Dear Timothy,

How do you know you transfection was successful? Even if you do get plasmid into the cells, this does not mean it was integrated into the genomic DNA... and thus will be stably expressed. You'll find that you loose the plasmid (&expression) over a few passages, because the few cells that do have it integrated do not have a survival benefit.

Exchanging the antibiotic resistance gene really is your best, quickest and cheapest option. I recommend you look at Gibson assembly for this.

An alternative would be to use HEK 293 F cells which are the HEK293FT parental cells that do not express the SV40 T antigen and are not resistant to G418. These are available commercially from Invitrogen, but will prob cost you more than doing the Gibson assembly to exchange the antibiotic resistance.

Your second alternative is to repeatedly cell sort your cells using flow cytometry. However, this only works if you have an antibody that works in flow cytometry and to recover live cells your protein would have to be expressed on the cell surface (you have to permeabilise for an intracellular antigen). Again, getting the antibody and paying for the flow cytometry core facility (if you have to pay?) will be more expensive than doing a little bit of quick cloning.

Hi Petra,

Just to clarify, my aim is to obtain the highest possible amount of protein by expressing and purifying my selected protein (SOST/sclerostin). So I have two options: 1) transfect cells every time I want to collect protein or 2) create stable cell line that will continuously express protein without transfection.

The reason I selected HEK293T cells: SV40 T antigen expression allows for episomal replication of transfected plasmids containing the SV40 origin of replication (the origin is present in SOSTpcDNA3.1+). This feature increases protein expression levels by permitting more plasmid copies to persist in the transiently transfected cells (would this still be true for a stable cell line? If not then HEK293 F cells should be considered).

I assumed that protein expression (detected by SOST ELISA) will indicate whether integration was successful. Now I realise that this is only an indication of transfection efficiency, not stable integration into the cell genome

With this in mind I have to agree that exchanging the antibiotic resistance gene is the best option. I've had a look at the Gibson Assembly Kit you suggested and am surprised that it's not as expensive as other options that I investigated. If both the cells and plasmid are resistant to G418, which antibiotic resistance gene do you suggest I use? I do have puromycin stock in the lab, therefore incorporating the pac gene (puromycin resistance) might be best.

This raises the question, will cells that were successfully transfected (but with no integration) with my plasmid also be resistant to the antibiotic? If only cells with integrated DNA are resistant then no other marker (such as GFP) is necessary? 

Flow cytometry is unfortunately not a possibility at this stage, but thank you for the suggestion.

Timothy,

Which of the options you should go for really depends on how much protein you need.

Making stable cell lines isn't that complicated but does take around 2-4 weeks. Yes, expression does not mean stable integration. However you should get at least a week or two of protein expression out of your SV40ori/HEK293T system.

You are entirely correct about HEK293T cells and SV40 origin. This is not the case in stable cell lines as the integrated plasmid should be replicated as genomic DNA. However, some people say that making stable cell lines using SV40 ori plasmid in T antigen expressing cells is not a good idea. Supposedly there will be genomic instability at the integrationg site as T antigen will lead to replication initiation there... A simple way of avoiding that is to linearize the plasmid with an enzyme cutting in the SV40 origin or get the 293F cells...

Puromycin is a good choice. Other option is hygromycin. Remember to do a titration curve on the required concentration for your cell line beforehand.

Yes, cells with your plasmid not integrated would initially be resistant, however plasmid (-SV40ori) should get diluted out very quickly (down ~1/2 every cell division as both daughter cells should get approximately equal halves). So in general you apply the antibiotic selection about 48 hours after transfection and continue for at least a week. By that time, non-integrated plasmid should have been lost. Usually you also apply a maintenance selection pressure (about one half of the conc of inital selection) to force cells to retain the integration.

You may find that some cells will be resistant but not express your protein - so best to use ELISA to check clones instead of relying on GFP. Although GFP is a good transfection control.

http://www.thermofisher.com/uk/en/home/references/gibco-cell-culture-basics/transfection-basics/transfection-methods/stable-transfection.html

Hi Petra,

Thank you for the quick reply, I really do appreciate it.

I need as much protein as I can get (it will be used in in vivo studies). Based on your comments I'm thinking that transfecting the HEK293T cells when protein expression is required might give the biggest yield. It will definitely give a larger yield than a stable cell line, as the stable cell line will contain genomic DNA and no episomal DNA (thus, overall DNA is less). In this case antibiotic resistance will not matter and I'll save the 2-4 weeks required to establish a stable cell line.

A stable cell line can also be transfected with the plasmid when protein expression is required, but I doubt that there will be a significant difference in protein expression between this and the proposed system in the previous paragraph.

Thus, a stable cell line appears to be unnecessary for my purposes. Would you agree with that conclusion?

Also, I contacted Invitrogen a while ago regarding the HEK293F (Freestyle 293 cells -  I didn't realise they were the cells you were talking about) and they confirmed that they cannot sell the cells to me due to Department of Health regulations. This was quite a disappointment as they were my first choice.

Well, if you need lots and lots of protein a stable cell line is better, as you can just grow as many flasks as you need, harvest and extract protein. Repeated transfections don't come cheap either (unless you use Calciumphosphate and then your transfection efficiency suffers, count in the amount of plasmid you need to make).

An alternative to the Freestyle HEK cells would be the original parental HEK293 cells, I believe you can get them from ATCC

http://www.lgcstandards-atcc.org/en/Search_Results.aspx?dsNav=Ntk:PrimarySearch|embryonic+kidney|3|,Ny:True,N:1000552-1000578&searchTerms=embryonic+kidney&redir=1

but not cheap...

I assume that you have a good way of purifying your protein (Tags or antibody?). Is there a reason why you are not making this protein in E.coli? That would be quickest - ah I see it is a glycoprotein, so needs to be made in mammalian cells...

I would normally suggest having a company make large scale protein as that tends to work out well price-wise here in the UK (considering my wages and the time it would take me to do this as well as the consumables needed), but you say your budget is really limited.

Hi Petra,

Oh, you are correct, in the long run (I was thinking of a single batch, which on its own will very likely not yield enough protein) a stable cell line will definitely be the best option. As you said, multiple transfections will be very expensive.

I've looked at the parental HEK293 cells and the only issue with them is that they're adherent. Both Freestyle 293 and the 293T cells are suspension cells adapted for growth in serum free medium, making them more efficient for protein production.

What I think I'll do: I'll add the puromycin resistance gene to the SOSTpcDNA3.1+ as per your initial suggestion. Then I'll linearise the plasmid by cutting within the SV40 origen and transfect the cells. New media (containing puromycin) will be added to transfected cells 48 hours after transfection. I will of course establish a kill curve to determine the antibiotic concentration that will be used. I will then use SOST ELISA to determine protein expression by the selected clones. Successful clones can then be expanded and used in all future protein production runs.

Regarding purification, we're using Heparin and Cation exchange columns (using HPLC).

Yes, it's a glycoprotein. The protein also needs to be correctly folded to reduce stickiness, therefore mammalian cells should be used. Other researchers have used bacterial cells for production (I've been in contact with them) and suggested mammalian cells should be used. 

Regarding someone else producing the protein for us, do you have any recommendations? I've requested a quote from Creative Biolabs and am still awaiting their response. 

There are lot's of companies making proteins now - my email inbox is spammed by advertising emails from companies that make proteins, but haven't needed to use any of them for years as I mostly work on the molecular biology/cell biology interface now.

I'd contact at least 5 different ones for quotes - they can vary wildly. Just google and you may even be able to find a South African one.

Wouldn't use Genscript again as we had a bad experience with them when we got them to make (or not make as was the case) an antibody.

Origene are selling lysate:

http://www.origene.com/lysate/LY403063.aspx

so they'll definitly be able to make protein for you.

Thank you so much for your assistance, you've really been very helpful.

Thank you for the link, that's definitely the type of company we need. I did some searching and found a couple of places that offer the service (unfortunately none of them were from South Africa) and they'll all receive emails from me before the end of the week.

As an other strategy you can make a stable line with the help of virus infection (off course it will require some cloning but the efficiency is way higher according to my experience) and you'll also got an integration in the genome right away. I'd also think of including some fluorescent protein in virus construct, then you'll be able to FACS cells after infection and in such case you do not need any antibiotic at all. Best, Eugene 

 Eugene, I don't understand why you would spend the time and effort of making a virus to get stable cell lines from easily transfected cells like HEKs? The higher efficiency is not needed as you kill off all the non-integrated cells with the antibiotic selection... not sure how this is meant to save time?

Dear Petra, according to my experience antibiotic selection will never give you 100% of positive cells in the population (due to cell background resistance insertion of resistance gene without gene of interest etc.) on contrary virus trunsduction will give you 80-90% of efficient genome integration right away and in combination with FACS this become even better. Also Starting from virus backbone plasmid virus particles are generated in 3 days, then transduction take 2-3 days (until target protein vill be expressed on reasonable level) so the whole procedure can be done in one week. Antibiotic selection will take 3-4 weeks I would say. But this is off-course my personal choice and can be considered as the other strategy as I wrote. Best, Eugene

Eugene,

Yes you can get clones that have the resistance but don't express your protein. That's why you select 10-20 clonal cells and grow them and check if they do express your protein. In my experience, you maybe get 1 in 10 clones that don't express your gene but are resistant (you of course have to do a control kill experiment to make sure you don't stop selection too quickly and that your selection antibiotic is at the right conc and not lost activity, you may have to change the medium a few times too).

I guess that according to what you are saying (done in one week) you don't bother to select clonal cell lines, but work with mixtures of integrations? How stable are they? You select based on EGFP expression - but cells with integrations that only contain EGFP will have a survival and growth advantage compared to those that do express loads of your protein. In a mixed population you will get overgrowth of the EGFP only cells...

Dear Petra, yes, with virus trunsduction I think it is possible not to bother about clone selection because virus usage guarantee the integration of the whole DNA seq that you've inserted in the backbone (at least I'm not aware that this in not the case). Best, Eugene