promoter mutagenesis question

Dear fellow Experts,

please give me advice on solving this problem:

I am about to assemble a mammalian expression vector. It contains the EF-1
(elongation factor 1) promoter. Unfortunateley, the promoter has a Xho I
site I need to remove.
I am thinking of two strategies:

1) cut with XhoI, do Klenow fill-in and re-ligate (in past experience,
there will be a lot of XhoI cleavable background)
2) perform site directed mutagenesis with a primer containing 6 wobble
bases (giving 4096 possibilities of new sequence). In this case, I'd
transfect cells with EGFP as reporter gene and screen them for expression
levels by fluometry of FACS.

Which is better in your opinion or what else would you suggest?

Best regards,

Wolfgang

Wolfgang Schechinger, Dr. rer. nat.
Med Clinic I, Endocrine Research Lab
Bochum University Hospital "Bergmannsheil"
- the world's first casualty hospital for miners and steel workes -
Germany

*^_^*

please reply to the group, the email address ends in auto trash
due to overwhelming spam caused by email harvesters


---

Wolfgang;

In a previous life I had a similar problem with a
prok expression vector. If you have not already done the hard
cloning part, take some time to do this right. Just filling in the XHO
site is one option, but you might want some useful site there after all.
Also, your vector grows by 4 bp and two restriction sites. Is that a
good thing?

How about this.....
Based on certain assumptions about the XHO I location....
Design the MCS/promoter site you should REALLY have, then cut out the
MCS that exists, clone back in the new one (two oligos). In this way
you can add/delete more than the XhoI site, and adjust spacings etc if
you need. In the ligation mix, you don't need to seperate or anything.
Overload a bit with the oligos.
You'll get parentals, recombinants, and bastards, but you just need to
screen a handful to get the right construct. Not so bad.

Or this:
Rather than Klenow, treat with a nuclease (eg Mugn bean) or exo/endo +
polymerase which can eat free ends in the absence of nucleotides (eg
T4). Let it run to completion, then add pol/nucleotide mix to fill in
scruffy ends.
You'll need to characterize the product carefully, but I have used this
approach a couple of times to reduce promoter-ATG spacing. OTOH, Mung
bean is a bit unpredictable.


It's more work at the beginning, but in the long run, you get a vector
you can actually use.

B.