transgene insertion and repetitive sequences

[Tom Anderson]

On Tue, 14 Aug 2007, Ed Siefker wrote:

> Now I know sometimes parts of other chromosomes get incorporated with
> the transgene, so I need to find the chromosome 3-X junction.

How about going about it a different way. Go back to FISH and try and get
a resolution higher than 'chromosome 3'. Conventional FISH should get you
to band resolution, and if you can get fibre FISH to work, and you have
suitable probes, you can take it to crazy resolution, potentially tens of
kilobases.

Once you've got that far, order up a genomic clone (or several) from
normal mice covering the area of the insert and use it as a probe on a
Southern blot. Comparing the pattern to that from normal mice will tell
you with, i think, a high degree of precision where your insert is.

If you want to go even higher than that, hmm. You need a clone or PCR
amplicon to sequence, and you're back to where you are already.

> Does anyone have any ideas? My best guess is to try something like this:
> http://www.open-access-biology.com/pcr/pcr-boulter.html
>
> But I worry about trying to clone large inserts from a mess of digested
> genomic DNA. There's at least 3kb of repeat sequence in my clone, so if
> I design primers to non-repeat sequence that's at least a 3kb insert to
> clone, and since size affects transformation efficiency I don't know if
> I'd get any clones from that ligation event. On the plus side though, Id
> be able to remove the normal X sequences by digesting them.

How about doing this protocol, but using an enzyme which cuts inside the X
chromosome repeats? You wouldn't get your transgene in the clones, but you
could get the X-3 junction, and the clone would be smaller. You'd need to
use a primer matching the repeats for the screening step. You run the risk
of just cloning repeats, but if you gel-purify the digest to remove all
the fragments smaller than the size of a repeat, you avoid that.
Alternatively, double-digest with an enzyme which cuts the repeats and one
which doesn't (which give different sticky ends), and clone into a vector
cut so it will only take inserts with one end from each enzyme.

Another way of getting rid of real X DNA would be to find some genomics
people and get them to flow-sort chromosomes from your mice. Once people
have things set up for doing this, it works very well.

Aha! Do FISH-and-Southern mapping, then make Boulter clones with an enzyme
that cuts the X repeats, then do the screening step with a primer that
matches the closest known bit of chromosome 3 sequence. This does require
that you map the location to within a few kb, though.

Disclaimer: i haven't had lunch yet, so some or all of these ideas may be
total bunk.

> I don't know how else to approach this, besides doing another library
> screen and hoping for the best.

Make a new mouse? :)

tom

-- 
Tom Anderson, MRC Laboratory for Molecular Cell Biology, UCL, London WC1E 6BT
(t) +44 (20) 76797264   (f) +44 (20) 76797805   (e) thomas.anderson from ucl.ac.uk