Skip to main content
Premium Trial:

Request an Annual Quote

FEATURE: The Protein Structure Race Has Begun

NEW YORK, Jan 26 – Protein structure determination may just be the next life sciences land grab.  

“It’s like we’re on a monopoly board, and it’s our intention to get the Park Place real estate,” said John Chiplin, CEO of GeneFormatics, a company that is aiming to capture thousands of protein structures using bioinformatics in combination with Nuclear Magnetic Resonance imaging.

Recently, venture capitalists have pumped nearly $200 million into GeneFormatics, Structural Genomix, and Syrrx, three San Diego-based companies that use different strategies to crack the protein structure problem. Several other companies have set their sights on determining structure as well.

While academic labs have been trying to determine protein structures for decades, the new flurry of commercial activity in this area stems from the belief that researchers will be able to harness DNA sequences from the completed genome to determine protein structure, then go from structure to candidate drug compound, facilitating faster drug discovery for pharma and biotech companies hungry for new blockbuster products.

Each company in this space, however, has its own recipe for combining computational and experimental methods to bridge the gap between sequence and structure.

“It’s quite clear that the most powerful package is where you combine both experiment and theory, both crystallography, NMR, and bioinformatics,” said Andrej Sali, an associate professor of molecular biophysics at Rockefeller University. “The question is how you combine them.”

Structural Genomix and Syrrx are both focused on streamlining the labor-intensive process of x-ray crystallography, in which proteins are isolated, then crystallized and zapped with x-ray beams.

The protein’s atomic structure causes the beam to scatter in a certain diffraction pattern, providing data on the various positions of atoms and the angles between them. Using bioinformatics, the companies can then take this structure information and model protein-compound interactions, which they can test in vitro .

“X-ray crystallography is the approach of choice,” said Tim Harris, Structural Genomix’ CEO . “But there’s a great deal needed to change it from an academic to an industrial strength exercise.”

Structural Genomix, which has raised a total of $85 million in private financing since its founding in 1999 and has 80 employees, has attacked this challenge by building a protein structure determination factory, outfitted with high intensity x-ray beamlines.

“We decided to be an industrial strength company that determines protein structure at the very large scale,” said Harris. “We believe in getting every part of the operation working, then sewing it together using appropriate automation at each step.”

Syrrx, which has banked $79 million in financing and has 55 employees, took the opposite route, building a protein structure determination robot to automate and speed up the multi-step process of preparing protein samples, crystallization, and x-ray crystallography.  

“If we just scaled the current process” of x-ray crystallography done in academic labs, “it would fail,” said Nathaniel Eames David, one of Syrrx’s founders and the company’s director of business development. “Structural Genomix‘ approach is like Celera trying to sequence the genome with slab gels.”

Syrrx has just completed its factory with three rooms of completed robots, and is now beginning the process of protein structure determination, according to David. The company also uses a bioinformatics program called Virtual Ligand Screening, which it licensed from Molsoft, a small San Diego bioinformatics company, to simulate the interactions between structurally determined proteins and small molecules. This way it can virtually determine which small molecules might be good drug candidates for targeting those particular proteins.

By spending so much time developing its technology, Syrrx has fallen behind Structural Genomix in the race to determine structures. Structural Genomix has already determined “many” structures, hopes to determine 150 protein structures in the next 12 months, and “over the next 18 to 24 months be firmly placed in the lead discovery business,” according to Harris.

Structural Genomix has also been able to make large amounts of hundreds of proteins—a major technological feat given that it is hard to take a mammalian gene sequence and splice it into bacterial cells to get proteins that fold correctly, Shapiro said. Syrrx, on the other hand, is using small amounts of protein in its crystallization trials. “Ultimately, they’re going to have to scale up,” said Shapiro.

David maintains that Syrrx will soon overtake Structural Genomix, because its robots can do “a million experiments per day.” He expects Syrrx to isolate its first drug candidates within the next 18 months, and said he is negotiating the final stages on two partnerships, one with a biotech company and one with big pharma.

Meanwhile, Structural Genomix is taking a page out of Syrrx’s book by adding robots and other technology to its protein structure assembly line. In January, the company hired Janice Culpepper, a senior technology executive at Millennium Pharmaceuticals, to oversee its efforts to purchase and implement robotics. The company is planning to build four robots at its facility, according to Shapiro.

Recently, the company also began a c ollaboration with the Department of Energy’s Argonne National Labs to build a dual-undulator beamline facility, an advanced x-ray machine. The company expects the beamline to be up and running in twelve months, and to be able to determine 2,000 protein structures a year. Even though the beamline will be located at Argonne's Advanced Photon Source , Structural Genomix will have the exclusive right to use it.

This x-ray crystallography-intensive approach counters the bioinformatics-heavy strategy taken by GeneFormatics, which is funded by a $30 million venture capital stake. While the company ultimately plans to include x-ray crystallography in the mix, GeneFormatics primarily uses bioinformatics modeling techniques, coupled with NMR, to determine protein structures.

Using x-ray crystallography on proteins “is like using a sledgehammer to crack a walnut,” said Chiplin. Instead, he said, the company uses bioinformatics to go from “sequence to function then to structure,” allowing it to develop protein models for over 10,000 sequences a week.   Then it does NMR spectroscopy to verify structure on select proteins.

NMR techniques exploit the way different atoms within a protein’s molecular structure resonate, use the measurements of these resonances to determine distances between atoms, and model protein structures based on these distances. But the models are not exact, said Shapiro.

The company has signed partnerships to determine novel targets with Bristol-Myers Squibb, Schering-Plough, and Astra Zeneca, and has just secured a Japanese distribution agreement with Takara Shuzo.

Based on this record, Chiplin calls GeneFormatics “the only structural genomics company with customers.”

But it’s not the only bioinformatics company doing structural genomics. Prospect Genomics, which Sali co-founded, is developing a database of protein structure models that use sequence homologies to known structures, and a database of ligands docked into the models’ binding sites.

Additionally, Structural Bioinformatics, which uses bioinformatics in a similar way to GeneFormatics, has just entered the experimental territory, penning a deal January 25 with the RW Johnson Pharmaceutical Research Institute at Johnson & Johnson to determine protein structures using x-ray crystallography.

Some of these companies may get gobbled up by big pharma—a possibility that Harris and David both allow. But experiment-heavy players might also be likely to partner up with bioinformatics companies in order to complement each other’s strengths.

“I’m quite sure that in the near future there will be mergers between different companies in the general area,” said Sali.

The exact structure of these consolidations, however, can only be determined once the market begins to crystallize.

This story has been altered from the original version, in which it was erroneously reported that Structural Genomix received public funding for its dual-undulator beamline facility.

The Scan

Shape of Them All

According to BBC News, researchers have developed a protein structure database that includes much of the human proteome.

For Flu and More

The Wall Street Journal reports that several vaccine developers are working on mRNA-based vaccines for influenza.

To Boost Women

China's Ministry of Science and Technology aims to boost the number of female researchers through a new policy, reports the South China Morning Post.

Science Papers Describe Approach to Predict Chemotherapeutic Response, Role of Transcriptional Noise

In Science this week: neural network to predict chemotherapeutic response in cancer patients, and more.