Team:Freiburg/Project/iRIf

""

Introduction to iRIf

Imaging Reflectometric Interference (iRIf) is a labelfree optical detection technology that can detect and visualize binding processes between biomolecules. For example, the binding of antibodies to their corresponding antigens can be analyzed. The detection method is based on the interference of light that is partially reflected at different (bio)layers. In this scenario, a biolayer would consist of antibodies that bind to antigens immobilized on the surface of a transparent material (see fig. 1). Adding a further biolayer changes the optical properties, resulting in a changed interference pattern of the reflected light. The intensity of the reflected light for each location of the observed area is detected by the CCD sensor of a camera and can be visualized in real-time.

One Method, Many Advantages

Label-Free: A huge number of analytical procedures for detection of binding between molecules depend on labeling with a fluorophore or enzyme (ELISA, Western Blot). Since iRIf is a label-free optical method, no expensive labeling reagents are required.

Fast and Simultaneous: It is possible to screen complex samples for several binding partners on a microarray within minutes. This, for example, allows us to screen a blood sample within 20 minutes for potentially hundreds of different pathogenic antigens or their corresponding antibodies.


Real-Time: With iRIf it is possible to observe binding events during a measurement in real-time!

Small Sample Amounts: The iRIf detection method is very sensitive and can detect even tiny target molecule concentrations. Therefore, usually only small sample volumes (in the µl range) are necessary to obtain reliable results. Hence, the iRIf detection is often coupled to a microfluidic device.

Applications

The number of applications for label-free reflectometric interference techniques like iRIf is as large as the number of specific binding processes. Applications include amongst others: drug discovery 1), kinetic interaction studies 2), food analysis 3), biomarker research 4) and (serological) diagnostics 5). Usually, the basis of each application is some type of biomolecule microarray (often proteins) which are arranged in distinct spots on a transparent glass side. Protein microarrays facilitate high-throughput screening with a small quantity of sample 6).

Our Focus

(Serological) diagnosis of infectious diseases
Concerning serological diagnostics, iRIf can be used to simultaneously detect potential antibodies for hundreds of diseases from one drop of blood. To this end an antigen peptide array (or antibody array) needs to be prepared and flooded with the blood of a patient (just like our DiaCHIP works). Different studies show successful detection of pathogenic antigens or the corresponding antibodies in the blood with this technique, e.g. for tuberculosis 5), Hughes syndrome 7), influenza 8) or pox 9).

Our DiaCHIP shows promising results for the detection of Salmonella and Tetanus antibodies within a complex mixture like blood.

Determining the Status of Vaccinaton
Another application is to decide whether the antibody levels of a patient still ensure sufficient immmunity or a vaccination boost is needed. Since iRIf measurements allow to quantify the amount of antibodies binding to the antigens, they provide a way of adequately quantifying the current vaccination status. Detection of none or only a small amount of antibodies would indicate that the vaccination has to be refreshed. An experiment comparing an immunized to a non-immunized blood sample can be found on our Results Page.

The iRIf Measurement

The iRIf Video

Video 1: Exemplary result of the binding of antibodies from three distinct solutions to the corresponding antigen spots. Antibody 1 and 2 bind their antigen specifically, antibody 3 shows slight cross-reactivity with antigen 2.


Figure 1: your mama

Video 1 shows an exemplary iRIf measurement with the binding of proteins from the sample to proteins fixed on the glass slide. Three different antigens are arranged in 3 distinct spots on the glass slide. One after another, three different antibodies binding to the corresponding antigen spot are flushed over the slide. As the video shows, antibody 1 binds to antigen 1, antibody 2 to antigen 2 and antibody 3 to antigen 3 (with a slight cross-reactivity to antigen 2). Upon binding, the optical thickness changes, resulting in an increased intensity of the reflected light at the spot. Using a look-up-table (LUT) the amount of bound protein is visualized by different colors.

From Reflected Light to Bright Spots

Figure 1: Light reflection and interference at iRIf. Each time the light of the LED travels from one medium to another, it is partially reflected. The reflected beams interfere and their amplitudes superimpose. The optical properties of the Ta2O5 layer are responsible for the generation of the first reflected beam R0. It interferes with the second reflected beam from the surface. In case of binding processes at the surface, the interference changes can be detected by the camera as an increase in intensity.

To understand the basic physics of iRIf it is sufficient to be familiar with only three propositions (P1-3) of ideal wave optics:

  • (1) Everytime light passes from one medium to another (non-absorbing) medium, it is partly reflected, while the rest is deflected to the next medium.
  • (2) If two waves of light superimpose they interfere with each other: They are considered as a new wave resulting from adding up the original two.
  • (3) The light intensity is proportional to the square of its amplitude. This also holds true for waves generated by interference.

During a measurement the following processes take place:
A green ray of light (wavelength: 520 nm) produced by an LED is directed to the iRIf slide.
The difference between an iRIf slide and a common microscopy glass slide is an additional layer of tantalum pentoxide (Ta2O5). This layer is located very close (nanometer scale) to the spotted slide (More details can be found in the physics section).
The ray of light will be reflected everytime the medium changes (P1): on the surface of the slide, at the Ta2O5-layer and on the spotted backside of the slide.
Finally, the remaining ray of light is mainly absorbed by the PDMS-flowcell on the opposite side of the flowchamber.
Neglecting the first reflection (see coherence length) there remain two reflected waves which are interfering (P2).

What happens if antibodies bind to a spot? Basically, the same processes take place as if the SiO2 layer (meaning the glass slide) would get thicker (see optical thickness) which results in a "delayed" reflection of the ray. This delay especially changes the resulting interfering waves' intensity.

Data Aquirement by Quotient Pictures

Data aquirement in iRIf is done by constantly taking images of the reflected light. Consequently, if this is done while the slide stays in contact with blood serum (or any solution containing antibodies) possible intensity changes (P3) due to antigen-antibody binding event at the spotted slide will be gathered, too. The change is not visible by naked eye (beyond 5% in relative intensity) but can be visualized by comparing pictures taken at a later time to initial ones using picture division. So to any time during a real time measurement the actual picture you see is a quotient picture. This quotient picture is calculated by dividing the grey values of your current picture (Pic1) by the grey values of an older reference picture (Pic2).

Quotient picture = Pic1 / Pic2

Since the grey values correspond to the amount of reflected light which is detected by the camera changes due to binding processes can be visualized. To illustrate how this works and looks like: four different quotient pictures (figure 2 to 5) are shown. All these quotient pictures were taken during a single iRIf measurement, but they differ due to different choices of either Pic1 or Pic2 that was taken to calculate the quotient picture. During the illustrated measurement two solutions flush over the chip; the first solution contains the specific binding partner for the positive control. The second solution contains antibodies against GFP. For a proper understanding let us define three points in time:

  • t = 0 ; no solution was flushed over the slide
  • t = 1 ; solution 1 flushed over the slide: this solution contains the binding partner for the positive control
  • t = 2 ; solution 2 flushed over the slide too: this solution contains GFP antibodies

Figure 2: Quotient picture = Pic1 (t = 0) / Pic2 (t = 0) No binding events are visualized. The quotient picture is calculated by dividing the picture at timepoint "0" by the picture at timepoint "0" - so literally by itself. Therefore, no differences in intensity can be seen.

Figure 3: Quotient picture = Pic1 (t = 1) / Pic2 (t = 0) The quotient picture is calculated by dividing the picture at timepoint "1" by the picture at timepoint "0" - every change in intensity that occurs during this period of time can be seen. During this period solution 1 flushed over the slide. Therefore, binding of the positive control binding partner is visualized


Figure 4: Quotient picture = Pic1 (t = 2) / Pic2 (t = 0) The quotient picture is calculated by dividing the picture at timepoint "2" by the picture at timepoint "0" - every change in intensity that occurs during this period of time can be seen. During this period solution 1 and 2 flushed over the slide. Therefore, the binding of the positive control binding partner and anti-GFP to GFP is visualized.

Figure 5: Quotient picture = Pic1 (t = 2) / Pic2 (t = 1) The quotient picture is calculated by dividing the picture at timepoint "2" by the picture at timepoint "1" - every change in intensity that occurs during this period of time can be seen. During this period solution 2 flushed over the slide. Therefore, the binding of anti-GFP to GFP is visualized.

Quantifications Visualized by Binding Curves

Moreover by observing the progression of relative intensity change binding curves can be aquirend, allowing quantification of the binding process. Figure 7 shows such a binding curve from a measurement during which two solutions were flushed over the slide successively. In the graph, the relative change of intensity (y-axis) at distincts spot (see figure 6) over time (x-axis) is shown. The lighter grey background indicates the time period in which a solution (that contains interaction partners for the immobilized proteins) is flushed over. The darker grey background represents the washing steps between each binding step. When the first solution, which contains Streptavidin - the binding partner for the positive control (biotinylated BSA), is flushed over the slide, the relative intensity at the positive control spot increases. Therefore, the red line, representing the relative intensity at the bBSA spot, rises during this step of the measurement. During washing steps the relative light intensity remains the same. When the next antibody (anti-GFP) is flushed over, the corresponding GFP spot (blue line) shows an increase in relative intensity due to binding. The binding curves are generated as follows: First you have to select regions of interests (ROIs) for which you want to plot the relative shift in intensity. Since it is a relative shift, one also needs to define reference spots (background - where no protein of interest is immobilized) to which the intensity values are normalized (see figure 6). Additionally, a reference point in time needs to be set: at this point in time the relative shift in intensity from your ROIs to the corresponding reference background spots is set to 1. Therefore, the relative light intensity at each following point in time is compared to its reference spot. If the intensity of your ROI increases more than the intensity of the reference spot the relative shift is greater than 1. If the intensity of your ROI decreases more than the intensity of the reference spot the relative shift is less than 1. According to this approach every picture of the measurement is evaluated. The resulting binding curve shows real-time binding kinetics for your marked regions (figure 7).

Figure 6: Selecting regions of interest (ROIs) and reference spots for the calculation of binding curves The shift in intensity for each ROI is normalized to its corresponding reference spot (background). Thus, ROI 1 is normalized to Ref 1, ROI 2 to Ref 2 and ROI 3 to Ref3.

Figure 7: Binding curves of the marked ROIs The relative shift in intensity (y-axis) of the selected ROIs (fig. 6) is plotted over time (x-Axis). The shift in intensity of the ROIs is normalized to their corresponding reference spots (Ref1-Ref3). As can be seen binding occurs during the Streptavidin flush at the positive control spot (red line) and during the anti-GFP flush at the GFP spot (blue line). No binding is detected at the negative control (grey line).

The Detection Device

The Basic iRIf Setup

Figure 4: Schematic iRIf Setup. The PDMS flowcell and the glass slide form a chamber. Solutions are flushed through the chamber using a microfluidic system. Proteins (antigens) are fixed on the glass slide before. Light constantly illuminates the glass slide from the bottom. If other proteins (like antibodies) within the sample bind to the proteins (antigens) on the glass slide, the interference pattern changes. This change is detected by the CCD sensor of a camera and visualized by a software.

Homogenous illumination of the glass slide with monochromatic light is necessary to achieve good iRIf results. This is best accomplished with a powerful LED light source shining on a lense which is positioned at the distance of one focal length, therefore parallelizing the light rays.
Since the purpose of an iRIf measurement is the observation of the minute changes of light intensity of light reflected at the slide, using a sensitive camera with a color depth of 12 bit is advisable. Images have to be stored lossless in the camera. Since compression methods such as JPEG or MPEG remove subtle changes in the picture, resulting the binding signal is lost otherwise. Although not mandatory, microfluidic systems are very well suited for use in conjunction with iRIf, since the iRIf measurement device and the microfluidic chamber can be positioned on opposite sites of the glass slide.

Building Our Device

Since the physics behind iRIf is well characterized and the parts necessary for building such a device are easily obtainable, we took it on ourselves to build our very own affordable iRIf device. The results of this endeavour, including a detailed manual on how to rebuild our iRIf device can be found here.

The Physics Behind iRIf

This chapter focuses on the physical theory behind iRIf in detail. After a short introduction to beam and lens optics, wave optics will be introduced to illustrate effects not describable with the former model. Finally, the manner of functioning of CCD chips will be explained to round up the theory.

1. Geometric optics

Refraction iRIf

A very intuitive way to describe light is in form of thin beams. Using this approach, effects around refraction can be illustrated by the following picture: A beam (B1) propagating through a medium (M1) confronting the surface of a different medium (M2) is partially reflected (B2) at the surface and partially refracted (B3) into the medium. The rates of reflection and transmission, R and T, depend on the angle of incidence θ as well as the medias' refraction indices n1 and n2 and are described by the Fresnel equations:



These coefficients depend on the lights' polarisation (linear in case of an LED). This polarisation can be seen as a combination of a perpendicular part (s) and a parallel one (p). Furthermore, continuity demands that transmitted and reflected light together return the original beam.

Geometric optics are also sufficient to describe the concept of lenses. A lens takes the light coming from an (illuminated) object and projects a sharp picture of it (here we will limit ourselves to only convex lenses). Its origin lies in the idea of refraction on a curved surface. A simple illustration is the following:

An object in distance o from the lens and height O is given an image of distance j and height J. The distance f describes the focal point of the lenses. To obtain a sharp image the lens equation must be obeyed:

The resulting image ratio A can be obtained using the intercept theorem:

One case of interest is the situation where the object is set at the focal point of the lense. This results in an infinitely large image in an infinite distance i. Therefore the refracted beams behind the lens can be seen as parallel ones. This is done with the LED light at iRIf.
But not every effect can be described by a beam concept. In order to understand interference, light has to be described as a propagating wave which is done in the next chapter.

2. Introduction to Wave Optics

Since the derivation of the Maxwell equations the wave character of light is accurately described. Light waves are propagating electric and magnetic fields and therefore can be described as such. For the sake of simplicity it's sufficient to describe the waves' electric field at position r and time t by:

Where E is the wave amplitude, k is the wave vector, ω the angular frequency and φ the wave phase. Regarding the given iRIf system this form can further be simplified: We only have to consider two waves from the same light source (LED), being in the same medium and the same position (fixed camera sensor). Therefore, we have the same angular frequency (ω1=ω2=ω) and can translate our coordinate system's origin unto the position of interest (where r=0). Also the periodicity of the two waves gives the freedom of choosing one of the phases to be zero and the other one to be the phase shift between them (Φ1 = 0, Φ2 = ΔΦ). Our system of interest therefore consists of two waves:

A property describing the power of light acting on a surface is its intensity. Physically the intensity is described by

where Z is a proportional constant and brakets mean the average over one period. The idea behind iRIf is the measurement of changes in the intensity of an interfered wave. What interference is will be explained now before returning to a proper phaseshift-depending form of I.

3. Interference of Two Waves

When two forces are acting on the same body the body will move as if only one force being the sum of those two would act on it. This principle of superposition is applicable to light waves too: If two waves are in the same place they can be summed up to one.

This effect is called interference and it leads to new waves with different properties, one of them being the amplitude. In our case this amplitude change is mainly depending on the phase difference between the waves. Using trigonometric identities [1] the resulting light intensity can be found:

Expecting the single wave amplitudes as well as Z to remain constant over the whole process (all are set to "1" in the following chart) makes the intensity a function depending solely on the phase shift between the interfering waves:

This function is not a monotoneous one which theoretically could lead to a problem if checking for changes. Yet iRIf is very sensitive and works in very small intervals in which monotony is given (only areas around minima and maxima would be problematic, yet are easily preventable by a clever multilayer system). Why is this important to know? Because the phase shift will change if antibodies bind to the surface. The phase shift depends on the additional distance of the second wave to return to the first one. Due to the described interference conditions it is necessary to select an LED with a suitable wavelength. The selected wavelength in combination with the optical properties of the used substrate determine the signal dynamics. For this project a substrate with a tantalum pentoxide layer in nanometer range was selected. The following sketch visualises the system:

Ray 1 (R1) is the reference ray (the one set to have no phase shift), R2 is the one reflected at the unbound side of the slide and R3 is the one that is reflected from the artificially longer end at bound spots. The binding can be seen as if the slide is actually thicker in this area (thanks to antibodies having almost the same refraction index as glass) and is physically described by the optical thickness, the product of refraction index and medium thickness. <7p>

How do we obtain the phase shift between R1 and R2/R3? By comparing the distances light has to overcome in order to return from the backside of the slide to the tantalum pentoxide layer (Ta2O5). We assume the light to have zero phase at this layer, so what phase does it have after returning? We only need the angular frequency to solve this:

The additional phase shift therefore depends only on the wavelength of the light as well as the thickness of the bound spot.

4. Coherence

Let us consider light as a statistical stream of wave packets. If one wave packet confronts a different medium, two new wave packets are created: a reflected and a transmitted one. Interference is only possible if the returning transmitted packet (after being reflected on the next layer) still interferes with the other packet and not a new reflected one. The concept behind this condition is called coherence. If it was not for coherence windows would be as colored as soap bubbles.
As long as the path between the two reflections is shorter than the coherence length l interference will occur. The coherence length depends on the spectral range of the light source and, for a 520nm LED, can be approximated by:

Fortunately, this is a very high limit as the tantalum pentoxid layer lies in nanometer distance to the last layer and proteins have a similar scale. At the same time this shows that standard glass slides could not be used for detection with iRIf.

5. Technical Details: CCD Sensors

Modern cameras use CCD chips to take images. A CCD sensor is a 2D-array made out of photo diodes. In these diodes an electric potential is induced by incoming light due to the photoelectric effect: Electrons in the diode material are excited and pushed out of the material leaving behind gaps. By closing a circuit at specific times (periodically) a current (due to the potential difference) can be measured. The more gaps were produced in a time interval the stronger is the resulting current (direct proportionality). Consequently, the sensored light strength is depending on the number of photons the light ray is transporting per time interval. Here we return to the definition of intensity. Intensity generally is defined as power acting on a surface. The power is the work done in a specific time. The work of light (seen as a ray of photons instead of waves) can be described as the total photon energy it is carrying. If the light is monochromatic the photons are all quanta of the same energy. Therefore:

The lights intensity is proportional to the average number of photons it carries, and the resulting current is proportional to the number of gaps generated due to the photoelectric effect which obviously is proportional to the number of photons of the light ray. Or in a short form: The registered light strength of the CCD chip is linearly dependent on the lights' intensity.

6. Summary

In order to understand iRIf one needs to know that the change in intensity of two interfering waves is measured by this system. The measuring tool is a CCD chip inside a camera. The intensity changes due to a lengthened distance between the two interfering waves as proteins bind to the glass slide make it virtually thicker. Other possibly disturbing reflections can be neglected due to incoherence. Lastly our iRIf system uses two lenses from which one is needed to produce parallel LED light while the other makes a sharp 1:1 image from the flowcell to the CCD sensor.

Legal Notice

The iRIf detection method is patented. Biametrics and associated persons own patents concerning the detection principle. These patents are:

References