About

Brian Kuhlman is an Oliver Smithies investigator and Professor of Biochemistry and Biophysics at the University of North Carolina at Chapel Hill. His research focuses on computational protein design, protein-protein interactions, antibodies, structural biology, and optogenetics. He is a co-developer of the molecular modeling program Rosetta, which is used to identify low energy sequences for target structures or interfaces, facilitating the creation of novel proteins and the manipulation of existing ones. Kuhlman's work includes designing proteins with enhanced stability, new structures, and specific interactions, as well as developing light-activatable proteins, therapeutic antibodies, and de novo protein structures. He employs techniques such as computer programming, molecular cloning, protein expression, biophysical analysis, directed evolution, X-ray crystallography, and live cell imaging. His research has contributed to the engineering of protein switches for controlling biological processes with light, the development of strategies for antibody engineering—particularly bispecific antibodies—and the de novo design of protein structures. Kuhlman has played a significant role in parameterizing the Rosetta energy functions and advancing computational strategies for protein interface design, combining these with high-throughput screening methods. His work aims to deepen understanding of protein structure determinants and to develop novel tools for research and therapeutic applications.

Research topics

• Computer Science
• Biology
• Biochemistry
• Chemistry
• Systems engineering
• Computational biology
• Cell biology
• Engineering
• Data science
• Programming language
• Biological system
• Software engineering
• Human–computer interaction
• Computational chemistry
• World Wide Web
• Immunology
• Mathematics

Selected publications

• De novo masking domains that gate TNF-family ligand assembly and activity

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-04-22

  articleOpen accessSenior authorCorresponding

  Tumor necrosis factor family ligands (TNFLs) are central regulators of immunity and promising agents for cancer therapy, but their clinical use is often limited by dose-limiting systemic toxicity. Conditional activation and genetic fusion with antibodies could improve safety and pharmacokinetics, yet these features are difficult to combine because TNFLs form obligate homotrimers that are structurally mismatched with antibody architectures. Here we use AI-enabled protein design to create de novo protein masks that conditionally control TNFL assembly. The masks are genetically fused to TNFLs through protease-sensitive linkers and prevent trimer formation by competitively binding the TNFL trimerization interface. We demonstrate with TNFα, OX40L, and 4-1BBL that protease-mediated release of the mask promotes trimer assembly, receptor binding and biological activity. These monomeric, switchable TNFLs are readily incorporated into full-length IgG formats, enabling plug-and-play construction of conditionally activatable antibody fusions. These fusions can be designed either to release activated soluble TNFLs upon protease treatment or, by reordering the fusion architecture, to trigger antibody multimerization and signaling by membrane-type TNFLs. When this multimerization strategy is applied to antibody-drug conjugates, conditional trimerization enhances cell killing by ~30-fold, thereby improving the therapeutic window and enabling multiple strategies for selective tumor microenvironment targeting.

  PublisherOA PDFDOI

• Perturbing the energy landscape for improved packing during computational protein design

  UNC Libraries · 2026-04-14

  articleOpen access

  The FastDesign protocol in the molecular modeling program Rosetta iterates between sequence optimization and structure refinement to stabilize de novo designed protein structures and complexes. FastDesign has been used previously to design novel protein folds and assemblies with important applications in research and medicine. To promote sampling of alternative conformations and sequences, FastDesign includes stages where the energy landscape is smoothened by reducing repulsive forces. Here, we discover that this process disfavors larger amino acids in the protein core because the protein compresses in the early stages of refinement. By testing alternative ramping strategies for the repulsive weight, we arrive at a scheme that produces lower energy designs with more native-like sequence composition in the protein core. We further validate the protocol by designing and experimentally characterizing over 4000 proteins and show that the new protocol produces higher stability proteins.

  PublisherDOI

• Shared PRAME epitopes are T-cell targets in NUT carcinoma

  Open MIND · 2026-01-01

  article

  Background NUT carcinoma is a rare but highly lethal solid tumor without an effective standard of care. NUT carcinoma is caused by bromodomain-containing NUTM1 fusion oncogenes, most commonly BRD4::NUTM1. BRD4::NUTM1 recruits p300 to acetylate H3K27 forming expansive stretches of hyperacetylated chromatin called “megadomains” with the overexpression of corresponding oncogenes, including MYC. We hypothesized that transcriptional dysregulation caused by BRD4::NUTM1 would lead to the generation of cancer-specific antigens that could be therapeutically actionable. Methods We integrated genomics, computational antigen prediction software, targeted immunopeptidomics using single-labeled and double-labeled peptide standards, and gain/loss-of-function genetic experiments on a panel of cell lines (N=5), a patient-derived xenograft, a tissue microarray (N=77), and patient samples from the Tempus AI Sequencing Database harboring evidence of NUTM1 fusions (N=165). We created an αPRAME425 T-cell receptor (TCR) × SP34 αCD3 bispecific molecule modeled after brenetafusp, an αPRAME425 TCR bispecific T-cell engager, as well as αPRAME425 TCR T-cells based on anzutresgene autoleucel and we applied these products to NUT carcinoma cells in vitro. Results We identified PRAME as the most commonly expressed cancer/testis antigen in patient samples harboring the three canonical NUT carcinoma fusions (BRD4::NUTM1, BRD3::NUTM1, and NSD3::NUTM1). Additionally, 56% (43/77) of NUT carcinoma tissue microarray samples stained positive for PRAME. BRD4::NUTM1 expression in HEK 293T cells enhanced PRAME levels and BRD4::NUTM1 knockout in NUT carcinoma cells reduced PRAME levels. Immunopeptidomics detected more PRAME-derived human leukocyte antigen (HLA) ligands (N=9) than all other cancer/testis antigens combined (N=5). Targeted mass spectrometry detected the HLA-A*02:01/SLLQHLIGL (PRAME425) epitope in 100% (4/4) of HLA-A*02+, PRAME+ NUT carcinoma samples at higher levels (>0.01 fM) than HLA-A*02:01/RLDQLLRHV (PRAME312) or HLA-A*02:01/YLHARLREL (PRAME462). The αPRAME425 TCR × SP34 αCD3 bispecific molecule and αPRAME425 TCR T-cells each exhibited potent, T-cell mediated cytotoxicity against PRAME+ NUT carcinoma cells. Conclusions PRAME is highly and frequently expressed in NUT carcinoma, and the most common oncoprotein causing NUT carcinoma, BRD4::NUTM1, contributes to these high PRAME levels. PRAME epitopes presented by HLA class I are a previously unrecognized therapeutic vulnerability for NUT carcinoma that warrants clinical trials testing PRAME-targeted immunotherapies in this neglected patient population.

• Engineered interfaces in Rac1 and Cdc42 biosensors enhance sensitivity and reduce cell perturbation

  Molecular Biology of the Cell · 2026-02-18

  articleOpen access

  Fluorescent biosensors are a valuable means to report the spatiotemporal dynamics of protein activities in live cells and animals. However, biosensors affect the activities they are reporting. This can be ameliorated by increasing sensitivity, using lower biosensor concentrations, or by choosing designs that minimize undesirable interactions. For biosensors in which fluorescent components interact to produce Forster Resonance Energy Transfer (FRET), perturbation is often due to interaction of biosensor components with nonfluorescent, endogenous proteins, rather than productive interactions that lead to FRET. Here, we engineer the interface between biosensor components using charge swap and "knob into hole" mutations to reduce all but desired interactions. Novel biosensors for Rac1 and Cdc42 showed reduced interactions with endogenous GTPases and effectors, normal activation by guanine nucleotide exchange factors (GEFs), and correctly reproduced previous reports of GTPase activation dynamics. Assaying concentration-dependent effects on cell motility showed substantially reduced perturbation of normal cell behavior. Computational models indicated that minimal perturbation could be achieved over a broader range of concentrations using the new "orthogonal" biosensors.

  PublisherOA PDFDOI

• Enhancing Enzymatic Bioconjugation Efficiency via Computer-Based Installation of a Substrate Recruitment Domain

  Bioconjugate Chemistry · 2026-03-17

  articleOpen accessSenior authorCorresponding

  Enzyme mediated bioconjugation provides a method for easy and rapid formation of protein-protein and protein-small molecule conjugates under mild conditions. Promiscuous enzymes are of particular interest because they can catalyze conjugation reactions on a broad set of substrates. However, this promiscuity carries the risk of undesirable off-target modifications. To mitigate this effect, we used computational design to install a substrate recruitment domain (SRD) onto the promiscuous enzyme, tyrosinase. The redesigned tyrosinase, called design42 (D42), preferentially modifies tyrosine residues adjacent to a 6-amino acid recognition motif/sequence (RS) that is bound by the SRD. Incorporation of the recognition sequence along with a neighboring tyrosine in peptides or proteins allows for rapid D42-mediated conversion of the tyrosine to an orthoquinone, which can be selectively modified with a variety of nucleophiles. We demonstrate the utility of our design system by rapidly installing cytotoxic molecules on a monoclonal antibody.

  PublisherDOI

• Intensity modulation of trichromatic split fluorescent proteins for live cell mapping

  Cell Reports Methods · 2026-03-26

  articleOpen access

  Current fluorescent protein-based multiplexed cell labeling techniques suffer from limited discrimination power due to stochastic color selection and large gene sizes from tandem repeats of multiple fluorescent proteins. We developed Caterpie, a rationally designed system using engineered split fluorescent proteins that enables deterministic identification of 20 distinct cell populations with 97% accuracy and reduced gene sizes. Through computational structure-guided design, we engineered enhanced split mNeonGreen3A and split sfCherry3C variants that achieve performance comparable to split CFP2, the best-performing split fluorescent protein. Our systematic library of trichromatic 11th β-strand tags with up to 12 tandem repeats enables predictable, high-fidelity labeling for precise cell targeting. This technology addresses critical limitations in simultaneous identification of multiple defined cell populations.

  PublisherDOI

• Automated Deep Learning‐Based Pipelines for Multi‐Objective De Novo Protein Design

  Current Protocols · 2025-10-01

  articleSenior authorCorresponding

  Computational protein design has been transformed by deep learning models that can accurately predict protein structure and generate sequences compatible with desired folds. Here we present a detailed protocol for EvoPro, an automated platform that uses a genetic algorithm along with iterative structure prediction (AlphaFold2/AlphaFold3) and sequence design (ProteinMPNN/LigandMPNN) to engineer protein-protein interactions with customizable properties. The protocol describes how to implement multistate design objectives to simultaneously optimize positive and negative design goals. We provide step-by-step instructions for setting up the genetic algorithm, configuring scoring functions for different design challenges, and analyzing results. The method builds on our previously validated approach, which successfully generated high-affinity binding domains without requiring experimental optimization. We describe key considerations for adapting the protocol to diverse protein engineering objectives, including binding site targeting, conformational specificity, and symmetric assembly. The complete computational protocol can be installed and executed in a week by a new user and provides a framework for leveraging deep learning models to address challenging protein design problems. © 2025 Wiley Periodicals LLC. Basic Protocol 1: Designing protein binders Basic Protocol 2: Engineering conformational switches Basic Protocol 3: Designing de novo homo-oligomers Support Protocol 1: Setting up the EvoPro code and environment Support Protocol 2: Input preparation for different design scenarios Support Protocol 3: Optimizing the scoring function and other parameters.

  PublisherDOI

• Enhancing enzymatic bioconjugation efficiency via installation of a substrate recruitment domain

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-10-27

  preprintOpen accessSenior authorCorresponding

  Enzyme mediated bioconjugation provides a method for easy and rapid formation of protein-protein and protein-small molecule conjugates under mild conditions. Promiscuous enzymes are of particular interest because they can catalyze conjugation reactions on a broad set of substrates. However, this promiscuity carries the risk of undesirable off-target modifications. To mitigate this effect, we used computational design to install a substrate recruitment domain (SRD) onto the promiscuous enzyme, tyrosinase. The redesigned tyrosinase, called D42, preferentially modifies tyrosine residues within the peptide core (core) linked to a 6-amino acid recognition motif/sequence (RS) specific for the SRD. Incorporation of the recognition sequence along with a neighboring tyrosine in peptides or proteins allows for rapid D42-mediated conversion of the tyrosine to an orthoquinone, which can be selectively modified with a variety of nucleophiles. We demonstrate the utility of our design system by rapidly installing cytotoxic molecules on a monoclonal antibody.

  PublisherOA PDFDOI

• Data Management and Sharing Plan for: Computational Design of Protein Structures and Complexes

  UNC Dataverse · 2025-12-10

  datasetOpen access1st authorCorresponding

  The Data Management and Sharing Plan describes the scientific data to be generated and/or used in the research and outlines a strategy for managing and sharing project data.

  PublisherDOI

• BPS2025 - Engineering a tumor-sensitive prodrug T cell engager bispecific antibody for safer immunotherapy

  Biophysical Journal · 2025-02-01

  articleSenior author
  PublisherDOI

Recent grants

• Computational Design of Protein Structures and Complexes

  NIH · $5.3M · 2019–2029

• Molecular and Cellular Biophysics Training Grant

  NIH · $5.6M · 1995–2023

• Computational Design of Protein-Protein Interactions

  NIH · $3.1M · 2005–2020

• Rosetta: An Integrated Macromolecular Modeling Suite

  NIH · $7.8M · 2005–2021

• Spatiotemporal Control of the Epigenome via Photoactivatable Nuclear Localization

  NIH · $2.1M · 2013–2019

Frequent coauthors

• David Baker

  University of Washington

  63 shared

• Daniel P. Raleigh

  University College London

  50 shared

• Andrew Leaver‐Fay

  University of North Carolina at Chapel Hill

  38 shared

• Steven M. Lewis

  Cold Spring Harbor Laboratory

  38 shared

• Klaus M. Hahn

  University of North Carolina at Chapel Hill

  35 shared

• Hayretin Yumerefendi

  University of North Carolina at Chapel Hill

  31 shared

• Richard Bonneau

  30 shared

• Jeffrey J. Gray

  Johns Hopkins University

  28 shared

Education

• Ph.D., Biochemistry

  University of North Carolina at Chapel Hill

  1994

• B.S., Chemistry

  University of California, San Diego

  1989

Awards & honors

• National Academy of Inventors Senior Member, 2025
• UNC Excellence in Basic Science Mentoring Award, 2020
• ASBMB DeLano Award for Computational Biosciences, 2019
• UNC Oliver Smithies Investigator, 2019
• W.M. Keck Foundation Distinguished Young Scholar in Medical…