Skip to main content
Back to timeline
AnthropicSource publication:

A phage-derived vault protein assembles into a dome inside bacteria: Anthropic solves the first prokaryotic vault structure with Claude

Synopsis

Anthropic's biology team reports that a phage-derived protein called L05 assembles into a vault-like particle inside E. coli, and solves the first 3D structure of a vault from prokaryotes: 96 copies of a single protein, shorter and wider than the human vault yet enclosing about 25% more volume. Vault-protein relatives were already known in prokaryote and phage genomes, but no one had shown they assemble into complete particles. Cryo-EM confirmed the vault shape for four candidates, and Claude handled sequence search, DNA construct design and data processing, cutting a purification that conventionally takes three days to about 90 minutes.

Source-provided article image: Investigating vault particles with Claude \ Anthropic
Anthropic

Interpretation

The prokaryotic side now has a complete vault structure: a phage-derived protein assembles inside bacteria into a closed, dome-shaped shell, showing that these large protein shells are not exclusive to eukaryotic cells. Distant relatives of the vault protein had been found in bacterial and phage genomes, but whether they actually form vaults was unclear, and the function of human vaults remains unconfirmed. Of 15 candidates that expressed solubly, 7 formed vault-like particles under the electron microscope, and cryo-EM confirmed the vault shape for 4 of those 7, two from phages and two from bacteria.

The particle is built from 96 copies of L05, 48 per half, with a combined mass of about 9.5 megadaltons and dimensions of roughly 54 nanometers long by 45 wide; that is shorter and wider than the human vault at 67 by 40, but being more spherical it encloses about 25% more volume. The human vault has 39 copies of the major vault protein per half, and its wall is built from nine repeating segments stacked end to end. The structure comes from a cryo-EM reconstruction of L05 particles expressed in E. coli and purified by size-exclusion chromatography.

The way the caps close is now visible: a wall 48 copies around narrows toward the pole in three steps, and in every group of three neighboring copies one ends at the first step, one at the second, and one continues to the tip, so identical chains along the wall take different conformations in the cap. No atomic-level structure of a prokaryotic vault was available to show how the caps close. In the cryo-EM density map, all 48 chains wind from the rim toward the tip and overlap, with two-thirds passing the first tier and one-third reaching the tip.

Changing the number of wall repeats changes particle size: duplicating two segments so each chain carries nine instead of seven, the same count as the human vault, produced larger particles with about 20% more enclosed volume. The ordinary L05 wall is built from seven repeating segments, versus nine in the human vault. Cryo-EM images at the same scale show the engineered particle with two extra segments per copy is longer and wider.

Perspective

The demonstrated scope is limited to L05 and other candidates expressed and purified in E. coli, with structural conclusions from cryo-EM reconstruction; whether these particles form inside the bacteria and phages where they occur naturally has not been tested.

Vault function remains unknown, and the authors say the work only deepens the mystery; a pocket partway up each L05 copy holds a small molecule that shows clearly in the map but has not been identified, and its shape resembles coenzyme A, which would be only an early hint at function. The team also reports that an early Claude-written protocol called for a medium additive that interferes with getting DNA into cells, setting the project back two weeks, and that early cryo-EM processing skipped a standard step and produced a flawed map blended with its mirror image.