Mapping GlycoRNAs on the Exosomal Surface: Methodological Innovation and Functional Discovery
For a long time, RNA has been viewed as a carrier of genetic information, with its core function restricted to protein synthesis. However, as epitranscriptomics research has advanced, it has become increasingly clear that post-transcriptional modifications confer upon RNA a chemical diversity and functional repertoire far exceeding traditional understanding. In recent years, a groundbreaking discovery has reshaped this field: RNA molecules themselves can serve as targets for glycosylation, forming a novel class of biomolecules termed glycoRNAs.
In 2021, Flynn and colleagues identified glycoRNAs carrying Sialic Acid and Fucose modifications on the cell surface for the first time. These glycosylated RNAs primarily include small nuclear (sn) RNAs, ribosomal (r) RNAs, small nucleolar (sno) RNAs, transfer (t) RNAs, and Ro-associated Y RNAs. This discovery is a milestone because it challenges the conventional view that only lipids and proteins can undergo glycosylation, demonstrating that RNA can also participate in biological processes such as molecular recognition and intercellular communication.
Exosomes are key mediators of intercellular information transfer, and their molecular composition directly reflects the physiological state of their parent cells. Unlike ectosomes that bud directly from the plasma membrane, exosomes originate from the endosomal system. Their surfaces are typically enriched in tetraspanin proteins, including CD9, CD63, and CD81. Previous studies demonstrated that following skin injury, bidirectional crosstalk between resident epidermal keratinocytes and wound-site macrophages is mediated by keratinocyte-derived exosomes (Exoκ) and macrophage-derived exosomes (ExomΦ), and that this crosstalk is central to inflammation resolution. Notably, Exoκ isolated from the wound edge (< 2 mm from the edge) are rich in N-glycan modifications and carry abundant small RNAs (< 200 bp) that are critical for resolving wound inflammation. Based on this background, the authors hypothesized that glycoRNAs, previously identified on cell surfaces, might also be anchored on the surface of endosome-derived exosomes. If validated, these surface glycoRNAs could play a crucial role in cell-cell communication.
Detecting glycoRNAs on exosome surfaces presents multiple technical challenges. Because of their nanoscopic dimensions, conventional optical methods cannot directly resolve surface molecular features. Moreover, existing glycoRNA detection technologies suffer from significant limitations. Metabolic chemical reporters (MCRs), such as N-azidoacetylmannosamine tetraacylated (Ac4ManNAz), rely on the incorporation of azido groups into sialic acid during biosynthesis in metabolically active cells, followed by copper-free click chemistry labeling. This approach is inapplicable to clinical samples or biological materials that cannot be metabolically engineered. Proximity ligation assays based on sialic acid aptamers and RNA in situ hybridization (e.g., ARPLA) offer high specificity but require prior knowledge of glycoRNA sequences, precluding their use for screening unknown glycoRNAs. Although mass spectrometry coupled with liquid chromatography (LC-MS) or capillary electrophoresis (CE-MS) can resolve Glycan Structures, these methods lack the specificity required to directly assign glycan structures to RNA scaffolds.
To circumvent these limitations, this study established a novel labeling strategy based on periodate oxidation and oxime ligation (OL). This method exploits the reactivity of the 7′-, 8′-diols in sialic acid, which undergo rapid oxidation to aldehydes under physiological or mildly acidic conditions in the presence of periodate. These surface aldehydes were subsequently conjugated to aminooxy-functionalized 640R molecules, forming stable oxime bonds. Concurrently, the authors employed the membrane-impermeant thiazole orange homodimer (TOTO-1) to label double-stranded RNA (dsRNA), together with the membrane-permeant SYTO RNASelect Green nucleic acid stain for RNA localization. This dual-labeling strategy bypasses the need for MCRs and enables direct, simultaneous visualization of surface glycans and RNA on isolated exosomes.
Methodological validation confirmed that periodate oxidation did not alter exosome morphology, surface charge, or size distribution. To assess RNA integrity following oxidation, the authors designed single-stranded and double-stranded Oligonucleotides with identical base sequences. High-resolution automated electrophoresis demonstrated that while single-stranded RNA (ssRNA) was substantially degraded, dsRNA remained intact. This finding indicates that the TOTO-1 fluorescence detected after periodate treatment can be attributed to dsRNA, supporting the notion that pre-miRNA is abundant on the exosome surface.
Quantitative bead flow cytometry analysis revealed that 15–35% of double-stranded nucleic acids, presumably pre-miRNAs, are localized on the exosome surface rather than encapsulated within the vesicle lumen. DNase I treatment caused only a marginal drop in TOTO-1 fluorescence, confirming that the exosome surface is predominantly composed of dsRNA rather than DNA. Using the dual-labeling strategy with aminooxy-640R and TOTO-1, further quantification showed that approximately 30% of the surface RNA was glycosylated.
To determine whether the RNA-glycan association involves noncovalent interactions, the vesicles were pretreated with guanidine hydrochloride (GdnHCl), a chaotropic agent that disrupts hydrogen bonds, ionic interactions, and hydrophobic associations without cleaving covalent bonds. The persistence of the dual-label signal following GdnHCl treatment strongly suggested that RNA and glycans are linked by covalent bonds rather than simple physical adsorption. Subsequently, ribonuclease A (RNase A) treatment provided more direct evidence. RNase A selectively cleaves the phosphodiester bonds in ssRNA adjacent to pyrimidine residues. Following RNase A treatment, the glycoRNA signal was abrogated by approximately 75%, whereas the free glycan signal remained largely unaffected. This result indicates that the glycans on the exosome surface are primarily linked to single-stranded regions or the loop structures of hairpin RNAs. Furthermore, proteinase K treatment failed to eliminate the glycoRNA signal, ruling out the possibility that glycans modify RNA through a protein bridge.
To precisely localize glycoRNAs spatially, the authors combined direct stochastic optical reconstruction microscopy (dSTORM) with tetraspanin markers (CD9, CD63, CD81). Super-resolution imaging revealed that glycan signals localized to the periphery of the exosomal membrane, distal from the core tetraspanin signals, confirming that glycoRNAs are indeed anchored on the exosomal surface.
miR-21 is a key regulator of macrophage biology and the inflammatory process. Previous work demonstrated that exosome-borne miR-21 originating from wound-site keratinocytes is delivered to wound macrophages and promotes inflammation resolution. However, an important technical detail often overlooked is that commercially available miR-21 primers used for reverse transcription quantitative PCR (RT-qPCR) also amplify the miR-21a-5p region of pre-miR-21. Consequently, most prior studies have not conclusively ruled out the presence of pre-miR-21, potentially confounding interpretations of mature miR-21 in exosomes. Sequence analysis revealed that pre-miR-21 contains an EXOmotif sequence within its stem-loop structure, whereas mature miR-21 lacks this motif. The presence of EXOmotif justifies exosomal loading, and pre-miR-21 is more stable as cargo than its mature counterpart.
Given that pre-miR-21 is less than 200 bp in size—consistent with the observation that cell surface glycoRNAs fractionate exclusively with small RNA populations—and that its hairpin structure represents a potential glycosylation substrate, the authors hypothesized that pre-miR-21 on the exosome surface may be glycosylated for effective cell-cell communication. To test this, they designed a molecular beacon (MB) complementary to the hairpin region of pre-miR-21. The MB was labeled with Cy5 at its 5′-end and hybridized to a partially complementary quencher oligonucleotide carrying a Black Hole Quencher 3 (BHQ3) at its 3′-end. This design ensured high selectivity for pre-miR-21 over mature miR-21.
Fig. 1 Glycosylated pre-miR-21 on the exosomal surface. (Yadav, et al. 2026)
Super-resolution dSTORM imaging revealed that glycan signals and MB signals were in close proximity, both localized distal to the core tetraspanin markers. This spatial distribution pattern directly demonstrates that pre-miR-21 exists on the exosome surface in a glycosylated form. The membrane-impermeable nature of the MB was validated using bead flow cytometry and XmiR-21-5p transfection experiments: no shift in MB fluorescence was observed in exosomes with encapsulated miR-21-5p, confirming that the MB cannot access luminal RNA. Additionally, RNase A treatment did not degrade the encapsulated miR-21-5p transcript, demonstrating that RNase A activity in this study was strictly a surface phenomenon and that the detected glycoRNA signal originates exclusively from the exosomal surface.
Previous work demonstrated that Exoκ exhibit functional impairments in diabetes, with surface modifications that reduce their uptake by wound macrophages. To investigate glycoRNA distribution under diabetic conditions, Exoκ were isolated from day 2 (D2) wound-edge tissue of diabetic db/db mice and their non-diabetic heterozygous littermate m+/db mice.
Phenotypic analysis showed that Exoκ production was significantly reduced in diabetic mice. Nanoparticle Tracking Analysis (NTA) revealed a multimodal size distribution in diabetic Exoκ, indicating probable aggregation. The ζ-potential of diabetic Exoκ was significantly higher than that of non-diabetic Exoκ, consistent with observations in human diabetic patients with chronic wounds. Quantitative bead flow cytometry using the OL strategy and TOTO-1 revealed that glycoRNAs accounted for approximately 62% of surface signals in non-diabetic Exoκ but only 7.2% in diabetic Exoκ. Molecular beacon detection further confirmed that glycosylated pre-miR-21 was highly enriched on non-diabetic Exoκ but almost absent on diabetic Exoκ.
This finding provides a new perspective for understanding persistent inflammation in diabetic wounds. The absence of surface glycoRNAs impairs recognition and internalization of diabetic Exoκ by macrophages, interrupting anti-inflammatory signaling and ultimately contributing to wound chronicity.
To directly assess the function of surface glycoRNAs in cellular uptake, Exoκ were treated with peptide-N-glycosidase F (PNGase F) or RNase A, fluorescently labeled with ExoGlow membrane dye, and incubated with murine macrophages. PNGase F cleaves N-linked Glycans between the innermost N-acetylglucosamine (GlcNAc) and asparagine residues, removing high-mannose, hybrid, and complex N-glycans while preserving vesicle integrity. Live-cell confocal microscopy revealed compromised uptake of PNGase F-treated exosomes compared to untreated controls. Similarly, RNase A-treated exosomes showed impaired cellular entry. Z-stack analysis confirmed that untreated exosomes were internalized to an intracellular depth of approximately 9 μm, whereas enzyme-treated exosomes remained primarily on the cell surface. These results demonstrate that the absence of glycoRNAs blocks endocytosis rather than merely affecting surface adhesion.
Collectively, these functional experiments establish that surface glycoRNAs are key molecular determinants mediating efficient macrophage endocytosis, likely operating through glycan-receptor interactions and/or charge-mediated mechanisms.
A recent study demonstrated that a distinct subset of mammalian small RNAs undergoes N-glycosylation at the hyper-modified base 3-(3-amino-3-carboxypropyl)uridine (acp3U), generating surface-exposed glycoRNAs that preserve immune homeostasis. These glycoRNAs traverse endosomal pathways and appear on the plasma membrane without activating innate immune sensors. Because exosomes also originate from endosomes, these observations align with the current findings and support the notion that endosome-derived exosomes incorporate glycoRNAs.
The enzyme DTWD2 catalyzes the conversion of uridine to acp3U, which serves as the covalent attachment site for N-glycans. Silencing DTWD2 in human keratinocytes using a double nickase plasmid completely abrogated the glycan-MB dual-labeling signal on exosomes. This result indicates that exosomal glycoRNAs achieve N-glycosylation through a DTWD2-dependent acp3U modification.
Fig. 2 DTWD2-mediated acp3U modification defines the structural basis of exosomal glycoRNAs. (Yadav, et al. 2026)
From an immunological perspective, the glycan moiety acts as a structural shield that masks the immunostimulatory potential of the acp3U-modified RNA backbone. Removal of N-glycans by PNGase F unmasks the RNA, converting these molecules into TLR3/7 agonists and activating innate immune responses. This mechanism explains why diabetic Exoκ, which lack glycoRNAs, fail to deliver anti-inflammatory signals: impaired macrophage uptake prevents cargo delivery, and any RNA that does enter the cell in the absence of glycan shielding may trigger aberrant immune activation.
The authors also examined the effect of MCPIP1 (Zc3h12a), an endoribonuclease that cleaves the terminal loops of pre-miRNAs. MCPIP1 treatment did not eliminate the glycoRNA signal, suggesting that the RNA duplex on the exosome surface may be embedded within the membrane or bound to other biomolecules, thereby maintaining the stability of the glycosylation complex.
The periodate oxidation-oxime ligation (OL) strategy established in this study overcomes the technical limitations of metabolic labeling, providing an efficient, sensitive, and universal tool for detecting glycoRNAs of unknown sequence without requiring live-cell pretreatment. This method not only confirms the existence of covalently linked glycoRNAs on the exosomal surface but also reveals the crucial role of glycosylated pre-miR-21 in wound inflammation resolution.
From a translational medicine perspective, the profound deficiency of glycoRNAs in diabetic chronic wounds identifies a clear therapeutic target. Restoring or engineering the glycoRNA composition on exosome surfaces holds promise for rebuilding the communication bridge between keratinocytes and macrophages, thereby promoting healing in diabetic wounds.
Future research should expand this labeling technology beyond sialic acid to systematically investigate glycoRNAs modified with other monosaccharides. Integrating advanced super-resolution imaging with next-generation sequencing to generate dynamic glycoRNA maps under diverse physiological and pathological conditions will greatly advance the convergence of glycobiology and RNA biology, opening new dimensions for deciphering the molecular code of intercellular communication.
Reference
