Translation is the process of turning basic scientific research into therapies that cure disease, new sources of energy that heal the planet, and other things that move the world forward. The Translation podcast takes a deep dive into scientific advancements with a huge potential to improve society. We talk directly with the people advancing the science with their own hands and minds, and focus on how we can translate the science from the bench to the benefit of all.
Initially centered on biology and synthetic biology, we’ll talk with the most promising young scientists in the field. We aim to demystify the science for a general audience and to shine a light on how great science turns into great business. We hope these discussions will inspire scientists, entrepreneurs, and investors to help commercialize breakthrough research.
Follow Fifty Years on Twitter to get notified when new episodes drop. If you’re an author of an upcoming paper in bio or know of any interesting papers dropping soon and want to hear from the authors, drop us an email at translation@fiftyyears.com.
Episode Contributors: Ayush Noori, Ashton Trotman-Grant, Linda Goodman
Episode Summary: Millions of people die every year from chronic diseases. Traditional drug discovery has failed in identifying solutions to many of these persistent health challenges. Functional genomics is offering a way forward by identifying gene networks and enabling the development of drugs with very specific targets. But, rather than just relying on gene targets within humans, Linda and her company, Fauna Bio, is casting a wider net across the animal kingdom. Extreme adaptation is common across many mammals, giving us an incredible pool of potential targets to go after. Whereas a single heart attack can kill a person, certain animals not only survive 25 heart attacks a year but also go on to thrive, living 2x longer than other mammals their size. By identifying and understanding the gene networks underlying these extreme adaptations, Fauna can identify novel targets across 415 different species, map them to human genes, and develop drugs that exploit our natural protective physiological mechanisms.
However, the process by which gene transcription is regulated is incredibly complex; thus, prediction transcriptional regulation has been an open problem in the field for over half a century. In his work, Eeshit used neural networks to predict the levels of gene expression based on promoter sequences. Then, he reverse engineered the model to design specific sequences that can elicit desired expression levels. Eeshit’s work developing a sequence-to-expression oracle also provided a framework to model and test theories of gene evolution.
About the Guest
Linda is the Co-Founder and CTO at Fauna Bio, a biotechnology company leveraging the science of hibernation to improve healthcare for humans.
She earned an MPhil in Computational Biology from the University of Cambridge and got her Ph.D. in Genetics and Genomics from Harvard University.
Key Takeaways
At Fauna Bio, Linda Goodman and her team are working to better understand the biological networks that underlie these adaptations, in hopes of developing therapeutics inspired by the adaptations of the animal kingdom.
Impact
Drawing on a completely new source of knowledge about the defense mechanisms of living organisms, Fauna Bio goes beyond the limitations of traditional drug development and looks for better, more effective drugs based on natural defense mechanisms.
Company: Fauna Bio
Episode Contributors: Michael Chavez, Alex Teng, Daniel Goodman
Episode Summary: Chimeric antigen receptors, or CARs, repurpose the build-in targeting and homing signals of our immune system to direct T cells to find and eliminate cancers. Although CAR-T cells have transformed the care of liquid tumors in the circulating blood, like B cell leukemia and lymphoma, CAR-T therapy has shown limited efficacy against solid tumors. To unlock the full potential of CAR-T therapies, better receptor designs are needed. Unfortunately, the space of potential designs is too large to check one by one. To design better CARs, Dan and his co-author Camillia Azimi developed CAR Pooling, an approach to multiplex CAR designs by testing many at once with different immune costimulatory domains. They select the CARs that exhibit the best anti-tumor response and develop novel CARs that endow the T cells with better anti-tumor properties. Their methods and designs may help us develop therapies for refractory, treatment-resistant cancers, and may enable CAR-T cells to cure infectious diseases, autoimmunity, and beyond.
About the Author
During his PhD in George Church’s lab at Harvard Medical School, Dan studied interactions between bacterial transcription and translation, built and measured libraries of tunable synthetic biosensors, and constructed a new version of the E. coli genome capable of incorporating new synthetic amino acids into its proteins. He also built a high-throughput microbial genome design and analysis software platform called Millstone.
As a Jane Coffin Childs Postdoctoral Fellow at UCSF, Dan is currently applying these high-throughput synthetic approaches to engineer T cells for the treatment of cancer and autoimmune disease. He is also working in the Bluestone, Roybal, and Marson labs.
Key Takeaways
By genetically engineering the chimeric antigen receptor (CAR), T cells can be programmed to target new proteins that are markers of cancer, infectious diseases, and other important disorders.
However, to realize this vision, more powerful CARs with better designs are needed - current CAR-T therapies have their restraints, including limited performance against solid tumors and lack of persistence and long-term efficacy in patients.
An important part of the CAR response is “costimulation,” which is mediated by the 4-1BB or CD28 intracellular domains in all CARs currently in the clinic. Better designs of costimulatory domains could unlock the next-generation of CAR-T therapies.
Since there are so many possibilities for costimulatory domain designs, it’s difficult to test them all in the lab.
Based on his experience in the Church Lab, Dan has developed tools to “multiplex” biological experiments; that is, to test multiple biological hypotheses in the same experiment and increase the screening power.
Dan and his co-author Camillia Azimi developed “CAR Pooling”, a multiplexed approach to test many CAR designs at once.
Using CAR Pooling, Dan tested 40 CARs with different costimulatory domains in pooled assays and identified several novel cosignaling domains from the TNF receptor family that enhance persistence or cytotoxicity over FDA-approved CARs.
To characterize the different CARs, Dan also used RNA-sequencing.
Impact
The CAR Pooling approach may enable new, potent CAR-T therapies that can change the game for solid tumors and other cancers that are currently tough to treat.
Highly multiplexed approaches like CAR Pooling will allow us to build highly complex, programmable systems and design the future of cell engineering beyond CAR-T.
In addition to new therapeutics, high-throughput studies will allow us to understand the “design rules” of synthetic receptors and improve our understanding of basic immunology.
Paper: Pooled screening of CAR T cells identifies diverse immune signaling domains for next-generation immunotherapies
First Author: Anastasia Ershova
Episode Summary: DNA is an ideal molecule for storing information in our genomes because it’s stable, programmable, and well understood. The same qualities make DNA a great building block or construction material for nanoscale biomolecular structures that have nothing to do with our genome, like molecular scaffolds created by folding DNA into 2D and 3D shapes. This technology is known as DNA origami.
However, the practical applications of DNA origami are limited by spontaneous growth and poor reaction yields. Anastasia developed a method that uses crisscross DNA polymerization of single-stranded DNA slats or DNA origami tiles to assemble DNA structures in a seed-dependent manner. This work may be useful to produce ultrasensitive, next-generation diagnostics or in programmable biofabrication at the multi-micron scale.
About the Author
Anastasia is a PhD candidate at Harvard University, currently working on DNA nanotechnology in William Shih's lab at the Wyss Institute and Dana-Farber Cancer Institute.
She received her bachelor’s degree in Natural Sciences from Cambridge University.
During her PhD at Harvard, she co-founded the Molecular Programming Interest Group, an international community of students in the molecular programming, DNA computing and related fields.
Impact
DNA Origami will provide us with a plethora of new information on biology and physics.
By manipulating that data on the nanoscale, we can get answers to a lot of questions in the future.
Quick diagnostics can enable people all over the world to quickly get diagnosis-related answers and seek targeted treatment.
Papers: Robust nucleation control via crisscross polymerization of highly coordinated DNA slats
Multi-micron crisscross structures from combinatorially assembled DNA-origami slats
First Author: Jocelyn Kishi
Episode Summary: Technologies like next-generation sequencing allow us to understand which RNA transcripts and proteins are expressed in biological tissues. However, it’s often equally important to understand how cells or molecules are positioned relative to one another! Whether it be a cell changing its shape, an organelle ramping up a metabolic process, or a DNA molecule traveling across the nucleus, understanding spatial context is critical. Current approaches for spatial sequencing are limited by cost, complicated equipment, sample damage, or low resolution. Recognizing this challenge, Josie and team developed Light-seq, a cheap and accessible method to combine sequencing and imaging in intact biological samples. Not only is the method inexpensive, but Light-seq can also achieve unprecedented spatial resolution by using light to add genetic barcodes to any RNA, allowing scientists to determine exactly where sequencing should occur with extreme precision. By helping researchers to understand spatial context, Light-seq-driven insights may illuminate cancer, neurodegeneration, and autoimmunity.
About the Author
Following her lifelong passion for computer programming, Josie studied Computer Science at Caltech and worked as a software engineering intern at Google. At Caltech, a biomolecular computation course introduced her to the field of biomolecular programming.
Josie quickly got excited about the intersection of computers and biology and its potential to bring about positive change in the world. She pursued this interest in her graduate studies in the Wyss Institute for Biologically Inspired Engineering at Harvard, where – as first a postdoctoral fellow, and then the Technology Development Fellow – she developed platform technologies for DNA-based imaging and sequencing assays.
Key Takeaways
Next-generation sequencing is a powerful technology to read the transcriptomic state of biological tissues by surveying the RNA transcripts present.
However, it’s important to understand not only what is being expressed but where this expression occurs! The spatial arrangement, structure, and interactions between molecules are critical to define the functions of biological systems.
By linking imaging with -omics profiling, the field of spatial biology seeks to understand molecules like RNAs in their 2D and 3D contexts.
Unfortunately, currently available spatial transcriptomics methods are limited in their ability to select individual cells with complex morphologies, require expensive instrumentation or complex microfluidics setups to the tune of several $100K, and often damage the samples.
Further, rare cells are often missed due to lower sequencing throughput, even though they may be critical for biological activity.
Recognizing this challenge, Josie and her collaborators developed Light-seq, a new, cheap, and accessible approach for single-cell spatial indexing and sequencing of intact biological samples.
Using light-controlled nucleotide crosslinking chemistry, Light-seq can correlate multi-dimensional and high-resolution cellular phenotypes – like morphology, protein markers, spatial organization) – to transcriptomic profiles across diverse sample types.
In particular, using the biological equivalent of photolithography, Light-seq can add genetic barcodes to any RNA by shining light on it, allowing scientists to control exactly where sequencing should occur with extreme precision – up to the subcellular level.
Light-seq can operate directly on the sample: the method does not require cellular dissociation, microfluidic separation/sorting, or custom capture substrates or pre-patterned slides.
Samples used for Light-seq remain intact for downstream analysis post-sequencing.
Josie evaluated Light-seq on mouse retinal sections to barcode three different cell layers and study the rare dopaminergic amacrine cells (DACs).
Impact
Josie created a cheap, accessible, and powerful tool for scientists to perform spatial sequencing at unprecedented resolution without requiring expensive or complicated setups.
By enabling new advances in spatial biology, Light-seq has the potential to help biologists discover biomarkers for disease, measure on and off target effects of therapeutic candidates, and illuminate poorly understood biological mechanisms where understanding spatial context makes all the difference.
Paper: Light-Seq: Light-directed in situ barcoding of biomolecules in fixed cells and tissues for spatially indexed sequencing
First Author: Christina Boville
Episode Summary
Commodity molecules are vital ingredients for everything important to our modern world, including food, energy, and medicine. However, creating these molecules still largely relies on old processes that suffer from low yield, laborious methods, and unsustainable inputs and byproducts. Tina envisions a world where all molecules are created quickly, easily, and sustainably through enzymes, biology’s chemical catalyst. Here, Tina describes how she used an extremely powerful method called directed evolution to build a novel enzyme that can create the non-canonical amino acid 4-cyanotryptophan, a fluorescent molecule that is extremely difficult to make with traditional chemistry.
About the Author
Tina performed this work as a postdoc in the lab of Nobel Laureate Professor Frances Arnold at Caltech. The lab is world renowned for developing the methods around directed evolution and applying them to create proteins that do unnatural chemistries.
Tina is now the co-founder and CEO at Aralez Bio whose focus is on developing efficient, sustainable alternatives to chemical manufacturing through enzyme engineering.
Key Takeaways
Enzymes are proteins that induce specific chemical reactions to occur. They can create molecules much more efficiently and sustainably than using traditional chemistry
One class of molecules, called non-canonical amino acids, are extremely important precursors to drugs and have specific properties that make them desirable for biotech.
Making highly pure non-canonical amino acids is difficult with traditional chemistry, requiring many time-consuming reactions and toxic byproducts. But nature has yet to generate an enzyme that can create these.
A process called directed evolution mimics nature’s process by heavily mutating a starting enzyme and sequentially pushing it to make a molecule of interest.
When using directed evolution, “you get what you screen for”. Said another way: the outcome of the process is highly dependent on how the experiment was run and what was optimized for.
With directed evolution, the non-canonical amino acid 4-cyanotryptophan is generated overnight with no harmful byproducts; something that would take a team of chemists months to do.
Translation
The evolved enzyme that creates 4-cyanotryptophan became the cornerstone technology of Aralez Bio.
Tina spent the last parts of her postdoc defining customers and building a team to launch the company.
Through enzyme engineering, Aralez Bio plans to replace many unsustainable and time consuming chemistries that currently plague commodity molecules.
Paper: Improved Synthesis of 4-Cyanotryptophan and Other Tryptophan Analogues in Aqueous Solvent Using Variants of TrpB from Thermotoga maritima. Journal of Organic Chemistry, 2018
First Author: Surojit “Surge” Biswas
Episode Summary
Protein engineering has been dominated by two opposing paradigms; directed evolution, a massive screening technique, and rational design, a completely computational approach. Surge has fused these two paradigms by developing a machine learning technique that discovers an optimal protein design by training on a low number of engineered proteins. Here, Surge discusses how this hybrid method works, how it enabled the creation of better fluorophores and enzymes, and what this method will unlock next.
About the Author
Surge performed this work as a graduate student at Harvard in the lab of Professor George Church. George is one of the founding fathers of synthetic biology and the lab is known for developing high throughput methods to design, build, and test bioengineered parts.
As CEO and co-founder of Nabla Bio, Surge is now focused on pointing the algorithms and methods developed in his academic work toward building proteins that can improve human health or protect the environment.
Key Takeaways
Methods from natural language processing algorithms (like Siri or Alexa) are adapted to understand how nature builds proteins.
These machine learning algorithms distill fundamental structural, as well as evolutionary and and biophysical properties about proteins.
Fusing these models with real world data enables us to make proteins with improved or novel functionality.
By checking how a few mutations (low-n) affect the function of a protein, Surge evolved proteins in a computer to make better fluorescent proteins and enzymes.
These models know a lot about proteins in general and can therefore be applied to a wide variety of tasks that improve human health and protect the environment.
Translation
This methodology dramatically reduces the time, cost, and labor of evolving proteins, making it a perfect tool to create commodity proteins.
Based on this technology, Surge co-founded Nabla Bio whose goal is to engineer supernatural proteins that enable biology to solve the world's biggest problems.
Paper: Low-N protein engineering with data-efficient deep learning. bioRxiV, 2020
