Secondment: Droplet Sorting in Cambridge

Today’s post was prepared by ES-Cat ESR Berndjan Eenink

A few weeks ago, I went to the Hollfelder lab for a week as part of my ES-Cat secondment. I plan to visit for a longer time (around a month) later in the Summer.

For my project, I use microfluidic droplets to screen large amounts of library variants. In our lab we can do most of the upstream and downstream work, but for microfluidics know-how and equipment I rely on our collaborators in the Cambridge group in the network.

The plan was to go for a short visit and to do some sorting on one of the libraries that I made to test if the whole process works, and if not, what I need to fix and/or improve before coming back for a longer visit in which I will sort through all my libraries.

In addition, we wanted to test a different method to express our cells for droplet sorting. Up until now we had encapsulated a single cell in each droplet. This works but has some issues. One of them is that cells seem to clump together after expression starts. This makes encapsulation tricky, as the cells have the tendency to stick together and form a biofilm, rather than moving neatly across the microfluidic channels as they ought to. Another problem is that after sorting the cells, the recovery is below 100% (not every cell that is sorted grows back.)

Growth and expression in droplets could potentially fix both of these issues. If no enzyme is expressed during the droplet formation, the sticking together is not an issue, and once expression starts the cells are free to stick together in the droplet however much they want. The recovery could also be helped, if each droplet contains (50-100) cells, 10-20% recovery still means getting back multiple copies of each variant. The reason to use microdroplets in the first place is to link the phenotype to the genotype (there is only one E. coli cell expressing a variant per droplet). Having many cells in each droplet may then sound counter-intuitive, but since a single cell is encapsulated, every cell that will grow inside the droplet is identical to the other cells in the droplet, thus phenotype and genotype are still linked. The advantage is that now we can oversample at a step where it is not that much extra time, instead of oversampling by sorting 5-10 times longer or with 5-10 times less coverage, now we can oversample by simply plating out more of the sorted cells (which is only a little bit of extra work in the evening).

Droplet sorting setup ready for action (as soon as the light goes out)

The visit as part of the secondment would started on a Tuesday. I arrived in Cambridge Monday afternoon, got to my airB&B and prepared a bit for starting the next day. Since we only had 6 days to make all droplets, do incubations and sort, the schedule was quite packed. I arrived on Tuesday morning to the lab and started preparing buffers, substrates, and whatever else was needed. I also got the samples that I had sent on dry-ice shipment a week ahead of me.

Immediately I (re)discovered the biggest problem of working someplace new. Where do these guys keep their Eppendorf tubes? Where is the medium? In which cupboard is this component of my buffer? Where are my tips? What is shared and what do I need to prepare myself? While I like to believe that as I get further along in my education and career I’m getting better at coming up with solutions to scientific or practical challenges, the answer to ‘where is this particular piece of material stored in this lab?’ remains as insurmountable as when I stepped into a lab for the first time.

The other thing that does not change when coming to a new lab (though this one pleasant, if at times a bit overwhelming) is meetings lots of new people and hearing about the stuff they are working on. I knew the people from the network, and some from previous visits during the network meetings and my previous sorting visit, but there were some new faces, as well as lots of faces I had to remember. ††

Droplets under the microscope

Since there was a limited amount of time, we set up both the envisioned new method and a repeat of the previous method as a comparison. This would also let us fall back to a method proven to be mediocre if the method thought to be great turned out to be dismal. We could look at some of the droplets under the microscope and confirm that something was actually growing. Over half were empty (they should be, as a side effect of preventing a lot of your droplets containing multiple cells), but the rest were all ‘teeming’ with cells. Some clumping (aforementioned problem) was observed, but since the cells are already in droplet this should not affect the encapsulation anymore.

A droplet being sorted

We did the screening at multiple timepoints, and using two different protocols for droplet growth and recovered the sorted cells, which I plated on LB-agar. The sorting seemed to work, and in all cases at least around 10 cells per droplet (based on calculations and estimations) were recovered. Due to the packed time-schedule I did not have time to do follow-up experiments to verify the recovery while in Cambridge, but luckily all the sorting work results in just a few Eppendorf tubes worth of material per sorted library, so these would be sent after me back to Munster, where I could do the follow-up experiments to verify that we actually sorted what we thought.

On Sunday, I managed to wrap everything up and get ready for the first train Monday morning to make it to the plane back to Germany. That Tuesday, I was scheduled for a regular research update for my group (either ideally or horribly timed, depending on which way you look at it). Since then, I did follow-up checks on the cells were recovered and found out that one of the protocols for the new method worked quite well, although there are still some issues that could be fixed to improve efficiency.

All in all, it was a short and busy but both productive and enjoyable stay.