Sagi Shapira: Experimental and computational interrogation of CoV protein functions (covid-19 symposium)

Summary by: Arooba Ahmed (CC '23)

     Dr. Shapira’s project started a few years ago, when a collaborating lab developed an algorithm called PrePPI. It takes protein structure, gene expression, anontions and functioning to determine whether the proteins interact or not. Using this they developed a method called P-HIPSTer that allows them to use structure and modelling to map interactions between pathogens and the host. 
     They used it to map viruses and what cells they infect, and ended up with a matrix of 1300 viral proteins by 20000 human proteins. Their experimental validation rates are close to 80% for true positives and true negatives, which is very high. Using the algorithm, they were able to discover novel biology about human-virus interactions. 
     One way viruses can manipulate the human host is through mimicry, which means things are structurally similar but sequentially divergent (only 5-10% sequence similarity). Dr. Shapira’s group studied the viruses to see if they use the structure spacers (or language) of the host. They found that the viruses infecting plants have proteins that look like plants, those infecting humans have proteins that look like humans, and those that infect bacteria have proteins that look like bacteria. Arboviruses (which infect both mosquitoes and humans) use structure spacers of both mosquitoes and people. The viruses then design novel cell circuits to manipulate cell behavior and cell functions. 
     They also looked more closely at human-infecting viruses and found that certain structures aligned really well even though the sequences were not similar at all. Coronaviruses use this strategy far more than any other human-infecting viruses. They have only 20-30 proteins (a small proteome) but these proteins mimic many human proteins.
     There is some interesting biology that is reflected in SARS, COVID-19, and MERS infections that are not present in the seasonal virus. For example, the protein NSP13 mimics DDX3 and DDX58 (involved in viral RNA recognition and the regulation of immune response). They believe that some structural mimicry is reflected on the functional level as these structural mimics inhibit nucleic acid sensing (which detects forgein DNA/RNA) in cells that are infected. 
     Two other structural mimics are PARP13 and PARP9, proteins which are known to have regulatory function on STAT1 (a protein  thought to be important for cell viability in response to different cell stimuli and pathogens). NSP5 also mimics the protein Complement C2. Many patients in SARS-1 and 2 many patients present complement and coagulation like phenotypes.      
     The complement and coagulation systems are two closely linked systems that serve a vital role in maintaining homeostasis, which includes immune response. Complement C2 specifically is known to be critical in the optimization of viruses. Based on Complement C2’s role, the group believes that some of this mimicry may help explain some of the enhanced susceptibility in older patients. 
     Dr. Shapira comments that with the focus on ACE2, the primary receptor which SARS-CoV-2 interacts with, we may be overlooking the underlying biology of the virus. An example he provides is that they predicted through P-HIPSTer that the spike protein interacts with the protein Siglec15. This protein may be worth studying in the context of SARS-Cov-2.