GST-fusion protein production in insect cells

From: Sebastiano Pasqualato
Date: 16 March 2012 12:57

Dear CCP4ers, Dr. Berger,

we have an accumulating series of GST-fusion proteins here that are all displaying the same behavior when expressed in Hi5 insect cells with the MultiBac system.

What we are experiencing is a massive production of free GST and only a limited amount of fusion protein.

Since the linker we have engineered between GST and our protein is the one that exists in the pGEX-6p vector series, with the preScission protease site, I was wondering if any of you had experience of this cleavage site being processed within these insect cells.

Thanks in advance for the help,

ciao,

Sebastiano

----------
From: Katya Heldwein


Dear Sebastiano,

what you see is not your cleaved GST tag but rather native insect glutathione-binding proteins that are produced in relatively large amounts (a few mgs per L). I am attaching a paper where recombinant GST-tagged proteins were successfully separated from these pesky insect GlutBPs under specific elution conditions. This has not worked for us, though. So, we simply do not express GST-tagged proteins in insect cells.

Good luck,

Katya


Ekaterina Heldwein, Ph.D.

----------
From: Sebastiano Pasqualato

Dear Imre, dear Katya,

thanks a lot for the prompt and insightful replies.

@Imre: well, I don't think it is a matter of linker of lack of Met/presence of stop codon after GST.

We use the exact same linker in bacteria and that works just nice, while wrt to the Met, we have cases in which we have kept the original Met (gene starting from aa 1) or in which we didn't include it (N-term deletion mutants) and we experience the same problem.

I tend to second rather Katya's opinion that what we purify in massive amounts are indeed native GSH-binding proteins (I dind't think at that at all!).

@Katya: I will have a look at the paper you sent me. However, the trouble is that the high amount of GHS binding protein is lowering the amount of GST-fusion protein we manage to recover from beads. So even being able to purify it form native GSH-binding proteins will not be ideal.

I think I will carefully consider the option of expressing MBP-fusion proteins, and move away from GST, as you both suggest.

Thanks a zillion for the tips,

best,

Sebastiano

On Mar 16, 2012, at 2:40 PM, Imre Berger wrote:

Dear Sebastiano -

I have actually no experience at all with GST in insect cells since I do not use this tag (mainly because I do not like the dimerization propensity of this tag and the glutathione elution step).

However, I do not think what you observe should be related to  insect cells...but I have no data on my own for GST fusions in insect cells.

May I ask: Do you have a starting methionine in your construction AFTER the GST (i.e. did you keep for example the native ATG start codon when you cloned your gene?). And: Did you verify the reading frame (basically an huge excess of GST would mean that there is no readthrough resp. a stop codon somewhere close to the fusion tag, that is very unlikely.

Is there a long unstructured linker between GST and your protein resp. doyou suspect there is a long unstructured region in the N-term of your proteins?

I am soprry that I can';t help you much, but as said, Idon;t like GST - I used it as an undergaduate many years ago and totally disliked it (i had a protein that dimerized and with the GST tag which also dimerized it became an agregating mess), kept this aversion until now.

How about MBP? That's EXCELLENT.

----------
From: Artem Evdokimov

Hi :)

In addition to excellent replies already posted, a few thoughts:

1. Do the whole-protein MS of your stuff. Assuming that you get a spectrum then
  a) If you are right and it is indeed your recombinant GST - you
will see a mass pattern that's interpretable via analysis of probable
cleavage sites. You will also glean some understanding of where the
undesired cleavage (or premature termination!) occurs.

  b) if the suggestions of previous replies are right, then these
protein(s) will be of the insect-cell origin. I have not seen this to
be a huge issue in the past, however it may be due to the experience
with Sf9, Sf21 cell lines and relatively low use of Hi5 in my past
work. Switching a strain might help (maybe you already tried, and it
still sucks, sorry).

If you can't get a spectrum of the whole protein I would potentially
suggest tryptic cleavage and peptide MS. You will get at least an idea
whether (a) or (b) is more likely.

In general (b) is tougher to fix because this means that your protein
isn't being expressed in large enough amounts. You don't mention - is
the insoluble fraction interesting? Is there a big fat band at the
correct m.w. in the insoluble material? Perhaps a change of lysis
conditions might help, or expression at lower temperature (yes, insect
cells do grow at somewhat lower temperature, however they're not very
happy there) or lower MOI.

Also: do you *need* MultiBac system in your case? It's an excellent
tool for large complexes but it's not clear from your post whether
you're working with one protein or many? MultiBac has Cre/Lox
leftovers in the bacmid which may be un-fun in your case for some
reason (unlikely but...)

Take a look at the DNA & RNA structure (predicted) in the junction
region. Is there any 'funny business' with hairpins, ribozymes,
unusual structures of any kind? A cryptic ribozyme can really mess up
one's life (it's not likely that you have one, much more likely an
early termination structure or inefficient read-through). Splicing can
also be difficult to control on occasion - any presence of strong
splicing sites? AG/GURAGU and YAG/RNN with a few additional parameters
thrown in. This would require a fair bit of RNA to be in between the
sites, too.

Cheers,

Artem