by Jeffrey M. Perkel
When John Kulman, a principal investigator at the Puget Sound Blood Center in Seattle, needed to put some antibodies to the test, he didn't reach for the fluorescence labeling kit. He didn't grab his chemiluminescence ELISA reagents, either. In fact, Kulman didn't use labels at all. Biochemically speaking, he went au naturel.
Labels, whether radioactive, colorimetric, fluorescent or other, have been the workhorses of biological research and especially of drug discovery. But labels need to be coupled to molecules to be useful. Therein, as they say, lies the rub.
"This labeling step," wrote Matthew Cooper of the University of Queensland, Australia, in a 2002 review, "imposes extra time and cost demands, and can in some cases interfere with the molecular interaction by occluding a binding site, which leads to false negatives. Fluorescent compounds are invariably hydrophobic, and in many screens, background binding is a significant problem, leading to false positives. Ideally, a biosensor-based screening platform should be label-free, sensitive and have sufficient throughput to be widely applicable in drug discovery." [1]
Label-free assay technologies provide a rapid, often real-time approach to both biochemical and cell-based assays. Most often, they enable direct measurement of the affinity, kinetics and thermodynamics of protein-protein and protein-small molecule interactions in vitro—information that is difficult if not impossible to collect by other methods. Exploiting any number of physical properties, from refractive index (surface plasmon resonance, also known as SPR) to resonant frequency (quartz crystal microbalances), these methods require a capital investment. Yet they provide perhaps the clearest—and most quantitative—biochemical characterization of molecular interactions available today.
Surface plasmon resonance
For Kulman, who next month will start a new job at a Boston-area biotech organization, the goal was "to evaluate the thermodynamics of interaction between antibodies and clotting factors" to identify potential immunotherapeutics for blood-clotting disorders. In all, he tested some dozen antibodies against three clotting targets. His tool of choice: surface plasmon resonance.
In SPR, a functionalized gold biosensor—the surface upon which the molecular interactions will be tested—sits atop a prism. When light is directed through the prism at the bottom of the biosensor, it mostly bounces off the gold sensor interface and toward a detector. But at certain angles, the reflected light energy can be absorbed by the gold—that is, the light is in resonance with the metal's electrons—causing a kind of shadow or eclipse in the reflected light illuminating the detector. The angle at which that eclipse occurs varies with the refractive index of the gold/sample interface above the sensor surface, which in turn varies with the amount of material bound to that surface.
As Phil Page of SPR vendor Reichert Technologies explains, "In SPR, you get an energy minimum within the reflected light, and this energy minimum shifts angle as mass binds and dissociates from the surface. And we monitor this angle shift as a function of time."
In a typical SPR experiment, one molecule, the ligand, is coupled to the sensor surface. A solution containing the second molecule (the analyte) then is flowed over the surface. This creates a mass change on the sensor surface as the two molecules interact, which is detected in real time. Researchers can use these data to extract a detailed picture of the intermolecular interaction, including its association rate, dissociation rate and affinity.
Kulman used his SPR instruments to probe the thermodynamics of antibody-antigen interactions, assessing binding under a range of temperatures. The instruments he used included a Biacore T100 from industry leader GE Healthcare (which has been supplanted by the newer, more sensitive Biacore T200), and a ProteOn XPR-36 from Bio-Rad Laboratories. The former has four flow cells (one of which is used as a reference); the latter can probe 36 reactions at once. "It's like an SPR array," says Laura Moriarty, ProteOn product manager at Bio-Rad.
For Kulman, that was the key differentiator. Though both systems performed more or less equivalently, throughput varied significantly. "Eventually, we got to where a three-week experiment on the Biacore T100 could be done overnight on the ProteOn," he says. “The Biacore range also offers a high throughput instrument; Biacore 4000, which can probe 20 reactions simultaneously, enables up to 4800 interactions to be processed within 24 hours” added Christina Burtsoff Asp, marketing manager at GE Healthcare.
Commercial SPR vendors include GE Healthcare, Bio-Rad Laboratories, Reichert Technologies, ICx Nomadics and Biosensing Instrument. Prices range from tens of thousands of dollars (Biosensing Instrument) up into the several-hundred-thousand-dollar range. The ICx Nomadics sensíQ costs $180,000; Bio-Rad's ProteOn XPR-36 costs $285,000.
Bio-layer interferometry
ForteBio's Octet line of label-free detection systems exploits another optical principle to read bimolecular interactions: bio-layer interferometry (BLI).
In BLI, the tip of a fiber-optic probe is coated with ligand and acts as the biosensor. That tip is then immersed in a solution of analyte, in the well of either a 96- or 384-well plate. Unlike SPR, the system doesn't use fluidics; instead, the plate is shaken during reading to create "orbital flow," explains Chris Silva, ForteBio's vice president of marketing. "It's like an ELISA-on-a-stick," Silva says.
To read the assay, white light is directed down the length of the fiber. Interference between light reflecting off the reference and immobilized surfaces of the tip creates a distinctive pattern of light returning up the fiber. As molecules bind to the immobilized sensor surface, that pattern changes in proportion to the extent of binding.
According to Silva, SPR and BLI generate more or less equivalent data. But BLI has one significant advantage, he says, in that it is relatively immune to fluctuations in the refractive index of the samples being tested. By comparison, because SPR specifically tests the refractive index, it is less tolerant of samples containing high levels of glycerol and DMSO. Another advantage, he says, is that BLI is microfluidic-free, "making Octet an easy-to-use benchtop instrument in a multiuser lab."
Both 96- and 384-well Octet systems are available, with sensitivities to support both protein-protein and protein-small molecule studies (down to 150 Daltons), Silva says. Prices range from about $50,000 to $300,000, and consumable costs run about 10 to 20 cents per sample.
Epic technology
Corning's Epic technology relies on the ability of specially patterned surfaces to reflect distinct wavelengths selectively when illuminated with broadband light.
Paul Butler, market development leader of imaging and detection at PerkinElmer, whose EnSpire Multimode Plate Reader has built-in Epic technology, explains that the system supports both cell-based and biochemical assays. In the case of cell-based assays, cells are seeded onto a surface coating under which is a resonant waveguide. "When that grating is illuminated with a broadband light source, it reflects a specific wavelength, which will vary with changing mass as molecules move in and out of a measurement zone above the sensor surface. This is known as 'dynamic mass redistribution.'" In the case of a biochemical assay, the wavelength change is related to increased mass in the measurement zone due to analyte binding to protein immobilized on the surface coating.
Epic technology can measure binding strength (but not on- and off-rates), which makes the technology "complementary" to SPR, Butler says, especially as the system's plate-based format offers higher throughput than non-plate-based SPR systems. And it can measure both in the context of biochemical assays (as SPR can) and in live cells. "Epic is one of only two systems on the market that can do both cell and biochemical assays," says Butler. (The other is SRU Biosystems' BIND technology.)
In cell-based assays, rather than requiring different assays for each GPCR pathway of interest, the Epic system reveals "the global response" of cells to stimulation, says Butler. Another advantage of the EnSpire platform in particular, he adds, is that it is multimodal: In addition to using Epic technology, it also can measure fluorescence, absorbance and chemiluminescence and use AlphaScreen assays. "You can use one platform to do all the assays," Butler says. "And because we know label-free is non-invasive, you can run label-free tests on a cell population and then run a labeled assay on that same population. That's the real power of the technology."
According to Butler, the Epic-enabled EnSpire costs "just over $100,000 for a fully loaded system."
Quartz crystal microbalance
TTP LabTech's resonant acoustic profiling (RAP) technology uses a quartz crystal microbalance (QCM).
According to TTP LabTech global product manager Helge Schnerr, QCM systems have greater flexibility in the nature of the samples they can use because the systems are acoustic rather than optical. "We have shown you can go as crude as 100% serum," Schnerr says. That "makes the technology interesting in terms of giving customers more biologically relevant data, because rather than looking at interactions in purified systems, you can learn how a protein may act under real conditions."
Like SPR systems, the RAP biosensor is a two-channel flow-based system. The heart of the system, Schnerr explains, is a quartz crystal: "a resonator." Applying an electrical field causes the crystal to resonate at a characteristic frequency, about 16.5 MHz, which changes as molecules bind to the sensor surface.
Independence of optical parameters and of proximity to the immediate sensor surface with RAP enables the user to study binding interactions on large objects such as whole cells. Furthermore, QCM systems offer a price advantage vs. optical systems, Schnerr says, because they don't require expensive optics. But, he adds, "We don't have the Holy Grail" —the RAP cannot detect analytes smaller than a kiloDalton, limiting its use to protein-protein (as opposed to protein-small molecule) analysis.
Electrical impedance
Roche's xCELLigence and Molecular Devices' CellKey systems are based on yet another physical principle: electrical impedance. Both systems work by measuring the change in electrical flow from one side of an electrode to the other caused by the proliferation, movement and morphology of cells growing on that electrode in response to, for instance, drug treatment.
"We do 'label-free cell monitoring,'" says Steven Hurwitz, marketing manager for cell analysis systems at Roche Applied Science, "and we do it in real time." Whereas labeled techniques capture a snapshot of a cell at a given time point, "xCELLigence allows you to look at the cells in their native state."
Indeed, the system keeps cells in an incubator and can monitor their growth over days and even weeks. And it can do that for cells that are "as close to a natural state" as possible, Hurwitz says—primary cells, for instance, as opposed to cell lines that have been genetically modified to express, say, a fluorescent reporter gene.
According to Hurwitz, xCELLigence systems are available in a range of configurations for both academia and industry, and they cost $50,000 to $500,000 depending on options such as throughput.
Microcalorimetry
GE Healthcare's MicroCal line of microcalorimeters profiles proteins' thermodynamic properties.
Two technologies are available, according to Burtsoff Asp. Differential scanning calorimetry (DSC) systems measure protein stability and homogeneity based on the amount of heat required to unfold the protein; isothermal titration calorimetry (ITC) systems measure molecular interactions based on the heat released as they bind.
The former, says Burtsoff Asp, "is a generic protein-stability indicator," whereas the latter addresses "the driving forces" governing molecular interactions, such as hydrogen bonding, van der Waals interactions and hydrophobic interactions.
Purchasing considerations
With so many different approaches available, researchers need to do their homework before making a purchasing decision. Among the key variables to consider are:
1. Cost. And not only the cost of the instrument itself, but the cost of consumables and maintenance contrasts. Complicating the decision, sensors often can be regenerated and reused multiple times, reducing the per-data point cost dramatically. For instance, a single biosensor chip for the ProteOn costs about $250, says Moriarty, or about $7 per data point. But if the chip is reused 10 times, that price drops to 70 cents apiece.
2. Applications. Do you need to analyze both protein-protein and protein-small molecule interactions, or just one of the two? Do you need what Moriarty calls "full kinetic information"—on-rates and off-rates—or just affinity data? And what about thermodynamics?
3. Flexibility. Do you have a specific application in mind, or do you want to be able to accommodate other kinds of applications as the need arises?
4. Throughput. Some systems are plate-based, and others use flow channels. So ask yourself, how many samples do you anticipate running at once? Is speed a concern?
5. Demo. Will the system work for you? Make sure to demo the system to be certain. If you cannot arrange a demo in your own lab, try sending samples to the company. The key is to make sure you can get the data quality you expect.
Reference:
[1] M.A. Cooper, "Optical biosensors in drug discovery," Nat Reviews Drug Disc, 1:515-28, 2002.
The image at the top of this article is GE Healthcare's Biacore T200.