Think of living cells as minicomputers that can, given sufficiently precise read-write “hardware,” record and process single-nucleotide mutations and DNA segments instead of bits and bytes. Such hardware has just been introduced by MIT scientists. At its heart is a CRISPR base editing system that can introduce cytosine (C) to thymine (T) mutations to stretches of DNA targeted by guide RNA.
CRISPR base editing systems can reliably overwrite DNA at the “bit level” because these systems, unlike ordinary CRISPR gene editing systems, don’t subject DNA to double-strand breaks. Ordinary CRISPR gene editing systems rely on the cell’s own DNA repair machinery, which can lead to uncertain mutational outcomes, limiting the amount of information that can be stored.
With their application of CRISPR base editing, the MIT scientists may bring about Cellular Computation 2.0. They have created a system—called DOMINO (DNA-based Ordered Memory and Iteration Network Operator)—that could be used to record the intensity, duration, sequence, and timing of many events in the life of a cell, such as exposures to certain chemicals. Cellular “memories” could then be processed so that one event, or series of events, would trigger another event, such as the production of a fluorescent protein.
“We need better strategies to unravel how complex biology works, especially in diseases like cancer where multiple biological events can occur to transform normal cell into diseased ones,” said Timothy Lu, MD, PhD, an electrical engineer and computer scientist at MIT and the Broad Institute. Lu and colleagues, including Fahim Farzadfard, PhD, a Schmidt Science Fellow, developed DOMINO to surpass the recording and scaling capabilities of existing strategies.
“With [DOMINO] we are using DNA as a memory tape to permanently record biological events that are occur in disease,” Lu elaborated. “This technology can give us deeper insights into what signals go up and down over time to drive disease development.”
Details about DOMINO appeared August 22 in the journal Molecular Cell, in an article titled, “Single-Nucleotide-Resolution Computing and Memory in Living Cells.” According to this article, DOMINO can be used to encode logic and memory in bacterial and eukaryotic cells.
“DOMINO operators enable analog and digital molecular recording for long-term monitoring of signaling dynamics and cellular events,” the article’s authors wrote. “Furthermore, multiple operators can be layered and interconnected to encode order-independent, sequential, and temporal logic, allowing recording and control over the combination, order, and timing of molecular events in cells.”
DOMINO is a modular system made of operator units. Each unit is composed of a base editor (a noncutting variant of Cas9 fused to a cytidine deaminase) and one guide RNA that binds to its complementary sequence on the genome and recruits the base editor to that sequence. Once recruited, the base editor can introduce C-to-T mutations.
The guide RNA in DOMINO operator can be designed in such a way that it can bind to its target sequence only after a certain mutation is first introduced into that sequence by a previous event. Thus, for the base editor to make a change, a previous mutation must have been made so that the next mutation can be made. If a particular sequence is mutated, then the next step can happen. If it is not mutated, then the next step cannot happen.
“You can design the system so that each combination of the inputs gives you a unique mutational signature,” Farzadfard explained. “And from that signature, you can tell which combination of the inputs has been present.”
The researchers used DOMINO to create circuits that perform logic calculations, including AND and OR gates, which can detect the presence of multiple inputs. They also created circuits that can record cascades of events that occur in a certain order, similar to an array of dominos falling.
Most previous versions of cellular memory storage have required stored memories to be read by sequencing the DNA. However, that process destroys the cells, so no further experiments can be done on them. In this study, the researchers designed their circuits so that the final output would activate the gene for green fluorescent protein (GFP). By measuring the level of fluorescence, the researchers could estimate how many mutations had accumulated, without killing the cells.
Currently, the team has used the technology to record events on the order of hours. But they are hopeful they can improve the temporal resolution and adapt it for recording cellular events that occur on faster timescales. They also plan on expanding DOMINO’s application to highly parallel computing and recording to process and interrogate more complex biological events.
“This type of biocomputing is an exciting new way of getting and processing information,” declares Lu. “It is part of a longer-term path to take advantage of the natural memory and computing capabilities in cells.”
Potential applications of DOMINO include creating mouse immune cells that produce GFP when certain signaling molecules are activated, which researchers could analyze by periodically taking blood samples from the mice.
Another possible application is designing circuits that can detect gene activity linked to cancer, the researchers said. Such circuits could also be programmed to turn on genes that produce cancer-fighting molecules, allowing the system to both detect and treat the disease. “Those are applications that may be further away from real-world use,” Lu noted, “but they are certainly enabled by this type of technology.”