De novo designed inhibitor confers protection against lethal toxic shock
Ragotte, R. J.; Liang, H.; Berman, J. M.; Gloegl, M.; Schramek, D.; Melnyk, R.; Baker, D.
Show abstract
Paeniclostridium sordellii causes a toxic shock syndrome with a mortality rate of nearly 70%, primarily affecting postpartum and post-abortive women. This disease is driven by the production of the P. sordellii lethal toxin, TcsL, for which there are currently no effective treatments. We used a protein diffusion model, RFdiffusion, to design high affinity TcsL inhibitors. From a very small set of 48 starting designs and 48 additional sequence optimized designs, we developed a potent inhibitor with <100 pM affinity that protects mice prophylactically and therapeutically (post exposure) from lung edema and death in a stringent lethal challenge model. This inhibitor, which can be lyophilized without any loss of activity, is a promising therapeutic candidate for this rare but deadly disease, and our results highlight the ability of deep learning-based protein design to rapidly generate biologics with potential clinical utility.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Structural basis for saxitoxin congener binding and neutralization by anuran saxiphilins 95%
- Decoding a cryptic mechanism of metronidazole resistance among globally disseminated fluoroquinolone-resistant Clostridioides difficile 95%
- Broadly neutralizing SARS-CoV-2 antibodies through epitope-based selection from convalescent patients 95%
Similar papers in this journal
Similar papers in this journal
- Analysis of the Diverse Antigenic Landscape of the Malaria Invasion Protein RH5 Identifies a Potent Vaccine-Induced Human Public Antibody Clonotype 95%
- Autoimmune mechanisms elucidated through muscle acetylcholine receptor structures 94%
- Butyrolactol A potentiates caspofungin efficacy against resistant fungi via phospholipid flippase inhibition 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.