NEW YORK – Proteomics firm Nautilus Biotechnology is aiming to have an initial version of its proteomic analysis platform up and running by the end of the year, with a commercial launch planned for 2023.
This week, the Seattle-based company made its debut on the Nasdaq after closing a business combination transaction with Arya Sciences Acquisition Corp III, a special purpose acquisition company sponsored by Perceptive Advisors.
Nautilus received gross proceeds of roughly $345 million from the transaction, which it initially announced in February. The proceeds comprised approximately $145 million of funds held in Arya III's trust account and a concurrent private investment in public equity financing of $200 million.
Cofounded in 2016 by CEO Sujal Patel, former cofounder and CEO of computer firm Isilon Systems, and Chief Scientist Parag Mallick, an associate professor at Stanford University, Nautilus has developed an array-based protein analysis platform that combines machine learning and iterative rounds of affinity reagents to enable single-molecule measurements at proteome scale.
The platform uses nanofabricated chips with individual wells that are functionalized with scaffolds that each bind a single protein. When a sample is introduced to a chip, individual proteins bind to individual scaffolds so that each well contains a single protein. These chips have room for billions of individual proteins, allowing the system to sample across a large dynamic range.
Once proteins from the sample of interest are bound to the chip, the system probes them iteratively with multiple affinity agents. In conventional protein analysis methods, the goal is typically to use a single affinity reagent known to be highly specific to a particular protein of interest. The Nautilus system takes a different approach, observing the binding patterns of many relatively non-specific probes across the proteins bound to the chip and then using machine learning to make protein identifications based on the different patterns of binding observed. Because the platform is a single-molecule technology, proteins are quantified simply by counting the different proteins following identification.
According to Nautilus, its machine learning software can identify greater than 95 percent of the human proteome with the data collected from around 300 cycles of probing. At this level of analysis, the system could process around one to two samples per day, though the company said it plans to introduce multiplexing capabilities that would allow for simultaneous analysis of six samples. At six to 12 samples a day, the approach would still be relatively low-throughput compared to a number of recently developed mass spec workflows, though it would be significantly more comprehensive than those approaches.
Nautilus is one of a number of firms — both relatively established ones like SomaLogic and Olink as well as emerging companies like Quantum-Si, NanoMosaic, Encodia, and Erisyon —that aim to provide an alternative to mass spectrometry for proteome-scale discovery and research.
Mallick cited a number of factors that he said could allow Nautilus to penetrate more deeply into the proteome than mass spec, including the fact that its platform operates on intact proteins as opposed to peptides and that it doesn't suffer from factors like ion suppression that bias what peptides are seen by the mass spectrometer. He said that to date, Nautilus has demonstrated the platform's ability to detect proteins across seven orders of magnitude dynamic range.
The system uses click chemistry to bind proteins from target samples to the chip scaffolds, labeling protein lysines and cysteines with crosslinkers that contain a tetrazine moiety that binds to a trans-cyclooctene molecule on the scaffold.
As for whether the scaffolding approach may bias the system toward capture of some proteins over others, Mallick said that while "it's certainly true that there are some proteins that are more likely to have their lysines functionalized, there are a lot of lysines on proteins, and the same with cysteines."
"From what we have observed, [the tetrazine labels] are modifying the vast majority of proteins in the sample," he said.
In addition to proteome-scale measurements, Nautilus is also positioning the system as a tool for more targeted analyses, in particular for looking at the presence of different variants of a protein, or proteoforms, in a sample. In this case, researchers can immobilize proteins of interest and then probe them repeatedly using affinity reagents specific to different modification sites — for instance, various types of phosphorylation. By probing the immobilized molecules with reagents to different modifications, researchers can generate profiles of the proteoform heterogeneity of a sample.
"Being able to probe [the proteins] over and over and over again allows you to say, for instance, we have six molecules that have this triple phosphorylation and we have 100 molecules that have this quadruple modification and a truncation," Mallick said. "In five cycles, you can elaborate 32 distinct proteoforms and get single-molecule quantification of how much is present in the sample."
This kind of targeted work is the current focus of Nautilus' product development and work with customers. The company aims by the end of 2021 to demonstrate the ability of its instrument to read out the result of experiments probing single molecules multiple times, as in the modification study Mallick described. The company signed a research collaboration agreement with Genentech in December 2020, under which it is using its system to map the proteoforms of one of the drugmaker's protein targets of interest and aims to submit the results of the collaboration for publication in late 2021.
Mallick said the company is working with several other collaborators but declined to name them.
Nautilus plans to demonstrate the ability to measure 2,500 proteins per run in early 2022, upping that to 10,000 proteins per run by late 2022, and to analyze the full proteome by the middle of 2023, with the commercial launch of an instrument targeted for 2023. The system is expected to sell for around $1 million, which would be comparable to the prices of high-end mass spectrometry systems used for proteomics.