# Computational design of specific binding proteins using Leave-One-Out

> **NIH NIH R01** · RENSSELAER POLYTECHNIC INSTITUTE · 2021 · $305,705

## Abstract

SUMMARY:
Computational design of specific binding proteins using Leave-One-Out
 The goal of this research is to design a receptor protein to bind any protein target.
Furthermore, the binding event will be signaled by the appearance of fluorescence. The novel
binding protein will be able to sense and report the presence of a specific protein or peptide in a
mixture of others, allowing the detection of any disease agent or protein of interest. The new
approach takes advantage of the green fluorescent protein (GFP) what we know about its
folding pathway. When GFP folds, it does in a specified order of events called a pathway. When
it finishes folding, its fluorescence activity is immediately turned on. If we leave out one small
piece of GFP so that the folding cannot finish folding, then it sits in an inactive state until the
missing piece appears. Using this Leave-One-Out strategy, partially folded proteins become
sensors for their missing pieces.
 Using computational design algorithms, a new amino acid sequence can be substituted for
the missing piece, making the designed GFP a sensor for the new sequence. But accurate
computational protein design is challenging because of inherent assumptions. Two approaches
are proposed to overcome the weaknesses. First, thousands of candidates will be designed,
synthesized in yeast and then screened for their biosensor function using high throughput cell
sorting technology. Second, more knowledge about the folding pathway will be generated by
pulse-labeling the protein as it folds, then finding out what parts of the protein fold first. This will
improve the computational model for folding, and therefore improve the ability to design partially
folded leave-one-out biosensors. A cautious, step-wise design strategy is proposed for
screening, so that every experiment tests a specific hypothesis about GFP folding and function.
 A known drawback of the leave-one-out method is the necessity of a having a partially
unfolded off-state protein that can aggregate an cause problems. To fix this, biosensor proteins
will be genetically fused to a fiber-forming protein to create robust and stable biosensor fabrics
that no longer have a problem with aggregation. The final product of this research will be is a
silk-like biosensor fabric that is computationally designed to sense any protein target and glow
green when the target is present.

## Key facts

- **NIH application ID:** 10224218
- **Project number:** 5R01GM099827-09
- **Recipient organization:** RENSSELAER POLYTECHNIC INSTITUTE
- **Principal Investigator:** CHRISTOPHER BYSTROFF
- **Activity code:** R01 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2021
- **Award amount:** $305,705
- **Award type:** 5
- **Project period:** 2012-09-20 → 2023-07-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10224218

## Citation

> US National Institutes of Health, RePORTER application 10224218, Computational design of specific binding proteins using Leave-One-Out (5R01GM099827-09). Retrieved via AI Analytics 2026-05-25 from https://api.ai-analytics.org/grant/nih/10224218. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*