iiinova — where your digital self keeps evolving

Jody Mou @moujo.bsky.social · 29/09/2026

Thank you to everybody who contributed to this project, especially my PI @nickpolizzi.bsky.social and collaborators at @danafarber.bsky.social @broadinstitute.org for their amazing support! (16/16)

001

Jody Mou @moujo.bsky.social · 29/09/2026

We hope LUCI-tag will be a helpful tool to the research community. Sequences are available in our preprint, along with a simple HPLC-free click chemistry protocol to generate imaging probes from commercially available reagents. (15/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Finally, to demonstrate the orthogonality of LUCI-tag, we performed simultaneous three-color imaging with LUCI-tag, SNAP-tag, and HaloTag7 fused to different targets. We're also working with collaborators to use LUCI-tag in other settings, and hope to share more soon! (14/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

We expressed LUCI-tag in HEK cells and performed live-cell imaging with an adagrasib-JF646 probe. Nuclear-localized LUCI-tag stained brightly and specifically within 30 minutes, without additional washing steps. (13/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Next, we wanted to confirm LUCI-tag is truly removed from biology. IP-MS of LUCI-tag with human cell lysates showed marked depletion of K-Ras(G12C) binding partners and no neointeractions. Overexpression in HEK cells also did not lead to downstream K-Ras signaling. (12/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Why do inhibitors react with LUCI-tag so rapidly? We provide a possible structural explanation with another structure of inhibitor-free LUCI-tag, showing the cryptic Switch-II pocket stays open without a ligand. Two mutations, M67E and Y71E, might primarily drive this. (11/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

We solved a crystal structure of LUCI-tag bound to adagrasib, showing it matches our design prediction exactly. Even though most of the sequence is redesigned, LUCI-tag retains the same fold and cryptic binding site as K-Ras(G12C). (10/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Due to its dramatically different binding site and probe, LUCI-tag is biochemically orthogonal to SNAP-tag and HaloTag7 - we also saw no covalent off-target labeling via intact mass spec. (9/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

LUCI-tag's speed is on par with existing SLPs and is agnostic to its fluorophore payload - unlike HaloTag7 and SNAP-tag, which are preferentially labeled with neutral rhodamine dyes. LUCI-tag is the fastest reported SLP to date for negatively charged payloads like AF488! (8/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

LUCI-tag is fast, labeling within seconds at nanomolar concentrations. Most surprisingly, we found that our sequence redesign led to a dramatic 27-fold increase in labeling kinetics compared to a control K-Ras(G12C) construct. (7/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Our top design, LUCI-tag, is a 19 kDa, monomeric, thermostable SLP with 54% seq ID to K-Ras(G12C). It rapidly reacts with K-Ras(G12C) inhibitor derivatives such as adagrasib-TAMRA, which we assembled from commercially available products using click chemistry. (6/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

We started with crystal structures of K-Ras(G12C) bound to inhibitors. We truncated the C-term tail that anchors K-Ras to the membrane, and used structure-to-sequence DL models to redesign the sequence while keeping the the drug pocket and GDP-Mg²⁺ binding site fixed. (5/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

But there's an obvious catch - K-Ras(G12C) drives oncogenic signaling. We sought to use deep learning guided protein design to make K-Ras(G12C) biologically inert while preserving its drug-binding pocket. (4/16)

110

Jody Mou @moujo.bsky.social · 29/09/2026

Instead of designing a covalent mechanism from scratch, we wanted to take advantage of K-Ras(G12C), a notorious oncoprotein with existing clinically approved covalent drugs. These inhibitors specifically target a cryptic pocket in K-Ras and irreversibly react with Cys12. (3/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

SLPs enable the covalent attachment of a fluorophore to a protein tag, and are indispensable in fields such as microscopy, single-molecule imaging, and biosensors. For almost 2 decades, SNAP-tag and HaloTag7 have been the only widely adopted SLPs, limiting multiplexing. (2/16)

100

Jody Mou @moujo.bsky.social · 29/09/2026

Excited to share a preprint of my main PhD work in @nickpolizzi.bsky.social's lab - introducing LUCI-tag, a new orthogonal self-labeling protein tag based on K-Ras(G12C) and its covalent inhibitors! www.biorxiv.org/content/10.6... (1/16)

163