Surface functionalization for fluorescence immunoassays and microsystems

Technical abstract

Mass parallel testing

Whether it is to find a vaccine, discover new drugs, diagnose patients, or study DNA, the motto is often the same: conduct more tests, more quickly. Mass testing enables biologists to try more molecules and vary experimental conditions, which helps them find the proverbial needle in the haystack of combinations.

The principle of biochips is similar to how progress in microelectronics has led to an increase in computing power, by decreasing the size of transistors and integrating more of them into a single chip. In biochips, laboratory functions and samples are miniaturized and thousands of experiments happen in parallel on a single chip. This kind of high-throughput testing enables scientists to conduct experiments and test many molecules at the same time, thus increasing efficiency.

In particular, microarrays involve printing a large number of biological samples using a technology similar to inkjet printing. Small droplets are deposited on a surface in an regular pattern, leading to a high density of biological probes, and small amounts of biological materials used in each.

A variety of biological molecules can be printed into microarrays, like DNA, proteins, and antibodies. To ensure regularity and consistency, the samples are printed using a machine called a spotter, or micro-arrayer. The biological probes are printed as micro-droplets onto a chemical layer on top of substrates like glass slides.

Biological molecules, and antibodies in particular, are fragile. Droplets are printed very gently in order to maintain their cohesion on the surface. Glycerol is often added to the sample solutions to increase their viscosity, help the droplets keep their shape, and limit evaporation. Once printed, the samples must be cleaned softly and kept in an environment that preserves their integrity until the experiment is complete.

Fluorescence in biochips

Microarrays often rely on fluorescence as a detection mechanism. Fluorophores are molecules that absorb light at a certain color and re-emit it at another color. They are combined with biological molecules of interest to detect their presence or their interaction. Because fluorophores emit light at a different color (wavelength) than what they receive, scientists can separate the two colors, and measure only the light emitted in return: it indicates how many biological molecules (DNA strands, antibodies, etc.) are in a specific spot.

The intensity of fluorescence emitted by the molecules is measured by a scanner or a microscope, leading to grayscale images where brighter zones indicate more fluorescence emitted than darker ones. A palette of false colors is then used to interpret the results.

Surface functionalization

Scientists from the LFCM laboratory have developed a chemical protocol to attach biological molecules to an inorganic surface like glass or metallic oxides.[1] Known as “CEA-2,” the protocol is based on a silane (a molecule that contains an atom of silicon) and therefore called silanization. The process involves successive steps that progressively modify the molecules on the surface using chemical treatments. In the final step, biological molecules of interest are spotted on the modified surface and bind to the chemical layer.

The CEA-2 protocol is an established way to attach biological molecules to surfaces, and is routinely used in the lab to print DNA microarrays on glass slides (as oligonucleotides). The slides are usually silanized in bulk to increase consistency and reproducibility of results.

Substance P

Substance P (SP) is a neurotransmitter from the neurokinin family, synthesized by neurons and able to excite nearby neurons. SP is involved in many physiological systems, including the transmission of pain information to the central nervous system.

Substance P was used as a model molecule in the development of a novel approach to detect biological warfare agents, led by Laure-Marie Neuburger of the Laboratoire d’Études et de Recherches en Immunoanalyse (LERI).[2] Laure-Marie had been developing the immunoassay in capillaries, and produced the antibodies and antigens, conjugated with fluorophores or other molecules. I adapted Laure-Marie’s immunoassay to planar microarrays using the CEA-2 protocol, traditionally used for DNA biochips.

Immunoassay protocols

I conducted several protocols all involving antibody microarrays, or immunoassays. They all involved a preliminary silanization to prepare the glass surface and coat it with a chemical layer. Biological probes (antibodies) are then spotted in droplets on that layer, and left to immobilize overnight in a high-humidity environment to prevent evaporation.

Once the probes are bound to the chemical layer, the surface is rinsed to remove excess molecules, and blocking proteins are attached to saturate free active sites on the surface. Blocking proteins make sure that nothing else can attach in areas that aren’t covered with antibodies, including fluorescent markers. Preventing this non-specific adsorption limits background noise during detection.

A solution containing the target molecules (antigens/peptides) is then deposited on the spotted surface, and left to incubate under a plastic cover slip. If the peptides are marked with a fluorophore, then detection is direct: after rinsing and drying the surface, the results are obtained directly from the fluorescence scanner.

Parameter study & protocol optimization

Biological tests, and immunoassays in particular, can be difficult to control because they depend on so many different parameters: duration and temperature of the successive steps, humidity, blocking proteins, buffers, etc. In order to increase the reproducibility of our tests, I worked with fellow engineer Isabelle Mingam to study those parameters and optimize them for the most consistent results.

The tests confirmed the need for blocking proteins to limit background noise, and a small amount of glycerol to limit the evaporation of droplets. We also found out that the drying step, done by centrifuge for DNA microarrays in the lab, might be too strong for antibodies: a softer drying method better preserved their integrity, in particular their antigen-binding site (paratope) needed to recognize and attach molecules of interest.

Reaction kinetics

All the fluorescence-based experiments conducted so far were done after a period of incubation between antibodies and peptides. I worked with Rémi Galland, from the CEA’s Laboratoire d’Imagerie des Systèmes d’Acquisition (LISA), to study the kinetics of that interaction in real time.[3] The principle of the experiment was similar to previous immunoassays, except that fluorescence wass measured continuously as the target peptides, marked with fluorophores, were introduced into the system.

Our exploratory work showed promising results: we were able to observe a rapid increase in signal during the first few minutes of the experiment, showing a plateau (indicating saturation) after about 30 minutes. The signal then decreased over time due to photobleaching (the gradual fading of fluorophores under the exciting light). These results prompted us to experiment with shorter incubation periods (described above).

Photothermal deflection spectroscopy

Photothermal deflection spectroscopy (PDS) is a technique used to characterize thin layers by measuring the change in refractive index of a sample due to heating by light. In other words, one laser heats a surface to different degrees depending on what’s on it; another laser is shone through the same surface, and the way it’s deflected by heat provides information on what’s there.

The principle of the immunoassay is the same as in fluorescence experiments, except the final detection step to visualize antibodies and antigens is done indirectly using gold nanoparticles rather than a fluorophore. For this experiment, I partnered with Violaine Vizcaino, from the CEA’s Laboratoire d’Ingénierie des Composants Photoniques (LICP).[4]

Although we admittedly used a highly concentrated antigen solution for this exploratory experiment, we were able to detect antigens on their specific antibodies, indicating that the interaction had taken place. No signal was detected on the control antibodies, indicating that the interaction was specific to our probes.

Neutron reflectometry

Towards the end of my time at CEA-Léti, I was offered the opportunity to visit the neighboring Institut Laue-Langevin (ILL), and to study my immunoassay layers using neutron reflectometry. I worked in collaboration with Giovanna Fragneto to prepare the samples, and subject them to the ILL’s intense neutron source inside its D17 reflectometer.[5]

Neutron reflectometry is a technique used to study thin films by shining a tight neutron beam from a high flux nuclear reactor onto a very flat surface, and measuring the intensity of the reflected radiation. It is particularly adapted to the study of stratified biological layers, because neutrons are highly penetrating and not as damaging as X-rays to delicate samples like ours.

Our results (see table below) were consistent with layers of native silicon oxide, silane, and antigens. The blocking proteins increased the density of the antigen layer, which was consistent with the hypothesis that they saturated active free sites. However, the results for antibody layers were unexpected, showing thinner layers than with just the antigens. One explanation might be that our sensitive biological molecules, usually preserved in chemical buffers, were denatured during the experiment, and couldn’t attach specifically.

Confirming the presence of the mixed layer of antigens and blocking protein would require deuterating one of those two substances, meaning replacing hydrogen by its heavier isotope, deuterium, to vary their contrast. To avoid the possible denaturation of antibodies, preparing buffers using D₂O and silicon-matched water would provide contrast while preserving a physiological environment adapted to biological molecules. Although I wasn’t able to conduct these follow-up experiments before the end of my contract, I still felt privileged to have been able to glimpse into this entirely different field of physics.

Adapting the protocol

All our experiments so far relied on attaching biological molecules to a chemical layer of silane, prepared on a flat surface like a glass slide, using the CEA-2 protocol. This liquid-phase silanization was done in organic solvents like toluene, which work well for glass and silicon substrates. However, they damage many other materials like polydimethylsiloxane (PDMS), a transparent and biocompatible polymer widely used in biological microsystems.

When such polymers are involved, another solution is to conduct the silanization in vapor phase: instead of diluting the silane in a solvent, the liquid silane is turned into a gas that attaches to the surface. My goal was therefore to adapt the regular, liquid-phase CEA-2 protocol to a vapor-phase method, by heating the silane in a closed container and depositing it on our surfaces.

Contact angle measurement

Measuring the contact angle of a droplet of water is a fast and easy way to characterize a surface. On a hydrophilic surface, which attracts water, the droplet spreads out and yields a low contact angle. On a hydrophobic surface, which repels water, the droplet bulges out and gives a higher angle.

This method only provides limited information, because different materials and layers can lead to the same angle. However, standardized chemical protocol like CEA-2 have well-known contact angles that correspond to the chemical functions present on the surface at each stage.

Therefore, I compared the contact angles of surfaces prepared with the CEA-2 protocol in vapor phase and in liquid phase, at two different stages of the process. The results indicated similar angles between vapor and liquid phase for both stages, which was encouraging, although not definitive.

Atomic force microscopy

Another method in the toolbox of the surface chemist is atomic force microscopy (AFM). By sweeping a microscopic tip very close to the surface, and measuring the interaction between the two of them, scientists can reconstruct an image of the surface at the nanometer scale.

I prepared two substrates with the vapor-phase and liquid-phase silanization protocols and observed them by AFM. The vapor-phase sample showed a smoother surface, with a rugosity (a measure of the small peaks and valley) of about 2 Å, close to that of a naked surface. By contrast, the rugosity of the surface prepared with the liquid-phase protocol was over 32 Å.

Analysis of surfaces functionalized with CEA-2 chemistry in vapor phase (top) and liquid phase (bottom), using atomic force microscopy.

On its own, this result might indicate that the vapor-phase silanization had failed. Taken individually, contact angle and atomic force microscopy don’t provide definitive proof of the success of the vapor-phase protocol. But since contact angles between protocols were consistent, the difference of rugosity might have been due to a more disorganized layer of silane deposited in liquid phase.

Antibody microarray on vapor-phase silane

I developed the vapor-phase protocol primarily for use in microsystems made with polymers that don’t hold up well with solvents, but it can also be used for immunoassays made on regular glass slides. Therefore, I printed antibody spots on surfaces coated with vapor-phase silanization, incubated them with fluorescent antigens, and compared the results to the same test prepared with the liquid-phase protocol.

The experiment was a success, with the antigens attaching to their specific antibodies and showing good fluorescence signal, similar to that on liquid-phase silanization. As expected, the antigens didn’t attach to the control antibodies, whose spots were barely distinguishable from background signal.

BioChipLab

The field of proteomics is dedicated to the study of proteins in the same way that genomics is the study of an organism’s DNA. But whereas DNA remains more or less the same, proteins vary widely between cells and change over time. Proteins are also much larger molecules than DNA strands.

One technique used by scientists to study proteins consists in cutting them into smaller fragments (peptides), and using mass spectrometry to identify those fragments by their charge and mass. The final piece of my work at CEA was to use vapor-phase silanization on a closed miniaturized device, in order to attach an enzyme that would cut break down proteins for analysis. For this work, I partnered with Frédérique Mittler, from the LFCM lab.

In technical terms, the goal of the BioChipLab project was to develop a microsystem coupled to a mass spectrometer for proteomics and pharmacology. It included microreactors, a digestion module, and an electrospray nozzle. My work focused on functionalizing the peptidic digestion module with vapor-phase CEA-2 chemistry in order to graft trypsin, an enzyme that catalyzes the breakdown of proteins into smaller peptides for analysis.

Fluorescence microscopy

The full analysis involved many steps: silanization, binding of the enzyme, digestion of the proteins, and analysis of the peptides by mass spectrometry. If the final spectrum results showed the peaks characteristic of the expected peptides, then we would have confirmation that the process was a success.

In order to experiment with the vapor-phase silanization and iterate more quickly, we first used devices with a transparent cover, and attempted to attach a fluorescent molecule on the silane. We were thus able to observe fluorescence in our device’s microchannel, confirming that the molecules had attached to a layer of silane.

Mass spectrometry

Finally, we attached trypsin, the enzyme, to the layer of silane in the microchannel of our module. We attempted the digestion of Cytochrome C, a small protein, and analyzed the results with mass spectrometry.

Initial results were promising, with a mass spectrum showing many of the expected peaks. The digestion might not have been complete, but this first result was an encouraging step towards further research.