Reappearing band shifts/smears after double digest and gel purification - (Apr/23/2013 )

I'm cloning a 850 bp insert (created by mutagenic PCR) into a 2.2 kb vector (from a miniprep).

I do an overnight double digest of vector and insert with SbfI-HF and NotI-HF (both are 8-cutters)
20 U of each enzyme (tried less but I often get incomplete digestion)
BSA
NEB buffer 4
H20 up to 50 uL
~1 ug PCR product or 3 ug plasmid
Enzymes are heat inactivated in 30 min at 80C

I put digestion products on a gel, and I always get a smear ABOVE the expected size (see the attached image). I figured that enzymes are still bound to DNA, that's why I see smearing/bandshift. I rerun the DNA on the gel, but first I add SDS to a final concentration of 0.3%. Now smears dissapear and the bands are of the right size.

I do gel extraction of the cut plasmid (I add SDS to the digest just before loading the whole thing on the gel). The band is clean (no smear), and I extract it from the gel using a Qiagen kit.
I column purify digested PCR product by adding SDS first.

Then I run both purified products on the gel. This time the smear reappears. If I add SDS just before loading on the gel, the band is clear.

Ligation from these purified products works, but the efficiency is horrible. Right now I get about 10^4 clones, but I'm aiming at 10^5-10^6. Cells are competent enough, so that shouldn't be a problem. I assume that leftover enzymes are blocking restriction sites, so ligase cannot efficienty ligate them.

Any ideas on how to get rid of this problem?

I attached a pic of the gel after purification (it looks the same before I purify)
1.gel purified digested vector
2. sample 1, but added SDS just before loading
3.column purified digested vector
4. sample 3, but added SDS just before loading
5. column purified digested PCR product
6. sample 5, but added SDS just before loading

Just curious, what made you to aim for 10^5-10^6 clones in ligation transformation? is it bcoz the clones you obtained are negative?

GNANA on Tue Apr 23 13:56:45 2013 said:

No, most of the clones I get are positive (~95-98%), but I need a larger library because it's a part of a directed protein evolution experiment. Small libraries = insufficient diversity