In the last post, we discussed just manipulating proteins with light is such an impactful field for research. Optogenetics can help us better understand the biochemistry of proteins, including proteins whose malfunction is associated with human health problems. For example, cancer, diabetes, and obesity are all connected in some way to the activity of a protein tyrosine phosphatase enzyme (PTP; in this paper, they look specifically at the PTP1B variant). Thus, there is a lot of interest in producing optogenetic tools for digging deep into how they work.
The first step of this experiment was to build and optimize a PTP1B protein so that it becomes light sensitive. Proteins that respond to light already exist; think, for example, plants. Phototropins are proteins that allow plants to respond to light (you know how those flowers on your kitchen windowsill always face the sun outside, and never towards you so you can see them? That’s because their phototropins sense where the light is coming from, and the plant grows towards that stimulus. There’s also a fun experiment you can set up to watch this in action.)
Phototropins respond to light because of their LOV2 domains, so the scientists here fused LOV2 to a PTP1B enzyme. Each of these proteins is composed of multiple parts (each part having a helix), including α-helices such as the Jα helix, α7 helix, and A’α helix. To bind the proteins, the scientists fused the α7 helix at the end of PTP1B to the A’α helix at the front of LOV2. This build, and its subsequent optimization, is depicted in Fig. 1 of the paper.
The scientists measured how well these chimeric proteins worked by measuring two things: reaction rate, and the dynamic range. The reaction rate (V) measures how fast the proteins work on some molecule (a phosphatase like PTP1B takes phosphate groups away from its substrate; we’ll return to this later.) In the case of their chimera, the protein functions in the dark. Turn a light switch on, and the A’α helix unravels, deactivating the protein. If you turn the light off again, the protein can re-fold and continue working.
Thus the dynamic range (DR) tells you how much the chimera changes in response to light; it tells you how light-sensitive it is. In this case, the scientists cared more about their chimera being more light-sensitive than how fast it worked, so they used that as their standard when picking a construct to build onto for round #2. (The winner for round #1? Construct 7.)
In their second build, they hypothesized that stabilizing the protein would optimize DR. To do this, they shortened the ‘bridge’ connecting PTP1B to LOV2. However, this hypothesis proved to be incorrect, and most new builds showed reduced DR. One build in this round, however (Construct 7.1), stayed at the same DR, but also showed an increased rate. Thus, they optimized this chimera further.
Another approach to stabilizing a protein is to introduce a few mutations and see if anything works better than before. A mutation on the A’α helix turned out to improve the DR the most, with an added modest boost in reaction rate. This kind of random mutagenesis is actually pretty common in bioengineering-based experiments: there’s often just so much we don’t know about how a genetic sequence relates to function, so it makes the most sense to just randomly insert mutations and check if anything happens to help.
After all this, they now had a chimeric, light-sensitive enzyme that they could test further. Chimeric because it was a hybrid between different proteins, and light-sensitive because they could measure differences in protein activity when exposed to light or darkness.
What’s the next step? Well, why not take your freshly engineering protein for a test-drive? Fig. 2 sums up the data explored in this part of the experiment, in which they test its biophysical properties.
First, they ran a couple experiments to see if any protein parts were extraneous. They fiddled with both the A’α helix and the Jα helix, the two major components of LOV2. They found that both parts are necessary if LOV2 is going to make PTP1B light-sensitive.
They hypothesized that the primary reason driving light sensitivity is a structural change between light and dark states. To test this, they took spectroscopic measurements of their enzymes when they were active (dark state) and when they were inactive (or supposed to be inactive; light state). This kind of spectroscopy can read the actual shape of the protein. In the world of proteins, shape is often a driving force behind activity.
Sure enough, enzymes with a high DR changed shape a lot between dark and light states – just as predicted. Of the low DR enzymes, however, some of them had no change (as expected); but some of them did! Thus, there must be something else that contributes to light sensitivity than just change in shape. LOV2 unravels under light, altering protein activity; what is the second factor?
This second factor turned out to be how quickly each half of the chimera recovers after light exposure. In the low-DR, high-shape-shifting enzymes, LOV2 recovers faster than PTP1B, and this appears to disrupt function and light sensitivity. The enzyme with the smallest different between LOV2 and PTP1B recovery? That would be their winning, most light-sensitive construct.
In other words, their chimeric protein included both LOV2 and PTP1B protein domains. If both change their shape when exposed to light, that’s good for their purposes. If both recover at the same rate, that’s even better. If they recover out of sync, the protein is not likely to actually work properly.
Bringing it into a Cell
Up to now, the scientists had been working with proteins in test tubes. In a living cell, PTP1B actually works in conjunction with the endoplasmic reticulum (ER). That means, in the cell, it has a tail of extra amino acids that binds to the ER.
The scientists wanted to check if this extension would reduce their chimera’s photosensitivity, so they tacked on this tail and measured V and DR. Although the enzyme remained photoswitchable, it did appear to reduce its activity (less dynamic range/DR was observed).
The Test: Can We Photocontrol the Protein In Vivo?
Now it’s time to translate learnings into some sort of tool. We have a chimeric enzyme that, in vitro, reacts to light. Will it do the same thing in a cell?
To do that, you’re going to need a biosensor, and biochemists often use FRET (Förster Resonance Energy Transfer) to test if some biochemical reaction is going on. A FRET sensor uses a substrate domain packaged in between two fluorescent proteins in which one activates the other. So if your substrate is phosphate, whenever phosphate binds the resonant energy goes down and the two fluorescent proteins stop glowing so much. This is a detectable output that tells you if there are enzymes in the cell attaching phosphate groups to substrates that bind them.
The scientists identified a biosensor that uses FRET to detect kinase activity. A kinase reaction is the opposite of a phosphatase reaction; phosphatases like PTP1B take away a phosphate group, but kinases tack one on. Thus, if a kinase biosensor reduces fluorescence when phosphate groups bind to it, that same biosensor will glow more when phosphate groups are taken away.
That is, if PTP1B is working, you get more glow.
This can be used to test if their chimera actually turns on and off under dark and light states. In Fig. 4, you can see the results confirming their hypothesis – the scientists successfully built a PTP1B enzyme variant that is light-sensitive and can be successfully photocontrolled, inside a cell.
Such an enzyme opens up a new tool for understanding PTP1B activity in cells, and thus, their role in human diseases.
Hongdusit, A., Zwart, P. H., Sankaran, B., & Fox, J. M. (2020). Minimally disruptive optical control of protein tyrosine phosphatase 1B. Nature Communications, 11(1). https://doi.org/10.1038/s41467-020-14567-8