Two papers landed this past week that both attack the same underlying headache — membrane proteins are miserable to work with — from completely different angles.
Detergent has always been the tax you pay for membrane protein work
If you've ever solubilized a membrane protein, you know the drill: pick a detergent, hope it doesn't strip the protein's native fold along with the lipids, then spend the rest of the prep managing micelles instead of studying your protein. It's been the standard approach for decades because there hasn't been a good alternative.
A team at the UW Medicine Institute for Protein Design, led by Ljubica Mihaljevic with David Baker's group, published a genuinely different option this week: computationally designed amphipathic proteins — they're calling them WRAPs — that wrap around a membrane protein's hydrophobic surface and hold it stable in plain water, no detergent required. The proteins are generated with RFdiffusion, the same design toolkit behind a lot of recent de novo protein work, custom-fit to the target's hydrophobic patch. The group demonstrated it on notoriously stubborn targets, including syphilis antigens that have resisted structural work for years, which is a decent stress test for a "keeps proteins intact in water" claim.
Published in Science on July 2, 2026 — read the EurekAlert summary or the paper itself, "Membrane protein solubilization and structure determination using de novo-designed amphipathic proteins" (DOI: 10.1126/science.adr3817).
Letting the cell do the aptamer screening for you
The second piece solves an adjacent problem: raising aptamers against G protein-coupled receptors, which is normally slow going because GPCRs are hard to present in a form that a SELEX screen can work with cleanly.
A group at the University of Toyama, led by Toshihide Tabata, noticed that activated GPCRs get naturally packaged into extracellular vesicles that cells shed into culture medium — and built a selection method, EV-SELEX, around harvesting aptamer candidates directly off those vesicles. Because the receptor is sitting in native membrane in a biologically relevant conformation, the selection converges in far fewer rounds than conventional approaches, and pulls in less noise from unrelated proteins. As a proof of concept, they generated Dapt-μR, a DNA aptamer that activates the μ-opioid receptor and produced measurable analgesic effects in mice.
Published in Communications Biology on July 7, 2026 — read the EurekAlert summary or the paper itself, "EV-SELEX: A novel platform for rapid G protein-coupled receptor drug discovery" (DOI: 10.1038/s42003-026-10525-0).
Worth a mention: seeing nerves without cutting them
One more, briefer: a University of Pittsburgh team combined light-sheet microscopy with a tissue-clearing protocol (c-Clear) to build 3D maps of nerve branching through intact joint tissue — no slicing required, which matters when the whole point is preserving how nerves actually route through the tissue. It's part of a larger NIH effort mapping nerve architecture to better understand pain signaling. Details in the EurekAlert writeup or the paper in npj Imaging, "Advanced tissue clearing and three-dimensional imaging approaches to visualize neural innervation in the rat knee joints" (DOI: 10.1038/s44303-026-00167-6).
Have a paper or technique you think belongs in next week's roundup? Let us know.