How emerging epitranscriptomic modifications are reshaping our understanding of miRNA biology.

Small RNAs used to look chemically simple: just a short strand with a 5′ phosphate, an unmodified sugar-phosphate backbone, and a 3′ hydroxyl. But in the past few years, newer techniques (mass spectrometry, immunoprecipitation, and direct RNA sequencing) have shown that many microRNAs carry extra chemical modifications, or "tags,". These tags modify the properties of each miRNA molecule after they are transcribed, adding an extra layer of complexity to miRNA regulation.

The following covalent modifications are confirmed on miRNAs. Each have unique biological role and can complicate small RNA-seq library prep.

3′-terminal 2′-O-methylation (2’-OMe)

Plants, piRNAs (piwi-interacting), and a growing list of mammalian miRNAs acquire a methyl group on the 2′-OH of the terminal ribose, usually catalysed by HEN1/HENMT1 homologues. The modification shields the 3′ end from attack by nucleotidyltransferases and exonucleases, stabilising the guide strand in vivo. Unfortunately, the same modifications interfere with the T4 RNA ligase-mediated 3′-adapter ligation used in most commercially available small RNA library prep kits, leading to underrepresentation of 2′-OMe species in conventional libraries¹. Randomised-splint ligation and in-vitro demethylation protocols (e.g., periodate oxidation plus β-elimination) recover many of these reads, but at the cost of extra hands-on time and potential non-specific products².

Terminal Uridylation and Adenylation

Terminal U- and A-adding enzymes (TUT4/7 and TENT2, respectively) remodel mature miRNAs by lengthening their 3′ tails. Uridylation often marks miRNAs for decay, whereas adenylation can either stabilise or destabilise depending on sequence context^3,4. From a library prep perspective, tailing increases insert heterogeneity. Longer species migrate into size fractions that can be discarded, and internal homopolymer runs boost adapter dimer formation. Increasing the upper gel cut-off to 40 nt (for those library preps that require gel excision) and using blocking oligos against adapter self-ligation which are included in the NEXTFLEX Small RNA sequencing kit v4 improves capture of tailed isomiRs.

Internal N⁶-methyladenosine (m6A)

Once thought exclusive to mRNAs, m6A is now mapped on dozens of human miRNAs. METTL3/14 deposits the tag, which weakens AGO2 binding and redirects modified guides into extracellular vesicles, affecting both intracellular repression and biomarker potential⁵. For sequencing, m6A causes modest reverse transcriptase (RT) drop-off and mis-incorporation. Even when using high processivity enzymes, studies report 10–15% under-calling of heavily methylated miRNAs.

5-methylcytosine (m5C)

NSUN2-mediated m5C modifications have been detected on tumour-associated miRNAs where they influence strand selection and decay⁶. m5C is RT-permissive but subtly increases C→T mis-calls, inflating isomiR counts unless SNP filters are applied. Studies using RNA with site-specific m5C have found that the RT used on the NEXTFLEX Small RNA sequencing kit v4 maintains fidelity, with error rates comparable to unmodified cytosine contexts. Bisulfite-based small-RNA-seq protocols can validate true m5C sites, although they lower complexity in low input libraries.

N¹-methyladenosine (m1A)

Less common than m6A, m1A disrupts Watson-Crick pairing and stalls standard RTs, producing truncated cDNAs that are lost during size selection⁷. Protocols that tether a thermostable group-II intron RT directly to the 3′ adapter recover these species with near-complete efficiency, while providing a built-in signature (mismatch + drop-off) to call the modification.

Internal N⁷-methylguanosine (m7G)

The METTL1/WDR4 complex introduces m7G inside G-rich loops of select pri-miRNAs, famously enhancing let-7 processing by destabilising an inhibitory stem⁸. m7G does not impede ligation yet induces RT stops like m1A. Using highly processive, thermostable RTs like the one used by NEXTFLEX Small RNA seq kit v4 and/or lowering Mg²⁺ can cut premature stops by half. Pseudouridine

Pseudouridine

High-resolution sequencing has now confirmed pseudouridine on both plant and mammalian miRNAs, where it correlates with germline transmission of epigenetic memory⁹. Pseudoridine slightly increases base-stacking, causing RT slippage rather and incorporation of incorrect nucleotides.

Oxidative adducts: 8-oxoGuanine (8-oxoG)

Reactive oxygen species convert guanine to 8-oxoG within miRNAs under stress. The lesion weakens AGO binding and skews target recognition. In libraries, 8-oxoG drives G→T mis-calls. A recently described biotinylation pull-down protocol both enriches and flags oxidised reads, but any oxidative signature still requires stringent duplicate collapsing to avoid artefacts¹⁰.

Atypical 5′ caps and phosphates

Beyond the canonical 5′ monophosphate, miRNAs have been found carrying NAD or FAD caps and, in antiviral settings, triphosphates. These modifications completely block T4 RNA ligase 1. Pre-treating RNA with recombinant decapping enzymes (NudC for NAD; RppH for triphosphate) restores ligation. When working with viral models, always include a phosphatase/kinase pre-screen to gauge cap diversity before committing to the library kit.

Conclusion

miRNA modifications continue to be characterized at a rapid pace, revealing a growing complexity in small RNA biology. However, these marks present a challenge for existing methods of library preparation, particularly during RT step.

Adopting workflows with robust RT enzymes, such as the NEXTFLEX Small RNA sequencing kit v4 , can help to read more effectively through modified nucleotides with limited bias. Features like reduced RNase H activity and elevated thermal stability allow the enzyme to traverse sites of 2′-O-methylation, internal methylations, pseudouridine, and other covalent tags, minimizing RT drop-off.