southern blotting

Frank O. Fackelmayer fof1 at
Thu Jul 2 05:12:59 EST 1998

Moens' lab wrote:
> I've tried to do mammalian genomic southerns before but all I get
> is a smear signal whose intensity correlates with the ethidium bromide
> intensity in an agarose gel.  Therefore, it is not due to repeat
> sequences or anything of the sort. 

Sounds to me like an alu sequence (you´re working on human DNA, I assume), or
any other highly repetitive sequence, is in your probe that gives you the
strong smear. The fact that the smear correlates with the EtBr staining is
actually the best argument FOR, not AGAINST repetitive elements in your probe.
I´d recommend you characterize your probe by restriction digestion, blot the
digested DNA to a nylon membrane, and hybridize with labelled (unfractionated)
genomic DNA. Hybridizing with total genomic DNA will give a signal only if
there is a repetitive element on the corresponding band. 
There are two possible results: Either all bands give a signal, or only a
subset of bands give a signal. If all bands give a signal it is likely that
your probe is very GC-rich, and you´ll have to compensate for that by higher
temperature: Increase your hybridization temperature to 70C and try again.

If you find only a subset of bands giving a signal, your conditions are ok,
but some of the bands contain a highly repetitive element. The bands that give
NO signal are then single-copy and well suited as hybridization probes.
Isolate one such negative band, label, and use as a probe. 

Hope this helps,

 I had PAC DNA on the same southerns
> and these worked fine.  Also, the lambda marker was clean.  I followed
> the protocol in Maniatis et al, 1989 "Molecular Cloning" to the letter
> (for what I can tell) but I keep getting these smears.  I've digested
> with BamHI to apparent completion, also.  What am I doing wrong?  The
> amount loaded was about 20 ug and I left hybridizing O/N.  I used
> Denhardt's solution without formamide.  Does anyone have any idea how I
> can solve this?
>         Thank you,
>         Rob Botelho
>         yu106715 at
>         or
>         moens at

More information about the Methods mailing list