De novo designed bright, hyperstable rhodamine binders for fluorescence microscopy

Yuda Chen, Klaus Yserentant, Kibeom Hong, Y. P. Kuang, Arghya Bhowmick, Arthur Charles‐Orszag, Samuel J. Lord, Lei Lü, Kai-Peng Hou, Samuel I. Mann, Jonathan B. Grimm, Luke D. Lavis, R. Dyche Mullins, William F. DeGrado, Bo HuangbioRxiv 2025 · DOI

TL;DR

Researchers have used de novo protein design to create "Rhobins"—a new class of small, hyperstable protein tags that bind rhodamine dyes. These tags are half the size of industry standards like HaloTag, remain functional at near-boiling temperatures (95°C), and enable high-fidelity super-resolution microscopy by allowing photobleached dyes to be constantly replenished.

What Problem Does This Paper Address?

Fluorescence microscopy is the workhorse of cell biology, but it is limited by the "tags" used to label proteins. Natural Fluorescent Proteins (like GFP) are often bulky, slow to mature, and dim. Self-labeling tags like HaloTag and SNAP-tag improved things by binding bright synthetic dyes, but they are still relatively large (~33 kDa) and thermally unstable.

Crucially, existing tags fail in extreme environments. If you want to study "extremophiles"—microbes that live in volcanic hot springs at 75°C—standard protein tags simply melt. Furthermore, most tags bind dyes covalently; once a dye molecule is "bleached" by intense laser light (a major issue in super-resolution microscopy), that label is dead forever.

Methods

The authors moved away from tweaking natural proteins and instead used computational de novo design to build a binder from scratch.

  1. Scaffold Generation: They used RFdiffusion, a generative AI tool, to create a protein backbone. They started with a stable four-helical bundle core and "guided" the AI to create a specific binding pocket.
  2. The "Anchor": They fixed a specific Arginine residue in the pocket to grab the carboxylate group of the rhodamine dye, ensuring the dye stays in its bright, fluorescent "zwitterionic" state.
  3. Sequence Optimization: They used LigandMPNN and Rosetta FastRelax to find the exact sequence of amino acids that would pack tightly around the dye.
  4. Experimental Validation: They synthesized nine designs (Rhobin1–9) and tested them using circular dichroism (to check heat stability) and various microscopy techniques (Confocal, STED, and PAINT).

Datasets

As a protein engineering paper, the "data" consists of the specific sequences generated and the experimental benchmarks.

  • Designs: Nine de novo sequences (Rhobin1-9).
  • Benchmarks: The Rhobins were compared against established tags: HaloTag7, reHaloF, and frFAST.
  • Dyes: A library of rhodamine derivatives, including Janelia Fluor 660 (JF660) and various Si-rhodamines.

Key Insights and Ideas

  • One-Shot Success: Unlike traditional directed evolution, which requires thousands of iterations, the authors designed these functional proteins in a "one-shot" manner, testing only a handful of sequences.
  • The Geometry of Brightness: Rhodamine dyes exist in an equilibrium between a dark "lactone" state and a bright "zwitterion" state. The Rhobin pocket is geometrically tuned to stabilize the bright state, making the system fluorogenic—the dye only "turns on" when it clicks into the protein.
  • Hyperstability by Design: Because the proteins were designed for ideal packing rather than evolutionary "good enough" fitness, they achieved melting temperatures (Tm) over 95°C.
  • The "Infinite" Fluorophore: Because Rhobins bind dyes reversibly (non-covalently), a photobleached dye molecule can fall out and be replaced by a fresh, glowing one from the surrounding buffer. This effectively bypasses the photobleaching limit that plagues super-resolution techniques like STED.

Key Results

  • Compactness: Rhobins are ~16 kDa, roughly half the size of HaloTag7, reducing the risk of the tag interfering with the protein's natural function.
  • Affinity: Rhobin9 binds JF660 with a high affinity of 57 nM.
  • Extreme Performance: The authors successfully imaged protein dynamics in Sulfolobus acidocaldarius (an archaeon) at 75°C and pH 2. This is a "first" for specific protein labeling in such harsh conditions.
  • Super-Resolution: Rhobins enabled PAINT (Point Accumulation in Nanoscale Topography) imaging, achieving sub-diffraction resolution of the endoplasmic reticulum by capturing the transient "blinking" of dyes binding and unbinding.

Limitations and Future Work

  • Background Signal: While fluorogenic, some dyes still have a slight "leakage" of fluorescence when unbound, which can create background noise in no-wash imaging conditions.
  • Promiscuity: Rhobin9 is somewhat "promiscuous," binding several different types of rhodamines. While useful for versatility, future work could focus on "orthogonal" Rhobins that bind only one specific dye for multi-color imaging.
  • Biosensors: The authors suggest that because the binding is so sensitive to the pocket's shape, Rhobins could be engineered into biosensors that change brightness in response to cellular forces or chemical signals.

Summary updated

Abstract

<i>De novo</i> protein design has emerged as a powerful strategy with the promise to create new tools. The practical performance of designed fluorophore binders, however, has remained far from meeting fluorescence microscopy demands. Here, we design <i>de novo</i> Rhodamine Binder (Rhobin) tags that combine ideal properties including size, brightness, and now adding hyperstability. Rhobin allows live and fixed cell imaging of a wide range of subcellular targets in mammalian cells. Its reversible fluorophore binding further enables live super-resolution STED microscopy with low photobleaching, as well as PAINT-type single-molecule localization microscopy. We showcase Rhobin in the extremophile <i>Sulfolobus acidocaldarius</i> living at 75°C, an application previously inaccessible by existing tags. Rhobin will serve as the basis for a new class of live cell fluorescent tags and biosensors.
BibTeX
@misc{chen2025i,
  author = {Yuda Chen and Klaus Yserentant and Kibeom Hong and Y. P. Kuang and Arghya Bhowmick and Arthur Charles‐Orszag and Samuel J. Lord and Lei Lü and Kai-Peng Hou and Samuel I. Mann and Jonathan B. Grimm and Luke D. Lavis and R. Dyche Mullins and William F. DeGrado and Bo Huang},
  title = {{<i>De novo</i> designed bright, hyperstable rhodamine binders for fluorescence microscopy}},
  year = {2025},
  doi = {10.1101/2025.06.24.661379},
  howpublished = {bioRxiv},
}
Figure