Current WRF Fellows

Below you will find the list of our current WRF Fellows as well as a little information about them and their work.

HUA BAI, PhD – Collaborating PI: David Galas, Pacific Northwest Diabetes Research Institute

To restore auto-antigen tolerance in autoimmune disease models, especially in Type 1 Diabetes model, Hua will work under the guidance of Dr. David Galas at the Pacific Northwest Diabetes Research Institute and Dr. David Baker at the Institute for Protein Design to design and validate therapeutic peptides (or small proteins) that can interfere with the auto-antigen*MHC*T-cells complex, and hence inhibit the activation of pathogenic T-cells. On the other hand, by exploiting self-assembling nanoparticles design methods, Hua will design tolerogenic MHC*peptide complexes which can promote the proliferation of regulatory T cells and then induce auto-antigen tolerance. This project can potentially create novel tolerogenic therapeutics, and it can lead to a better understanding of immune tolerance mechanisms as well.

Hua obtained a B.S. degree in Biological Sciences in Peking University, China. Later, he came to the University of Wisconsin-Madison pursuing a Ph.D. degree in physiology. After two years of working on “metabolism” using animal models and five years of studying “membrane trafficking”using various in vitro biochemistry and biophysics techniques, Hua’s career objective to be an outstanding protein engineer is getting clearer and more determined. Hua will be dedicated to addressing the metabolic syndrome and related disorders by utilizing the powerful protein design and engineering methods.

RALPH CACHO, PhD – Collaborating PI: Michael Gelb, Department of Chemistry

The anticancer agents salinosporamide A and rebeccamycin, the antifungal griseofulvin and the antibiotic vancomycin all contain carbon-halogen bonds. The presence of the halogen atom in these compounds is
vital to the activity of the respective molecules. A lack of specificity and regioselectivity (a process that favors bond formation at a particular atom) has hampered classical synthetic chemistry methods for carbon-halogen bond formation. But enzymes evolved a variety of elegant mechanisms to catalyze this synthetically challenging reaction. Hence novel methods for the introduction of carbon-halogen bonds in advanced synthetic or biosynthetic intermediate with complex molecular scaffolds can be developed by engineering these halogenating enzymes.

My project aims to apply computational protein design to redesign a chlorinase enzyme with a modified substrate scope. To demonstrate the potential of computational protein design to efficiently change the substrate scope of the enzyme, I will design the chlorinase to catalyze the conversion of dechlorogriseofulvin to griseofulvin. Griseofulvin is an antifungal compound used in the treatment of skin and scalp infections caused by dermatophytes like tinea capitis and tinea pedis. While previous engineering approaches have been performed on chlorinase enzymes, previous examples use indole-containing substrates (such as tryptamine) or aryl rings with nitrogen substituents (such anthranilic acid or kyneurine). While these examples highlight the plasticity of the substrate scopes of this class of enzymes, they also highlight the deficiencies of the current approaches used to modify the specificity of flavin-dependent chlorinases. My project aims to supplement these current methods with computational design by targeting a reaction involved in the synthesis of a medically relevant compound.

zachary-crook227x227ZACHARY CROOK, PhD – Collaborating PI: Jim Olson, FHCRC

Zach is interested in novel therapeutic applications for cysteine knotted peptides. While the Olson Lab has a strong interest in their potential uses for cancer treatment, Zach comes from a research background in neurodegeneration, and wishes to investigate the means for getting these natural drug-like peptides across the blood-brain barrier. Efforts towards this will make use of Rosetta and several known structures. Both the knotted peptides that the Olson Lab routinely produces, and several receptor proteins that facilitate transcytosis of natural signaling proteins, will be used to identify potential knotted peptides that can bind these receptors and efficiently transport into the brain. This has the potential to make the blood-brain barrier a little less imposing for therapeutic applications, including neurodegenerative diseases and brain cancer.

Zach attended college at the University of Colorado in Boulder, graduating with honors in Molecular, Cellular, and Developmental Biology. He went right into the PhD program at MIT, where he studied mouse models of Huntington’s Disease under David Housman, and developed assays to rapidly and accurately determine the effect of test therapeutics. The idea of drug design using naturally stable, bioavailable knotted peptides and their potential utility in diseases of the central nervous system drew him to the lab of Jim Olson, where he brings his past knowledge of screening assays in mouse disease models to mix with the biochemical and drug-design expertise of the Olson Lab and collaborators at the FHCRC.

CourbetALEXIS COURBET, PhD, PharmD – Collaborating PIs: Joshua Smith and Luis Ceze, Computer Science and Engineering

Computers have revolutionized our understanding and relation to the world. Automating the manipulation of information, they transfer human labor to machines and augment human capabilities. While science and engineering have placed increasing demands on computation, miniaturization of silicon-based electronics has been the main driving force behind its enhancement. However, physical limits of CMOS technology are announcing the end of Moore’s law. Further advancing computers to achieve ever-higher densities of useful computational work under specified quantity of time, material, space, energy and cost, remains a critical challenge in the 21st century. Novel approaches relying on biological substrate (i.e. biocomputing) have the potential to outperform conventional silicon. Indeed, living systems are incredibly efficient three dimensional computers capable of solving hard computational problems. Synthetic biologists are thus considering the possibility of engineering computing systems where input, output, software, and hardware are made of biological molecular-scale machinery to store and perform operation on data. This approach holds promising advantages: high density of data storage, massive parallelization and ultra-low power signal processing. Biocomputers are biosynthesized and self-assembling, ensuring a low cost and high scale of production. By nature biocompatible, they could support the monitoring, control and electronic interfacing of biological systems. Therefore, the engineering of biological computing machines could pave the way towards unprecedented scientific opportunities, offering powerful solutions for computer sciences, biotechnology and molecular medicine. Yet, developing functional biocomputers with a scalable architecture remains elusive since tools are lacking to perform precise assembly of biomolecular components at nanoscales.

In this project, we identified proteins as versatile and modular components constituting a vast engineering playground. Proteins self-assemble in a sequence dependent way and are capable of information processing, which we intend to exploit for the rational design of complex three dimensional biocomputers. Although proteins rely on complex folding, allosteric mechanism and interfaces of noncovalent interactions, recent advances in the development of Rosetta software allowed computational design of protein nanomaterials with unprecedented accuracy and design space. De novo protein components can now be generated in silico to self-assemble into specified symmetric scaffolds, which we suggest could support high-order biocomputing architectures. This project thus proposes to harness advances in computational protein design as a systematic methodology to engineer universal nanoscale biocomputers. We propose to investigate how massively parallel computing architectures can be built using self-assembling 3D arrays of protein logic gates, ultimately implementing any given Boolean function. There are two engineering challenges to realizing such systems: first, methods for designing protein nanomachines need to be developed; and second, because of non-conventional substrate, mechanism of information processing, programmability of the computer and interfacing will need to be overcome. The proposed research intends to i) investigate machine architectures and programmability in silico ii) explore design rules, experimental and computational methods for designing 3D self-assembling protein logic gate arrays iii) develop methods for designing digital protein information carriers (i.e. switches). Such computers designed with 10 nm nodes could theoretically accommodate 1015 layered logic gates within a single microliter, while achieving energy requirements orders of magnitude below those of silicon computers.

GerardDGERARD DANIEL, PhD – Collaborating PI: Thomas Spiro and Karen Goldberg, UW Chemistry

Naturally occurring metalloproteins, were adaptively evolved to use earth abundant metals as cofactors to effectively tackle their needs and therefore are limited in their function. On the other hand, non-biological late transition metals possess unparalleled potential to catalyze desirable organic transformation for modern synthetic needs. Today, these paradigms can be combined through de novo design of enzyme active sites that incorporate non-biological metals. Gerard will be working along with Dr. Thomas Spiro and Dr. Karen Goldberg to design a metalloenzyme for photocatalytic reduction of CO2 at a (Ru,Zn)-bimetallic active site. Gerard will strategize design techniques to incorporate Ru ion by itself and incorporation of Ru complex to a specific binding pocket in the protein to form the active site. Ultimately, this project will accelerate the merger of the fields of organometallic catalysis and de novo protein design.

Gerard learnt the basics of electronic structure computation while working with Dr. M. Balakrishnarajan during his Master’s degree in Chemistry at Pondicherry University, India. Later, during his Ph. D. at Virginia Commonwealth University under Dr. Nicholas Farrell, he studied novel zinc binding environments in proteins using computational and biophysical techniques. His interests in artificial photosynthesis and the origin of life directed him to his current interest in de novo design of enzymes with non-biological metals for in vitro catalysis. As a WRF Innovation Fellow, he will be working with Dr. Spiro and Dr. Goldberg on the design of a metalloenzyme for photocatalytic reduction of carbon dioxide.

GLENNA FOIGHT, PhD – Collaborating PI: Dustin Maly, UW Chemistry

Cellular signaling pathways are complex networks of protein and enzymatic interactions. Understanding their contributions to disease states is an important goal. The development of protein inhibitors through computational design and directed evolution is a powerful approach for specifically targeting individual elements of signaling pathways. Small-molecule inhibitors offer temporal control and cell permeability, but the development of small molecules that specifically target only one protein in the cell is a difficult and lengthy process. The Maly lab has combined the power of protein and small-molecule inhibitors in a system known as a chemical genetic switch. This involves the fusion of two components of a proteinprotein interaction to a protein of interest such that a small molecule that disrupts the interaction will allow activation of the protein. The first part of my research will involve creating a new chemical genetic switch system by designing a protein interaction partner for an existing protein and small-molecule inhibitor. The new system will use components that are foreign to mammalian cells, thus offering compatibility with future studies of mammalian signaling pathways. My second project will focus on developing inhibitors of the oncogenic protein Ras. Ras is an important signaling protein in numerous cellular processes, and the Ras family is the most frequently mutated protein family in human cancers. I will design protein inhibitors that specifically target individual oncogenic mutants and family members of Ras. These inhibitors will aid in dissecting the complex functional differences between different mutants and variants of Ras. Furthermore, I will use the successfully engineered chemical genetic switch developed in my first project to confer small-moleculebased, temporal control over these Ras inhibitors in cells. I did my undergraduate degree in biochemistry at North Carolina State University where I performed Xray crystallography research, coincidentally, on structures of oncogenic Ras mutants in the lab of Dr. Carla Mattos. My interest in protein structure led me to the lab of Dr. Amy E. Keating at MIT, where I studied the determinants of protein-protein interaction specificity in my graduate work. My interest in studying signaling processes controlled by protein-protein interactions led me to the lab of Dr. Dustin Maly at the UW. Designing specific protein interactions and using my designs to disrupt cellular interaction networks will be an exciting combination of my expertise and interests.

HANNAH GELMAN, PhD – Collaborating PI: Doug Fowler, Department of Genome Sciences

Guiding protein design with comprehensive maps of mutant function
Computational protein design — in which new protein sequences are developed to perform a specific function — promises efficient generation of biological molecules that can carry out novel functions. Designed proteins could be ideal for the treatment of rare or emerging diseases and potentially mitigate the side effects of more broadly acting classes of drugs. This method is hampered by the difficulty of predicting the effect of mutations on protein function, especially if the mutation affects protein stability or structure, and by our incomplete understanding of how protein physical properties like thermodynamic stability affect protein function.

The efficiency of a designed protein can be significantly enhanced with deep mutational scanning (DMS), in which a library of mutants based on the designed sequence is expressed and subjected to weak functional selection (e.g., binding to the targeted ligand). The change in the distribution of sequences over the selection — enrichment of some sequences and depletion of others — is measured and used to determine the relative function of every sequence. This strategy refines the sequence found by computational design to find a better performing variant, but does not address the underlying mismatch between the performance predicted by the design algorithm and that measured in the real world. Integration of DMS and computational design can be pushed even further so that data from DMS is used to improve design algorithms at the outset instead of to refine the algorithm’s output.

Incorporating the large-scale DMS data into the development of protein design algorithms will require us to quantitatively and accurately measure specific biophysical and biochemical properties rather than rely on the more easily obtained relative measurements of mutant function that are currently used. A limited number of assays have been developed to measure these properties in a high throughput manner, but each is highly targeted to a specific protein target. We will expand and combine them in a unified platform that can reproducibly characterize libraries of protein mutants across many of the properties that may be correlated to overall function. For each mutant we will analyze how the measured physical properties correlate with each other and with mutant function. We can then compare these measurements to the predictions of the design algorithm and use the quantitative data obtained to improve the algorithm’s ability to accurately predict physical properties and how these properties affect protein function.

Our method can guide the implementation of additional design constraints that better represent the key
contributions to a protein’s ability to function. A more complete and accurate design algorithm will streamline the design process as it will reduce the need for multiple rounds of computational design and functional screening. In addition, more accurate design algorithms will open the door for more ambitious targets for protein design — for example, the design of novel functions or modes of action — that are currently out of reach.

Jason_GilmoreJASON GILMORE, PhD – Collaborating PI: Michael MacCoss, UW Genome Sciences

Jason is a 2014 recipient of a WRF Innovation fellowship and will be working jointly with Michael MacCoss (UW Genome Sciences) and David Baker (UW Biochemistry and IPD) to develop high-throughput, highly sensitive screens to accelerate experimental validation for de novo protein design. The first aim of this project will use liquid chromatography coupled tandem mass spectrometry (LC-MS/MS) to simultaneously quantify hundreds of protein designs expressed together in pooled cultures, as opposed to the traditional methods that require individual expression and confirmation of new designs. Additionally, this project will extend software and experimental techniques for MS-based protein chemical crosslinking experiments to rapidly evaluate disulfide linkages in small de novo designed proteins. Together, these mass spectrometry methods will improve the productivity of protein design protocols and accelerate the development of novel protein-based therapeutics.

During his undergraduate summers, Jason worked at the Pacific Northwest National Laboratory in Richland, WA where he helped to implement automated quality control metrics for the proteomics core facility. After graduating from the University of Pennsylvania in 2007, Jason returned to PNNL for one year and wrote a software tool for predicting protein-protein interaction probabilities. His dissertation work at Dartmouth College, in the proteomics laboratory of Dr. Scott Gerber, focused on the detection and quantification of phosphopeptide species by mass spectrometry-based shotgun sequencing. This included a publication on sequential digestion by complementary proteases to survey previously inaccessible regions of the proteome in complex biological mixture. Subsequently, he developed a computational technique to improve the sensitivity and precision of peptide quantification in cases where isotopically labeled standards failed to fully complement endogenous peptide profiles.

HerpoldtKARLA HERPOLDT, PhD – Collaborating PI: Patrick Stayton, Bioengineering

Protein-based nanoparticles have been designed and used for a variety of drug-delivery systems. These drug carriers are based on naturally self-assembled protein subunits which form a cage that can be used to trap pharmaceutical compounds. The use of naturally derived proteins offers benefits in terms of their biocompatibility, biodegradability, low toxicity and relative abundance. Despite these advantages, they remain limited in their use, being repurposed from their original biological application. In contrast, the computational design of protein nanomaterials has created the ability to design self-assembling cages which incorporate additional synthetic functionalities into their structure.

Working with the Baker/King labs Karla is working on using computational design to develop ‘smart’ protein cages that exhibit a strong response to environmental pH. In collaboration with the Stayton group these materials can then be loaded with polymer-prodrug carriers. It is hypothesized that the polymer therapeutics can be loaded via pH-dependent assembly of the cages, and subsequently their higher molecular weight will lead to intra-cage retention. In this way she hopes to develop targeted drug delivery vehicles which release their cargo inside the tumor microenvironment, minimizing chemotherapy dosage levels.

Karla obtained her MPhys in physics from the University of Oxford where she studied laser-plasma interactions and space dust formation before discovering a love of biological physics. She then carried out her PhD research in the lab of Prof. Molly Stevens at Imperial College London where her main focus was on the study of phage-derived peptides for use in diagnostics and therapeutics for HIV. She is also interested in peptide-protein interactions and the rational design of protein ligands. During her PhD Karla held a sabbatical fellowship at the UK Parliament’s Office of Science and Technology and was heavily involved with educational outreach.

HosseinzadehPARISA HOSSEINZADEH, PhD – Collaborating PI: Michael Gelb, Chemistry

As a WRF fellow, my goal is to develop new computational tools to design cyclic peptides and to use these peptides as specific inhibitors/binders to target enzymes/proteins. Cyclic peptide binders have the advantages of both proteins and small molecules: they can offer specificity through providing more contacts and they are small and usually more stable. Design of cyclic peptides is an exciting new addition to the field of protein design.

In particular, I am working with Dr. David Baker and Dr. Michael Gelb to design new cyclic peptide binders for specific inhibition and study of different members of secreted phospholipase A family of enzymes. These enzymes are known to be important in inflammatory disease states including asthma; however their exact roles remain elusive. While there are small molecule inhibitors for some members of these enzymes, similarly specific small-molecule based inhibitors have not been identified for all members of this family of enzymes hampering their study. This project is aimed to use the designed cyclic peptide to address this issue and provide better understanding of secreted phospholipases.

I was trained as a molecular biologist in my undergrad. My interest in proteins led me to do my graduate research on rational design of metalloproteins in the lab of Dr. Yi Lu. I was mainly focused on altering second shell interactions to tune the activity of proteins. My research provided a general guideline for tuning the redox potential of metal centers.


MARC LAJOIE, PhD – Collaborating PI: Nora Disis, UW Medicine/Medical Oncology

As a WRF Innovation Fellow at the Institute for Protein Design, Marc is working with the Baker, King, and Disis labs to develop next generation vaccines for cancer treatment and prevention. Antigen-presenting cells determine how the immune system responds to antigens. To exploit this control point, Marc is developing protein nanorobots capable of delivering immunogens directly to sub-compartments of dendritic cells in order to control how the immunogen is presented to naïve T cells. Cytosolic targeting could specifically activate CD8+ T cells that destroy cancer cells, and endosomal targeting could specifically activate CD4+ Th1 cells that promote a strong, sustained immune response.

After receiving a B.A. in biophysical chemistry from Dartmouth College, Marc then completed his PhD in chemical biology under the mentorship of George Church at Harvard Medical School. During his dissertation research, Marc developed genome engineering technologies to reassign the genetic code. His graduate research has implications for enabling virus resistance, improving biocontainment of recombinant organisms, and expanding the amino acid repertoire for industrial organisms. He received a NDSEG fellowship in 2009 and was named to Forbes 30 under 30 in Science in 2012.

Roy_picANINDYA ROY, PhD – Collaborating PI: David Rawling, Seattle Children’s/UW Immunology

As a WRF fellow, Anindya will be working with Dr. David Rawlings (Seattle Children’s Hospital) and Dr. David Baker (UW Biochemistry and IPD) to develop APRIL and BAFF specific inhibitors for the development of therapeutics for autoimmune diseases and cancer. Both these ligands have been shown to play critical roles in maintaining humoral immunity and signaling network pertaining to this ligand axis is highly exploited in cancer and autoimmune diseases. Current therapeutic approach relies on utilizing the soluble extracellular domain of the receptor for these ligands to block the signaling network. However, initial clinical trial results show that extracellular receptor decoy might not be a safe therapeutic agent presumably due to the complexity and shared ligand space of this signaling network under normal conditions. This project under the WRF fellowship aims to design orthogonal binding partner for BAFF and APRIL with high affinity and specificity.

Anindya started his scientific career with a special interest in chemical biology and protein biochemistry. He received his MSc from the Indian Institute of Technology (IIT-Kharagpur, India) studying protein-small molecule interactions using biophysical techniques. In 2008, Anindya moved to the US and received his PhD from Arizona State University under the guidance of Dr. Giovanna Ghirlanda where he worked on de novo design of artificial metalloproteins. In this work, he laid the foundation for designing multi-cofactor redox proteins that go beyond naturally existing systems.

SahtoeDANNY SAHTOE, PhD – Collaborating PI: Andrew Scharenberg, Seattle Children’s Research Institute

Plant pathogens cause significant economic damage since they often inject toxic small molecules and proteins into the cells of agriculturally important plants, such as tomato and bean plants, that can lead to crop loss. One of such pathogens is the bacterium Pseudomonas syringae. In his project*, Danny’s aim will be to introduce novel proteins, that are designed to bind and neutralize a subset of Pseudomonas syringae toxins, into plant cells in order to protect plants from such pathogens. The genetic accessibility of plants and the long standing GMO tradition in plant biotechnology, makes plants suitable candidates for computational protein design approaches. We hope that in the future, our designer proteins will complement the strategies that are currently used to combat plant disease in agriculture.

Danny obtained his PhD in structural biology and biochemistry in Titia Sixma’s lab at the Netherlands Cancer Institute in Amsterdam, the Netherlands. There he worked on elucidating the regulatory mechanisms of a special class of proteases called deubiquitinating enzymes using X-ray crystallography and a variety of biochemical and biophysical techniques.

*Danny Sahtoe is an EMBO long term fellow and was also awarded a WRF innovation fellowship that supports his non-salary research expenses.


FRANZISKA SEEGER, PhD – Collaborating PI: Mohamed Oukka, Seattle Children’s

The ability to therapeutically modulate protein-protein interactions bears enormous potential for the treatment of human diseases. Traditional experimental tools are limited by naturally occurring scaffolds to target proteins central to disease development. Thus, the capability to engineer entirely novel proteins de novo – with a particular structure to fulfill a desired therapeutic function – would be a major advancement in the field of drug development. Franziska is working with David Baker (UW Biochemistry and IPD) and Mohamed Oukka (Seattle Children’s Hospital) to computationally design high-affinity binders to the cytokines IL-23 and IL-17 for the treatment of autoimmune diseases. Autoimmune diseases such as Multiple Sclerosis and Crohn’s Disease have posed a major challenge – elucidating their molecular mechanisms as well as finding effective therapies have been formidable. Current treatments have severe side effects and merely delay disease onset. Small, rationally designed protein therapeutics may have multiple advantages over currently available antibody therapeutics. In this proposal, we aim to design stable and effective IL-17 and IL-23 cytokine binders with minimal immunogenicity and favorable biodistribution to test the role that computationally designed proteins can play as novel therapies.

Franziska has always been fascinated by biochemical pathways and protein-protein interactions and will apply her expertise in protein biochemistry, structural biology, and protein-protein interactions to designing novel interfaces between de novo designed binders and the IL-17 and Il-23 cytokines. In her graduate work with Elsa Garcin at the University of Maryland Baltimore County and John Tainer at Lawrence Berkeley National Lab, Franziska determined the first human heterodimeric wild-type structure of the catalytic domain of soluble Guanylate Cyclase (sGC) – an important drug target for the treatment of cardiovascular diseases. Her work furthered our understanding of sGC heterodimerization and activation and opened new drug discovery routes for targeting the NO–sGC–cGMP pathway in acute heart failure and pulmonary hypertension.