GLORI-seq Optimization: Five Key Technical Steps Involved

With the unique principle of enzymatic transformation, GLORI-seq technology has realized the detection of m6A with single-base resolution, which is significantly superior to the traditional antibody-dependent method. However, the experimental process of this technology involves many links, such as sample processing, enzymatic reaction, library construction, quality control, and data analysis. The deviation of any step may lead to false positive results or signal loss, which seriously affects the reliability of data.

At present, the optimization of GLORI-seq still faces many challenges, such as the difference in transformation efficiency caused by the fluctuation of ALKBH5 enzyme activity, the lack of detection sensitivity of low-abundance RNA samples, the introduction of bias in the construction of the sequencing library, and the threshold setting of high-confidence sites in data analysis. These problems limit the wide application of technology in complex samples (such as clinical tissues and individual cells), and also restrict the comparability of results between different laboratories.

This paper focuses on five key optimization steps of GLORI-seq technology, and systematically expounds the core parameters and standardization scheme from sample preparation to data analysis. By clarifying the key points of experimental design, optimizing the enzymatic reaction conditions, controlling the deviation of library construction, selecting quality evaluation indicators and statistical models, we can provide operable technical guidance for researchers, aiming at improving the accuracy and repeatability of m6A detection and promoting the standardized application of GLORI-seq in basic research and clinical transformation.

Experimental Design and Sample Preparation

The reliability of GLORI-seq technology begins with rigorous experimental design and a standardized sample preparation process, and its core goal is to reduce technical deviation and ensure the repeatability of results. At this stage, we should focus on three key elements: RNA input quantity and quality control, sample type-specific treatment, and control setting.

Optimization of RNA Input Quantity and Quality

The selection of RNA input should balance the detection sensitivity and experimental cost. The research shows that:

• For cell line samples, 1-5μg total RNA can meet the needs of routine detection.
• Tissue samples contain more impurities (such as protein and polysaccharide), so it is suggested to increase the input to 5-10 μg.
• Low-abundance RNA samples (such as exosome RNA and unicellular RNA) should be enriched before amplification, and the initial amount should be controlled in the range of 100ng-1μg.
• RNA amplification kits (such as SMART-seq v4) should be used to avoid information loss.

The quality of RNA directly affects the efficiency of the enzymatic reaction. The RIN value (RNA integrity score) should be detected by Agilent Bioanalyzer:

• The RIN value of cell lines and fresh tissue samples should be ≥8.0, and that of frozen tissues should be ≥ 7.0.
• If the sample is partially degraded (RIN 6.0-7.0), it is necessary to shorten the fragmentation time and increase the enzyme dosage to compensate for the loss of activity.

In addition, the ratio of A260/A280 (which should be between 1.8 and 2.1) should be detected by an ultraviolet spectrophotometer to eliminate the inhibition of protein pollution on the activity of the ALKBH5 enzyme.

Specific Treatment Strategy for the Sample Type

There are significant differences in the pretreatment processes of different sample types:

• Cell line samples: TRIzol method is used for direct lysis and extraction, without additional impurity removal step, but the accuracy of cell counting should be ensured (1× 10^6-5× 10^6 cells/sample is recommended).
• Solid tissue sample: It needs to be ground into powder by liquid nitrogen first, and then treated with pyrolysis liquid containing β-mercaptoethanol to destroy the protein junction between tissues.
• Low-abundance samples: If circulating RNA is used, the poly (A)+RNA should be enriched by magnetic beads (such as Thermo Fisher Dynabeads), and RNase inhibitors should be added to prevent degradation.

Contrast Setting

• Biological Repetition: Set up at least 3 independent repeated experiments to distinguish biological variation from technical error. Statistical analysis showed that three repetitions could reduce the false positive rate of the m6A locus to less than 5%.
• Enzyme-free control group: the sample was treated with the same buffer but did not contain ALKBH5 to correct the background of spontaneous demethylation.
• Blank control group: only containing reaction buffer, used to eliminate reagent pollution.

These controls can be verified by PCA analysis: the treatment group and the control group should show obvious clustering separation, while the repeated samples should be closely clustered.

Overview of m6A writers, erasers, and readers (Zhang et al., 2024)

Core Enzymatic Reaction: Optimization and Validation

The demethylation of m6A catalyzed by ALKBH5 is the core step of GLORI-seq, and its efficiency directly determines the detection sensitivity. The optimization of the reaction system should focus on three variables: enzyme concentration, reaction time, and buffer conditions, and ensure the completeness of the reaction through multi-dimensional verification.

Enzyme Concentration and Reaction Time

• ALKBH5 concentration: It is recommended that the initial concentration be 0.5-2μM, and the concentration gradient should be set in increments of 0.5 μM. The experiment shows that more than 90% of m6A sites can be transformed at 37℃ by 1μM ALKBH5, while the high concentration (> 2μM) will lead to the degradation of non-specific RNA.
• Reaction time: According to the length of the RNA fragment, it takes 30 minutes for a 100-200nt fragment and 60 minutes for a longer fragment (> 300nt). The kinetic curves were drawn by time gradient experiments (15min, 30min, 60min, 90min), and the minimum time point (usually 30-60 min) when the reaction efficiency reached the plateau period was selected.

Optimization of Buffer and Auxiliary Factors

• A. Basic buffer system: 20mM Tris-HCl (pH 7.5) and 100mM NaCl are the most suitable reaction environment for ALKBH5. Key cofactors include:
  • a) Fe2+: It is added in the form of (NH4) 2Fe (SO4) 2, with a final concentration of 0.2mM, which needs fresh preparation to avoid oxidation.
  • b) α-ketoglutaric acid: The final concentration is 1mM, and it participates in the oxidation reaction as a co-substrate.
  • c) Ascorbic acid: The final concentration is 2mM, which keeps the reduction state of Fe²+ and prevents the enzyme active center from being inactivated.
• B. The pH of the buffer should be strictly controlled in the range of 7.2-7.8, and deviation from this range will lead to a decrease of enzyme activity by more than 50%. The stability of the enzyme can be improved by adding 10% glycerol, which is especially suitable for low-temperature and long-time reactions.

Verification Method of Reaction Completeness

• Mass spectrometry verification: LC-MS/MS was used to detect the change in the ratio of m6A to A before and after the reaction. In the sample with complete reaction, the ratio of m6A/A should be reduced to less than 10% of the initial value. Typical detection conditions are C18 column (2.1×150 mm), mobile phase of 5mM ammonium acetate-methanol (95:5), flow rate of 0.2ml/min.
• Dot blot verification: Use an anti-m6A antibody (such as Synaptic Systems 202003) to detect the signal changes before and after the reaction. The signal intensity of the completely reacted sample should be reduced by more than 80% compared with that of the untreated group, and there is no obvious RNA degradation (the consistency of RNA sample size was verified by ponceau staining).
• In addition, qPCR can be used to detect the transformation efficiency of known high-abundance m6A sites (such as the 3'UTR of the ACTB gene), which can help to verify the completeness of the reaction.

Partial illustration of the association between m 6 A methylation and inflammatory responses in metabolic syndromes (Ye et al., 2024)

Library Construction and High-Throughput Sequencing

The library construction of GLORI-seq needs to be optimized based on the standard RNA-seq process to adapt to the characteristics of RNA after enzyme treatment, and focus on solving the problems of adapter connection deviation and PCR amplification preference.

• A. Adaptive adjustment of RNA-seq library protocol
  • a) Fragmentation strategy: RNA treated by ALKBH5 does not need secondary fragmentation, and it is directly purified with 1.8 times of AMPure XP magnetic beads (recovery range is 100-300nt) to remove unreacted small molecules.
  • b) Terminal repair: Because enzyme treatment may cause a small amount of RNA terminal damage, it is necessary to extend the terminal repair time to 30 minutes (the standard process is 15 minutes), and use a mixed enzyme system of T4 polynucleotidase and T4 DNA ligase (such as NEB Next End Repair Module).
  • c) Reverse transcription: Random hexamer primers (instead of oligo (dT)) were used to avoid 3'-terminal preference. It is recommended to use SuperScript IV reverse transcriptase and react at 50℃ for 60 minutes to improve the efficiency of cDNA synthesis.
• B. Deviation control strategy
  • a) Adapter connection: A Y-shaped connector (such as Illumina TruSeq) is adopted, and the connection temperature is lowered to 16℃, and the reaction time is extended to 16 hours, so that the connector connection is more uniform. After connection, 0.8 times the volume of magnetic beads was used for purification to reduce the residue of free connectors.
  • b) PCR amplification: High-fidelity DNA polymerase (such as KAPA HiFi) is used, and the number of cycles is controlled at 12-15 (15-18 for conventional RNA-seq) to reduce the amplification bias. Every 25μl of the reaction system, 1 μl of 50 μm dUTP was added to prepare for the subsequent molecular tag redo.
• C. Sequencing parameter recommendation
  • a) Sequencing depth: For human genome samples, 30-50M paired-end reads per sample is recommended to ensure that the average coverage of each m6A locus is ≥20×. Low-abundance samples can be increased to 60-80M reads.
  • b) Selection of reading length: 150bp paired-end is superior to single-ended sequencing, which can improve the genome alignment rate (especially in repeated sequence regions). For small RNA samples (such as miRNA), PE50 reading length can be used to balance efficiency and cost.
  • c) Sequencing platform: Illumina NovaSeq 6000 is suitable for large-scale samples, while NextSeq 2000 is suitable for small and medium-sized experiments, and the base quality Q30 of both should be kept above 90%.

Functional characterization of the most pan-cancer-associated m6 A methylome (Xia et al., 2024)

Key Controls and Quality Assessment Metrics

A strict quality control system is the guarantee of the reliability of GLORI-seq results, and the stability of the whole experiment process needs to be evaluated through multi-level control settings and quantitative indicators.

• A. Core control system
  • a) Mock treatment control: the sample was treated with inactivated alkbh5 (heated at 65℃ for 30 minutes) to evaluate the specificity of the enzyme. If the control showed an obvious demethylation signal (> 10%), it suggested that there was interference of a non-enzymatic reaction.
  • b) Spike-in standard: Add exogenous RNA with a known m6A site (such as a synthetic GFP mRNA fragment containing m6A), and its concentration should be 0.1-1% of the total RNA. By detecting the transformation efficiency of spike-in (should be > 90%), the differences between samples can be calibrated.
  • c) Negative control gene: A gene (for example, HIST1H1D) without m6A modification was selected as the internal reference, and its sequencing signal between the treatment group and the control group should have no significant difference (Fold change <1.2).
• B. Quantitative evaluation of conversion efficiency
  • a) Global demethylation rate: The m6A/A ratio of the treatment group and the control group was calculated by LC-MS/MS, and the conversion efficiency was =(1-treatment group ratio/control group ratio) ×100%. The global transformation efficiency of qualified samples should be ≥85%.
  • b) Site-specific transformation rate: For each candidate m6A site, the ratio of untransformed reads (still m6A) to total reads in the treatment group was calculated. The transformation rate of high confidence loci should be > 70%, and the coefficient of variation between biological repeats should be < 15%.
  • c) Visual evaluation: In the IGV genome browser, the m6A peak signal of the treatment group should be reduced by more than 70% compared with the control group, and the peak shape should be consistent (no obvious shift).
• C. Frequently asked questions
  • a) The transformation rate is less than 60%, and the possible reasons include:
    • i. Insufficient enzyme activity: Re-measure the enzyme concentration by the BCA method to ensure that the activity unit is enough.
    • ii. Lack of cofactor: Check whether α-ketoglutarate is freshly prepared (it needs to be packaged and stored at-20℃).
    • iii. RNA secondary structure hindrance: The reaction temperature can be increased to 42℃, or 10% DMSO can be added to help dissolve.
  • b) There are also some technological problems researchers encounter:
    • i. Linker dimer: The main peak of the library (which should be 300-500bp) is detected by Agilent Bioanalyzer. If there is a linker peak of about 120bp, it is necessary to increase the purification times of magnetic beads.
    • ii. PCR repeat: Use molecular tag (Unique Molecular Identifier) deduplication tools (such as molecular tag based tools) to remove repeated reads introduced by amplification (repetition rate should be less than 20%).

Schematic outline of the two generations of m1A sequencing methods (Zhang et al., 2018)

From Raw Reads to High-Confidence m6A Sites

The data analysis of GLORI-seq should give consideration to the standardization of sequencing data and the accurate identification of m6A sites, and reveal the biological significance of modified sites through a statistical model and functional annotation.

• A. Pretreatment and comparison of original data
  • a) Quality filtration: Trimmomatic is used to remove low-quality sequences and joint pollution.
  • b) Genome alignment: STAR alignment software (version 2.7.10a) was adopted, and unique alignment reads were reserved in parameter setting, and GRCh38 (human) or GRCm39 (mouse) was selected as the reference genome. After comparison, PCR duplication (reads marked as duplicate) should be removed by Samtools.
  • c) Data standardization: The TPM (Transcripts Per Million) method is used to correct the sequencing depth of different samples to ensure comparability between samples.
• B. Statistical identification model of the m6A locus
  • a) Detection of non-transformation sites: A binomial test was used to evaluate the non-transformation probability of each site. The null hypothesis is "there is no difference in base distribution between the treatment group and the control group". By calculating the P value and carrying out FDR correction (Benjamini-Hochberg method), the sites with FDR less than 0.01 and an untransformed reads ratio greater than 20% are selected as candidate m6A sites.
  • b) Enhanced model: Combined with the DESeq2 algorithm to control the variation between samples, the difference analysis of biological repeated samples was carried out, and the non-transformation sites significantly enriched in the treatment group were identified.
  • c) Batch analysis tool: It is recommended to use the dedicated analysis pipeline of GLORI-seq (such as GLORI-pipeline on GitHub), which integrates the whole process from comparison to calling, and provides visual reports.
• C. Visualization and downstream function analysis
  • a) Genome browser display: The data of the m6A locus is converted into bed format, and displayed in IGV or UCSC Genome Browser superimposed with gene structures (exon, intron, UTR) to observe the regional preference of modified sites (for example, m6A of human mRNA is mostly concentrated near 3'UTR and stop codon).
  • b) Motif analysis: Use HOMER software to predict the motifs of 50nt sequences upstream and downstream of the m6A site. The typical m6A motif is RRACH (R=A/G, H=A/C/U). If the frequency of this motif in the sample is more than 3 times the background value, the detection result is reliable.
  • c) Functional enrichment analysis: The gene containing m6A modification was introduced into the ClusterProfiler tool, and enrichment analysis of GO (Gene Ontology) and KEGG pathway was conducted to reveal the biological processes (such as RNA metabolism, cell cycle regulation, etc.) regulated by m6A.

Analyses of m6A methylomes of multiple cell lines (An et al., 2020)

Conclusion

The experimental success of GLORI-seq technology depends on the optimization of the whole process from sample preparation to data analysis. By strictly controlling RNA quality, optimizing the enzymatic reaction conditions of ALKBH5, and setting up a multi-level control system, the detection of m6A modification with single-base resolution can be realized.

In the data analysis stage, it is necessary to combine a statistical model with functional annotation, so as to mine the highly confident m6A locus and its biological significance from the sequencing data. The standardized experimental scheme and analysis process will make GLORI-seq a new generation of epigenomics research tool after MeRIP-seq, and provide more accurate technical support for revealing the role of RNA modification in physiological and pathological processes.

References

1. Zhang J, Tong L, Liu Y, et al. "The regulatory role of m6A modification in the maintenance and differentiation of embryonic stem cells." Genes Dis. 2023 11(5): 101199.
2. Ye D, Zhang Y, Zhang B, Liu J, Wei T, Lu S. "The regulatory role of m6A methylation modification in metabolic syndrome pathogenesis and progression." Front Physiol. 2024 15: 1271874.
3. Xia R, Yin XY, Huang JM., et al. "Interpretable deep cross networks unveiled common signatures of dysregulated epitranscriptomes across 12 cancer types." Molecular Therapy: Nucleic Acids. 2024 12: 35.
4. Chi Zhang, Guifang Jia. "Reversible RNA Modification N1-Methyladenosine (m1A) in mRNA and tRNA." Genomics, Proteomics & Bioinformatics. 2018 16: 3.
5. An S, Huang W, Huang X, et al. "Integrative network analysis identifies cell-specific trans regulators of m6A." Nucleic Acids Res. 2020 48(4): 1715-1729.