regarding the benchmarking of our eukaryotic cell expression facilities I would like to start discussing the details mainly with the 4 of us (Sabine, Tsafi, Yvan and myself) who volunteered in Copenhagen to organize it but everybody is invited to comment on the ongoing planning.
Topics for discussion are following (please add more if I missed anything critical):
what kind of proteins? how many proteins? which baculovirus generation system (do we have the constructs for the same protein e.g. for two different systems)?
criteria for analysis and assessment of yield, solubility, activity, etc.? timeline?
plan in such a way that we can publish results?
Thanks for any input and please let us know your ideas and (although we are still at an early stage) if you are willing to participate..
Dear Huseyin Sabine and Yvan,
In order to benchmark the baculovirus expression system, I think we should each express 3 types of proteins: Secreted, Trans membrane (polytopic) and Cytosolic.
Maybe, so it won't take too much work, each can suggest a proteins they alredy work on, one for each type, and after we pick the appropriate proteins from this list, each of us will end up testing at least one target they already work on.
As for the systems:
Bac to Bac is not compatible with the FlashBac/BestBac/BaculoGold. I think it would be good to compare the 2 different systems ( transposition in bacteria vs. homologous recombination). I think many people would like to find out whether there is a difference between the 2 types of viruses, in terms of virus stability and expression profile.
There are also the FlashBac Gold (dChit/dCath) and the FlashBac Ultra, that has more deleted genes. OET sells those as better choice for secreted proteins. Do we wish to test the Ultra too? (I would like to test the dChit/dCath virus in this comparison).
Will we test cell types? SF9, SF21 and H5 or T.ni? H5 are usually better for secreted proteins, but not for membrane or cytozolic. However, I think that there might be variations between the cell lines we are using. Maybe we can all use the same SF9 and H5 cells? so we won't have to comapre too many vartiables?
I am sure this work can be published if we stick to "Standartization of Procedures" (SOP).
Please let me know if you thought of something more simple. maybe stick to secreted proteins only? we might have too many variables here.
Dear Tsafi, Hüseyin, Sabine and Yvan.
I am interested to join in this benchmarking. From our point of view we have to test different proteins which could be crystallized easily to show that our new mHOST-XS Baculosystem allows the production of material that will be successfully crystallized. Björn told that he had a benchmarking pool of around 10 protein/genes. Does anyone have a suggestion which genes should be taken for this type of benchmarking.
We will stick to our bac-to-bac based system using derivatives of pFastbac, pMultibac and Acembl.
This will be a major effort if you consider that question have been raised for the expression system (bac to bac/flashbac) Media (Excell, Hyclone, SF900 III) and cell types (Sf2/1Hi5). Even temperature might be an important factor.
Wo is interested and how do we set this up.
Joop van den Heuvel (HZI-RPEX)
Dear all (and Joop, welcome!!!)
I do agree that there are too many variables.
Maybe before we agree on the general targets we should start with a small pilot on proteins we already work on, and test 1 or 2 of our own targets in several media formulations, 2 cell lines , 2 viral vectors, 2 temp's, and the most important variable in my opinion: same cells. I think many discrepancies result from the fact that our cell lines are not the same, although they are called the same.
Maybe we should start by testing each others cell lines with 1-2 of our current targets, and decide from there?
Also: I think this mode of communication is way too slow. Years might pass by the time we will agree on something ...
I suggest we should we should all schedule a skype conference within the next couple of weeks to discuss these issues on line.
Dear P4EU partners,
before we go deeper into the benchmarking discussion, we need to know if we do have the critical mass of labs joining this effort in order to obtain meaningful results and in order to publish. Not all of us have the capacity to join, we need to know who would.
I would like to ask all partners for a yes or no commitment to join a benchmarking of insect cells and or mammalian cells. No answers are as important. You can either post the answer in this forum or send an e-mail to Hüseyin, Tsafi, Yvan or me.
Thank you very much in advance
It isn't very clear at what point the experiment finishes-westerns, small scale screens/purifications or full litre scale-ups and purifications. Ideally it should be the latter-you know you get Xmg/litre of final product and not a vague score from a gel/Western.
Keeping it simple is also very important-anyone remember the first SPINE benchmark paper?-too many variables make comparisons impossible unles you just compare 'systems'.
Dear Sabine, Tsafi, Hüseyin and Yvan,
I am interested to join this benchmarking of eukaryotic cell expression. For baculovirus/insect cell expression, we use Sf9 and Hi5 cells, bac-to-bac or flashBac (various versions), and our protein targets can be cytosolic, secreted or membrane proteins. We usually go with flashBac as it's much quicker than bac-to-bac. Previously, we have also investigated genetic stability of recombinant baculoviruses during serial passaging and various signal peptides for efficient secretion from insect cells. We find temp an important factor too and is starting to routinely test expression at temp lower than 27-28 oC.
here's my comment regarding the benchmarking that I have sent to Sabine, Tsafi, Joop and Yvan by email:
It's true that there are many variables in every step that will make the
benchmarking a bit difficult. I don't see this as a big issue because one
goal of the benchmarking could simply be that, whatever differences there
are in the parameters of each lab, we could just have a panel of e.g. ~10
proteins and see which protocol would give the highest success rate or
yields or both.
The more labs participate the easier it would be to identify those
parameters that critically influence the performance of the expression
system. If success rate or yield correlate clearly with e.g. the medium,
cell line, virus type (Bac2Bac or others), then it is easier for each lab
to modify their protocols. In a second round each lab could independently
decide whether they're happy with their overall performance or would like
to change parts of their protocols. Maybe we find out that most protocols
have very similar results in terms of success rate and yields then nobody
would need to change anything (of if need be we could compare the costs for
I believe that the main challenge for the insect cells will be the
availability of identical constructs for Bac2Bac and the other systems.
So, yes, in my opinion we should just compare systems at this stage.
To the question of how far we need to go, one possibility would be collecting pellets from e.g. 20-30 ml cultures and sending them to one of our colleagues with a HT purification system to process and analyze them in small scale in parallel. By this we would exclude variation due to different processing protocols.
I'm not sure if we should compare the extraction and purification protocols in addition to the expression systems.
This purification benchmarking could be done in another benchmarking test where we have a defined set of cell pellets that should be processed by different labs to have the comparability of their systems.
I would also be interested in participating. We have Sf9, Sf21, and Hi5 cells and are using BactoBac. As we have just started offering service we would not be able to contribute many targets. Most of our targets will be cytosolic.
Hi there, we are not using the baculovirus system in our lab and we are therefore not able to participate in this benchmarking activity isf that is what you decide to go for first. We will maybe come along in the mammalian cell evaluation later on. I guess that one system will be evaluated first and then the other?
We could however offer to analyse the proteins that the different groups have produced, if it desirable that one lab performs this analysis. I believe that the evaluation should be done on the prodcued proteins directly and not after purification, otherwise there is a risk that you don't know in the end what you are evaluating; production or purification. We could do this analysis by Western blot. Perhaps you also want to go for purification and quantification, but I think a quick Western analysis (and/or Coomassie?) could tell us quite a lot. /Malin