Northern Blot: RNA Transfer, Probing and Quantification

The Northern blot detects a specific RNA transcript within a complex RNA sample by size-fractionating it on a denaturing gel, transferring it to a membrane, and finding it with a labelled probe. James Alwine, David Kemp and George Stark described it in 1977, in a paper titled Method for detection of specific RNAs in agarose gels by transfer to diazobenzyloxymethyl-paper and hybridization with DNA probes, published two years after Edwin Southern’s DNA method — the name is a direct pun on his surname, and Western blotting (protein) completed the set later.

It has been largely superseded for routine expression quantification: RT-qPCR handles a handful of known transcripts faster and with far less input RNA, and RNA-seq now profiles the whole transcriptome cheaply enough that it is the default choice even for small gene sets. What Northern blotting still provides, that neither of those does directly, is a size — you see the transcript’s actual length on the same blot as its abundance, which is why it persists for confirming alternative splicing, precursor-to-mature processing, degradation products, or an unusual transcript that a Ct value or a read count alone cannot show you.

The sequence, and why each step exists

Every step exists to solve a problem specific to working with RNA rather than DNA. Skipping one is the most common reason a blot fails to size cleanly.

1. RNA extraction and quality assessment. Northern blotting starts from intact total or poly(A)-selected RNA — see RNA extraction protocol basics for RNase-free technique, and check integrity beforehand with the RNA Integrity Number (RIN) rather than discovering degradation on the blot itself. A degraded input sample produces a smeared lane with band sizes that misrepresent true transcript length, regardless of how carefully the rest of the protocol is run.
2. Denaturing gel electrophoresis. RNA is single-stranded and folds into secondary structure — hairpins and intramolecular base-pairing — that changes how fast a given transcript migrates unless it is kept denatured throughout the run. This is the step with no real Southern-blot equivalent: linear double-stranded DNA fragments run true to size without pre-denaturation, so a Southern blot only denatures after electrophoresis, immediately before transfer. A Northern blot has to denature during electrophoresis itself, because the size read off the gel is only accurate if the RNA never folds. Two chemistries dominate: formaldehyde-MOPS gels, with roughly 2.2 M formaldehyde present in both the gel and the MOPS/EDTA running buffer; and glyoxal-DMSO gels, where RNA is denatured with glyoxal and DMSO before loading and run in a low-ionic-strength buffer (commonly a sodium phosphate or BPTE system) with no formaldehyde in the gel itself. Glyoxal gels require the running buffer to be recirculated during the run, because the glyoxal–RNA adduct is unstable above roughly pH 8 and will reverse — letting the RNA refold — if the buffer is left to drift alkaline over a long run.
3. Capillary transfer to a membrane in high-salt buffer, the same upward-wicking principle as a Southern blot. Positively charged nylon is the common choice for RNA, since it binds it more efficiently than plain nitrocellulose.
4. Immobilisation by UV crosslinking or baking, so the RNA survives hybridisation and repeated washing.
5. Prehybridisation and hybridisation with a labelled probe — DNA, RNA (a riboprobe), or an oligonucleotide. Riboprobes form RNA:RNA duplexes, which are more thermally stable than a DNA:RNA hybrid, so hybridisation and wash temperatures for a riboprobe typically run higher than for a DNA probe against the same target, and strand-specific riboprobes let you distinguish sense from antisense transcripts, which a double-stranded DNA probe cannot do.
6. Stringency washes, then detection.

How this differs from a Southern blot

The two techniques share a transfer-and-probe skeleton, but the differences are not cosmetic. A Southern blot asks about genomic context — copy number, integration site, rearrangements — from double-stranded DNA that survives ordinary handling and only needs denaturing once, after the gel has already separated it accurately by size. A Northern blot asks about a transcript’s identity, size and abundance from single-stranded RNA that must be kept denatured during the run itself, because its own secondary structure would otherwise distort the sizing the whole technique depends on. RNA is also far less forgiving upstream: ubiquitous, stable RNases mean the RNA extraction and handling stage carries more of the real failure risk than it does for a genomic DNA prep. The two blots are read differently as a result — a Southern blot’s band count tells you about the genome; a Northern blot’s band size tells you about the transcript.

Where Northern blotting still earns its place

For routine “is this gene up or down” questions, RT-qPCR and RNA-seq are the default tools now, and for good reason — both need far less input RNA, both scale to many transcripts or the whole transcriptome, and RNA-seq in particular has become inexpensive enough to use even when only a handful of genes are actually of interest. Northern blotting remains genuinely useful in narrower situations where transcript size is itself the answer, not just abundance: confirming that a gene has multiple isoforms from alternative splicing or alternative polyadenylation, distinguishing an unprocessed precursor from its mature transcript, sizing an unusual or novel RNA that sequencing data alone leaves ambiguous, or independently validating an RNA-seq differential-expression result at the level of the physical transcript rather than a read count. It is a low-throughput technique next to RT-qPCR or RNA-seq, and nobody profiles a transcriptome with it — but for the specific question “what size is this RNA, and is there more than one form of it,” it still answers directly what neither higher-throughput method shows without extra analysis.

Quantification and interpretation

• Band size vs. transcript length. Run an RNA ladder alongside your samples; the transcript’s migration distance against that ladder gives its size directly, the way it would for a Southern blot’s restriction fragments — this is the information RT-qPCR and standard RNA-seq counting do not hand you without extra work.
• More than one band is often real, not a failure. Multiple bands from a single probe can reflect genuine alternative splice variants, differential polyadenylation, or a precursor-to-mature processing series — the same logic as an unexpected extra band on a Southern blot, where the first move is to check whether it is a real biological signal before assuming a technical problem.
• Loading normalisation. The traditional control is the 18S and 28S ribosomal RNA bands, visualised by staining the gel or membrane directly (they are vastly more abundant than any single mRNA and track total RNA loaded), or a probe against a stable housekeeping mRNA. Housekeeping-transcript stability is not universal across conditions or tissues, the same caveat that applies to reference gene selection for qPCR — pick and justify a normalisation control for the specific experiment rather than defaulting to one by habit.
• Degradation looks like a smear, not a band. A smeared lane with no distinct bands almost always traces back to input RNA quality rather than the blotting procedure itself, which is why checking RIN before running the gel saves diagnosing the wrong step afterward.

Common failures and their causes

• Smeared lanes, no distinct sizing — degraded input RNA (check RIN before blotting) or RNase contamination introduced during the procedure itself.
• Bands at an unexpected size — incomplete denaturation letting residual secondary structure distort migration, or a genuine biological signal (isoform, unprocessed precursor). Rule out incomplete denaturation before treating an odd-sized band as real.
• Faint or absent signal — a single transcript is often a small fraction of total RNA, so underloading is a common cause; also check that the probe was labelled efficiently, and, on a glyoxal gel specifically, that the running buffer did not drift alkaline and reverse the denaturing adduct partway through the run.
• High background across the membrane — insufficient prehybridisation blocking, too much probe, or washes that were too gentle, the same causes as on a Southern blot.
• Uneven transfer or bubbles — introduced at assembly and not fixable afterward; they read as patches of missing signal that are easy to mistake for absent expression.

Related methods

RNA quality going in determines whether the blot is readable at all — see RNA extraction protocol basics and RNA Integrity Number (RIN). For routine expression quantification, qPCR and RT-qPCR and RNA-seq are the higher-throughput alternatives this technique has been largely displaced by. The DNA analogue of this same transfer-and-probe logic is the Southern blot; the protein analogue is Western blotting.

Frequently asked questions

What does a Northern blot actually detect?

A specific RNA transcript within a complex RNA sample, reported as a band whose position tells you its size. Multiple bands from one probe usually mean multiple real transcript forms — splice variants, differential polyadenylation, or a precursor and its mature product — not necessarily a failed experiment.

Why does the RNA have to be denatured during the gel run itself, not just before transfer?

Because RNA is single-stranded and folds into secondary structure that changes how it migrates. A Southern blot can denature DNA after electrophoresis because linear double-stranded DNA runs true to size unmodified; a Northern blot’s sizing is only accurate if the RNA is kept denatured throughout the run, which is why the denaturant is built into the gel and buffer system itself.

Formaldehyde or glyoxal — how do the two denaturing gel methods differ?

Formaldehyde-MOPS gels carry the denaturant (roughly 2.2 M formaldehyde) in the gel and running buffer throughout the run. Glyoxal-DMSO gels denature the RNA before loading and run it in a low-ionic-strength buffer with no formaldehyde in the gel, but that buffer must be recirculated, since the glyoxal–RNA adduct reverses above roughly pH 8.

Is Northern blotting obsolete?

Not obsolete, but no longer routine. RT-qPCR and RNA-seq have taken over most expression-quantification work because they need less RNA and scale further. Northern blotting stays in use specifically where transcript size itself is the question — isoforms, processing intermediates, degradation, or validating an unusual transcript — because it shows size and abundance on the same blot.

What is the difference between Southern, Northern and Western blotting?

The target molecule. Southern detects DNA, Northern detects RNA, Western detects protein. Only Southern is named after a person — Edwin Southern; Northern and Western are puns built on that name.

How do you normalise a Northern blot for loading?

Most commonly against the 18S/28S ribosomal RNA bands, stained directly on the gel or membrane, or against a probe for a stable housekeeping mRNA — chosen and justified for the specific experiment, the same discipline reference-gene selection requires for qPCR.

References

• Alwine JC, Kemp DJ, Stark GR (1977). Method for detection of specific RNAs in agarose gels by transfer to diazobenzyloxymethyl-paper and hybridization with DNA probes. PNAS 74(12):5350–5354 — pnas.org
• Northern Blotting: Protocols for Radioactive and Nonradioactive Detection of RNA — PubMed 39535701
• Modified Northern blot protocol for easy detection of mRNAs in total RNA using radiolabeled probes — PMC8772191, BMC Genomics 2021
• Strategies for Detecting mRNA — Thermo Fisher Scientific