Coronaviruses’ visual hallmarks, those nubby protrusions sticking out in every direction, are the keys they use to enter cells. These so-called spike proteins bind to cells—in the case of SARS-CoV-2, human cells—to launch infection. To prevent that from happening, scientists around the world are focusing on the spike to reveal how it works, and find potential weaknesses to exploit.
The spike structure itself is actually made up of three proteins. At the top lies the point at which the viral particle grasps an enzyme on the surface of human cells known as the ACE2 receptor.
This point must be in an “open” or “up” position, flexed and ready to attach itself to the host cell receptor, says Rommie Amaro, a biophysical chemist at the University of California, San Diego. An animation posted online by Greg Bowman, a biophysicist at Washington University School of Medicine, reveals that this looks like a creature opening its jaws.
But in many illustrations that have been made so far of the SARS-CoV-2 spike structure, an important feature is largely missing, says Amaro.
The structure is covered in sugars known as glycans. They’re thought to disguise the virus to the human immune system, making it seem just like a harmless cell, given that these are also often coated in sugars. In diagrams and 3-D models of the coronavirus’s spike structure, the glycans are usually represented as small, stubby nodules, but they’re actually fuzzier and more obstructive than that, says Amaro.
“They protect it literally like a physical shield,” she explains. The glycans are so protective, in fact, that the spike protein might have to flex up and out simply to reach through them and bind to ACE2 on human cells. Any antibody targeting the spike structure will have to slot in between the glycans and attach to the spike protein itself.
“If you know where the holes are, that’s where you want to target,” says Amaro.
In order to model what sort of defensive coating glycans give to the SARS-CoV-2 spike, Amaro has used mass spectrometry data on the spike protein, which was recently published in a preprint published on bioRxiv March 28. This revealed the location of the glycans in more detail than was previously available.
An animation Amaro posted online recently shows how the glycans coat the spike itself and also jiggle, which may further affect their ability to keep antibodies off the structure. In the video, the spike sticks up from the virus’s lipid membrane, shown in pink.
Similar work is being carried out by Chris Oostenbrink, an expert in molecular modeling at the University of Natural Resources and Life Sciences in Vienna. He explains that his method involves using a database of glycan shapes and matching them to what is known about the shape of the coronavirus’s spike structure—a molecular jigsaw puzzle.
“Basically, we put them all on, we look at which ones fit and we take the best-fitting model as a representative example,” he says. An illustration by his graduate student, Jan Walther Perthold, shows the glycans as fuzzy pink blobs covering the spike structure. Like Amaro’s model, this reveals just how prevalent the glycans are and that they present a serious obstacle to any antibody that might otherwise bind with the spike.
By mapping the glycan shield, scientists should find it easier to find the right antibody peg to slots through the holes in it, says Amaro. A vaccine could be designed, for example, to provoke a person’s immune system into generating antibodies that would successfully latch on to the spike structure and disrupt its opening mechanism or otherwise prevent it from attaching to ACE2. Think of it like jamming a wrench into a machine to stop it working as intended, suggests Amaro.
Our first complete model of the SARS-CoV-2 Spike protein in complex with the human ACE2 receptor, with full complex glycosylation added. @PertholdJW made this model, using methods developed by @AysegulTurupcu pic.twitter.com/R1tqi9t9Qc— Chris Oostenbrink (@mms_boku) March 27, 2020
A study published in Science on April 3 reveals that a particular antibody, CR3022, can bind to the SARS-CoV-2 spike protein. This antibody was isolated back in 2006 from a patient who had recovered from SARS and is much better able to target the SARS-CoV virus, which caused the SARS outbreak of 2003, instead.
In lab tests, the researchers mixed the antibody with either SARS-CoV or SARS-CoV-2 in vitro. The antibody failed to neutralize SARS-CoV-2, suggesting it doesn’t bind as well to the new virus. It’s an older weapon, after all, not specifically adapted to the SARS-CoV-2 target.
The authors of the Science paper write that they think the antibody might still be effective against the new coronavirus in vivo but more experiments are needed to show that.
It’s interesting work, says Jeremy Rossman, a virologist at the University of Kent who was not involved in the study. He notes that the data show how the CR3022 antibody binds at a location slightly below the point where the SARS-CoV spike protein binds to host cells. This means it clearly does not function by physically blocking binding.
“It’s not one-hundred percent clear how this antibody neutralizes and stops the virus,” he adds.
More promising, perhaps, is work from researchers who found that antibodies isolated from a llama could neutralize SARS-CoV-2 in vitro when those antibodies were fused together with human antibodies—a sort of two-pronged attack. In a preprint published on bioRxiv March 28, the team behind the work write that the specially engineered dual antibody works by grabbing hold of the binding site at the tip of the spike structure that targets host cells, effectively blocking it from infecting them.
The authors also propose that a treatment could be administered via a spray for patients to inhale. That way, the antibodies could be inhaled “directly to the site of infection.”
There are yet other approaches. A number of pharmaceutical firms have launched projects to develop laboratory-cloned antibodies. GlaxoSmithKline, for example, is using antibodies isolated from a SARS patient to find out if they could be effective against COVID-19.
Antibodies or other molecules could also hamper SARS-CoV-2 in other ways that target the spike structure. They could, for example, prevent the protein furin in the human body from interacting with the virus. This would be useful because researchers have suggested that furin helps the spike structure’s two subunits to separate from one another—a process that allows the virus to break open and enter host cells. Furin happens to be abundant in the human body, meaning that we provide an ideal environment for SARS-CoV-2 to infect us. A molecule that separates furin from the virus could stop the pathogen in its tracks and some teams are currently evaluating whether a furin inhibitor could do this.
Whatever we use to target the spike protein, we need to be careful, says Akiko Iwasaki, an immunobiologist at Yale School of Medicine. Scientists must determine which antibodies bind to the spike structure but also ensure that they don’t also trigger a negative immune response. She points to a paper published in JCI Insight last year that showed how antibodies in macaques infected with SARS-CoV, the virus that caused the SARS outbreak of 2003, could sometimes exacerbate the disease rather than quell it.
In the macaques, an anti-spike antibody stimulated blood cells called macrophages to cause inflammation in the primates’ lungs. The authors noted that patients who died from SARS had similarly inflamed lungs.
“I fear that might be what’s going on [with SARS-CoV-2],” says Iwasaki. “The real bad types of disease don’t occur for about two weeks. That’s when the antibodies come up.”
If scientists identify antibodies that don’t trigger a dangerous immune reaction, it might be possible to provide them to infected patients to help them overcome COVID-19, says Rossman. But it would be even better if we could find, for example, a peptide that prompts the production of such antibodies to immunize individuals before they catch the disease.
Chris Baraniuk is a freelance science journalist based in Northern Ireland. Find him on Twitter at @chrisbaraniuk.