November 1, 2012 (Vol. 32, No. 19)
Greg Crowther, Ph.D.
Producing high yields of pure, high-quality proteins remains a goal of many academic and industrial labs. Their efforts can be thwarted by problems ranging from incomplete control over cells’ growth conditions to inadequate post-translational modifications. However, progress is being made on many fronts.
Bioreactors with rocker platforms, which literally rock the cell cultures back and forth, were introduced about a decade ago and have since become widely accepted for the development of seed cultures, says Loe Hubbard, applications manager of Pall’s cell culture systems. However, “because you are working with rocker shaking, the oxygen transfer that you’re able to achieve isn’t really high enough to do actual production for most processes.”
To address this limitation, “we’ve added an extra dimension to the shaking,” Hubbard notes. “Instead of just rocking back and forth, the XRS 20 Bioreactor also rocks side to side, so you almost have a hula motion to the platform.”
Two-dimensional rocking appears superior to one-dimensional rocking in several respects, Hubbard said. Thanks to the augmented oxygen transfer, higher cell densities can be supported. In addition, “We see less cell debris at the end of the run,” she says. “We think that the additional dimension gives a tumbling characteristic to the flow, so that the cells stay nice and round and are not getting pulled apart by extensional shear.”
Another innovation in bioreactor design—also exemplified by the XRS 20, made by Pall in conjunction with New Brunswick Scientific—is the transition to disposable plastic bioreactor vessels.
Though this trend might not sound environmentally friendly, “the people who are doing the calculations are showing that it’s more green overall,” says Hubbard. “The amount of energy that’s being used and the kinds of chemicals that are being used to sterilize and clean bioreactors after each use add up to much larger waste and energy usage than incinerating the bioreactors at the end.”
“It’s really counterintuitive, but it’s a really exciting thing,” Hubbard concludes. “You could get to where you could run a process in a conference room. You don’t need the infrastructure; you just plug it in and go.”
ArrayXpress’ general strategy is to study cell cultures by collecting samples at multiple time points, performing RNA-Seq to quantify gene expression, mapping the expression data onto biochemical pathways to see how the cells are responding to the task of making a recombinant protein, and analyzing this data at the pathway level to determine what is limiting the production of the protein.
As an illustration of this strategy, Len van Zyl, Ph.D., CEO, gives the example of a pharmaceutical company that was manufacturing a commercial therapeutic protein in E. coli, but was only achieving about half of the expected yield. To diagnose the problem, ArrayXpress grew the production strain alongside a control strain (which contained an expression vector lacking the open reading frame for the target protein) and compared their expression patterns after 11, 18, 24, and 28 hours of growth.
“Based on this, we could very quickly pinpoint at which hour there was a real issue, and what the issue was,” says Dr. van Zyl. “What we could see from this experiment was that there was a very significant shutdown of purine metabolism in the production strain.”
Mining of the peer-reviewed literature then led to the hypothesis that excess adenine in the medium might be causing the formation of PurR-hypoxanthine complexes, which sensitize the cells to growth arrest by adenosine and inosine, thus reducing protein yields.
This hypothesis, in turn, led ArrayXpress to propose several chemical and genetic means of correcting the imbalance in nucleotide metabolism. One possible solution, adding guanosine to the media, improved the yield by over 150%.
Cevec Pharmaceuticals’ protein and vaccine production work is based on its CAP cell line of immortalized human amniocytes, which are especially useful for ensuring that human proteins receive post-translational modifications correctly and efficiently—via glycosylation, for example. “Most biopharmaceuticals on the market or in development are glycoproteins,” notes Hartmut Tintrup, Ph.D., director of marketing and business development at Cevec.
According to Dr. Tintrup, a couple of breakthroughs have allowed Cevec to make glycoproteins very effectively. “A significant step forward for us was the recent introduction of an optimized CAP growth media that we developed together with a German company in order to meet a demand of our industrial customers,” he says. This has increased cell densities to twice their previous levels, leading to higher protein yields.
A second breakthrough, Dr. Tintrup says, was the creation of the CAP-T cell line for transient protein production.
“CAP cells are designed for stable protein production, so you have to put the cells under selection pressure and go through a cloning procedure to end up with a final producer cell line, a procedure that still represents a considerable investment of time and effort,” he explains.
“By introducing the large T-antigen of the SV40 virus into our CAP cells, we generated our CAP-T cells, a cell line optimized for transient protein production. Within less than two weeks you can now generate significant protein amounts for early characterization—for example, animal studies or crystallization.”
The isogenic background of CAP cells and CAP-T cells ensures a smooth transition from short-term production of a protein to long-term production, Dr. Tintrup says. This is in contrast to cases where, for instance, HEK 293 cells are used for transient production and CHO cells are used as the stable cell line. When you switch expression hosts, “you’re at high risk to recharacterize your CHO-derived material versus the 293 material,” he says. “This can be avoided, since CAP and CAP-T reflect one integrated protein expression platform.”
Improving Yields in Yeast
VTU Technology specializes in protein expression in Pichia pastoris. This methylotrophic yeast is a good expression host in several respects: it grows on simple defined media, with methanol or glycerol as its sole source of carbon and energy; it stably integrates cloned genes into its genome, rendering antibiotics unnecessary after initial transformation and selection; and it possesses a strong, inducible promoter for alcohol oxidase (AOX1) that can be used to drive expression of other genes.
VTU has built upon these advantages in creating a library of AOX1 promoter variants.
These variants—generated via rational and random genetic insertions, deletions, and mutations—display different patterns of derepression and induction. The basic VTU strategy is to clone the gene of interest into numerous expression constructs, each with a different promoter, followed by simultaneous transformation. A high-throughput screening and cultivation regimen of up to 25,000 clones per week allows rapid determination of which construct or combination of constructs generates the best yield for that particular gene.
Though growing Pichia on methanol is convenient in many cases, some labs wish to avoid this potentially explosive substrate. In these situations, growth on glycerol with AOX1 promoters optimized for methanol-free expression enables protein yields superior to those obtained by conventional alternatives in yeast, such as GAP promoter-driven expression, says Thomas Purkarthofer, Ph.D., head of business development at VTU.
A final feature of VTU’s expression platform is that all proteins are tagged with a cleavable signal peptide that causes secretion of the proteins into the media. This allows for convenient harvesting of the protein.
“When you look at the recombinant protein in the culture supernatant, purity is often around 70 to 80 percent,” says Dr. Purkarthofer. “If you compare that, for example, to intracellular production with E. coli, where you have a completely crude lysate, it’s a very attractive starting point for downstream processing. It’s really relatively pure before you start the first column for further purification.”
Easing the Transition into Bioreactors
BioSilta’s flagship technology was inspired by problems in translating results from unregulated small-scale cultures (such as those found in shake flasks and 96-well plates) into bioreactors, where oxygenation, nutrient availability, and pH are generally regulated carefully, says Craig Fuller, Ph.D., business development and product manager.
To make small cultures more bioreactor-like, a proprietary polysaccharide along with a glucose-releasing agent promote steady, controlled growth rates that will not outpace oxygen delivery or cause severe pH changes. As a result, cultures can be grown for longer, with up to ten-fold gains in cell density, according to Dr. Fuller.
“We have optimized conditions to ensure that glucose is being provided on a continual basis from the start,” he explains. “The next day, because you have been able to control the growth of the microbes, you have viable cells at a higher density.” The bottom line for customers, he says, is “significantly higher yields of soluble, active recombinant protein.”
After originally distributing its technology in gelatin or liquid formats, BioSilta now offers EnPresso growth systems for E. coli and yeast as presterilized tablets, which Dr. Fuller considers even more convenient to use.
“You get a package that is sterilized, you open the package and drop the tablet into sterile water, and that is your growth system, right then and there,” he notes. “You can go ahead and start your incubation; you don’t even have to wait for it to completely dissolve.”