AI Charge Fix Lets Antibodies Work Inside Brain Cells

University of Essex researchers converted 672 ordinary antibodies into stable intracellular fragments called intrabodies that bind disease proteins inside living human cells, using AI protein redesign guided by a simple electrical charge rule. The molecules target drivers of Alzheimer’s, Parkinson’s, Huntington’s and motor neurone disease (MND) and the sequences are free for any lab to use.

Ordinary antibodies work outside cells. These redesigned single-chain variable fragments (scFvs) stay soluble and functional in the cytoplasm, where many of the damaging processes begin.

The Charge Rule That Kept Antibodies Outside Cells

Antibodies evolved for the bloodstream and extracellular space. Their variable domains usually carry a net positive charge that makes them aggregate when forced into the reducing environment inside a cell. Dr Caitlin O’Shea, lead author and specialist in MND and Parkinson’s at the School of Life Sciences, said the team compared millions of antibodies with human cytoplasmic proteins and found the charge mismatch.

“From this we figured out that antibodies usually have the wrong charge to exist inside cells without sticking together,” O’Shea said. “We used software developed by Nobel Prize winner David Baker and his group to redesign our antibody fragments, so they had the right charge and are super stable.”

The group measured solubility for 45 scFv intrabodies relevant to neurodegeneration. Whole-molecule net charge at physiological pH showed a clear negative linear correlation with solubility (R² 0.75) across the range from +3 to -20. High solubility (over 70 percent in the soluble cell fraction) required net charge below -15. Only 0.02 percent of unmodified Fv sequences met that bar. Adding standard linkers and tags barely helped.

New electronegative interdomain linkers such as (G₂E)₇ or (G₄E)₄ add negative charge without destroying binding. Framework mutations outside the binding loops and domain-order swaps (VLVH versus VHVL) further tune solubility. The team packaged the rule into a free web tool that predicts the soluble fraction from any antibody or scFv sequence.

672 Redesigned Fragments Reach Tau, Alpha-Synuclein and SOD1

Applying the charge filter and AI inverse-folding redesign produced 672 non-redundant scFv sequences. They cover roughly 60 cytoplasmic proteins and post-translational modifications, including tau, α-synuclein, SOD1, TDP-43, p53, HIF-1α, histones, ubiquitin, phosphorylation, citrullination and acetylation sites. Linear, conformational, oligomer and modification-specific binders are included.

Binding was validated experimentally for p53, α-synuclein, SOD1, polyQ, FUS/TLS, UCHL1 and GFP. Dr Gareth Wright, who directed the research, said: “We’ve made intracellular antibodies that stick to proteins that cause neurodegenerative diseases such as Alzheimer’s, Parkinson’s, Huntington’s and motor neurone disease. These diseases can lead to cognitive impairment, forgetfulness, loss of muscle control and death. They affect over one million people in the UK alone, so they are a big public health concern.”

• Validated interactions: p53, α-synuclein, SOD1, polyQ, FUS/TLS, UCHL1, GFP
• Key disease proteins covered: tau, TDP-43, SOD1, α-synuclein
• Specificity types: linear epitopes, conformational, oligomers, PTMs
• Open-source principle: all sequences released with the paper

Eighty-five percent of the human proteome already has at least one known interacting monoclonal antibody. Until now almost none of those could be trusted inside a cell.

Sequences Go Free the Day the Paper Lands

The full set appears with the Nature Communications paper on scFv solubility. Any laboratory can download the sequences, order synthetic genes or further redesign them with the same charge logic. The scFvright solubility prediction tool lets groups test their own candidates before synthesis.

Wright and O’Shea argue this turns decades of existing antibody work into an immediate intracellular resource instead of forcing every lab to invent new binders from scratch. The University of Essex announcement of the work stresses that the molecules will serve first as research tools and later, potentially, as the basis for treatments.

Quick numbers from the redesign

• 672 non-redundant intrabody sequences released
• 45 experimental scFvs used to establish the charge rule
• R² 0.75 correlation between net charge and intracellular solubility
• 0.02% of unmodified Fvs predicted highly soluble before redesign

Gene Therapy Becomes the Delivery Route

Antibodies themselves cannot cross cell membranes in useful amounts. Intrabodies expressed from DNA or mRNA can. The MND Association’s chief scientist, Dr Brian Dickie, pointed to that combination. “Their research findings provide optimism that a combination of this novel ‘intrabody’ science with emerging gene therapy techniques may lead to new therapeutic strategies that can hit specific molecular targets within neurones,” Dickie said.

Viral vectors such as AAV already deliver genes to neurons in clinical trials for other neurological conditions. An scFv sequence small enough to fit in those vectors could be expressed continuously inside affected cells, binding toxic protein species or post-translational marks that ordinary drugs struggle to reach. The same fragments can first be used in cell and animal models to choose the best targets and epitopes.

MND Association Backing and the One-Million UK Burden

The MND Association funded the core work. Its project page lists £332,188 for October 2025 to September 2028 under principal investigator Dr Gareth Wright. The brief covers new intrabodies against small changes in TDP-43, FUS, SOD1 and C9orf72 proteins, testing whether they keep the proteins in the right place, block harmful clumps and protect cell health, then advancing the best candidates into mouse models.

1. October 2025, funding period begins for the MND Association grant
2. January 2026, Nature Communications paper published (online date)
3. March 2026, University of Essex public announcement
4. August 2026, wider ScienceDaily coverage of the free resource
5. September 2028, current grant end date

There are still no cures for these conditions. Wright called finding molecules that interact with the causal proteins in their native cellular environment “a major challenge in the medicine discovery process.” The charge solution and the open library attack that challenge directly. The MND Association funded project page frames the work as part of a broader pipeline that moves promising molecules toward patients.

What Any Lab Can Do With the New Tool

A group that already owns an antibody against a cytoplasmic target can now:

• Feed the sequence into scFvright to score baseline solubility
• Apply the published charged linkers, tags and framework adjustments
• Run AI redesign with Baker-group tools (the same software used for the 672)
• Order the gene, express it inside cells and test binding or functional rescue
• Compare multiple redesigned versions in parallel before committing to animal work

Because 85 percent of the proteome already has sequenced antibodies, the starting material is already sitting in freezers and databases worldwide. The bottleneck was intracellular behaviour; that bottleneck is now largely computational and open.

David Baker shared the 2024 Nobel Prize for computational protein design. His group’s software supplied the redesign engine that made bulk conversion practical once the charge rule was known.

The Next Constructs Will Target Post-Translational Marks

The released set already includes binders to phosphorylation, citrullination and acetylation. Future rounds can focus on disease-specific marks on tau, TDP-43 or α-synuclein that appear early and may be more selective than total protein knockdown. Parallel work can fuse the best scFvs to degradation tags or other effectors once binding is confirmed.

The immediate output is practical: 672 ready sequences, a public prediction tool, a transparent design rule and an explicit path toward gene-therapy delivery. Labs that order the genes this month will be the first to test whether these fragments can shift the balance inside diseased neurons.

Disclaimer: This article is news reporting and analysis of published scientific research for general information only. It does not constitute medical advice, diagnosis, treatment recommendations or investment guidance of any kind. Readers should consult a qualified physician or specialist neurologist before making any decisions related to neurodegenerative disease care or participation in clinical studies. All figures, funding amounts, sequence counts and project statuses reflect the cited primary sources as of August 2026 and may be updated by the researchers or funders.