September 15, 2014 (Vol. 34, No. 16)

Demands for protein at early stages of drug discovery and development have upped the ante for transient expression systems, the method of choice for producing protein for characterization and other studies.

Vendors have responded with systems that generate ever-higher yields without sacrificing speed, simplicity, and ease of use.

Recognizing this, a group at Thermo Fisher Scientific led by Henry Chiou, Ph.D., associate director for cell biology, has redesigned the company’s transient expression system from the ground up by “synergizing” the performance of the culture medium, cells, and transfection agents that constitute the company’s Expi293™ transient expression system.

Expi293 achieves yields of up to one gram per liter in about seven days, an increase of up to 10-fold over typical transient expression systems, he points out.

“Companies are trying to complete more upfront screening work to identify molecules with the highest likelihood of success in clinical studies,” Dr. Chiou observes. That means more target protein to complete screens and in vivo assays, and more reagent proteins such as standards, detection antibodies, and proteins involved in cell-based assays.

A great deal of transient transfection work occurs in HEK293 cells, the expression system in Expi293. HEK293 cells transfect easily and at high efficiency. They are robust protein producers.

Yet many therapeutic proteins originally expressed in HEK will eventually be manufactured in CHO. “Since HEK and CHO are different species, there can be some differences in glycosylation, as well as structural and conformational properties,” Dr. Chiou notes. “Some groups initiate more difficult transient expression in CHO once they have whittled down their candidate list, right before they turn to stable transfection.”

According to Dr. Chiou, demand for more protein at early stages relates to another aspect of scale: “Companies prefer to test 200 or 300 therapeutic candidates rather than 100. The efficiency of Expi293 means that instead of culturing in 1L vessels, the  companies can go with one-fourth that volume. They have more room in their cell culture facility for generating more and different proteins.”

In reality, companies can take advantage of Expi293’s volumetric productivity by producing more of a few different proteins, or many more different proteins, or anything in between. “They have that flexibility,” Dr. Chiou says.

Due to their high growth rates, CHO cells present difficulties for transient expression. Because cells divert more resources to expansion from protein expression, productivity is poor. Yet stably transfected CHO is highly desirable. Dr. Chiou believes that a similar redesign of expression technology exclusively for CHO could improve transient expression in those cells as well.

Cell-Free Transfection

Sutro Biopharma’s scalable cell-free protein synthesis platform produces homogeneous therapeutic proteins including full-length IgGs and antibody-drug conjugates, bispecific antibodies, cysteine-knot peptides, and proteins. Protein variants are expressed in research quantities within hours, according to the company. Then, within just a few days, the variants are scaled for toxicology work and even early clinical studies.

“Shorter preclinical design and development timelines could result in new protein therapeutics reaching the clinic 18 months faster than with conventional cell-based expression and production methods,” says Aaron Sato, Sutro’s vp of research.

The company’s approach does allow direct insertion of nonnatural amino acids, through which drug conjugates connect to the protein through standard coupling chemistry.

In what has almost become a requirement for transient expression, Sutro’s process is scalable. “It currently works at up to 100 L, and we’re aiming for 2,000 liters,” Sato remarks. He hopes also to reach a titer of about 1 g/L, which not too long ago was a benchmark even for stably transfected CHO cells.

What’s interesting about titers, scale, and time is that Sato and colleagues reach this productivity point in 10 hours, compared with a two-week or longer cell culture. And since the fermentation is based on Escherichia coli, capital investment, process parameters, and other economy-determining factors are lower or simpler than with CHO.

The exact process, based on work at Stanford University, remains under wraps as Sutro scientists work at improving robustness. The company has a collaboration with Celgene and is pushing through its own pipeline molecules. So far they have produced materials for exploratory toxicology studies, and they anticipate quantities suitable for IND-enabling studies by the end of 2014.


Sutro Biopharma’s cell-free protein synthesis platform can allow direct insertion of non-natural amino acids (nnAAs), through which drug conjugates connect via standard coupling chemistry. The platform, which is based on Stanford professor James R. Swartz’s patented Open Cell-Free Synthesis technology, can produce homogeneous therapeutic proteins including full-length IgGs and antibody-drug conjugates, bispecific antibodies, cysteine-knot peptides, and proteins.

Eliminating Plasmid Transfection

Bacteria are the go-to protein expression systems in academia and drug discovery. Even products eventually produced in mammalian cells may be characterized as E. coli fermentation products. Drawbacks of bacteria include expression system instability, poor control over yield, plasmid loss, and the need for antibiotics.

Most E. coli expression systems employ plasmid vectors. Joseph Kittle, Ph.D., a professor at Ohio University and principal at startup Molecular Technologies Laboratories, has developed a method that inserts genes directly into the bacterial chromosome, avoiding plasmids and antibiotics. The technique works well for proteins that are either toxic to host organisms or difficult to secrete.

E. coli is a workhorse, and wonderful if you need to get your hands on a protein quickly,” Dr. Kittle says.

Instead of making plasmids, Dr. Kittle uses a proprietary process to insert a synthetic gene directly into the chromosome at desired locations. Protein yields are, in many instances, as high as with high-copy-number plasmid insertion, but the directly-transfected cells are much more stable. Plus the approach is fast. Dr. Kittle claims “a couple of weeks” elapse between DNA design and a confirmatory protein gel.

A preferred insertion point for the foreign gene is on the arabinose operand, which produces genes for catabolizing its eponymous sugar. More relevant here: This region is easily turned off and on, allowing control over protein production. Plasmids are notoriously difficult to turn off.

“We’ve broken the process down into an engineering problem,” Dr. Kittle explains. “We ean identify and construct the genes we want to insert, and we can control location and production levels. We’ve taken things people are comfortable with and put them together in new way, so it’s not hard to understand that this approach works.”


In a process developed by Molecular Technologies Laboratories, sequence-verified synthetic DNA fragments are assembled into a linear fragment. The assembly of the fragment brings together the regulatory elements and gene open reading frame required for high-level protein expression at single or low copy number. After the fragment is site-specifically inserted into the chromosome, the testing of bacteria and the scaling of production can begin.

Induction from the chromosomes is much cleaner than through plasmids, and it eliminates concerns overexpressing too much of a protein toxic to the expression system.

Dr. Kittle originally cloned genes into a plasmid, and transferred that to the chromosome, but found that process was slow. Now he simply takes the synthetic gene fragments, assembles them, and inserts them into the chromosome in one shot. In addition to being faster, the technique is also more versatile with respect to types of expressible protein.

Another benefit is the ability to use strong promoters at low copy number where these same promoters at high copy numbers overwhelm the cell. Even where such efforts to overexpress proteins do not kill the hosts outright, Dr. Kittle says, high-copy-number expression systems can be disabling. “You wind up not getting very much protein because all the cell’s resources are diverted toward saving the cell.”

The method is also linearly scaleable, from bench to fermenter, while with plasmid-based transfections, the processes or plasmids are often redesigned at each scale. Well-designed direct integration should scale easily to mega-fermentation volumes. This, Dr. Kittle believes, may ultimately be the crowning achievement of this work. He has, in fact, begun investigating the process in Gram-positive organisms better suited to very large scale production.

Producing “Undruggable” GPCRs

Techniques exist for expressing G protein-ßcoupled receptors (GPCRs), but outside their natural milieu these proteins lose stability and become unsuitable for structure-based drug design. That is why small molecule drug discovery for GPCRs has lagged behind efforts based on kinases and proteases.

A group led by James Errey, Ph.D., associate director for biomolecular structure at Heptares Therapeutics, has now discovered how to express many of the more difficult, “undruggable” GPCRs and employ them in traditional characterization and screening studies.

Heptares’ solution consists of StaR® proteins engineered for improved thermostability and a preselected conformation through the introduction of point mutations. Thermal stability allows purification through standard protocols; conformational stability provides regions where small molecule drugs may bind.

“Previously, instability made GPCRs unsuitable for any kind of biophysics characterization, whether by means of structural analysis or high-throughput screening,” Dr. Errey says.

He introduced point mutations at every amino acid position from the second residue onward. Dr. Errey’s group tested each new protein for stability through a radioligand binding assay, specifically applying heat until the complex breaks apart. Mutations that improve binding stability are further optimized.

Heptares has several in-house programs built around the StaR platform, all involving previously intractable drug targets. One lead molecule is in Phase I trials; several others are under development with collaborators.

“This has opened up multiple avenues in discovery,” Dr. Errey asserts. “We can now tackle previously inaccessible targets such as family B receptors. We also have the opportunity to rationally design drugs based on structure, as demonstrated, for example, by our M1 program, which has led to the first truly selective muscarinic receptor agonist. StaRs are also amenable to antibody generation—we can purify highly stable, conformationally locked receptors that can be used to generate antibodies.”


To target challenging G protein-coupled receptors and thereby advance drug discovery and devel-opment efforts, Heptares Therapeutics uses a structure-based approach. Here, mGlu5, a family C receptor, is bound to mavoglurant, a developmental small molecule drug for treating fragile X syndrome.

Light-Scattering Toolbox

In discussions about rising titers, the significance of quality over quantity is sometimes lost. The light-scattering “toolbox” is therefore preeminent among the tools of biophysical characterization of expressed proteins. For example, in quality assessments of purified proteins, investigators rely on two complementary forms of light scattering: multiangle static light scattering (MALS) and dynamic light scattering (DLS).

MALS is used online, downstream of a sized-base separation method such as size-exclusion chromatography (SEC). The hyphenated technique, known as SEC-MALS, measures the molar masses of the eluting molecules, absolutely and in solution.

Because it is completely independent of retention time, SEC-MALS overcomes errors associated with column calibration due to proteins with conformations that differ from those of reference molecules, or because of nonideal (usually hydrophobic) column interactions. Therefore it is ideal for determining the oligomeric state of a purified protein and its aggregates. When used with a differential refractive index (dRI) concentration detector, SEC-MALS also determines the molar masses of host-cell proteins, unidentified fragments, and other degradants.

Thanks to upstream separation by SEC, SEC-MALS provides a complete and accurate molar mass distribution of the purified protein solution, according to Daniel Some, principal scientist at Wyatt Technology.

“As an added benefit, a triple-detection system consisting of UV, MALS, and dRI detectors can analyze conjugated protein and determine the degree of glycosylation or the degree of PEGylation,” he says.

Contrary to popular misconception, DLS does not measure molar mass but rather a size-related metric known as hydrodynamic radius. Once the size is known, assumptions may be made regarding conformation to convert hydrodynamic radius to molar mass. However, unless the conformation and requisite model are known with good reliability, this conversion will probably be inaccurate.

“A very nice feature of DLS is the ability to determine size distributions without fractionation,” Some adds. These distributions do not approach the resolution and accuracy of SEC-MALS, as DLS cannot resolve monomer, dimer, trimer, tetramer, etc. It can resolve only populations that differ in radius by a factor of three to five, which translates to molar mass differences of up to 100.

DLS’s advantage is speed: 30 seconds vs. 30 minutes over SEC-MALS, which is why Some calls it a “quick-and-dirty prescreen” suitable for quality assessment.

Both MALS and DLS are highly sensitive to large aggregates, which means that they detect a 1,000-mer at 1/1,000th the concentration of a monomer. “For typical proteins, the sensitivity of MALS and DLS to large aggregates is much better than that of concentration detectors,” Some says.

Previous articleHuman Stem Cell Clock Reset to Zero
Next articleSix Websites You Need to Bookmark: September Picks