SDS Page and WB: Wrong size bands - (Nov/02/2012 )

Hi all,
New to posting in this forum but I have been actively reading for over a year. Now I find myself in need of an answer for a question.

I am expressing VEGF protein in E.coli (special from NEB that forms disulfides). VEGF is a homodimer and I have modified in via cloning too include a 6x HIS tag, a Cystein "tag" (one Cys amino acid) and 2x GGGS linker at the N terminal (C-term is active end).

I am getting good expression but odd SDS-Page and WB results.
My dimer should be at 50kDa (calculated) and unmodified VEGF runs at about 40kDa dimer.
My dimer is showing up at 31kDa and a monomer just under 20kDa.
This is very obvious in a western blot and not so obvious in SDS Page (dealing with purification).

Is this common?

I am not using reducing conditions for the SDS Page. Boiling for 5 mins at 95 in sample buffer.
Ladders are biorad broad range (not prestained) and I am using biotinylated ladder for the WB.

The unmodified VEGF monomer is 19kDa, with the dimer between 38 and 40kDa, but I have added the 6xHIS, thrombin cleavage site, a cys residue and 2x GGGS linker prior to the VEGF.

Checked sequencing and it is correct and has no frameshift mutations.
Antibody is monoclonal from R&D.
Does not detect reduced dimers in presence of DTT

It is very common for proteins to run at odd sizes in PAGE systems, this is because the running of the proteins doesn't just depend on the SDS wrapping around the protein, but also the charge on the protein and a few other things, so I am not at all surprised to find that your protein is not running at the calculated size.

As you are not using reducing conditions, it is even more likely that the proteins will not run at the correct size. Note that the ladder is in a reduced state so it is running at the correct size. Comparing your non-reduced protein to this will not give the correct size.

Check the antibody data sheet and look for the migration size if they show a ladder, and preferably run a reduced purified VEGF standard so that you can see a band on a membrane stained with Ponceau S.

Thanks for your response. (Forgot to check back).

What I do not understand is that the commercially derived protein runs at the expected sizes even under non reducing conditions.
What can I do to fix this?
Whenever I add DTT it has no effect that I can see on the SDS Page. The band at 32kDa stays the same. VEGF has 7 disulfides, two of which form bridges between the monomers to form the dimer. It is a cysteine knot motiff if that helps. I add 5ul of 1.25M DTT to a solution of 35ul containing my protein and the loading buffer.
I did a proliferation assay and my protein is active (nearly the same as "standard" vegf) and is detected by ELISA, but Im having a hard time convincing my PI (and myself) that the wrong size is ok.

I just did another round of purification using a slightly different tag on the VEGF (it is a fibrin binding domain with 7 amino acids). So this raises the MW by less than 1kDa. However I am seeing the same thing.

I do have one question that may help me but is not directly related. I am using a pET-28a vector and when I use clone manager it has to transcribe and translate in "complementary". Does this have any effect on protein expression?

Again any help on forming this dimer correctly is greatly appreciated. (I have tried refolding from IB's with multiple methods with no results unfortunately).