WO1997043631A1 - Sensor for detecting proteins and process for its production - Google Patents

Abstract

The invention relates to a sensor for detecting proteins on the key-lock reaction principle, consisting of a sensor body, one surface of which is coated with a polymer layer, where the receptor molecules on the key-lock reaction are bonded to said polymer layer. It is the aim of the invention to design the sensor in such a way that it can be easily and highly reproducibly made. This is achieved in that the bond between the polymer and the receptor molecules is provided by a photoreactive molecule which is covalent to the lysine of a receptor molecule and inserted into a C-H bond of the polyimide.

Description

Sensor for the detection of proteins and method for its production

The invention relates to a sensor for the detection of proteins according to the preamble of patent claim 1, as known from DE 44 18 926, and a method for its production.

So far, there have been some established analytical methods for analyzing biomolecules, which, however, mostly use complex and expensive laboratory methods (HPLC, GC-MS). In contrast, the so-called immunoassays represent a similarly inexpensive alternative to the method described here. These are also based on the binding of an analyte to an antibody, but inevitably use an indirect method for the detection of this binding. A radioactive, fluorescent or enzyme-labeled analyte-analog molecule is added to the sample, which competes with the analyte for the antibody binding sites. A method is therefore required for the evaluation, which consists of several reagent feeds, incubations and washing cycles; the total time per assay is typically one hour. An online measurement is therefore impossible.

In addition, other immunosensor principles have been described in the literature. In the opinion of many authors, they still deviate significantly from the ideal idea of an inexpensive, sufficiently sensitive future immunosensor: In several review articles, such methods have already been mentioned because of their high costs (surface plasmon resonance, grating coupler, differential interferometer) or because of them ¬ rer low sensitivity (potentiometric immunosensor) criticized, while the immunosensor based on surface waves (OFW) has been favored so far.

EP 0 361 729 A2 discloses a method of the type described above for producing a sensor which has a protective layer for the spatial separation of substrate and aqueous analyte solution - 2 - has. At an operating frequency> 100 MHz, this sensor has attenuations between 30 and 40 dB, which causes a high susceptibility to interference with strong noise.

The disadvantage of the sensor described in DE 44 18 926 is that the production requires a lot of work with a high risk potential and that the sensor properties have a large spread.

The object of the invention is now a sensor of the e. G. Design in such a way that simple manufacture is possible with good reproducibility.

This object is achieved by the features of claims 1 and 3. The dependent claims describe advantageous embodiments of the method.

The sensor enables the specific measurement of the presence or concentration of various biomolecules such as proteins, enzymes and more complex macromolecules - parts of the genetic material (DNA, RNA) or various pathogens (e.g. viruses or bacteria) - by directly detecting the binding to specific antibodies in aqueous solutions. With this method, time-consuming procedures based on competition between labeled and unlabeled analytes are no longer necessary (indirect detection procedure in immunoassays).

The sensor is a mass-sensitive sensor which uses the change in the speed of sound of acoustic surface waves (SAW) caused by the sorption of the analyte to determine the sorbed mass of the analyte and thus its presence or concentration in the solution to conclude.

In order to obtain an analyte-specific sorption reaction, selective coatings on the SAW substrate are necessary. Special flexibility is guaranteed here if these coatings consist of immunochemically active molecules such as antibodies or antigens. The sensor according to the invention is therefore a real immunosensor which determines its data in-situ and thus enables a real on-line measurement method for bioanalytics.

Compared to conventional bioanalytics, the sensor described offers a number of advantages:

Inexpensive: 4 - 10 DM per piece

Sensitivity like bio-assay

- transferable to any biosystems

very good long-term stability

The invention is explained in more detail below using an example with the aid of the figures.

Fig. 1 shows the course of an enzymatic decomposition of glucose on a sensor and the

2 shows the course of immunoreactions for different antibody concentrations.

3 shows the TRIMID content of the receptor protein and

4 shows its enzyme activity on the basis of absorption spectra.

5 shows the receptor connection to the polymer polyimide.

The sensor in our example works on the basis of surface acoustic waves. Such a sensor is described in DE 43 19 215. The sensor body is first coated on one side with a polymer, in this example an aromatic polyimide. The surface of the sensor body is coated with polyimide as described in DE 44 18 926, p. 3, lines 11 to 55. 10 μl of the suitably diluted, modified receptor molecule are then added to the polyimidized sensor surface.

As can be seen in FIG. 5, only the photo-labeled enzyme is immobilized on the polyimidized sensor surface if the water content in the protein matrix is kept low. If the water content in the protein is too high, the carbene reacts with water on exposure, so that the insertion reaction between the polyimide and the protein is pushed into the background and the connection to the sensor surface is absent. For this reason, the sensors are treated for 20-40 minutes in a vacuum drying cabinet at room temperature and a pressure <10 mbar. A 30 minute drying at 1 mbar is optimal. The resulting dried enzyme film was then exposed to light using a mercury vapor lamp. To generate the triplet carbene, exposure of the enzyme film at 348 nm and 0.7 mW / cm ² for 30 min is necessary.

For the immobilization, TRIMID-modified glucose oxides (T-glucose oxidase, T-GOD) with a protein content of 1.92 mg / ml were used. The TRIMID content was 6.5 mol TRIMID per mol glucose oxidase.

Preliminary investigations showed that the T-GOD solution could not be deposited undiluted on the sensor. If the protein concentration is too high, there are high input attenuations and no acoustic wave can be observed.

As a result of these preliminary investigations, the T-glucose oxidase solution was diluted 1: 125 with phosphate buffer (1: 100), 10 μl of this solution was applied to the sensor and the Enzyme immobilized as described above (vacuum-treated, exposed).

The amount of enzyme deposited on the sensor surface could be determined spectroscopically using an enzymatic assay.

The following solutions are necessary to carry out the enzyme activity check:

Solution 1

53 mg of 3, 5-dichloro-2-hydroxy-benzosulphonic acid are dissolved in water (bidistilled) and brought to pH = 7 with 1 M NaOH. Then 3 mg peroxidase (horseradish) are added and made up to 10 ml with water.

Solution 2

16.2 mg of 4-aminophenazone are dissolved in 10 ml of water.

Solution 3

1.4 g Na ₂ HP0 ₄ * 2 H ₂ 0 and 700 mg NaH2Pθ4 and 37.2 mg EDTA are dissolved in 100 ml water.

Solution 4

1.8 g of -D-glucose are dissolved in 10 ml of water.

Then 1.55 ml of solution 3, 0.2 ml of solution 4, 0.2 ml of solution 1, and 50 μl of solution 2 are placed in a cuvette

given and after thorough mixing (vortex) measure the background at 520 nm. 50 μl of an appropriately diluted GOD solution are then added and the change in absorption at 520 nm (formation of a red dye) is measured within the first minutes. With high glucose oxidase concentration It takes only 2 minutes, the amounts of antigen deposited on the sensor were measured for 10-20 minutes.

The extinction coefficient of the resulting dye is 13300 (M cm) ^-1 . It is thus possible to determine the specific enzyme activity of the glucose oxidase solution. This is given in "units / mg", a "unit" being defined as follows: A "unit" oxidizes 1 μmol -D-glucose per minute at pH = 5.1 (T = 35 ° C).

Since 2 mol of hydrogen peroxide (from glucose cleavage) are necessary to produce 1 mol of the red dye, the measured increase in chromophore corresponds to 0.5 units per minute.

The increase in absorption at 520 nm is given as an example for three different sensors in FIG. 1. This gives an average of the slopes of 0.0021 absorption units per minute. A comparison with the enzyme activity of the T-GOD stock solution shows that this increase in absorption corresponds to a protein mass of 18.5 ng on the entire sensor.

The sensor sensitivity and the detection limit for polyclonal antibodies against glucose oxidase were determined by varying the amount of antibody in the analyte stream.

For this purpose, the sensors were coated with T-glucose oxidase via the photoinitiated reaction described and first rinsed with bovine serum albumin (4 mg / ml) in order to block the non-specific binding sites. Since the sensors were sampled in flow with the protein, they showed different deposition rates, but all of them then reached a frequency change of 35 kHz (within an error range of 10%).

The sensors pretreated in this way were then individually sampled with anti-body solution, the analyte - polyclonal antibody against glucose oxidase - was passed in a circuit via the sensors. Different amounts of antibody were dissolved in 5 ml of phosphate buffer and passed over the sensors. The concentration series, which is shown in extracts in FIG. 2, comprised a range of 2-200 μg / ml antibody (corresponds to 10-1000 μg).

Different frequency decreases and different initial speeds of the immunoreactions can be seen.

The change in the resonance frequency of the oscillator was plotted against the amount of antibody in the analyte stream. The slope of the straight line reflects the sensor sensitivity, which was determined to be 58.8 Hz / μg.

In order to determine the correlation between the amount of antibody and the reaction rate, the initial rate of the immunoreaction was determined for each measurement. The frequency decrease per unit of time could be calculated from the measuring points within the first minute after addition of the analyte by linear regression. The correlation coefficients were significantly greater than 0.98 in all evaluated curves.

The initial speeds determined in this way correlate with the amount of antibody added. There is a clearly linear relationship between the two quantities (r = 0.9822). The slope of the regression line is 4.92 Hz / (sμg).

To determine the detection limit for polyclonal antibodies against glucose oxidase, the procedure is as follows. The sensitivity is 58.8 Hz / μg, the intercept was determined to be 27.1 kHz. With these values, taking into account the triple noise signal from the sensors (120 Hz), the detection limit for polyclonal antibodies against glucose oxidase can be determined to be 2 μg or 13.6 pmol. Since the respective amount of antibody was weighed into 5 ml of phosphate buffer, this corresponds to - 8 - _. This value of a minimum detectable concentration of 2.7 nmol / 1.

The coating method described can also be applied to other sensor or measuring principles. For example, the sensor chips of the optical grating coupler can be coated in the same way with modified receptor molecules.

Enzymes, antigens and antibodies as well as nucleic acids can be used as possible receptor molecules.

An example of a coating is shown on the antigen glucose oxidase, which was reacted directly with 3-trifluoromethyl-3- (m-isothiocyanophenyl) diazirine (TRIMID) in order to obtain a photoreactive protein. In principle, it is also possible to modify the antibody itself with TRIMID, but it is more expensive.

The synthetic route could not be transferred directly from the literature-known T-BSA to T-GOD, since glucose oxidase contains the coenzyme FAD, which is not covalently bound to the enzyme. By reacting GOD with TRIMID at pH 11.4 by means of incubation at 50 ° C., glucose oxidase was modified with TRIMID, but the coenzyme was separated, which was demonstrated by the UV spectrum. A shoulder at 348 nm (TRIMID) was found, the FAD peaks at 375 and 450 nm were no longer recognizable.

Investigations have shown that the incubation at 50 ° C. leads to the separation of the coenzyme, but the treatment of the enzyme at pH = 11.4 does not bring about this dissociation.

T-GOD was therefore manufactured according to the following regulation:

18 mg glucose oxidase and 1.27 g -D-glucose were mixed in 0.1 vol% TEA (in water, pH = 11.4) dissolved and the resulting solution adjusted to a pH of 10.4 with pure TEA.

Add 170 μl TRIMID (29 μmol / 1) in chloroform

Sonicate 30 seconds in an ultrasonic bath, with a milky yellow

Suspension was created

Incubate for 2 hours at 37 ° C in a water bath

Chromatograph over a Sephadex G-25 column in 1.5 mM NaCl, 0.05 mM sodium phosphate buffer (pH = 7.4).

The protein concentration can be determined using the Lowry method. For this purpose, a specially prepared BSA standard was measured as a reference and the absorbance of the T-GOD related to it.

For this investigation, a stock solution of 0.5 ml 2% CuS0 ₄ solution, 0.5 ml 2% tartrate solution and 49 ml 2% Na2CO3 solution in 0.1 M NaOH was necessary. In order to determine the protein concentration of fractions 12 and 13, 100 μl were diluted to 1 ml with PBS (1: 100) and 6 samples were prepared by adding 10 μl, 20 μl and 30 μl to 200 μl with PBS (1: 100) were diluted (double determination). 1 ml of the stock solution was added to each of these samples, and the mixture was left to stand for 10 min.

Then 100 μl of 0.5 N folin solution was added, which produces an extinction maximum at 578 nm, which could be measured after 30 min. The protein amount of T-GOD could then be calculated using the BSA standard; the protein mass concentration was 4.22 mg / ml.

The trimide content of a protein can be checked by the difference in extinction of a sample before and after exposure at 348 nm. The problem with glucose oxidase is that the FAD molecules in this area also absorb light. beer that covers the TRIMID peak. When determining the TRIMID content in T-GOD, the fact was used that the FAD is separated from the enzyme after an incubation at elevated temperature. Since the FAD peak overlaps the absorption band of TRI-MID, the FAD was first separated to detect TRIMID by incubating the protein (500 μl in each case) at 50 ° C. for 2 h and then chromatographing the solution on a PDIO column. After that, only the protein peak at 280 nm was visible in the enzyme fraction.

The protein fraction was then exposed twice in the cuvette for 10 min each and an absorption spectrum was recorded after each exposure. 3 clearly shows a change in the absorption between 340 nm and 400 nm, that is to say in the region of the TRIMID band.

The covalently bound TRIMID content was 8 + 2 mol TRI-MID per mol glucose oxidase.

The enzymatic activity of the modified protein could be determined spectroscopically using the enzymatic assay described above. For this purpose, 50 μl of T-GOD, which contained 8.73 μg protein, were examined.

The enzymatic activity of the glucose oxidase solutions used was determined with this method. 4 shows the kinetics of the enzymatic catalysis of the stock solution (GOD) and that of the modified enzyme (T-GOD).

A straight line was determined through the linear part of the curve by means of linear regression. This served as a calibration line for the detection of the enzymatic activity of the respective glucose oxidase on the sensor.

The determination of the increase in absorption by means of linear regression gave a value of 0.989 min ^-1 for T-GOD. With the help of the Lambert-Beer's law, the specific enzyme activity of the modified glucose oxidase can be determined at 34.92 units / mg. In comparison, the unmodified glucose oxidase has a value of 87.59 units / mg.

In order to observe the binding of T-GOD on a surface and in particular on polyimide, the T-GOD produced was marked with [ ¹⁴ C]. For this purpose, the glucose oxidase was first reductively methylated and then immobilized on a polyimidized surface.

In reductive methylation, amino functions are reacted with formaldehyde and the resulting Schiff bases are reduced. This highly specific reaction only attacks the -amino groups of the lysine units in the protein and the N-terminus in the amino acid sequence. In order to avoid denaturation by destroying the disulfide bridges and the immediate reduction of the formaldehyde, sodium cyanoborohydride is used as a mild reducing agent, which leads to N, N-dimethyl derivatives. The result is an almost complete conversion when using a six-fold excess of formaldehyde, based on the protein mass.

Since the reductive methylation takes place in HEPES buffer (pH = 7.5), 1 ml of T-GOD (4.22 mg / ml) was buffered over a PDIO column and the protein fraction used for radioactive labeling. The procedure was as follows:

800 μl of this protein fraction were placed in a light-protected reaction vessel (protection of the TRIMID group) addition of 6.5 μl [ ¹⁴ C] -formaldehyde (= 200 nmol), with a radioactivity of 11.6 μCi addition of 100 μl NaCNBH ₃ , (= Make up 240 mmol / 1) with HEPES buffer to 1 ml and stir for 3 h at room temperature. The product was chromatographed on a PDIO column in order to remove excess formaldehyde. The T-GOD fraction of the radiolabeled enzyme was examined for protein content as described above. The protein concentration was 0.915 mg / ml.

In order to determine the degree of radioactive modification of T-GOD, 5 μl of the individual fractions of the [ ¹⁴ C] -T-GOD synthesis were each treated with 5 ml of scintillation solution (a mixture of 1080 ml of toluene pa, 5.4 g 2,5-diphenyloxazole (PPO), 0.2 g 2,2'-p-phenyl-bis-5-phenyloxazole (POPOP), 920 ml Triton and 40 ml glacial acetic acid) mixed (vortex) and the radiation with a scintillation counter certainly.

An activity of 2793 dpm (dpm = decompositions per minute) was found in the protein fraction. The correlation of these values with the results of the Lowry protein determination of the radioactively labeled protein gave 597 dpm / μg protein.

The enzymatic activity of the radioactively labeled protein fraction was again determined using the enzymatic assay. For this purpose, 20 μl of the protein fraction were diluted to 200 μl with HEPES and the enzymatic activity was determined from 50 μl of this solution (= 4.6 μg).

The evaluation showed an increase in absorption of 0.27 min ^-1 , which corresponds to a specific enzyme activity of 18.19 units / mg. This means that even after two modifications to the protein, the enzymatic activity is still present.

In order to observe the coupling of the protein to a polyimidized surface, glass cover plates were polyimidized, then coated with [ ¹⁴ C] -T-GOD and the radioactivity was measured. The polyimidization of hydrophilic surfaces is, among other methods, an important basis for immobilization on acoustoelectric components.

30-50 μl of the radioactive labeled enzyme were applied to the polyimidized glass plate, according to a defined one Exposure time exposed and then washed. As a control experiment, analog treated but unexposed plates were carried along.

The washing procedure consisted of rinsing the platelets five times with a solution of 50 mM PBS, 150 mM NaCl and 0.02% by volume TWEEN 20. To determine the radioactivity, the cover platelets were placed in the scintillation tubes, covered with 5 ml scintillation solution and then short mixing (vortex) determines the radiation of the polyimidized glass slides. In FIG. 5 it can be seen that the exposure after 30 minutes of exposure has no effect on the polyimide. The radiation from these glass plates is approximately of the same order of magnitude as that from the unexposed control samples. Only after significantly longer exposure times before exposure or partial drying by applying a vacuum, which has resulted in a completely dried-up protein film, does the protein become covalently bound to the polyimide, which can be recognized by the significantly higher radioactivity of the corresponding glass carriers is.

A photoimmobilization of T-GOD on polyimide is consequently possible if only a little water is present in the protein matrix. In the presence of water, the degree of modification of T-GOD with TRIMID is too low, so that all carbenes produced by exposure react with water and there is no measurable connection to the surface. Furthermore, these investigations show that the selected washing procedure is suitable for removing non-specifically adhering protein molecules from the polyimized surface and for washing out residual radioactivity.

Claims

Claims: 1. Sensor for the detection of proteins according to the principle of the key-lock reaction consisting of a sensor body, of which a surface is coated with a polymer layer, the receptor molecules of the key-lock reaction being bound to the polymer layer, characterized in that the bond between the polymer and the receptor molecules is mediated by a photoreactive molecule which is covalently bound to the receptor molecule and inserts into a CH bond of the polymer. 2. Sensor according to claim 1, characterized in that the photoreactive molecule is 3-trifluoromethyl-3- (m-isothiocyano-phenyl) diazirine (TRIMID). 3. Method for coating sensors according to one of claims 1 or 2, with proteins, the surface of the sensor being modified and a silane-containing promoter layer being applied to the sensor surface and a polymer layer being bonded thereto, characterized by the following method steps: a) modify the proteins by attaching carbene-generating molecules to the lysine units of the amino acid sequences of the proteins, b) applying the modified proteins to the polymer layer c) partially drying this layer and d) covalently linking the modified proteins to the polymer layer by exposure to UV light. 4. The method according to claim 3, characterized in that the carbene-generating molecules are TRMID. 5. The method according to claim 3 or 4, characterized in that the degree of dryness of the layer is achieved by applying a vacuum.