[Christopher Bahl]

hi everyone, we recently started thinking about a problem that I’m sure lots of folks here have already thought about.

We’re growing E. coli expression cultures in 96-well deep well plates, and (unsurprisingly) we’re having a tough time getting enough oxygen into the cultures. We’ve reduced the volume of liquid in each well, and we have a shaker with a high RPM and small orbit.

Still, we’d like to get better aeration and would prefer to avoid using 24 well plates. We were wondering whether bubbling air into the wells of the plate – possibly while they’re in the shaker – would be a helpful approach. Has anyone tried this before, and if so, was it successful?

I spent some time on google but wasn’t able to find anything for sale to do this, so we’ve been considering designing and 3D printing our own 96-well bubbler manifold to sit on top of a 96-well deep well plate.

Any advice or suggestions would be greatly appreciated!
-Chris

[Eric Geertsma]

Hi Chris,
we manage to get quite okay densities (>20) in 96 well plates with 700 µl cultures in TB, a gas-permeable seal, and 400 rpm with an orbit of 12.5 mm.
I have experimented with bubbling (home-made) and it was a total nightmare to get equal aeration. So please reconsider the shaking again:-)
Best of Luck,
Eric

[Renaud Vincentelli]

Hi Chris,
like Eric we had no main issues to get high OD in DW96 (900-1000 rpm in small orbital Infors), but we switched back to DW24 years ago (I can fit 48 in the Infors) as I was not always happy with the reproductibility when scaling up and as this gives more biomass and therefore more proteins and better sensitivity. We still can process several thousands cultures/week. Do you purify afterwards?

Renaud

[Christopher Bahl]

thanks Eric and Renaud!

We’re also able to get to high OD with the DW96 in our infors shaker too, but we recently got an interesting result that has me thinking that our cultures aren’t getting enough oxygen.

We are benchmarking our protein production methods using 48 different (mostly) random proteins that we pulled out of the PDB (the proteins were all made in E. coli by the crystallographers). When we prepped these proteins in DW96 format, we got a wide array of yields following purification. Some of the proteins gave a good yield (~150 ug protein per mL of culture), others gave little to no material following purification, and some were somewhere in the middle. Next, we picked 2 different proteins from these “high”, “medium”, and “low” bins and re-prepped the protein from 50 mL cultures using baffled erlenmeyer flasks. To our surprise, we got great yields for all six proteins – even the two “low” yield proteins. This led us to conclude that we need more oxygen in our cultures in the DW96. What to you think?

thanks again for the advice!
-Chris

[Renaud Vincentelli]

Hi Chris,
I think the answer to your problem could be aeration but I think that it could be many other reasons around the purification.

We do now mostly culture in DW24 (2 ml to 24 ml/targets, see turchetto et al, 2017 for instance) or flasks (mostly 350 ml/2 litres flasks) and our purifications on Nickel are either 384 (for PPI), 96 or 24 all this on Tecan or AKTA for the 350 ml scale.
When we have “few” targets (24 or so) we do first 350 ml cultures in our default best (because we can Nickel purify this in one day on the AKTA), at the end of the culture we harvest the equivalent of 24 ml and purify these 24 aliquots in our Nickel 24 to see which ones are worth purifing in the big scale and check if the proteins are in the wash or elution.
In most cases we ge the same yield and gel profiles (pure protein at the good MW) at 24 ml purification and on the AKTA Xpress BUT in some cases we get very different yields/not the same gel profile (degradation of the main band for instance in one case and not the other one…)/not the protein in the same fraction (wash versus elution). This is from the same culture and it usually comes from variables that are hard to control; time of purification/ local concentration on the proteins on the beads that could lead to agregation for instance, time of contact between the proteins and the beads which palys a role in the final yield… This mostly come from what I call the “sick” proteins; we usually realise afterward that either the construct/ the buffer… Was not the best and this led to agregation/proteolysis. usually chnaging the constructs/buffer for these proteins make the work but we sometimes stick to multiple analytical purification that works and pool them, this is quicker.

My 2 cents.

I guess that you did validate as well your 96 cultures/purifications with a single protein 96 times. People use GFP/mCherry for that, people that do that realise that depending on the choice of buffer, beads, culture volume versus volume of beads/trademark you sometimes get very “bad” results (the beads are still colored and you can’t elute the whole protein without GnHCl, you have you colored protein in various fractions… This tells us that things are not as simple as we think they are…

Renaud

[Nick Berrow]

Hi Chris,
For sure aeration can make a difference but there could be many more explanations (I’m sure you have most of them covered already) :
Starter culture ‘age’, density and dilution factor.
Media buffering (lack of) although I’m pretty sure that you are already using TB or similar.
Plasmid/antibiotic stability.
Protein toxicity (full pre-induction repression e.g. with glucose especially with non lacIq plasmids can give you all or nothing results in terms of expression).

We always found cultures in 96-well format inconsistent so we use 5ml in square well 9ml/well capacity 24DWP in 25mm orbit Kuhner shaker and then re-format as Renaud does (although we don’t all necessarily have the luxury of robots to do the re-formatting : )). The majority of the time this gives us fairly reliable data wrt scale up but there are always exceptions. In most cases we are changing lysis method, affinity media etc. that can still have an impact on extraction and purity when we move to large scale.

What orbit and speeds does everyone else use for 96 DWPs?

I would avoid aeration by bubbling if possible as preventing cross contamination is really difficult and unless you are a completely phage-free lab it can just turn your incubator shaker into a phage factory.
Do they still make the HiGro shakers that have stacked small orbit shakers and air/O2 supply over the wells? Air should be enough but if you wanted to use 02 in most places it also requires extra health and safety measures that can make things even more complicated.
Best
Nick

[Jan-Erik Hoffmann]

We use a Multitron shaker, 25 mm orbit, 225 rpm. For expression tests we use 24 well plates (3 mL culture), but for plasmid preperations we also use 96 deep well plates with 1.1 mL cultures. So far we didn’t observe problems with these conditions.

[Christopher Bahl]

thanks everyone for the help and troubleshooting!

for all the samples, we’re using autoinducing TB for the media, kanamycin selection with a pET vector, and Lemo21(DE3) cells for the expression strain

I hadn’t thought about the potential for an aeration manifold to turn our shaker into a phage factory – yikes! thank you Nick for saving us from that

We’re currently running an experiment to check and see if culture aeration is impacting soluble protein yield with several of our test proteins (via reduced culture volumes and larger wells / flasks). I’ll post an update once we have some results to share

thanks again!

[Author]

Posts