XSEDE Allocations on Stampede2 and Simulation of Comet Speed ​​of Protein Oligomers – ScienceDaily

Red blood cells are extraordinary. They take oxygen from our lungs and carry it throughout our body to keep us alive. The hemoglobin molecule in red blood cells carries oxygen by changing its form in an all or nothing way. In hemoglobin, four copies of the same protein open and close like flower petals, structurally coupled to meet each other. By using supercomputers, scientists are just beginning to design proteins that bademble to combine and resemble vital molecules such as hemoglobin. Scientists say their methods could be applied to useful technologies such as pharmaceutical targeting, artificial energy recovery, "smart" detection and building materials, and so on.

A scientific team did this work by overloading the proteins, which meant that they changed the protein subunits, the amino acids, to give the proteins an artificially high positive or negative charge. Using jellyfish-derived proteins, scientists were able to bademble a complex structure of sixteen proteins consisting of two stacked octamers by supercharging only, results reported in January 2019 in the journal. Nature Chemistry.

The team then used supercomputer simulations to validate and inform these experimental results. Researchers have received supercomputer allocations on Stampede2 at the Texas Advanced Computing Center (TACC) and Comet at the San Diego Supercomputer Center (SDSC) through XSEDE, the Extreme Science Discovery Environment and the University of California. Engineering funded by the National Science Foundation (NSF).

"We found that by taking proteins that did not normally interact with each other, we could make copies with a very positive or very negative charge," said Anna Simon, co-author of the report. study, postdoctoral researcher at Ellington Lab of UT Austin. . "By combining heavily positively and negatively charged copies, we can make the proteins bademble into very specific structured bademblies," Simon said. Scientists call their strategy "supercharged protein badembly", in which they induce defined protein interactions by combining modified supercharged variants.

"We have exploited a well-known fundamental principle of nature, which attracts opposite charges," added Jens Glaser, co-author of the study. Glaser is an badociate scientist at the Glotzer Group of the Department of Chemical Engineering at the University of Michigan. "The group of Anna Simon discovered that when they mix these charged variants of green fluorescent proteins, they get highly ordered structures – it was a real surprise," Glaser said.

The stacked octameric structure looks like a braided ring. It is composed of 16 proteins – two rings of eight intertwined that interact in very specific and discrete areas. "The reason it is so difficult to design synthetically interacting proteins is that it is very difficult to bademble and align them properly so that proteins can bademble into structures. bigger and more regular, "explains Simon. They circumvented the problem by adding many positive and negative charges to technical variants of green fluorescent protein (GFP), a well-studied protein from "laboratory mouse" derived from the jellyfish Aequorea victoria.

The positively charged protein, which they called cerulean fluorescent protein (Ceru) +32, further provided opportunities for interaction with the GFP-17 negatively charged protein. "By giving these proteins all these opportunities, these different places where they could potentially interact, they have been able to choose the right ones," said Simon. "There were certain models and interactions present, available and favored with energy, that we had not necessarily predicted in advance, which would allow them to bademble in these specific forms."

To obtain genetically engineered engineered fluorescent proteins, Simon and his co-authors Arti Pothukuchy, Jimmy Gollihar and Barrett Morrow coded their genes, including a chemical tag used for purification on portable DNA fragments called plasmids in E. coli, and then collected the labeled protein. E. coli grew up. The scientists mixed the proteins together. At first, they thought that proteins could simply interact to form large clumps of irregular structure. "But what we kept watching is this weird and fun pic of about 12 nanometers, much smaller than a big bunch of protein, but significantly bigger than that of a single protein" said Simon.

They measured the size of the particles formed using a Zetasizer instrument at Texas Materials Institute at UT Austin, and verified that the particles contained both cerulean proteins and GFP Förster Resonance Energy Transfer (FRET) proteins. ), which measure the energy transfer between different colored fluorescent colors. Proteins fluoresce in response to different light energies to see if they are close to each other. Negative staining electron microscopy identified the specific particle structure, led by the David Taylor group, badistant professor of molecular bioscience at UT Austin. He showed that the 12-nm particle consisted of a stacked octamer composed of sixteen proteins. "We discovered that it was beautiful shaped flower structures," Simon said. Taylor's Taylor group co-author, Yi Zhou, formed by Taylor's UT Austin, further increased the resolution using cryo-electron microscopy to reveal atomic-level details of the stacked octamer.

According to Jens Glaser, computer modeling has refined the measurements of how proteins have been organized to give a clear picture of the beautiful flower-shaped structure. "We had to come up with a model complex enough to describe the physics of charged green fluorescent proteins and present all the relevant atomistic details, while being efficient enough to allow us to simulate this scenario on a realistic time scale. would have taken more than a year to get a single computer simulation, no matter how fast, "Glaser said.

They simplified the model by reducing resolution without sacrificing important details of protein interactions. "That's why we used a model where the shape of the protein is exactly represented by a molecular surface, just like that measured from the crystallographic structure of the protein," added Glaser.

"What really helped us turn the tide and improve what we've learned from our simulations is cryo-EM data," said Vyas Ramasubramani, a graduate student in chemical engineering at the University. from Michigan. "That's what really helped us find the optimal configuration to integrate into these simulations, which then allowed us to validate the stability arguments we were putting forward and, hopefully, predict how we could destabilize or modify this structure, "said Ramasubramani.

Scientists needed a lot of computing power to perform the calculations at the desired scale.

"We used XSEDE to take these huge systems, in which you have a lot of different elements interacting, and calculate all of that at the same time, so when you start moving your system forward for a while, you you can get an idea of ​​how it would evolve over fairly real time scales, "Ramasubramani said. "If you had tried to do the same kind of simulation that we did on a laptop, it would have taken months, if not years, to actually understand if some sort of structure would be stable or not. able to use XSEDE, where you could use basically 48 cores, 48 ​​compute units at the same time to make these calculations very parallel, we would have done that much more slowly. "

The TACC Stampede2 supercomputer contains 4,200 Intel Knights Landing compute nodes and 1,736 Intel Skylake X compute nodes. Each Skylake node has 48 cores, the base unit of a computer processor. "The Skylake nodes of the Stampede2 supercomputer have been instrumental in achieving the performance needed to compute these electrostatic interactions effectively acting between the oppositely charged proteins," Glaser said. "The availability of the Stampede2 supercomputer was at the right time to allow us to perform these simulations."

Initially, the scientific team tested its simulations on the Comet system at SDSC. "When we began to determine the type of model to use and if this simplified model gave us reasonable results, Comet was a great place to try these simulations," said Ramasubramani. "Comet was an excellent test bench for what we were doing."

Scientists hope that this work will help to better understand why so many proteins in nature will oligomerize or come together to form more complex and interesting structures.

"We showed that it was not necessary to have a very specific and pre-distinguished set of plans and interactions for these structures to form," Simon said. "This is important because it means that we can perhaps, and most likely, take other sets of molecules that we wish to oligomerize and generate positively and negatively charged variants, combine them and specifically order them."

Natural biomaterials such as bones, feathers and shells can be both strong and lightweight. "We think that supercharged protein badembly is a simpler way of developing the kind of materials that have interesting synthetic properties without having to spend so much time or knowing exactly how they will be formed before," he said. Simon. "We believe this will accelerate the ability of synthetic materials engineering, discovery and exploration of these nanostructured protein materials."