Protocol for identifying glycosylated RNAs using solid-phase chemical conjugation and glycosidase release

Summary

RNA glycosylation plays a critical role in cellular signaling. Here, we present a protocol for enriching glycoRNAs from cultured cells and clinical tissues using a solid-phase chemoenzymatic technique. We describe steps for employing galactose oxidase (GAO) to selectively oxidize terminal sugar residues, enabling covalent capture on hydrazide resin, followed by enzymatic release using PNGase F. This scalable and robust assay efficiently enriches N-linked glycoRNAs, facilitating the systematic characterization of disease-specific RNA glycosylation patterns.

Graphical abstract

Highlights

• •Steps for glycoRNA enrichment using GAO-mediated oxidation and solid-phase chemistry
• •Optimized workflow for isolating glycoRNAs from cultured cells and tissue specimens
• •Steps for profiling glycosylated small RNAs by downstream molecular analyses

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

RNA glycosylation plays a critical role in cellular signaling. Here, we present a protocol for enriching glycoRNAs from cultured cells and clinical tissues using a solid-phase chemoenzymatic technique. We describe steps for employing galactose oxidase (GAO) to selectively oxidize terminal sugar residues, enabling covalent capture on hydrazide resin, followed by enzymatic release using PNGase F. This scalable and robust assay efficiently enriches N-linked glycoRNAs, facilitating the systematic characterization of disease-specific RNA glycosylation patterns.

Before you begin

Small RNAs (sRNAs), defined as non-coding transcripts under 200 nucleotides (nt), are important regulators of gene expression, genome integrity, and physiological development across all domains of life.^1,2,3,4,5 By directly targeting over 30% of cellular genes, sRNAs act as critical modulators of disease pathogenesis and serve as valuable clinical biomarkers.^6,7,8 The functional repertoire of both stable and transient RNAs is vastly expanded by over 100 known chemical and post-transcriptional modifications (PTMs).^9,10,11 Dynamic and occasionally reversible modifications—such as methylation (e.g., m⁶A, m⁵C, m¹A), pseudouridylation, and terminal adenylation or uridylation—directly govern sRNA biogenesis, stability, nuclear export, and target repression efficiency.^{12,13,14,15,16,17,18}

Distinct from these canonical PTMs, recent discoveries have identified abundant, glycosylated RNAs (glycoRNAs).^19,20 Specific small non-coding RNAs (sncRNAs) are modified with sialoglycans and are subsequently trafficked to the cell surface, where they directly engage immune receptors.²¹ Subsequent studies have expanded the functional landscape of glycoRNAs. In human monocytes, two glycoRNA populations, termed glycoRNA-L and glycoRNA-S, mediate monocyte–endothelial adhesion through interactions with Siglec-5.²² GlycoRNAs also associate with cell-surface RNA-binding proteins to form nanoscale domains that facilitate the cellular uptake of the TAT cell-penetrating peptide.²³ Furthermore, N-glycosylation masks immunostimulatory acp³ U-containing endogenous RNAs from recognition by TLR3 and TLR7, thereby promoting the non-inflammatory clearance of apoptotic cells.²⁴ In addition, neutrophil surface glycoRNAs bind endothelial P-selectin to promote immune cell recruitment in vivo, a process that is strictly dependent on the RNA transporter SIDT1.²⁵ Collectively, these findings establish glycoRNAs as functional regulators of immune recognition, cell adhesion, extracellular cargo uptake, and immune homeostasis, underscoring the need for robust sequencing approaches to comprehensively define glycoRNA composition across diverse cell types and physiological contexts. To systematically profile this emerging class of modified RNAs, we previously developed a chemoenzymatic approach for the specific enrichment and identification of N-linked glycoRNAs.²⁶ This protocol utilizes galactose oxidase (GAO) to selectively oxidize terminal galactose and N-acetylgalactosamine (GalNAc) residues, enabling stable covalent conjugation to a hydrazide resin under mildly basic conditions while preserving RNA integrity.²⁷ Following capture, N-linked glycoRNAs are released from the solid support via PNGase F cleavage (Figure 1).²⁶

Figure 1.

Procedure for the capture and enzymatic digestion of glycosylated RNAs for N-glycoRNA analysis

(A) Enrichment: Total RNA is first oxidized by galactose oxidase (GAO), which targets both D-galactose and D-GalNAc residues. The resulting aldehyde groups on glycoRNA-associated glycans are covalently coupled to a hydrazide resin, allowing immobilization of glycoRNAs, while non-glycosylated molecules are washed away to achieve enrichment. N-glycoRNAs can then be selectively released from the resin by PNGase F digestion.

(B) Enzymatic digestion: Enriched glycoRNAs can be further processed using specific glycosidases. PNGase F selectively cleaves N-linked glycans from RNA, whereas other glycosidases can be employed to release or further characterize different classes of glycoRNAs.

This highly specific strategy provides several key advantages: it facilitates the direct identification of RNAs bearing Gal, GalNAc, NeuAc-Gal, or NeuAc-GalNAc residues, and it establishes a robust platform for comparative biological studies of glycosylation dynamics. While this assay efficiently enriches total cellular glycoRNAs, distinguishing between cell-surface and intracellular populations requires prior subcellular fractionation. Ultimately, this optimized enrichment methodology provides a critical foundation for building comprehensive glycoRNA databases and elucidating their diverse biological functions.

Innovation

This protocol introduces a novel solid-phase chemoenzymatic framework for the highly specific enrichment and identification of N-linked glycoRNAs. Overcoming key limitations of existing techniques—which often compromise RNA integrity through harsh chemical oxidation or lack subtype specificity—our method employs galactose oxidase (GAO) to achieve mild and selective oxidation of terminal galactose residues. Coupled with hydrazide resin conjugation, this enzymatically driven strategy markedly enhances capture efficiency while minimizing nonspecific background. By integrating enzymatic specificity with solid-phase chemistry, the workflow represents a significant advancement over traditional approaches, allowing efficient enrichment while preserving RNA stability. Consequently, this protocol provides a robust, sensitive, and reproducible platform for accurately decoding the glycoRNA landscape across diverse biological samples.

Institutional permissions

Cancer tissue and matched adjacent paracarcinoma (non-tumor) tissues were obtained with approval from the Ethics Committee of Soochow University and all experimental protocols were approved by the Second Affiliated Hospital of Soochow University.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples
Cancer and paracarcinoma tissues	Cancer and paracarcinoma tissues were obtained with approval from the ethics committee of Soochow University. The protocol was approved by the Second Affiliated Hospital of Soochow University.	N/A
Chemicals, peptides, and recombinant proteins
RNAlater™	Thermo Scientific	cat. no. AM7021
TRIzol™ Reagent	Beyotime Biotechnology	cat. no. R0016
Chloroform	Sinopharm	cat. no. 10006818
Isopropanol (IPA)	Macklin	cat. no. I811927
Ethanol	Macklin	cat. no. E809056
Diethylpyrocarbonate-treated Water (DEPC water)	Beyotime Biotechnology	cat. no. R0022
Dimethyl Sulfoxide (DMSO)	Sinopharm	cat. no. 30072418
Horseradish Peroxidase (HRP)	Meilunbio	cat. no. MB6017
Galactose Oxidase (GAO)	Sigma	cat. no. G6125
UltraLink™ Hydrazide Resin	Thermo Scientific	cat. no. 53149
Ammonium Bicarbonate (NH4HCO3)	Sinopharm	cat. no. 20002760
1 M Tris-HCl, pH 6.8	Beyotime Biotechnology	cat. no. ST768
Peptide-N-Glycosidase F (PNGase F)	New England Biolabs	cat. no. P0704S
α2-3,6,8,9 Neuraminidase A (NEU)	New England Biolabs	cat. No. P0722L
Gel-Red	Beyotime Biotechnology	cat. no. D0139
10× loading buffer	Takara	cat. no. 9157
100 bp DNA Ladder	Abclonal Technology	cat. no. RK30193
Agarose	Abclonal Technology	cat. no. RM02852
RNase, RNA and DNA Remover	Vazyme	cat. no. R504-01
Proteinase K	Sigma	cat. no. 1.24568
StcE	Sigma	cat. no. SAE0202
Experimental models: Cell lines
MIA PaCa 2	PDAC cell line	CRM-CRL-1420™
hTERT-HPNE		CRL-4023™
SW480	ATCC	CCL-228™
Software and algorithms
MSConvert	This paper	http://proteowizard.sourceforge.net/downloads.shtml
Python	This paper
Anaconda 4.2.0	This paper	https://www.continuum.io/downloads
GIG Tool	This paper	http://www.biomarkercenter.org/gigtoo
GlycoWorkBench	This paper	https://code.google.com/archive/p/glycoworkbench/downloads
Trimmomatic	This paper	http://www.usadellab.org/cms/index.php?page=trimmomatic
FastQC (V0.12.1)	This paper	https://www.bioinformatics.babraham.ac.uk/projects/fastqc/fastqc_v0.12.1.zip
bowtie2	This paper	https://bowtie-bio.sourceforge.net/bowtie2/index.shtml
miRDeep2	This paper	https://github.com/rajewsky-lab/mirdeep2/archive
Blast	This paper	https://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/
Other
Microcentrifuge tube(RNase-free, nuclease-free)	Invitrogen	cat. no.AM12480
Pipette tip (RNase-free, nuclease-free)	ABW	AH-LT10B-S for 10 μL; AH-LT1250B-S for 1250 μL
Snap-cap spin column (SCSC)	Thermo Scientific	cat. no. 69725
Mechanical Pipette	Eppendorf AG	cat. no. 3124000121
Hand Grinder	Huxi	cat. no. HMR-1
NanoDrop	Hangzhou Alllsheng Instruments Co., Ltd.	cat. no. Nano-300
Incubator Shaker (constant temperature mixing instrument)	Hangzhou Alllsheng Instruments Co., Ltd.	cat. no. MS-100
Illumina HiSeq 2000 Sequencing System	Illumina, Inc.	Output range: 95-600 Gb, Max read Length: 2x 100 bp
MGI2000 Sequencing System	MGI Tech	Output range: 55 M∼1440 Gb; Effective read count: 300-1800 M
Tube Rotator	Yooning Instrument	cat. no. VM-100
Vacuum Freeze Dryer	Ningbo Scientz Biotechnology	cat. no. SCIENTZ-10

Materials and equipment

Oxidation system (GAO solution)

Reagent	Final concentration	Amount
GAO	1 U	1000 U
DEPC water	N/A	1000 μL
Total	1 U/μL	1000 μL

Note: Thaw the aliquots at 4°C prior to use on the day of the experiment. Aliquot 10 μL of this solution into each 0.5 mL RNase-free, sterile tube. Store at −20°C for up to 6 months.

GAO stock solution (HRP solution)

Reagent	Final concentration	Amount
HRP	8 U	1000 U
DEPC water	N/A	125 μL
Total	8 U/μL	125 μL

Note: Aliquot 10 μL of this solution into each 0.5 mL RNase-free, sterile brown tube. Store aliquots at −20°C for up to 6 months. Avoid using aliquots that have been thawed and refrozen multiple times.

Oxidation buffer

Reagent	Final concentration	Amount
100% DMSO	20%	20 μL
GAO	1 U	1 μL
HRP	8 U	1 μL
RNA	N/A	40 μg
DEPC water	N/A	up to 20μL
Total	N/A	100 μL

Note: To prevent degradation by RNase in distilled water (DI water), DEPC water must be used when preparing this solution, as the reaction time at room temperature is 2 h.

Elution buffer (PH 7.0)

Reagent	Final concentration	Amount
NH4HCO3	20 mM	79.06 mg
DEPC water	N/A	50 ml
Total	20 mM	50 mL

Note: Prepare NH₄HCO₃ buffer freshly prior to use.

Step-by-step method details

CRITICAL: To ensure a contamination-free and RNase-free environment, all procedures must be performed in a clean lab workbench. First, wipe all surfaces with an RNase decontamination solution, followed by 75% ethanol. Finally, UV-irradiate the workspace for 30 min before use.

Note: The rationale for using different RNA extraction procedures for cell culture and tissue samples lies in their inherent differences.²⁸ Cell culture samples, generated under controlled conditions and typically containing fewer contaminants, can be processed using relatively straightforward extraction methods. In contrast, tissue samples present additional challenges, including high levels of RNA-degrading enzymes, variability in collection and storage conditions, and a greater abundance of interfering substances. As a result, tissue RNA extraction protocols require additional precautions—such as stringent temperature control, the use of protective storage buffers, more rigorous purification steps, and comprehensive quality assessment measures—to ensure optimal RNA yield and integrity.

Total RNA can be extracted from either cell lines (steps 1-9) or tissues (steps 10-19). If using tissue specimens, begin directly at step 10.

Total RNA extraction from cultured cells

Timing: 2 h

Here, we describe steps for extracting RNA from cultured cells.

Note: Isolate total RNA using the TRIzol-chloroform method. Perform critical phase separation and rigorous ethanol washes to ensure high purity by effectively eliminating protein and DNA contaminants.

• 1.For cell harvesting and lysis, remove the existing cell culture medium; wash the cells (10 cm²) three times with 1 ml 1× PBS.
• 2.Add 1 mL of TRIzol reagent, and lyse the cells at room temperature (20-25°C) for 5 minutes

Note: Lysis time on ice can be set to 10 min

CRITICAL: Orient the cell flask horizontally to ensure uniform coverage by TRIzol

• 3.Using a cell scraper, detach the adherent cells and transfer the resulting TRIzol cell lysate into a sterile, RNase-free centrifuge tube.

Pause point: Although RNA is protected in TRIzol and can be stored at −80°C for up to 12 months, we recommend completing the steps of cell collection, total RNA extraction, and glycoRNA enrichment on the same day to minimize degradation.

• 4.Add 200 μL of chloroform to the cell lysate and mix thoroughly. Incubate the mixture on ice for 5 min, then centrifuge at 12,000 rpm for 15 min at 4°C to separate the phases and precipitate the RNA.

CRITICAL: Centrifugation separates the mixture into three distinct layers: an upper aqueous phase (containing RNA at pH < 7), a middle interphase, and a lower organic phase (containing proteins and DNA). The aqueous phase should be carefully aspirated along the tube wall using a pipette tip, taking care to avoid disturbing the interphase and organic layer to prevent contamination.

• 5.After aspirating 400 μL of the upper aqueous phase, add an equal volume of isopropanol (IPA), mix thoroughly by pipetting or vortexing, and incubate for 15 minutes at room temperature.

Note: Centrifuge at 12,000 rpm for 15 min at 4°C will pellet the RNA, which can then be collected.

• 6.Wash precipitation: Wash the RNA pellet with 1 ml 75% ethanol three times. Centrifuge at 8,000 rpm for 5 minutes at 4°C after each wash.

CRITICAL: While the tube is inverted, gently flick the side to resuspend the pellet as thoroughly as possible; however, avoid using excessive force, as the resulting mechanical shear could fragment the delicate RNA molecules. Utilizing a controlled touch ensures the pellet is fully integrated into the solution without compromising the structural integrity of the genetic material.

• 7.After inverting the tube (ensure the tube walls are free of moisture), place it on lint-free paper and allow the RNA precipitate to air dry for approximately 5 minutes.

CRITICAL: After drying, the RNA pellet should appear transparent.

• 8.To resuspend the RNA pellet, add approximately 20-50 μL of DEPC water and incubate until the RNA pellet is completely dissolved.
• 9.Measure the RNA concentration (typically 600–1200 ng/μL) and purity using either a Spectrophotometer or a Qubit instrument.

Total RNA extraction from tissues

Timing: 2.5 h

Next, we describe steps for extracting RNA from tissues.

Note: Total RNA is extracted from RNAlater-preserved clinical tissues using TRIzol. To minimize degradation and ensure high purity from complex solid samples, low-temperature homogenization and triple ethanol washes are strictly applied. Cancerous and adjacent noncancerous tissues (e.g., colon and pancreas) are collected from patients who provided written informed consent (Ethics approval no. JD-LK2023001-R01). After collection, tissues are rinsed with 1× PBS prepared using DEPC-treated water to remove residual blood, then promptly transferred into enzyme-free sterile centrifuge tubes containing RNAlater™ at a recommended tissue-to-reagent ratio of 1:20 (v/w). Samples are stored at 25°C for up to 24 h, at 4°C for up to 14 days, or at −20°C for long-term storage (up to 6 months).

CRITICAL: Paired paracancerous tissues should be obtained from mucosal samples located within 3 cm of the tumor margin. Surgically excised specimens must be immediately immersed in RNAlater solution and transferred to −80°C storage within 30 min of removal to preserve RNA integrity.

Pause point: While RNAlater can preserve tissue samples at −80°C for up to a year, storage exceeding 3 months is not recommended.

• 10.To isolate RNA, homogenize 50 mg of RNAlater-stored tissue with 1 mL TRIzol using a manual homogenizer on ice until the tissue is completely homogenized.

Note: Typically require 3–5 min of continuous homogenization, use short bursts with intermittent cooling to prevent overheating and ensure RNA integrity until no visible tissue fragments remain.

• 11.Following homogenization, incubate the mixture at room temperature for 5 minutes to ensure complete lysis.

Pause point: TRIzol can preserve tissue fragments at −80°C for up to 12 months.

• 12.
  After homogenization, add 0.2 mL of chloroform to the TRIzol lysate.

  • a.
    Vortex the mixture thoroughly to ensure complete homogenization.
  • b.
    Incubate at room temperature for 2 to 3 min to allow phase separation.
• 13.Centrifuge the sample at 12,000 rpm for 15 minutes at 4°C to separate the aqueous, organic, and interphase layers.
• 14.After centrifugation, carefully aspirate 300–400 μL of the upper aqueous phase and transfer it to a new tube. Add equal volume of isopropanol (IPA) and gently mix by inversion.

CRITICAL: To prevent contamination and RNA degradation, ensure that pipette tips do not come into contact with the tube walls or the lower liquid surface during transfer.

• 15.
  Incubate the sample for 10 minutes on ice to precipitate the RNA.

  • a.
    Centrifuge at 12,000 rpm for 10 minutes at 4°C.
  • b.
    Following centrifugation, discard the supernatant, as the RNA will be present as a pellet at the bottom of the tube.
• 16.After removal of the supernatant, wash the RNA pellet with 1 mL of 75% ethanol 3 times to remove contaminants.
• 17.Centrifuge at 8,000 rpm for 5 minutes at 4°C. Carefully aspirate the supernatant, leaving approximately 30 μL of solution containing the RNA pellet.
• 18.
  Repeat centrifugation for 1 minute at 8,000 rpm for 5 minutes at 4°C.

  • a.
    Air-dry the RNA pellet until it appears transparent.
  • b.
    Resuspend RNA in 30 μL of DEPC water.

Pause point: For storage, RNA in DEPC water remains stable at −80°C for up to 6 months. Avoid repeated freeze-thaw cycles to preserve RNA integrity.

Optional: To remove residual protein contaminants from samples requiring additional de-proteinization, RNA is subjected to proteinase K (1 mg/mL) digestion at 37°C for 30 min.

CRITICAL: RNA re-extraction using TRIzol after decontamination procedures is likely to significantly reduce recovery efficiency. Strict adherence to the previously outlined protocol effectively minimizes protein contamination, making this additional step generally unnecessary.

Total RNA quality assessment

Timing: 2.5 h

We describe how to assess the RNA purity and integrity.

RNA purity and integrity were validated using spectrophotometry and non-denaturing agarose gel electrophoresis.

CRITICAL: RNA purity is assessed by the ratios of A260/A280 and A260/A230 measured using a spectrophotometer. High-quality RNA sample should have an A260/A280 ratio between 1.8 and 2.1, and A260/A230 ratios no lower than 1.8.

• 19.Evaluate RNA integrity using 1% non-denaturing agarose gel electrophoresis.
• 20.To prepare the gel, dissolve 1 g of agarose in 100 mL of 0.5× TAE buffer by microwaving for 2-3 minutes. Allow the solution to cool slightly before adding 10 μL of Gel-Red

CRITICAL: Gently swirl to prevent the formation of bubbles.

• 21.Pour the mixture into the gel mold with the sample comb inserted. Once solidified, place the gel in the tank and load RNA samples, and then run the electrophoresis.

CRITICAL: Load approximately 500 ng RNA per lane.

• 22.Place the gel in the electrophoresis tank filled with 1× TAE buffer and run at 150 V for 20 minutes.

CRITICAL: For optimal results, decontaminate the electrophoresis tank with RNA decontaminant the night before running the samples. To assess total RNA integrity, visualize the RNA on a non-denaturing agarose gel. Ideally, the intensity of the 28S rRNA band should be approximately twice that of the 18S rRNA band, with both bands significantly brighter than the 5S rRNA band (Table 1; Figure S3).

Table 1.

Comparison of key parameters and characteristics for glycoRNA enrichment from cells and tissues

Parameter	Cell line	Tissue
Quantity	≥ 1 x 106 (one dish, 10 cm2)	∼50 mg
Preparation	Cells often need to be harvested and washed	Tissue needs to be finely minced or homogenized
RNA Protection	Not required	Required (RNAlater)
RNA Yield (μg)	1-10 per dish	0.5–1 per 1 mg
Total RNA Quality	28S >> 18S >> 5S	28S > 18S >> 5S
Initial RNA Amount for glycoRNA (μg)	40–60	40–60
Homogenization	Not necessary	Essential for efficient tissue lysis and RNA release
Processing Time (hour)	Generally faster, 1-2	Longer, 2-4
Sample Complexity	Homogeneous	More heterogeneous with complex matrix
Storage	RNAlater or frozen	RNAlater or snap-frozen
Contaminants	Fewer contaminants (e.g., proteins, lipids)	More potential contaminants, including fibrous material
RNA Integrity	High integrity if processed quickly	Subject to faster degradation without proper handling

Optional: Assess RNA integrity using the Agilent 2200 ScreenTape system. First, prepare the RNA samples by diluting them to 5–500 ng/μL.

• 23.
  Load 1 μL each of the RNA sample and RNA Sample Buffer into the sample plate, along with the RNA Ladder.

  • a.
    Heat the samples at 72°C for 3 minutes to denature them.
  • b.
    Cool on ice immediately to prevent secondary structures.
• 24.Place the sample plate and RNA ScreenTape into the TapeStation instrument for automated electrophoresis and fluorescence detection.
• 25.Upon completion, the software will display an electropherogram and provide an RNA Integrity Number (RIN) for each sample.

CRITICAL: To assess total RNA integrity, carefully examine both the peak shapes on the electropherogram and the RIN values provided by the Agilent 2200 ScreenTape system. For high-quality RNA, the RIN should ideally be greater than 8.

Total RNA oxidation

Timing: 30 min

The following steps describe RNA oxidation.

Note: Total RNA (40 μg) undergoes controlled oxidation using galactose oxidase (GAO) and horseradish peroxidase (HRP) in the dark. HRP is essential to decompose toxic hydrogen peroxide, preserving RNA integrity while targeting specific glycoRNAs.

• 26.
  Total RNA oxidation (Table 1)

  • a.
    (Optional) Treat total RNA with 10 U NEU in DEPC water, and then incubate it at room temperature for 30 minutes.

    Note: Recommend NEU treatment when the target glycoRNAs are expected to carry sialylated glycans or when broader detection of glycoRNAs with diverse glycan structures is desired.

  • b.
    In a light-protected environment, add DMSO oxidation buffer, 1 U GAO, and 8 U HRP sequentially to the 40 μg total RNA sample.

    Note: (Ensure a 20% final DMSO concentration in the solution.), targeting specifically glycoRNAs with galactose (Gal) or N-acetylgalactosamine (GalNAc) moieties.

  • c.
    Vortex the mixture briefly, then incubate the reaction at room temperature for 30 minutes.
  • d.
    Protect the mixture from light during incubation.

    CRITICAL: For optimal glycoRNA enrichment, we recommend using 40 μg of total RNA. Using lower amounts may result in insufficient glycoRNA yield, limiting the sample quantity available for small RNA sequencing. If the enriched glycoRNA yield is still low, the input can be increased to 60 μg of total RNA.

    CRITICAL: For optimal RNA oxidation, it is critical to maintain a 1:8 ratio of galactose oxidase (GAO) to horseradish peroxidase (HRP) in the DMSO-based reaction system. GAO-mediated oxidation generates hydrogen peroxide (H₂O₂), which can cause oxidative RNA damage. To mitigate this, HRP is included to catalytically decompose excess H₂O₂ into water and oxygen (2 H₂O₂ → 2H₂O + O₂). This enzymatic regulation effectively minimizes oxidative stress, thereby preserving RNA integrity during the oxidation process.²⁹

Hydrazide resin preparation

Timing: 10 min

The following steps describe hydrazide resin preparation for conjugation of oxidized glycoRNAs.

Note: Equilibrate UltraLink™ Hydrazide resin to room temperature and homogenized thoroughly before use. Aliquot 50 μL of resin slurry into spin columns and wash twice with DEPC-treated water to remove storage buffer.

• 27.Before using the UltraLink™ Hydrazide resin, allow it to equilibrate to room temperature.
• 28.Thoroughly vortex the resin slurry to ensure homogeneity before pipetting it into the tubes.
• 29.Pipette 50 μL of UltraLink hydrazide resin into each snap-cap spin column (SCSC).
• 30.Briefly vortex the column following the addition of 500 μL of DEPC water, then centrifuge and discard the resulting wash solution. Repeat this process once.

Oxidized RNA immobilization

Timing: 1 h

The following steps describe the conjugation of oxidized glycoRNAs to hydrazide resin.

Note: Incubate oxidized RNA samples with hydrazide resin (50 μL slurry/40 μg RNA) for 1 h at 25°C. This covalent immobilization efficiently captures target glycoRNAs on the solid phase.

• 31.
  Immobilization of oxidized total RNAs

  • a.
    Transfer the oxidized RNA samples to the SCSC using RNase-free pipette tips.
  • b.
    Vortex the sample-resin mixture for 30 s.
  • c.
    Incubate the mixture at 25°C for 1 hour.

CRITICAL: The amount of resin required depends on the sample size. For optimal immobilization, use approximately 50 μL of resin slurry per 40 μg total RNA input.

CRITICAL: Ensure complete transfer of the sample, including any residual material on the pipette tip, rinsing with DEPC water if necessary. The total sample volume should not exceed 500 μL (as the SCSC capacity is 600 μL). If the volume approaches 500 μL, gently vortex and centrifuge the column for 5 s at 2,000 x g to collect all the resin.

Resin washing

Timing: 1 h

The following steps describe washing hydrazide resin.

Note: Wash resin extensively with DEPC water to remove non-specifically bound molecules. Repeat washing steps until the flow-through RNA concentration falls below 0.5 ng/μL, ensuring zero background contamination.

• 32.Briefly vortex the sample-resin mixture after adding the sample to ensure proper mixing.
• 33.Centrifuge the spin column for 5 seconds at 2,000 x g to remove the reaction buffer.
• 34.Wash the column three times with DEPC water to remove non-specifically bound RNA and other molecules.
• 35.Wash the resin six times with DEPC water.

CRITICAL: After the washing steps, add 100 μL of DEPC-treated water to the spin column and centrifuge. Collect the eluate and measure the RNA concentration. If the concentration is below 0.5 ng/μL or no RNA is detected, proceed to the next step. If the RNA concentration exceeds this threshold, perform additional washes with DEPC-treated water to remove any unbound RNA from the resin until the eluate meets the specified criteria.

GlycoRNA release

Timing: 2 h

The following steps describe glycoRNA release from hydrazide resin.

Note: Release N-glycoRNAs using PNGase F in a pH-controlled elution buffer (pH 6.8–8.0). Precise incubation (2 h, 25°C) and agitation ensure optimal cleavage efficiency while minimizing protein contamination.

• 36.N-glycoRNA: To enrich the N-glycoRNA, transfer 80-200 μL (We recommend using 100 μL) of Elution buffer to a SCSC column containing 100 μL of resin and 1 μL of PNGase F.
• 37.Secure the column with the rubber cap provided by Thermo Scientific.
• 38.Gently vortex the mixture, then incubate it at 25°C for 2 h with shaking at 600 rpm.
• 39.After incubation, gently vortex again, centrifuge the column (remove the rubber cap first), and use a fresh tube collect the eluate in RNase-free, nuclease-free tubes.
• 40.Store the eluted RNA at −80°C for future small RNA sequencing.

CRITICAL: For optimal enzyme activity, ensure that the freshly prepared elution buffer has a pH between 6.8 and 7.0. The recommended incubation time is 2 h.

Optional: Following steps 40, to ensure complete elution of glycoRNA, you can collect the eluate and repeat the collection step once.

Note: However, if the RNA concentration is insufficient for RNA-seq analysis, the sample may be concentrated by vacuum freeze-drying (lyophilization). After concentration, reconstitute the sample in a minimal volume (20–50 μL) of RNase-free DEPC-treated water to achieve the desired RNA concentration for downstream analysis.

CRITICAL: For optimal enzyme activity, ensure that the freshly prepared elution buffer has a pH of 6.8–7.0.

CRITICAL: It is crucial to avoid using excessive enzyme. Insufficient enzyme will limit glycoRNA recovery, while excessive enzyme can introduce protein contamination, compromise RNA purity, and potentially prevent the sample from meeting downstream quality requirements.

Comprehensive workflow for small-RNA sequences

Timing: 120 h

The following steps describe the analysis of glycosylated small RNAs.

Note: Small RNA libraries are constructed and sequenced on Illumina platforms (PE150). Following rigorous trimming and quality control, clean reads are mapped to custom databases. Specialized pipelines quantify glycoRNA classes (miRNA, tRNA, snoRNA) with strict statistical significance thresholds.

• 41.Measure the concentration and purity of the eluted RNA using either a spectrophotometer or a Qubit.

CRITICAL: Ensure that the RNA purity is greater than 1.8, which can be verified by an A260/A280 ratio of 1.8 or higher.

• 42.Use Agilent 2100 Bioanalyzer to detect the size distribution of glycoRNAs (20-200 nt).

CRITICAL: When analyzing the peak intensity and length distribution plots, carefully assess whether the main RNA length range falls within the analyzable window. For example, miRNAs typically range from approximately 18 to 50 nt.

• 43.
  Small RNA library preparation and sequencing

  • a.
    RNA input and adaptor ligation

    • i.
      Use 50 ng to 1 μg of purified glycoRNA for library preparation.
    • ii.
      Ligate the 3′ SR Adaptor for Illumina using 3′ ligation enzyme.
    • iii.
      Hybridize any excess adaptor with the SR RT primer to prevent dimer formation.
    • iv.
      Ligate the 5′ SR adaptor using 5′ ligation enzyme.
  • b.
    cDNA synthesis: Generate first-strand cDNA from the adaptor-ligated RNA using ProtoScript II Reverse Transcriptase.
  • c.
    PCR amplification and purification

    • i.
      Amplify cDNA with P5 and P7 primers.
    • ii.
      Purify the PCR products using DNA clean beads.
  • d.
    Size selection and validation

    • i.
      Recover 140–160 bp fragments via PAGE and clean them up.
    • ii.
      Subsequently validate the library size distribution using an Agilent 2100 Bioanalyzer.
  • e.
    Multiplexing: Pool libraries that have distinct indexes to enable multiplexed sequencing.
  • f.
    Sequencing: Perform paired-end sequencing (PE150) using Illumina HiSeq Xten, NovaSeq, or MGI2000 platforms.

Note: The PE150 read length is determined by the final library size after adapter ligation.

• 44.Small non-coding RNA analysis

Note: Current miRNA analysis employs standardized workflows,³⁰ including specialized tools like miRDeep2 and mirTarBase. However, significant technical bottlenecks persist in the bioinformatics analysis of related small non-coding RNA family members, such as snoRNA and tRNA derivatives. Existing tools are primarily designed for individual RNA types, lacking a unified framework for cross-category integration. To address this challenge, our research team has developed an innovative analytical system through novel methodological advancements. Notably, certain modules in this RNA analysis pipeline are presented with differential descriptions, while complete algorithmic implementations and parameter optimization details are provided in the Supplementary Information. Please refer to those steps for small non-coding RNA analysis.

• 45.Install required analysis software: Before starting the analysis, ensure the following software is installed: FastQC, Cutadapt, Trimmomatic, BLAST+, Bowtie, and miRDeep2. Perform the sequencing data analysis on an Ubuntu 20.04.6 system.

CRITICAL: Because software environments can vary across different computers and may lead to technical issues, this protocol does not include standardized installation instructions. Users are encouraged to consult the official websites and GitHub repositories of the relevant software packages for detailed and up-to-date installation guidance.

• 46.Perform quality control on the raw sequencing data using FastQC, resulting in the generation of HTML-formatted reports.

CRITICAL: Special attention is given to evaluating base quality scores and checking for adapter contamination within the small RNA sequencing data.

• 47.
  Perform small RNA sequencing data quality control and preprocessing using Trimmomatic software (version 0.30) to ensure high-quality data for downstream analysis.

  Note: This process removes adapter sequences and low-quality reads, resulting in clean data. The quality control pipeline consists of the following steps:

  • a.
    Adapter sequence removal: Remove any adapter sequences that may have ligated to the small RNA fragments during library preparation.
  • b.
    Read length selection:
    Option A:Retain reads of 17-34 nt as potential miRNA sequences.
    Option B:Retain reads ≥35 nt as potential snoRNA, lncRNA, tRNA, and other sequences.
  • c.
    Low-quality base trimming: Remove bases with a quality score below a defined threshold (e.g., Phred score less than 20) and any ambiguous nucleotides (N) from either the 5′ or 3′ ends of the reads.
  • d.
    Sliding window quality filtering: Scan each read using a sliding window of a defined size (e.g., 4 base pairs). If the average base quality score within this window falls below a set threshold (e.g., less than 20), the entire read is discarded.
  • e.
    Minimum read length filtering: Remove any reads in the dataset that are shorter than a specific length (e.g., 18 base pairs) after the trimming process.

    Optional: Perform a secondary quality control check on the processed data to verify that adapter sequences were correctly removed.

• 48.
  Small non-coding RNA database construction

  • a.
    Download the Rfam database (version 14 or later), which contains a collection of non-coding RNA family sequences.
  • b.
    Download the reference genome for the species of interest (e.g., Homo_sapiens.GRCh38.ncrna.fa for human), which includes known non-coding RNA annotations.
  • c.
    Combine these downloaded resources to create a comprehensive non-coding RNA reference database.
• 49.
  Convert the clean data (FASTQ format) of each sample into FASTA format using scripts.

  • a.
    Ensure that quality information is omitted and the sequence itself is used as the header. (e.g., >AGCT).
  • b.
    Input these FASTA files into “collapse_reads_md.pl” for de-duplication.
  • c.
    Use a three-character sample ID derived from the first two letters of the group name (co/tr) and the biological replicate number (1/2/3).

CRITICAL: This process consolidates all identical reads into a single representative sequence.

• 50.To analyze the distribution of small non-coding RNAs within the obtained small RNA sequencing data, the high-quality, clean data (in FASTA format) align to the comprehensive small non-coding RNA reference database using BLASTN.
• 51.
  Process the ncRNA alignment data using Python to consolidate duplicate sequence expression information.

  • a.
    Import sequence details and calculate read counts.
  • b.
    Use the curated dataset for subsequent steps, including filtering low-abundance small RNAs and performing statistical differential expression analysis.
    Glycosylated snoRNA analysis (Option A).
    The snoDB resource, accessible at https://bioinfo-scottgroup.med.usherbrooke.ca/snoDB/, can be utilized for snoRNA sequence analysis.
    Glycosylated tRNA analysis (Option B).

    Note: tRNAscan-SE 2.0 is a powerful tool for predicting tRNA structure, function, and anticodon isotype. To ensure analytical rigor, differential expression profiling was performed using a three-tiered filtering approach.

  • c.
    Expression threshold: Remove low-abundance transcripts with Counts Per Million (CPM) less than 1 in at least 50% of the samples.
  • d.
    Statistical significance: Only transcripts with Benjamini-Hochberg adjusted False Discovery Rate (FDR) values below 0.05 are considered significant.
  • e.
    Biological relevance: Molecules with an absolute log₂(Fold Change) greater than 1 are selected as biologically relevant.
    Glycosylated miRNA analysis (Option C).
    To filter out non-miRNA contaminating sequences, organize small RNA classification information through the following steps.

    • i.
      Align using BLASTN-short and remove un-annotated sequence names.
    • ii.
      Execute the “find_fasta.py” script to retrieve the original sequences from the FASTA file.

      CRITICAL: This optimization offers two key benefits: it accelerates downstream alignment and reduces the occurrence of false-positive results.

  • f.
    Perform genome alignment using the mapper module within the miRDeep2 software.
  • g.
    Identify and predict both known and novel miRNAs using the miRDeep2 module.
  • h.
    Quantify the expression levels of novel miRNAs using the quantifier module.

    Note: This concludes the upstream analysis. Further downstream analysis can be performed using other statistical software packages.

Expected outcomes

Successful glycoRNA enrichment will first be validated by confirming the high-quality integrity of total RNA extracted from clinically obtained colorectal cancer (CRC) and normal adjacent tissues (NAT) (Ethics Approval: JD-LK2023001-R01) using an Agilent 2200 BioAnalyzer and 1% agarose gel electrophoresis. Post-PNGase F digestion, small RNA integrity is expected to be verified via an Agilent 2100 BioAnalyzer and 2% agarose gel electrophoresis. We anticipate observing consistent, size-based separation of glycoRNAs (20–200 nt) within the 5S rRNA region—encompassing species such as miRNAs, snoRNAs, and tRNAs—across biological triplicates (Figure 2A).

Figure 2.

Comparative analysis of N-glycoRNAs in colorectal cancer (CRC) and normal adjacent tissue (NAT)

(A) Gel electrophoresis of total RNA and N-glycoRNA enriched using SPCgRNA from three NAT and CRC tissue samples, showing consistent glycoRNA length profiles across replicates.

(B) A heatmap illustrating the abundance of different RNA types (miRNA, rRNA, snRNA, snoRNA, tRNA, and others) in NAT and CRC, with a Venn diagram highlighting the significant overlap of 352 miRNAs shared between the two tissue types. Volcano plots depict the differential expression of glycosylated miRNAs, lncRNAs, and snoRNAs, identifying specific miRNAs (miR-21-5p, miR-19a-3p, miR-503-5p, miR-27a-3p, and miR-3458-3p) that show notable changes in CRC compared to NAT.

Downstream sequencing and differential analysis are projected to reveal distinct glycosylation profiles. We expect to identify approximately 352 glycosylated miRNAs shared between NAT and CRC, alongside tissue-specific signatures (e.g., 12 NAT-specific and 22 CRC-specific miRNAs; Figure 2B). Visualized via volcano plots, the differential expression analysis is anticipated to highlight significant dysregulation. Specifically, we expect to observe the upregulation of glycosylated oncogenic drivers like miR-503-5p, miR-21-5p, and miR-27a-3p in CRC, concurrent with the downregulation of potential tumor suppressors such as miR-19a-3p and miR-3158-3p.

These findings are expected to suggest that these modifications dictate direct functional consequences. For instance, the reduced glycosylation of miR-19a-3p may dampen its known tumor-inhibitory regulation of the FOXF2-mediated Wnt/β-catenin pathway,^25,26 whereas the increased glycosylation of the established oncogene miR-21-5p 27 may synergistically drive tumor progression. Ultimately, alongside the anticipated identification of broader modified non-coding transcripts (e.g., lncRNAs and snoRNAs; Tables S3 and S4), these outcomes support the critical role of RNA glycosylation in CRC tumorigenesis and provide a robust foundation for subsequent mechanistic studies.

Limitations

While solid-phase chemoenzymatic methods offer notable advantages, researchers must acknowledge their current limitations. Specifically, these methods selectively target glycans containing Gal, GalNAc, NeuAc-Gal or NeuAc-GalNAc, potentially excluding glycoRNAs with alternative structures such as high-mannose, sialic acid, fucose, or complex O-glycans. Thus, this method cannot enrich glycoRNAs without Gal, GalNAc, NeuAc-Gal or NeuAc-GalNAc, resulting in potential data loss. Future research should focus on developing strategies to overcome this limitation. One possibility lies in engineered GAO variants. While naturally oxidizing Gal residues, recent research using directed evolution has yielded GAO variants³¹ that accept sugars like N-acetylglucosamine (GlcNAc) and mannose present in other glycans and potentially glycoRNAs. Protein engineering has generated variants with activity towards secondary alcohols like glucose (Glc) and fructose (Fru).³² Another potential option is sodium periodate (NaIO₄) for glycoRNA oxidation. Periodate oxidizes cis-diols in terminal sialic acid residues but can also react with the cis-diol of the 3′-terminal ribose of RNA and other susceptible cis-diol-containing sugar residues. Because these substrates exhibit different reactivity, the periodate concentration, reaction time, temperature, and pH must be carefully optimized to maximize glycan labeling while minimizing nonspecific oxidation. Although the broader reactivity of periodate may increase glycan coverage, it can also reduce glycan selectivity and increase the risk of unintended RNA oxidation or degradation. Notably, an RNA-optimized periodate oxidation and aldehyde ligation method, termed rPAL, has been developed for the detection and characterization of native glycoRNAs.³³ More recently, Ge et al.³⁴ reported a periodate-based rPAL-seq workflow that combines capture-and-release chemistry with sequencing for the analysis of sialoglycoRNAs. Together, these studies demonstrate the feasibility of periodate-based glycoRNA detection and sequencing; however, further optimization and validation are needed to achieve an optimal balance among glycan coverage, reaction specificity, and RNA integrity.

An additional limitation arises from the use of conventional 5′- and 3′-adapter ligation during sequencing library preparation. T4 RNA ligase-based workflows preferentially capture RNA molecules with ligation-compatible 5′-phosphate and 3′-hydroxyl termini. Consequently, glycoRNAs bearing alternative terminal chemistries may be inefficiently ligated or excluded from library construction. In addition, RNA sequence, secondary structure, and steric hindrance introduced by glycan conjugation may further influence ligation efficiency, leading to biases in the observed relative abundance of glycoRNAs. Therefore, the sequencing results generated by this workflow should be interpreted as representing the ligation-compatible subset of enriched glycoRNAs rather than the entire glycoRNA population. Future developments in ligation-independent library preparation strategies or terminal-repair-based approaches may improve glycoRNA coverage and reduce these technical biases. The other important consideration is the susceptibility of RNA to degradation. To minimize RNA loss, researchers should exercise extreme caution throughout the glycoRNA analysis process. This includes conducting the entire procedure in an RNase-free environment. It is recommended to flash-freeze the samples in LN2 and store them at −80°C or in LN₂. Additionally, freeze-thaw cycles of both the tissue sample and the extracted RNA should be minimized.

Troubleshooting

Problem 1

Tissue blackening (step 10).

Potential solution

To minimize tissue damage, collect paracancerous and cancerous tissue away from areas affected by ultrasonic scalpel. Alternatively, use dissection tools that don’t cause thermal damage.

Problem 2

Low RNA integrity (step 11).

Potential solution

Grind tissue samples with a hand-held homogenizer for short intervals (e.g., 5 seconds). After each grinding interval, immediately place the homogenizer on ice for 10 seconds to minimize heat generation. Repeat this cycle for the desired total grinding time.

Problem 3

A white and cloudy suspension in the supernatant (step 14).

Potential solution

Increase the volume of chloroform by 0.2 mL. Following thorough mixing, perform repeated centrifugation cycles until a transparent supernatant is obtained.

Problem 4

Low RNA yield (step 18).

Potential solution

Increase the amount of paracancerous tissue collected to improve RNA extraction yield.

Problem 5

Low glycoRNA yield (step 28).

Potential solution

Ensure sufficient HRP is present to support RNA integrity during oxidation.

Problem 6

Low glycoRNA yield (step 32).

Potential solution

Verify GAO enzyme expiration date and storage conditions before use. Replace if compromised.

Problem 7

No glycoRNA (step 37).

Potential solution

Employ RNase-free SCSC, pipette tips, and DEPC water throughout the experiment.

Problem 8

Presence of RNA in flow-through (step 41).

Potential solution

Perform a gentle vortex to ensure thorough mixing before each centrifugation step. Repeat the washing cycles until RNA is no longer detectable in the flow-through.

Problem 9

Shortened glycoRNA length (step 45).

Potential solution

To maximize glycoRNA yield, perform enrichment immediately after total RNA extraction.

Problem 10

Some snoRNA sequences not matched in snoDB (step 53).

Potential solution

Perform a comprehensive quality check of the snoRNA database FASTA file. Eliminate any entries identified as non-snoRNA sequences.

Resource availability

Lead contact

Further information and clarification should be directed to and will be fulfilled by the lead contact, Dr. Shuang Yang (jake.yang@gmail.com).

Technical contact

Technical questions on executing this protocol should be directed to and will be answered by the technical contact, Dr. Xiaodong Yang (yangxiaodong2366@suda.edu.cn).

Materials availability

No unique materials are generated in this study.

Data and code availability

All specific sequencing analysis code is available in the supplemental information file.

Acknowledgments

This work was funded by the Shantou University Medical College (SUMC) Scientific Research Initiation Grant, a start-up fund from the First Affiliated Hospital of SUMC; the Jiangsu Province-Suzhou Science and Technology Planning Project (SL T201917); the Priority Academic Program Development of Jiangsu Higher Education Institutes (PAPD); Jiangsu Science and Technology Plan Funding (BX2022023); Jiangsu Shuangchuang Boshi Funding (JSSCBS20210697); the Suzhou Health Youth Talent Project (GSWS2022087); Suzhou Science and Technology Plan Funding (SYW2024037); the Key Medical Research Project of the Jiangsu Provincial Health Commission (DZ2021045); and the Provincial-level Talent Program for the National Center of Technology Innovation for Biopharmaceuticals (NCTIB2024JS0101).

Author contributions

S.Y. conceptualized and developed SPCgRNA method, designed the experiments, prepared the figures and tables, and drafted the manuscript. J.D., X.M., X.W., and J.L. worked on cell- and tissue-based glycoRNA analysis. X.W. cultured cell lines. J.D., Z.Z., and X.Y. collected tissues and performed data analysis. J.D., X.M., X.W., S.W., and Y.G. conducted experiments. X.W., X.Y., and S.Y. secured funding for the research. All authors approved the final manuscript.

Declaration of interests

The authors declare no competing interests.