Science – it works! Or does it?

“Science – it works!” – Every time I see somebody wearing a T-shirt with a print like this I don’t know how to feel. In times of increasing scepticism towards established and well-studied matters ranging from global warming to vaccination, it is great to see people standing up for science and advertising its successes on their chest. At the same time, I jokingly think to myself that surely anyone who has ever done research can confirm that science does, in fact, not work. Sometimes it may even feel like the most reproducible thing about science is its ability to lead young researcher to the brink of despair. In this blog post, I would like to share my own experiences after one year in grad school, and how my greatest challenges so far were those I expected the least.

Adapted from: http://www.twisteddoodles.com/image/86414780702

DNA origami exploits the predictability of Watson-Crick base pairing where A pairs with T, and G pairs with C. We use hundreds of short single-stranded DNA (ssDNA) snippets to fold a long ssDNA loop into complex 3D structures. It’s really cool – until you have to admit that most DNA nanostructures created so far do little more than looking pretty under the electron microscope. To change that, I started my PhD at the Dietz Lab in Munich with the ambitious goal of giving function to existing DNA nanostructures. As a biotechnologist I am naturally interested in manipulating biological systems, and thus I have set the goal of using DNA origami to manipulate lipid vesicles which are essentially hollow bubbles of grease similar to biological cells, just a lot simpler.

Naïve as I was, I initially gave myself half a year to finish this endeavour. After all, I am using an existing nanostructure, and binding DNA to lipid membranes had also been achieved before. The plan was clear: Modify the structure a little bit so it can bind to lipid membranes, and then use imaging techniques to check if it worked. The only challenge would be getting it to do something with the membrane aside from just binding to it. Needless to say, my first experiments failed.

When I looked at my samples under the transmission electron microscope, I did not only not see what I had wanted to see, but what gave me an even greater headache was that I couldn’t tell for sure what it was that I was seeing. Some spots were bright and crisp, others were dark and blurry. Sometimes the DNA nanostructures colocalised with the bright spots so one could think they represent lipid patches with DNA origami bound to them. On the other hand, I had no proof that this was the case also in solution, before imaging. Confronted with these problems, I began to see increasingly more shortcomings in my experiments. Aside from the many things that could go wrong during my experiments, the contrast generating staining procedure prior to imaging or the high vacuum inside the microscope can distort results as well. The more I thought about it, the more I realised how many things needed to be optimised or validated. The devil lies in the detail, as they say.

The internet is full of comics covering these classic hardships of a graduate student. One that I find particularly fitting portrays research as a hike through the mountains, just that all the ropes, bridges and ladders have to be set up by yourself.

Image source: https://www.kdnuggets.com/2016/10/big-data-science-expectation-reality.html

There is no point in trying to construct a ladder near the peak if you’re still stuck in a ravine at the foot of the mountain. In an attempt to organise the chaos inside my head I decided to tackle the most basic problem I could think of, and that was validating my vesicle production protocol using fluorescence microscopy. By keeping it as simple as possible, I soon after had a direct proof that, indeed, my sample contained giant lipid vesicles. Woohoo! I proceeded in a similar fashion and step by step ensured that everything behaved the way I wanted. This included learning how to use a confocal microscope as it is superior to electron microscopes when it comes to imaging giant lipid vesicles in solution. Of course, learning and applying a new technique came along with another series of failed experiments, but each of these gave me hints that let me understand what was going wrong.

After a few sessions at the confocal microscope I realised that the fluorescent dyes I was using to visualise the lipids and DNA origami independently were not detected completely independent of each other. I figured that the dyes I had chosen may be interacting, so I decided to use different dyes for my next experiment. At the same time, I hypothesised that the DNA strand that acts as a lipid membrane anchor to bind the origami to vesicles may be dysfunctional, so I decided to replace it for a new one. And voilà: Suddenly, the problems that had kept me busy for so long during my first year of PhD study were solved. I had a direct proof of specific binding of DNA origami to lipid vesicles, and I could also proof that the membrane anchor is mediating the binding. This by itself is not a ground-breaking scientific discovery, but for me it marks an important milestone on the way towards my goal. Albeit annoying, I feel like hurdles like these made me a better scientist by teaching me how to systematically overcome them. Sometimes all it needs is dedication, patience and a little bit of luck. Sooner or later things will turn in your favour, and perhaps I will now advance faster with my project. I’ll give it my best. 🙂

Science – it (finally) works (at last)!

Michael, PhD student at TUM