Southern blot
Detect a specific sequence in genomic DNA with a labelled probe; largely replaced by PCR.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
1Seeing a whole locus at once
A Southern blot asks a sequence question about genomic DNA without ever amplifying it. Genomic DNA is cut with a restriction enzyme so that the region you care about ends up inside one fragment of a defined length, the fragments are separated by size on an agarose gel, the separated DNA is denatured and transferred onto a membrane, and a labelled probe is left to find its complement there.
Structure
The size of the fragment the probe lies in. A shift means the locus is not what you expected: insertion, deletion, rearrangement or a lost site.
Integration events
One band per distinguishable allele or integration site. Two bands where you expected one means a second, independent event.
Copy number
Signal is proportional to the mass of that fragment, so against a known single-copy equivalent the darkness becomes a copy number.
2Hybridisation stringency is the whole specificity argument
A probe does not bind only its perfect complement. Every partly matched sequence in the genome forms a duplex too, just a less stable one, and what decides whether that duplex survives to the film is the stringency of the hybridisation and the washes.
3The transfer is a physical process that can simply fail
In classic capillary transfer, buffer is drawn from a reservoir up through a wick, through the gel, through the membrane and into a stack of dry paper towels, carrying the DNA out of the gel and onto the membrane, where it sticks. Three things routinely go wrong.
4How it fails, and what each failure looks like
- Incomplete digestion
- Extra bands above the expected size, at the positions of the next sites out, plus a haze of uncut DNA under the well. Genomic DNA is viscous and slow to cut: digest overnight, and do not put more DNA in than the enzyme units can handle.
- Too little DNA
- A blank or barely visible film. A single-copy locus in a mammalian genome needs roughly 5–10 µg per lane for a conventional probe; at 1 µg there simply are not enough copies of the target fragment on the membrane.
- A probe that hits a repeat
- The same ladder of extra bands in every lane, including wild type. Repeat-containing probes have tens of thousands of partial targets. Re-make the probe from a repeat-free part of the region; raising stringency helps but does not fix it.
- Transfer that dried out
- Weak, patchy signal across the blot, often worse in the middle. Nothing distinguishes it from an absent target, which is why a lane of known positive material belongs on every membrane.
- Stringency too low
- A grey film with bands everywhere. The probe stayed bound to partly matched sequence and to the membrane itself.
Key terms
Practise these as flashcards →A labelled single-stranded nucleic acid complementary to the sequence you want to detect; usually a few hundred bp to a kilobase, random-primed so it carries many labels.
How much mismatch a hybridisation or wash tolerates, set together by salt concentration and temperature. Low salt and high temperature are stringent.
Saline–sodium citrate, the standard hybridisation and wash buffer. Concentrations are quoted as multiples of the 1× stock (0.15 M NaCl, 15 mM sodium citrate).
A digest in which only some copies of a site were cut, so signal appears at the size defined by the next site out as well as at the expected size.
A probe lying outside the sequence that was introduced, so it detects only the endogenous locus. It is the standard way to prove a targeted insertion is where it should be, since a randomly integrated construct carries no probe sequence.
A lane of carrier genomic DNA spiked with a known molar equivalent of the fragment of interest, loaded so band density can be converted into copies per genome.
Moving DNA from gel to membrane by letting buffer wick through the stack into dry towels; alternatives are vacuum and electroblotting.
Above roughly 20–50 kb all DNA moves at the same speed in a constant field, so very large fragments cannot be sized on a conventional gel.
When to use it
- Confirming a gene-targeting event: an external probe proves the cassette is at the intended locus and that the flanking DNA was not rearranged, which a junction PCR cannot do.
- Counting transgene copies in a founder line, where a tandem array is common and PCR gives a product whatever the copy number.
- Detecting a structural change (a large deletion, insertion or duplication) in a locus where a short amplicon would read as normal.
- Sizing an expanded repeat, where the expansion is longer than any amplicon and amplification is biased against the long allele.
- Reading methylation at a specific site, by digesting with a methylation-sensitive enzyme and probing across it.
Compared with alternatives
Faster, needs nanograms, and the right tool when you only need to know whether a short, known sequence is present. It cannot see outside the amplicon, so it misses rearrangements, copy number and large insertions.
The separation step a Southern is built on. Run the gel first for the digest itself; the blot adds sequence specificity to it.
Choosing the enzyme is most of the experimental design: the probe must sit inside one fragment that spans the region you care about, with no site of its own.
The same transfer-and-probe logic applied to RNA, to measure transcript size and abundance rather than genomic structure.
The usual modern route to copy number: faster and quantitative over a wide range, but it still only reports on a short amplicon, so it cannot distinguish a tandem array from separate insertions.
Long-read sequencing now answers most structural questions a Southern was used for, at higher resolution and with the breakpoint sequence, at much higher cost per sample.
A methylation-sensitive Southern was the original way to read methylation at one site; bisulfite methods replaced it for anything needing per-base resolution.