HELP! contaminants do not separate from my protein during purification - (Jun/18/2007 )

Hi,

I'm working on 118 kDa protein with pI 5.5. It was fused with maltose binding protein (MBP) and expressed in E. coli. Because of its large size, the level of expression is not quite great but I still get some amounts of protein from a large volum of culture (~10L). Fusion protein elutes well from the amylose column with some smaller bands on an SDS-PAGE gel. Fusion protein has an additional 6xHis tag at the N-terminus but it doesn't bind the Ni-NTA column. So I stick to the amylose column for the first purification step. Then, I cleaved MBP away using protease to obtain the protein without tags.

Here's the problem:

I've tried to further purify this protein using Q, Heparin, or gel filtration.

My protein doesn't bind Q at pH 7.4.

It binds Heparin but MBP doesn't, which is great. However, all the other small proteins which can be seen on the gel co-elute in the same fraction.

I've seen the same from gel filtration. Small bands still come together with my protein despite a huge size difference.

Questions:

Q1. Why are small bands always there? Is my protein too sticky? I tried high salt (500 mM NaCl), beta-mercaptoethanol (14 mM), or Tween 20 (0.05%) to reduce non-specific binding but nothing worked.

Q2. What kind of columns can I try to remove such small bands?

Q3. I'm also suspicious with premature translation termination, which mgith be responsible for such small bands. Is there any way to improve expression and purification of such a big protein?
Thanks in advance for your help,

J

---

First of all, you might have to rethink your pI. Unless you have some experimental data that says it is 5.5, don't just trust calculations. Try a Q column at pH 4, try an S column at pH 7.4.

Best of all might be to try a broad range of pHs with both types of beads. Get some loose beads, and equilibrate them at pH 4 to pH 9, going up in 0.5 pH units steps. Prepare your sample in the same buffer. Add some of this to the beads and let them stand for 10 minutes or so. Let the beads settle or centrifuge them (gently!), then run supernatant and beads on SDS-PAGE. You'll find the pH that gives you best binding (and hopefully avoids most of the contaminating proteins).

Next, at your optimal binding pH, prepare some more beads an protein and mix together. Now separate this slurry into several tubes. Add different salt concentrations to the tubes, settle the beads and run the gels again. This will give you the salt concentration at which your protein will elute. With any luck, the contaminants won't bind as strongly as your protein, so they will be seen in the low-salt supernatants. However, if they still co-elute, you might have to try something like hydrophobic interaction chromatography, which starts with high salt (so they should be disassociated from your protein) and goes down the salt gradient.

Finally, use the information you've just found to develop your large-scale purification in columns. (Also good for padding out the methods section of theses ...)

Yes, it seems like a lot of work, but it could just save you a great deal of heartache later on when you scale up.

As for the expression, have you checked the codon usage of your protein in the E coli cells? If you have low-abundance codons, your expression rate will suffer. If all is well in that department, you can try slowing down the expression rate by dropping the temperature to 30, 25 or even 15 C; expression time will extend to overnight up to 24 hrs (but beware of proteases chewing up your precious protein...)

---

Did you add some protease inhibitor in your crude extracts of harvested E Coli?

Your smaller bands may be derived from degraded your protein, and your protein would be in dimeric form.

---

Swanny,

The pI was not experimentally determined but simply calculated.

I will try to optimize the binding conditions as you suggest.

Approximately 10% of amino acids have rare codon usage in E. coli. Therefore, Rosetta strain is used to compensate for it. Good expression is achieved at 18C overnight.

Any other suggestions would be much appreciated.

Thanks a lot,

J

---

I tried protease inhibitors but couldn't see a difference. Also, cleared lysate and amylose column-purified protiens are fairly stable for quite a long time (>a week).

---

Calculated pI can be really good or really bad. I once compared two small domains that folded in almost identical structures due to high sequence identity; both had the same theoretical pI. One bound to Q and not S at a certain pH, the other refused to bind to Q and only bound to S.

Proteins can be really stoopid some times!

In order to see if anything can be done with the expression levels, could you post a picture of an SDS-PAGE gel? I presume you have most of the protein in the soluble fraction.

M

---