Who: Darryl Bornhop
Position: Professor, chemistry, Vanderbilt University, 2003 to present
Background: Professor, chemistry and biochemistry, Texas Tech, 2002 to 2003; PhD, analytical chemistry, University of Wyoming, 1987
Darryl Bornhop and his team of researchers at Vanderbilt University have developed new label-free technique for studying molecular interactions, including protein-protein interactions. The approach, which they call back-scattering interferometry, involves shining a red laser into a microscopic chamber filled with liquid with two kinds of molecules mixed in. The interferometer then measures the strength of the reaction between the molecules.
The work appears in the Sept. 21 issue of Science.
Can you describe your method and how it works?
The method [is based] on an interferometer, which depends on the interaction of light with itself, basically. I guess the simplest way to understand this is if you can imagine two sine waves and then you impose them on each other. If they’re perfectly aligned they will constructively interfere and destructively interfere in such a way so that where the energy is going positive, they’ll enhance the energy, and where they’re going negative, they will decrease the energy.
And so what you would get, if you just took two sine waves and interfered them with each other, two beams of light, you would get some bright and dark spots. What we’ve been able to realize is that when a solute reacts or interacts with another species in solution, the phase of that fringe pattern, or the position of one of those sine waves, changes and that leads to a phase change or a positional change in the bright and dark spots.
Interferometry is used in astronomy, in communications, it’s used in testing. Probably the most common example of interferometry is that it’s used to test the smoothness, or flatness, of optics.
So you take rings, Newton’s rings … if you take a set of Newton’s rings and shine them on a mirror, if the mirror is perfectly flat, the rings will not be distorted in any way, shape, or form. And if the mirror has some curvature on it, at the points where it has curvatures, those rings are going to be distorted a little bit.
How did you apply this to molecular interactions?
The interferometry that we have devised, we’ve been working on for a number of years, more than a decade now. Basically, what an interferometer measures is a change in index of refractions. When we shine the light on the channel, some of the light is reflected and refracted off the channel … and then some of the light gets bounced around inside of the channel.
As a consequence of interacting with the molecules in solution, or as a consequence that the density of the sample is different than the density of the surrounding medium, some of that light is retarded or it slows down a little bit. Or the sine wave gets shifted a little bit, and then when you multiply the two you get a set of fringes that looks different than before there was a more dense solution in the channel.
From my perspective, the interferometer is a really cool trick. It’s an extremely unique way to do interferometry because we simply shine a laser on this little microfluidic channel, and it turns out that the chip is actually the optics.
The microfluidic chip is the optics, and then all of the interferometry is done in a really small volume. Then when the fringes leave, other light is scattered off that the chip, [and] we just display that onto a [charge-coupled device].
Basically, we realized years ago that we could measure the change in refractive index of solutions in very small volumes with this form of interferometry. And it turns out that you can do capillary electrophoresis, or on-chip electrophoresis, or a number of other things with it. A handful of years ago, we started to look at molecular interactions at the surface, or surface immobilized interactions, much in the same way that one would use a [surface plasmon resonance] instrument, for example … or some sort of wave guiding technique where you immobilize one of the antibodies and then you capture the antigen out of solution, or vice versa.
And then you look at the change, the optical property change, at the surface of the channel.
That’s how we got started. But it dawned on me one day that if we are able to measure the bulk refractive index change in solution, maybe we can measure the bulk change in index that occurs when a molecular interaction happens.
In the end, this is really the most important part of the observation. It’s not the widget. The technology is pretty cool, it’s infinitely commercializable, it’s miniaturizable, it’s compatible with high-throughput. It’s all of these things [and] can be made inexpensively.
But what we’re most excited about, at least in my research groups, is that we can now begin to study molecular interactions over a very broad range in terms of avidity or affinity, without any kind of label or any kind of surface immobilization — free solution molecular interactions.
And this is actually what we’ve been struggling with for a number of years. And the biggest question … in the reviewers’ minds when this paper went through the [peer-review] process [was]: ‘What is the mechanism? Can the folding of a protein and consequently injection of water molecules, waters of hydration, or change in the waters of hydration, change in the ion composition around the molecule, and/or the intrinsic molecular dipole, molecular quadrupole of that molecule — does that change the index of refraction? Or does it react with the light enough to change its property to produce a signal that’s large enough so that you can actually measure it?’
And that’s what’s really at the heart of what we propose the mechanism to be because there is another free solution interaction methodology, and that’s called isothermal titration calorimetry.
In that case, you measure the heat taken in as a consequence of two species reacting. There, you do that in basically a calorimeter, a miniaturized bomb calorimeter.
How does your technique compare to other types of approaches for measuring protein-protein interactions like SPR or yeast two-hybrid?
I think the thing that’s most important from our perspective is that this is the only label-free technique that has been found that is truly compatible with high-throughput analysis, can give end-point and kinetic data, so we can actually mix the two solutes offline and then measure the amount of change in signal and do that in a stepwise fashion and generate a saturation-binding isotherm that will tell us the binding affinity.
Is the data that you can get with your method different, you can get more data? Or is it the process that’s different, or both?
Both. I would say that we can either do end-point or kinetic assays. And we can do it in a much simpler format. The other thing is that we’re intrinsically more sensitive, so as compared to SPR, we’re comparable in terms of sensitivity, but SPR requires immobilization of at least one of the two species.
That means you need to do some chemistry, you need to know what you’re going to screen against.
In terms of ITC, which is the technique that doesn’t require immobilization, we’re between three and four decades more sensitive and require — to do an entire analysis — anywhere from 10 to 5,000 [fewer] grams, milligrams, or moles of the species to do the assay.
Have you done any research with this method looking specifically at protein-protein interactions?
Oh yeah. We’re working on a paper right now that we hope to send to Nature where we’re comparing function versus structure for a series of proteins that are called crystalline proteins or heat shock proteins.
[But] this is a tool to measure interactions between molecules. And in fact that we were able to show in the paper that you can measure or watch calmodulin being folded up by calcium ions. If you look at that calmodulin figure [in the Science article] that shows that we can measure protein-ion interaction, protein-protein interaction, protein-small molecule interaction, and protein-peptide interaction.
Besides the protein A and IgG system [that] we showed in that paper, we’ve looked at a number of protein systems.
What were the other methods that you used to validate your results?
We’ve compared and/or validated, or are in the process of validating, our technique versus fluorescence, versus ITC, and versus SPR. I should tell you in some cases SPR just doesn’t work for some of the systems that we’ve started to study. And in some cases ITC doesn’t work for some of the systems that we’ve been trying to study.
In the case of SPR, on occasion, when you tether the molecule to the surface, it extinguishes one mode of binding that is interesting in terms of studying the biophysics of the system.
And in terms of ITC, there are systems out there where the reaction doesn’t produce very much heat. Consequently, there isn’t very much energy evolved and you have to use very, very large quantities of the solute, and in fact, we’re working on a paper right now where we have found for at least one particular protein-protein interaction, at the concentrations you need to get some sort of an answer from ITC, the protein is in some form of aggregate and that just makes the measurement useless.
Does sample size affect the results?
As it turns out, no. Actually, we used a couple of hundred nanoliters in sample total to do all of the analyses. Normally, we just mix a microliter with a microliter, and that’s easy to handle. One could easily reduce the total consumption amount in volume by building a more well-designed microfluidic network and handling the fluids on the fluidic network.
Do you have any idea why this method works so well for molecular interaction measurements?
Because it’s one of the most sensitive measurement techniques known to humankind. As it turns out, there are some defraction techniques that are based on putting the grading on the surface or trying to do interference measurements on the surface as well.
Those techniques are out, they’re kind of analogous to SPR, only they’re not surface plasmon techniques, they’re defraction or interference techniques where you look at the change of interference or defraction at the surface.
What’s the next step?
The next step is to further validate this with a broad array of systems that have even lower and even higher binding affinities. We’re in the process of transferring the technology to a commercial entity.
They plan on developing a bench-top version of the device for basic research. Then going forward, we hope to begin to investigate the potential for this technique in the arena of diagnostics.
[Editor’s note: A company, Molecular Sensing, has been created to commercialize the technology. Bornhop is one of the founders of the company and serves as its chief scientific officer. The company plans on completing a prototype system by the end of the year.]