Welcome to Science with Shrike! Today we’re going to discuss the recent Project Veritas video of a Pfizer employee. We’ll use this to talk about what directed evolution is, and general approaches one might take to altering the SARS-CoV2 Spike protein. There are a few other updates to be covered in another post.
Directed Evolution
Everyone is about to become an expert on gain-of-function research and directed evolution, thanks to Pfizer employee Jordon Walker and Project Veritas. The most salacious clips are posted by Project Veritas:
On one hand, the guy could be trying to impress a date (or a PV plant). On the other hand, directed evolution is a reasonable approach to improving the efficacy of proteins. So what is directed evolution?
Directed evolution is a cool way to say that we will be lazy and let biology do the optimization work for us. In directed evolution, you take your biological target, expose it to something that mutates it, and then screen for the mutants that have your desired phenotype. After several rounds of this, you look at your biological target and see how it has changed.
The idea is that the mutagen does the work for you, and you just select the winners. It’s way faster and easier than trying to mutate every amino acid in a protein that has 300+ amino acids.
One example of directed evolution is the creation of new fluorescent proteins not observed in nature. The reason for doing this is that we have different needs for these fluorescent proteins than the organism they were taken from. The lab of the late Roger Tsien took the fluorescent protein DsRed and used directed evolution to generate new proteins mCherry, mPlum and others.
Aside from being able to paint cooler bacterial art, directed evolution helped improve the fluorescent properties of DsRed for research purposes.
This example also serves to show that directed evolution does not automatically equate to ‘gain of function’ research. ‘Gain of function’ is modifying a pathogen in a way that improves its infectivity or virulence. In the case of mCherry, this protein was never introduced back into the original organism, nor was the original organism (a Mushroom Anemone) a pathogen in the first place. The pathogens to which we add mCherry do not gain any advantage in virulence or infectivity. They just glow in the dark and are easier to see with our microscopes.
Gain of function can also be done without using directed evolution. For example, the Nature Medicine article from Zhengli Shi, Ralph Baric and others did gain of function research without directed evolution. Gain of function concerns with that work were discussed in 2015, following on the US moratorium on gain of function research. Thus, gain of function and directed evolution are two separate things. Gain of function adds to a pathogen’s capability, while directed evolution is a method of improving a biologic.
Directed evolution as a method could be used to perform ‘gain of function’ research. However, in this case, one of two things would need to happen. Option 1, the whole pathogen would be mutated, and variants selected for improved infectivity, virulence, immune evasion, or mortality. Option 2, one part of a pathogen (e.g. one protein) would be mutated, selected for improved binding, function, altered immunogenicity, AND put back into a pathogen.
If neither of these options apply, it’s not gain of function research, even if directed evolution is being used to change a biologic, even if that biologic came from a pathogen. The biologic modified needs to be modified to make the pathogen worse, and go back into a pathogen. Thus, work that deletes key virulence genes in a pathogen is not ‘gain of function’ research. Similarly, taking a protein out of a pathogen, and mutating it isn’t enough to be gain of function. One example of pulling a protein out of a pathogen, mutating it, and it not being gain of function is CRISPR. CRISPR technology relies on the protein Cas9, which was originally isolated from flesh-eating bacteria. But modified Cas9 does not go back into the flesh-eating bacteria, and modified Cas9 does not improve virulence.
Directed Evolution to Fight SARS-CoV2
With this understanding in hand, let’s look at how Pfizer might go about ‘directed evolution’ to improve the SARS-CoV2 vaccines. Recall that the vaccines only target the Spike protein of SARS-CoV2. That means the whole virus is not needed, just the Spike protein. The Spike protein is what lets the virus enter the cells. Since SARS-CoV2 is a Risk Group 3 pathogen (=more expensive and more nuisance to work with), initial screening will be done with a harmless virus, or virus-like particles. A relatively harmless virus (vesicular stomatitis virus (VSV) or the Moloney murine leukemia virus (MMLV) would be Shrike’s picks) has its version of a Spike protein replaced with the SARS-CoV2 Spike protein. Now this new virus relies on the SARS-CoV2 Spike to enter cells instead of its normal protein. This is called a pseudotyped virus. This is not considered gain of function because enabling infection of human cells alone is not enough to make a super-virus (plus VSV already can infect human cells).
Autist note: Some things that a lay person might logically call ‘gain of function’ also do not fall under the ‘gain of function’ rubric. For example, giving bacteria antibiotic resistance is normally not considered gain of function, even though it would make them harder to treat. The reason for this is that giving antibiotic resistance to bacteria is part of how we force genes into them, we have 20+ classes of antibiotics with which we can target bacteria, and most of the bacteria used are lab strains unable/unlikely to cause a pandemic if they escape into the environment. Similarly, adding toxins to relatively benign bacteria is often ok because it’s not enough to make them into a pathogen of concern. Requiring labeling of this work as ‘gain of function’ would add lots of cost in red tape and drown out the true gain of function research (let’s make a super-flu). In general, we need to reasonably expect the outcome is going to create something that would be a major problem for humans if it escaped… like a helping a new bat virus infect human cells, or adding immune evasion or other key virulence factors to clinical isolates of a human pathogen. There are also some pathogens that get more scrutiny than others due to their risk.
Once you have your new virus with Spike, you can screen for mutants that can escape vaccine control. The way Shrike would do this would be to start with the Omicron Spike protein, mix a lot of the pseudotyped virus with vaccinated human patient serum (which has neutralizing antibodies), and infect tissue culture cells like HEKs. Isolate the virus produced by those HEKs, mix with more patient serum and repeat ~10x. Then sequence the Spike protein to see how it has mutated to avoid neutralization. Compare these mutations to the ones arising in nature to see how well the directed evolution is working. In the case of RNA viruses like VSV or MMLV, their replication process is sloppy and introduces mutations during normal replication.
The challenge with this approach would be the in vivo testing to show the new vaccine works. One option would be to have most of the legwork (structure, mRNA production, expression, etc) done, and wait for the new variant to emerge. Once a new variant with the same Spike protein emerges, it’s not gain of function work because the virus already gained that function in nature. A second option is to immunize mice with the old vaccine and see if the new purified Spike protein (without virus) stimulates an immune response. A third option would be to infect mice with a pseudotyped virus, and measure immune responses to the pseudotyped virus. A fourth option would be to see if the new purified Spike protein (or a pseudotyped virus) is recognized by antibodies from human blood and/or stimulates human blood cells from vaccinated patients. In these scenarios, Pfizer could put it on the FDA to decide if they need to put the modified Spike back into SARS-CoV2 for full approval or not. One final option would be to put the mutant Spike back into SARS-CoV2 and test the new virus in mice. This option is when the work may become potential gain of function research.
One could argue that adding a new Spike that can evade antibodies back to SARS-CoV2 is potential gain of function work. The counter argument would be if the mutant Spike does not improve infectivity, it is not ‘gaining’ infectivity or virulence, and that people with immune responses to other SARS-CoV2 components are still expected to be resistant. Generating an attenuated virus (which is how SARS-CoV2 is now escaping immune control) is not gain of function because the virus is now weaker, but may confer immunity to the more dangerous versions. Which way it gets decided will come down to the regulatory body’s preference for the arguments. Expect if it gets decided that this is NOT gain of function for many to claim Pfizer bribed or unduly influenced the regulatory body. However. The arguments above show that a regulator could fall on the side of ‘not gain of function’ even without Pharma leaning on them.
Wrapping it up
As mentioned at the beginning, directed evolution is a reasonable approach to changing the Spike protein, regardless of the motivation behind making these statements. However, directed evolution on its own is nothing scary. Instead, it is a method widely used to optimize protein function.
With any protein biologic bringing in lots of money, a company is going to mutate every amino acid in the protein to learn how it changes protein structure/function. Normally, each position would also get additional modifications (add lipids, sugars, etc). Since this is intended for an mRNA vaccine, additional modifications are not on the table because the mRNA cannot encode those. New mutations represent potential new intellectual property that could become new patents. So directed evolution or not, assume every Pharma company with a vaccine is modifying (or has modified) the SARS-CoV2 Spike protein to learn how it changes, and to see if they can make a better vaccine target.
Changing Spike protein alone is not gain of function research. It would have to be put back into the virus, and the change would need to help out the pathogen for it to be gain of function. Given the cost, nuisance, and safety risk, creating a new SARS-CoV2 is not something that would be undertaken lightly. Most likely, changes to the Spike protein (whether by directed evolution, or human-planned mutagenesis) are being studied in isolation, and/or with pseudotyped viruses. A lot of the initial work can be done without a full virus, and that would be cheaper, faster and easier to do. No gain of function needed.
Another consideration is that if the SARS-CoV2 vaccines go the same route as the flu vaccines and become an annual deal, the CDC will be responsible for choosing the virus strains to target. The CDC will distribute those strains to the Pharma companies for testing and development. That will render some of the discovery moot, but not stabilizing/other structural changes to improve vaccine efficacy.
Notice that this discussion does not require a debate on the pros and cons of gain of function research. It also does not require assuming good or bad about Pfizer. Overall, the video engages your emotions. Those engaged emotions made the video a successful fundraiser for Project Veritas. No hate to Project Veritas for raising money, but from the videos, Shrike would not expect any hearings, etc to uncover misconduct by Pfizer. Though if you’re a Pfizer employee, expect training on how to avoid Project Veritas stings.