Programmable DNA Enables Rational Design of Protein Crystals, Overturning Decades-Old Assumption
Key Insights
Northwestern University chemists developed a method using flexible DNA (search) strands as programmable "molecular glue" to direct proteins into diffraction-quality crystals with atomic-level order.
The approach replaces the traditional trial-and-error crystallization process with intentional design, successfully growing over 1,000 crystals and determining 28 distinct protein-DNA (search) structures.
The work overturns the long-held assumption that flexible building blocks cannot produce crystals with atomic-level order, demonstrating that DNA (search) bonds represent a fundamentally new class of chemical bond.
For decades, persuading proteins to form crystals suitable for X-ray crystallography has been one of structural biology's most frustrating bottlenecks—a process defined by painstaking trial and error. Now, researchers at Northwestern University have transformed this unpredictable art into a programmable science by repurposing flexible DNA (search) strands as both a blueprint and molecular glue, directing proteins to assemble into diffraction-quality crystals with atomic-level precision.
The study, published in Science Advances, was led by Chad A. Mirkin (search), the George B. Rathmann Professor of Chemistry at Northwestern and founding director of the International Institute for Nanotechnology. Zhenyu Han, a graduate student in Mirkin's laboratory, served as the paper's first author.
"The implications of this research are profound," Mirkin said. "Proteins are the building blocks of life, and their structure determines their function. When we intentionally determine those structures, we gain powerful new insights into how proteins recognize other molecules, catalyze chemical reactions and interact with living systems. That knowledge ultimately helps us identify new drug targets, design new medicines and engineer new materials."
DNA (search) as a Programmable Bonding Element
The approach builds on foundational work from Mirkin's laboratory dating back to 1996, when his team first demonstrated the use of complementary DNA (search) strands to assemble gold nanoparticles into ordered structures, published in Nature. Over the subsequent three decades, DNA has been used as a programmable bonding element to organize nanoparticles, polymers, and other nanoscale building blocks.
"Over 30 years ago, the Mirkin group proposed the idea to take nanoparticles and modify them with DNA (search) to create programmable atom equivalents," Han said. "Nature's nanoparticles are proteins. The difference between synthetic nanoparticles and proteins is that proteins are naturally uniform and molecularly precise. So, if we modify proteins with DNA and assemble them, we can form high-quality, atomically precise crystals, overcoming a decades-old challenge in the field of DNA-programmable assembly."
How DNA (search) Directs Protein Assembly
Traditional protein crystallization relies on weak, fortuitous chemical interactions between protein surfaces. Mirkin's team instead exploited DNA (search)'s predictable base-pairing rules—A with T, C with G—to create programmable bonds. By attaching short DNA strands to each protein and varying key design features such as DNA strand length and placement, the researchers could dictate precisely which molecules connect and how they orient within the crystal lattice.
"The well-defined DNA (search)-DNA interactions drove the assembly and crystallization process," Han explained. "As a result, the proteins not only adopted specific positions and orientations, but the atoms within each protein are also aligned in the exact same way throughout the crystal. When that happens, you can use single-crystal X-ray diffraction methods to determine their structure and see atomic-level details within the protein."
Validation Through Systematic Testing
To validate the method's reliability, the team grew more than 1,000 protein crystals and determined the atomic structures of 28 distinct protein-DNA (search) designs. Despite systematically altering DNA length and position, the resulting crystals consistently assembled into the designed architectures. Direct visualization of DNA double helices linking neighboring proteins within the crystals confirmed that DNA—not chance—directed the assembly process.
"These numbers show that we didn't just get lucky," Mirkin said. "We systematically demonstrated this approach across literally hundreds of possibilities and showed over and over again that it's a reliable method. The DNA (search) bond is a fundamentally new class of chemical bond, and this work shows how it can be used in ways that conventional chemical bonds cannot—to create a new type of material that yields new fundamental knowledge as well as technological breakthroughs that could significantly benefit society."
Broader Implications
Beyond simplifying one of structural biology's most difficult challenges, the work overturns a long-held assumption that flexible building blocks cannot produce crystals with atomic-level order. The resulting crystals are unusually soft and flexible while maintaining the high structural order necessary for atomic-resolution structure determination. The approach also opens avenues for a new generation of flexible, customizable biomaterials with potential applications in biosensing, drug delivery, bioelectronics, and robotics.
The study, titled "Diffraction-quality, ultraflexible protein single crystals engineered with DNA (search)," was supported by the Air Force Office of Scientific Research and the National Science Foundation.