, 2H), 1.65–1.58 (m, 4H), 1.42–1.31 (m, 4H), 1.18 (t, J = 7.3 Hz, 12H); ³¹ P NMR (162 MHz, CDCl₃) δ: 147.9.

4.2. Solid-phase Oligonucleotide Synthesis

The synthesis was carried out with a DNA/RNA synthesizer using phosphoramidite method. After the synthesis, the CPG beads were treated with 10% dimethylamine in CH₃CN for 5 min followed by a rinse with CH₃CN to selectively remove cyanoethyl groups. Then, the oligomers were cleaved from CPG beads and deprotected by treatment with concentrated NH₃ solution/40% methylamine (1:1, v/v) for 10 min at 65 °C. The oligonucleotides were purified by 20% PAGE containing 7 M urea to give highly purified oligonucleotides.

4.3. MALDI-TOF/MS Analysis of ONs

The spectra were obtained with a time-of-flight mass spectrometer equipped with a nitrogen laser (337 nm, 3 ns pulse). A solution of 3- hydroxypicolinic acid (3-HPA) and diammonium hydrogen citrate in H₂O was used as the matrix. Data of synthetic ONs: RNA 1 m/z = 6843.65 (calcd for C₂₁₄H₂₇₁N₆₈O₁₄₇P₂₀F₄ [M-H]⁻, 6843.94); RNA 2 m/z = 7532.94 (calcd for C₂₃₇H₃₁₀N₇₃O₁₆₂P₂₂F₄ [M-H]⁻, 7531.55); RNA 3 m/z = 8400.13 (calcd for C₂₆₆H₃₆₂N₇₈O₁₈₁P₂₅F₄ [M-H]⁻, 8399.32); RNA 4 m/z = 8070.65 (calcd for C₂₅₄H₃₄₆N₇₄O₁₇₄P₂₅F₄ [M-H]⁻, 8071.00); RNA 5 m/z = 10855.59 (calcd for C₃₄₈H₅₁₈N₈₈O₂₃₅P₃₅F₄ [M-H]⁻, 10855.50); RNA 6 m/z = 9120.99 (calcd for C₂₉₀H₄₁₄N₇₈O₁₉₇P₂₉F₄ [M-H]⁻, 9119.96); RNA 7 m/z = 10671.67 (calcd for C₃₃₉H₄₉₆N₉₁O₂₃₁P₃₄F₄ [M-H]⁻, 10672.30); RNA 8 m/z = 6935.41 (calcd for C₂₁₅H₂₆₉N₇₈O₁₃₃P₂₀F₆S₄ [M-H]⁻, 6936.03).

4.4. Thermal Denaturation Study

The solution containing 3.0 μM duplex in a buffer of 10 mM sodium phosphate (pH 7.0) containing 100 mM NaCl was heated at 100 °C and then cooled gradually to room temperature and used for the thermal denaturation study. Thermally induced transitions were monitored at 260 nm with a UV/vis spectrometer fitted with temperature controller in quartz cuvettes with a path length of 1.0 cm. The sample temperature was increased by 0.5 °C/min.

4.5. CD Spectroscopy

All CD spectra were recorded at 20 °C. The following instrument settings were used: resolution, 0.1 nm; response, 1.0 s; speed, 50 nm/min; accumulation, 10.

4.6. Reverse-phase HPLC Analysis of ON

The purity of each synthesized ONs in this study was evaluated by reverse-phase HPLC (C18G 5 mm, 150 4.6 mm SS). Elution started from 100% buffer A, followed by a linear gradient to 60% buffer B in 30 min at a flow rate of 1.0 mL min^–1 (buffer A: 5% MeCN in 0.1 M TEAA (pH 7.0)); buffer B: 50% MeCN in 0.1 M TEAA (pH 7.0).

4.7. Cellular Uptake of siRNA

K562 cells (4 × 10⁵) were incubated with 3.0 μM fluorescein-labeled M6P ligand in 0.2 mL of Opti-MEM I Reduced Serum Medium (Thermo Fisher Scientific) supplemented with 2% fetal bovine serum. After 24 h of incubation, the cells were washed twice with phosphate-buffered saline (PBS), resuspended in 0.8 mL of FACS buffer (0.5% BSA, 2 mM EDTA in PBS), and analyzed by flow cytometry using a FACSCalibur (BD Biosciences).

4.8. Electroporation of siRNAs

K562 cells were transfected with siRNAs (100 nM) using the Neon Transfection System (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Briefly, 1.5 × 10⁵ cells were electroporated in a 10 μL Neon tip under the following conditions: 1000 V, 50 ms, and a single pulse. After transfection, cells were incubated for 48 h in antibiotic-free medium. Total RNA was then extracted and reverse-transcribed into cDNA. Quantitative real-time PCR (qRT-PCR) was performed to determine relative KNTC2 mRNA levels, which were calculated using the ΔΔCT method. Data are expressed as the mean ± standard deviation (SD) from three independent experiments.

4.9. Reverse Transcription-Quantitative Polymerase Chain Reaction

Total RNAs were isolated from the cells using an NucleoSpin RNA Plus Kit (Macherey-Nagel, Düren, Germany). cDNAs were synthesized from total RNA using a PrimeScript RT Master Mix (Takara Bio, Shiga, Japan). Real-time RT-PCR was performed on a Thermal Cycler Dice Real Time System (Takara Bio) using TB Green Premix Ex TaqII (Takara Bio). All reactions were run in duplicate, and the relative expression levels were calculated by the ΔΔCT method using beta-actin (ACTB) as a reference. The primer sequences used were as follows: KNTC2 Primer 1 forward 5′-CCTCTCCATGCAGGAGTTAAGA-3′ and reverse 5′-GGTCTCGGGTCCTTGATTTTCT-3′; KNTC2 Primer 2 forward 5′-GGTCTCAATGAGGAAATTGCTAGA-3′ and reverse 5′-GGTTGTCAATGATATTCTGTAGTCG-3′; ACTB Primer forward 5′-GGAGCAATGATCTTGATCTT-3′ and reverse 5′-CCTTCCTGGGCATGGAGTCCT-3′.

All data supporting the findings of this study are available within the article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c08046.

• MALDI-TOF-MS analysis of RNAs 2–7; characterization of M6P-modified RNAs 2–7 by reverse-phase HPLC; UV melting profiles of siRNAs 1–7; flow cytometry histograms depicting uptake of M6P-conjugated siRNA constructs in K562 cells (summaries of two independent experiments, A and B); KNTC2 gene silencing in K562 cells using M6P–siRNA conjugates under transfection-free conditions (summaries of three independent experiments, A–C); suppression values of KNTC2 gene expression in the absence of lipofection reagents; KNTC2 gene silencing under electroporation conditions (summaries of three independent experiments, A–C); and copies of NMR spectra (¹H, ¹³C and ³¹P) (PDF)

The authors declare no competing financial interest.