WO1988008875A1

WO1988008875A1 - Cycling biochip sensor

Info

Publication number: WO1988008875A1
Application number: PCT/US1988/001501
Authority: WO
Inventors: Clement Eugene Furlong, Jr.; Edward Bakewell Weiler; Peter Wing-Poon Cheung
Original assignee: The Washington Technology Center
Priority date: 1987-05-05
Filing date: 1988-05-04
Publication date: 1988-11-17
Also published as: CA1316212C

Abstract

A cycling biochip sensor adsorber produced by immobilizing a ligand binding protein having high ligand affinity and specificity and capable of reversible denaturation on a substratum. The immobilized binding protein is denatured by heat which induces ligand release and renatured by cooling to ambient temperature regenerating an active biochip sensor. The substratum may be selected from among various solid support structures including semiconductor materials. When the binding protein is immobilized on the insulated gate region of an IGFET, the component so formed is suitable for use as an element of a sensor, when electronic circuitry is provided.

Description

CYCLING BIOCHIP SENSOR Field of the Invention This invention relates to an apparatus for selectively adsorbing a ligand from aqueous solution in which the adsorption apparatus may be used as an element of a sensor. Specifically this invention relates to substratum having biologically active molecules immobilized on the surface. The biologically active molecules are capable of forming a complex with a ligand in aqueous solution, the formation thereof causing a variation in a physical property of the substratum which may be detected by electronic means. Subsequent to forming the complex the biologically active molecules are reversibly denatured to recover the ligand and regenerate the active adsorber.

Background of the Invention For a variety of reasons, it is often desirable to selectively complex small molecules from aqueous solution. First, selective complexing may provide a method for isolation of unique ligands from solution. Biologically active molecules can selectively bind to ligands with extremely high selectivity. For example, biologically active molecules such as proteins can easily differentiate chemical isomers. This high selectivity can provide a means for isolation, purification, or sensing unique small molecule ligands, if the biologically active molecule is ligated to a substratum and can be selectively induced to release its bound ligand, and if the biologically active molecule does not transform the small molecule ligand into another species. It is also often desirable that release of the small molecule ligand from the biologically active molecule be accomplished without the introduction of other substances such as competitive inhibitors, ions or denaturing agents that would complicate the isolation, purification or sensing procedures.

Second, selective complexing of small molecule ligands can provide a means for accurately detecting and quantitating specific small molecule ligands in aqueous solution. For example, it is often desirable to qualitatively and quantitatively measure substances dissolved in an aqueous medium, without the necessity of withdrawing an aliquot and measuring the included substances by standard techniques. For example, rapid and continuous measurement of substances dissolved in body fluids such as serum is often critical to making a proper diagnosis of a patient's condition. Present techniques often involve withdrawing a sample of blood or urine and sending it to a clinical chemistry lab for analysis. It would be a great advantage to be able to measure levels of dissolved substances without withdrawing these fluids or waiting for analysis.

Similarly, continuously monitoring solute levels is often important to the chemical and food processing industries in order to control levels of important constituents, and to make sure that effluent discharge requirements are met.

Accordingly, there exists a need for a sensing device that selectively responds to the presence of specific small molecule ligands dissolved in an aqueous medium. Furthermore, it is important to have a sensing device that produces an output that varies according to the concentration of the dissolved substance. It is also important that the sensor be able to respond to a wide range of concentrations and that it be capable of functioning in a continuous mode. Finally, the device should be long-lived and be capable of operating under relatively harsh conditions. It should be small, easily manufactured and it should retain its selectivity and sensitivity over the span of its useful life.

Summary of the Invention The invention consists of a cycling biochip sensor having a semiconductor substratum, and a binding protein capable of reversibly complexing a ligand in aqueous solution when immobilized on the semiconductor substratum. The binding protein is attached to the semiconductor substratum such that complexing the ligand with the binding protein perturbs the semiconductor substratum in a detectable manner. The binding protein is capable of forming a high affinity binding complex with a small molecule ligand upon contacting the small molecule ligand in aqueous solution. The binding protein is denatured upon heating to temperatures of about 75°C, releasing the small molecule ligand into solution. The binding protein then renatures upon cooling, regenerating the biochip sensor. The cycling biochip sensor also is equipped with a signal processor coupled to the sensor for producing a signal that varies according to the small molecule ligand and binding protein forming a complex.

The cycling biochip sensor has a protein immobilized on the semiconductor surface wherein the protein is selected from appropriate ligand- binding proteins, protein fragments or synthetic analogs thereof. In a preferred embodiment of the invention, the protein is selected from periplasmie binding proteins, and mutant periplasmie binding proteins having an altered amino acid sequence. Examples of suitable periplasmie binding proteins include those isolated from Escherichia coli and Salmonella typhimurium, which form a high affinity binding complex with arginine, ornithine, lysine, oligopeptides, cystine, glutamine, glutamate, aspartate, histidine, leucine, isoleucine, valine, threonine, arabinose, galactose, glucose, maltose, ribose, xylose, β-methylgalactoside, citrate, phosphate, glycerol-3-phosphate, sulfate, vitamin B12, thiamine, and cadmium.

The cycling biochip sensor has a semiconductor substratum made from Si, Ge, Ga As, SiC, and silicon on saphire, at least a portion of the surface being oxidized or covered with a metal and/or metal oxide. Alternatively, those skilled in the art will appreciate that the substratum can be a metal or metal/metal oxide system, to which a biologically active molecule can be immobilized. Examples of such metal or metal/metal oxide systems include antimony or antimony/antimony oxide, or palladium or palladium/palladium oxide. The cycling biochip sensor further consists of a linking compound covalently bonding the binding protein to the semiconductor substratum. The linking compound reacted with the oxidized semiconductor substratum consists of a compound with the structural formula:

H₂N(CH₂)_n-Si(O-(CH₂)_m-CH₃)₃

where: n is an integer from 2 to 7 m is an integer from 0 to 4

and a homobifunctional cross -linking compound or a heterobifunctional cross-linking compound, selected from dimethyl adipimidate, dimethyl-3,3'-dithiobispropionimidate, dimethyl pimelimidate, dimethyl suberimidate, bis-[2-(succinimidooxycarbonyloxy)ethyl]sulfone, disuccinimidylsuberate, disuccinimidyltartrate, dithiobis-(succinimidylpropionate), 3,3'-dithiobis(sulfosuccinimidylpropionate), ethyleneglycolbis-(succinimidylsuccinate), bis-[2-sulfosuccinimidooxycarbonyloxy)ethyl]sulfonate and the heterobifunctional cross-linking compound is selected from those having functional groups independently selected from N-hydroxysuccinimidyl, maleimidyl, pyridyl, N-hydroxy-sulfosuceinimidyl, alkyl α-keto halides, benzyl halides alkyl or aryl α-haloamide, aryl isothiocyanate and azidophenyl. Also provided for in the present invention is a kit for detecting a small molecule ligand in aqueous solution made by covalently linking a cycling biochip sensor to a semiconductor substratum. The binding protein is capable of reversibly complexing a small molecule ligand in aqueous solution, which perturbs the semiconductor substratum in a detectable manner. A sterile transparent package containing the cycling biochip sensor is also provided.

Detailed Description of the Invention A class of proteins termed "binding proteins" resides in the space between the inner and outer membranes (periplasmie space) of gram-negative prokaryotes. These proteins are involved in transporting nutrients across the plasma membrane. These binding proteins bind their respective nutrients with high specificity and affinity. In addition, these binding proteins are resistant to proteases and heat. Further, these proteins bind their respective substrates through a broad range of pH values and ionic strengths. These special properties make this class of proteins especially adaptable for use in cycling column adsorbers for removal of specific small molecules from aqueous environments, and as binding components of "biochip" sensors.

X-ray crystallographic studies of several of these binding proteins show that they are bilobate in shape, the space between each lobe defining a concave cleft where the ligand binds to the protein. The molecule is flexible and upon ligand binding, the two lobes close down on the ligand in a "Venus fly-trap" type binding mechanism. Thus the protein ligand binding causes a eonformational change in the protein (Ames, G.F.L., Ann. Rev. Brochure 1986 55:397-425) resulting in a high affinity binding complex. The binding affinity (K_D) for many saccharide ligands is from about 0.1 μ M to about 1.0 μ M, for amino acids is from about 0.01 μ M to about 50 μ M, and for anions about 0.02 μ M to about 30 μ M. Furlong, C.E., Methods In Enzymology 125: 279-289 (1986), hereby incorporated by reference.

The small K_DS make these binding proteins excellent candidates for small molecule ligand adsorbers as well as for elements of sensors when they are immobilized on the surface of an insulated-gate field-effect transistor (IGFET). However, the high ligand affinity is directly related to a low dissociation rate requiring that these proteins be "encouraged" to release their ligand by external stimulus to prevent the sensor from remaining saturated. It is known, for example, that isolated binding proteins often carry ligands complexed with their binding sites (Amanuma, H.M. et al., J. Biol. Chem., 79, 1167 (1976)). It has been discovered in accordance with the present invention that prokaryote periplasmie binding proteins which have formed a high affinity binding complex with a ligand will release the ligand upon heating to temperatures above 75°C and that the binding protein will renature to its original conformation upon cooling to ambient temperature. In addition to high temperature stability, prokaryote substrate binding proteins are stable to drying conditions, protease treatment, and other denaturing conditions such as guanidine-HCl treatment (Miller, P.M., Olson, J.S., Pflugrath, J.W. and Quiocho, F.A., J. Biol. Chem., 258: 13665-13672 (1983)). Thus, these binding proteins may be suitably employed in a sensor for detecting low concentrations (near the K_D) of small molecule ligands in aqueous solution, and they may be cycled by heating and cooling or other reversibly denaturing conditions such as organic solvents and chaotropic agents to release the bound ligand and regenerate the active sensor. These sensors being useful for both quantitative and qualitative measurement of ligands in aqueous solution. Alternatively, those skilled in the art will appreciate that the substratum can be metals or metal/metal oxide materials such as antimony or palladium as metals and antimony/antimonyoxide or palladium/palladium oxide as metal/metal oxides. In accordance with the present invention a cycling biochip sensor has a semiconductor base or substratum. Suitable semiconductors include, Si, Ge, Ga As, SiC, and silicon on sapphire. A particularly useful semiconductor is the insulated-gate field-effect transistor (IGFET) having one or more insulated gate regions. When the amount of charge near these insulated gate regions either increases or decreases, a physical property of the FET, for example its conductance, varies accordingly. This variation in a physical property of the IGFET can be used to produce a signal that varies according to the amount of ligand binding complex formed. This signal may be processed according to techniques well known in the electronics art. In the case of an IGFET, binding of a ligand to a binding protein immobilized near the surface of the insulated gate region is the event that causes the conductance to vary according to the amount of substrate binding that has occurred.

In one configuration of the present invention, the cycling biochip sensor is a remote sensor connected by wire leads to electronic circuitry suitable for processing signals produced by the remote biochip sensor, and providing information to the user regarding the presence and relative amounts of ligands in aqueous solution. In this configuration, these cycling biochip sensors may be replaced from time to time with new sensors having the same or different binding proteins immobilized on the insulated gate regions. These remote cycling biochip sensors and wire leads are provided in kit form contained in sterile transparent packages.

In another embodiment of the present invention the substrate is a charged species. Periplasmie binding proteins having specificity for anions such as citrate, phosphate, glycerol-3-phosphate sulfate can be immobilized upon the insulated gate region of the IGFET. As the number of these anions bound to their respective binding proteins increases, the current from source to drain varies due to the increased charge sequestered near the insulated gate region. The increased charge sequestered near the insulated gate is due in part to the charged ligand and in part to the conformational change of the protein that occurs upon ligand binding.

In still another embodiment of the invention, uncharged or neutral ligands complexed with binding proteins immobilized at the insulated gate region of the IGFET produce a similar change in conductance according to the number of ligand molecules bound. It is believed that conductance change in the IGFET upon binding neutral ligands is due to conformational change induced upon ligand binding which alters the surface charge of the binding proteins. Examples of suitable periplasmie binding proteins having high specificity for neutral ligands include proteins which bind arabinose, galactose, glucose, maltose, ribose, xylose, β-methylgalactoside, vitamin B12, and thiamine.

Zwitterions complexed with binding proteins on the insulated gate region of an IGFET are also capable of producing a change in conductance proportional to the number of ions bound. Examples of suitable periplasmie binding proteins having high affinity and specificity for zwitterions include binding proteins of: cystine, glutamine, glutamate, aspartate histidine, leucine, isoleucine, valine, and threonine.

Other binding proteins obtained from Escherichia coli and Salmonella typhimurium which are suitable for use in accordance with the present invention include those described by Furlong, C.E., in Methods in Enzymology 125: 279-289 (1986), herein incorporated by reference. Binding proteins from these prokaryotes are particularly useful since they can be easily produced in large quantities in a bioreactor. Still other proteins suitable for use include those described by Copeland, B.R., et al., J. Biol. Chem., 257: 15065-15071 (1982), herein incorporated by reference. In addition, two cadmium binding proteins from E. Coli have been described by Khazaeli, M.B., Appl. Environ. Microbiol. 41: 46, (1981). These latter binding proteins are suitable for use in a cycling cadmium adsorber.

In general, for a biologically active molecule to be useful as a ligand adsorber or sensing element of a biochip sensor it need only bind its ligand with an affinity and specificity appropriate to the particular application. A general procedure for identifying and isolating proteins is described by Copeland, B.R., et al., vida supra. The entire protein may not be necessary if an active binding domain of the protein can be isolated and if the domain can be reversibly denatured.

For a binding protein or other protein to be useful as a sensing element for a biochip sensor, it is preferred that the binding protein have a K_D near the concentration of the solution being measured. If the concentration of ligand in solution is higher than that of the K_D for a given binding protein, the measuring system can employ a feedback controlled dilution system to adjust, calculate, and control the necessary dilution of the original solution to bring it with measurement range for a given binding protein attached to the IGFET. In general, a binding protein is suitable for use in detecting the corresponding ligand at a concentration of from about ^KD/20 to about 2K_D. For example, the phosphate binding protein from E. Coli has a K_D of 0.8 μ M, therefore, this periplasmie binding protein would be suitable for use in a biochip sensor for measuring phosphate concentrations from about 40 μ M to about 1.6 μ M. In many instances it is useful to measure phosphate concentration above this range. For example, normal ranges for serum phosphate are between 0.74 and 3.1 mM (Teitz, N.W. Textbook of Clinical Chemistry, W.B. Saunders Co. (1986)) or about 1000 times higher than the useful range of the "wild type" E. Coli phosphate-binding protein. Mutant binding proteins having an altered amino acid sequence and therefore K_DS different from the wild type K_D can be produced by site specific oligonucleotide-mutagenesis (see for example Schultz, S. C. and Richards, J.H., Proc. Natl. Aead. Sci USA 83, 1588-1592 (1986), and references cited therein). In this way a spectrum of phosphate binding proteins can be produced each with its own K_D. A mutant phosphate binding protein can then be selected for a given application. In an alternative embodiment of the invention, a biochip sensor effective in measuring a broad range of ligand concentrations is produced by immobilizing binding proteins having different K_Ds selected from among the various mutant binding proteins or other proteins described above, onto different insulated gate regions of the IGFET. By selectively choosing binding proteins for a particular ligand with K_Ds separated from one another by a factor of about 100 or less, a broad range cycling biochip sensor is produced.

In another embodiment of the invention binding proteins for different ligands are immobilized on different insulated gate regions of the IGFET to produce a multiple substrate cycling biochip sensor. For example, normal serum sulfate and citrate concentrations for adults are 25 to 40 mM and 80 to 160 mM respectively. (Montgomery, R., et al., Biochemistry 3rd Ed., C.V. Mosby Co. 1980.) Thus selecting mutant binding proteins for sulfate, citrate and phosphate with K_DS of about 30 mM, 100 mM, and 1.5 mM respectively and immobilizing these proteins on different insulated gate regions of a IGFET would produce a sulfate-citrate-phosphate cycling biochip sensor suitable for continuously measuring serum levels of these three anions. Alternatively, the "wild type" binding proteins from E. Coli or S. Typhimirium could be used if a feedback controlled dilution system as previously described is employed. It will be appreciated by those skilled in the art that a broad range multiple substrate biochip sensor can be produced by immobilizing an array of binding proteins having a broad spectrum of K_Ds for each ligand on separate insulated gate regions of an IGFET.

The use of periplasmie ligand-binding proteins or other appropriate proteins as elements of a biochip sensor has the advantage of forming a stable high affinity complex with a substrate without altering the ligand, unlike the case with enzymes. However, in order to make a sensor from a binding protein useful for more than one concentration measurement, these binding proteins must be encouraged to release their ligand under conditions that do not alter either the specificity or affinity for a particular ligand.

A variety of methods and substances are suitable for reversing ligand binding to a binding protein. These include the use of "denaturing" substances such as guanidine-HCl, urea, organic solvents, and chaotropic agents. Other methods suitable for inducing ligand release from these binding proteins include incubation in a substrate free aqueous solution.

The preferred method for inducing the release of ligand from an immobilized binding protein is to heat the protein to a temperature sufficient to cause the protein to release its ligand in a relatively short period of time, the time and temperature determined by the particular application. In the case of periplasmie binding proteins, it is critical that the proteins be immobilized on a surface prior to heating, otherwise the protein coagulates irreversibly. Thus, heat is used as a "denaturant" only when periplasmie binding proteins are immobilized.

There are many advantages of using heat instead of ionic or molecular denaturants to facilitate ligand release. First, the use of denaturants such as guanidine-HCl to remove ligands from binding proteins necessarily requires removal of the sensor from the test medium. This may be undesirable for continuous concentration measurements. Furthermore, guanidine-HCl must be completely removed from the sensor prior to its reuse, a time-consuming pro cedure. Secondly, heat is easily generated by a number of methods including microwave irradiation, resistive heating from the semiconductor itself, and infrared irradiation. The removal of ligands by heating the protein to a point that it may be reversibly renatured allows the ligand to be released into a stream of desired composition, such as deionized water, for example, where the ligand may be concentrated in the absence of salts or other molecules. The use of heat avoids the introduction of other materials, which may be difficult to remove. The procedure is general, in that many different proteins may be used in this process. The only requirement is that the protein has the required specificity and affinity for the ligand of interest and that the protein be reversibly denatured by heat.

Thirdly, the temperature and its duration can be accurately and automatically reproduced or varied according to a particular application without the introduction of time-consuming, nonstandard procedures, the reproducibilϊty of which may depend on the skill of the technician.

In applications where the biochip sensor is to be used intermittently, simply immersing the sensor in hot water is effective to induce release of ligand from the protein. For example, a biochip sensor of periplasmie phosphate binding protein immobilized on a silicon semiconductor exposed to 10 μ M K₃PO₄/Tris-HCl buffer pH 7.4 becomes saturated with phosphate within 10 seconds. Immersion of the biochip sensor in 80°C water for 15 to 30 seconds causes the release of greater then 85% of the bound phosphate without measurable loss of binding affinity or specificity.

In applications where the biochip sensor is to be used continuously or where removal of the sensor is impractical, heat is applied to the sensor by resistive means, i.e., passing current through the semiconductor. The amount of current and its duration in order to effect ligand release will depend on the particular application and can be empirically determined.

The means for immobilizing a binding protein on a semiconductor insulated gate surface depends on the composition of the semiconductor. In the case of a doped silicon semiconductor, a portion of the surface is cleaned and a thin silicon dioxide layer approximately 500-1000 Å thick is thermally grown on the cleaned surface under an atmosphere of dry oxygen by the method described in Sze, S.M., Physics of Semiconductor Devices, 2nd Ed., John Wiley & Sons 1981). The silicon dioxide surface is then preferably derivatized using vapor phase deposition of silane compounds having structural Formula I:

H₂N(CH₂)_n-Si(O-(CH₂)_mCH₃)₃

where: n is an integer from 2 to 7 m is an integer from 0 to 4.

A preferred compound of structural Formula I is 3-aminopropyltriethoxysilane (APTES). Vapor phase deposition of a compound of structural Formula I deposits a chemically active layer of the silane compound covalently bonded to the SiO₂ surface. A binding protein, selected from those previously described, is then immobilized on the derivatized surface of the semiconductor by a covalent bond. Bonding the protein to the derivatized SiO₂ surface must be conducted under conditions that preserve the biological activity of the binding protein. This may be achieved in a number of ways. For example, the derivatized semiconductor base has a free primary amino group which can form an adduct with a binding protein which has been itself derivatized with a Michael-type acceptor, e.g., maleimide. Alternatively, the free amino group of the derivatized glass can be derivatized by converting the amino group into a maleimide which in turn may form an adduct with a nucleophilic group on the protein. Still another method of linking the binding protein to the derivatized semiconductor is to bond the binding protein through bifunctional cross-linkers, of the type disclosed in the 1986-87 catalog of the Peirce Chemical Company, Rockfork, Illinois, pages 312-340, herein incorporated by reference. These cross-linkers include but are not limited to: homobifunctional cross-linking compounds selected from dimethyl adipimidate, dimethyl-3,3'-dithiobis-propionimidate, dimethyl pimelimidate, dimethyl suberimidate, bis-[2-sulfosuccinimidooxycarbonyloxy)-ethly]sulfone, disuccinimdyl suberate, disuccinimidyl tartrate, dithiobis-(succinimidyl propionate), dithiobis-(3,3'-dithiobis-(succinimidylsuccinate), bis-[2-(sulfosuccinimidooxycarbonyloxy)ethyl] suflonate and heterobifunctional cross-linking compounds selected from those having functional groups independently selected from N-hyrdroxysuccinimidyl, maleimidyl, pyridyl, N-hydroxy-sulfoxysucinimidyl, alkyl α-keto halides, benzyl halides alkyl or aryl β-haloamide, aryl isothiocyanate and azidophenyl. For other semiconductors such as Si₃N₄, immobilization of the binding protein is achieved by the same procedures described above, since Si₃N₄ forms a thin surface oxide layer of several A providing sites for attachment for compounds of structural Formula I. Alternatively, avidin or streptavidin may be used to couple biotinylated binding protein to a biotin derivatized surface. It will be appreciated that the sensitivity and effectiveness of the biochip sensor will depend on the number of binding proteins immobilized per unit area or density of the cycling biochip surface. In the case of a silicon based IGFET, it has been found that suitable binding protein densities are on the order of 10¹³ protein binding sites/cm². Quiocho, F.A. and Pflugrath, J.W., J. Biol. Chem., 255: 6559 (1980). The maximum density being determined by the protein molecular diameter. The lower limit is determined by minimum signal detection. The preferred protein densities range from 10 to 10 protein binding sites/cm².

Those skilled in the art will recognize that the embodiments disclosed herein are exemplary in nature and that various changes can be made therein without departing from the scope and spirit of the invention. Numerous configurations of the sensor can also be provided to enhance the utility of a measuring instrument with respect to a particular application. Because of the above and numerous other variations and modifications that will occur to those skilled in the art, the following claims should not be limited to the embodiments illustrated and discussed.

Claims

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A cycling biochip sensor comprising: a semiconductor substratum, a protein capable of forming a binding complex with a ligand in aqueous solution, and means for attaching the protein to the semiconductor substratum such that complexing of the ligand with the protein forms a reversible heat-labile complex.

2. A cycling biochip sensor of Claim 1, wherein the protein denatures upon heating to release the ligand and renatures upon cooling to regenerate the biochip sensor.

3. A cycling biochip sensor of Claim 2, wherein the denaturation temperature is above physiological temperature.

4. A cycling biochip sensor of Claim % wherein the renaturation temperature is above physiological temperature.

5. A cycling biochip sensor of Claim 1, further comprising electronic input and output means coupled to the semiconductor substratum.

6. A cycling biochip sensor of Claim 5, further comprising signal processing means coupled to the input and output means.

7. A cycling biochip sensor of Claim 1, wherein the protein comprises a protein, protein fragment, or synthetic analog thereof.

8. A cycling biochip sensor of Claim 7, wherein the protein comprises a periplasmie binding protein, fragment, or analog.

9. A cycling biochip sensor of Claim 7, wherein the prokaryote is Escherichia coli or Salmonella typhimurium.

10. A cycling biochip sensor of Claim 7, wherein the attached protein is capable of forming a reversible heat-labile binding complex with a ligand selected from among arginϊne, ornithine, lysine, oligopeptides, cystine, glutamine, glutamate, aspartate, histidine, leucine, isoleucine, valine, threonine, arabinose, galactose, glucose, maltose, ribose, xylose, β -methylgalactoside, citrate, phosphate, glycerol-3 -phosphate, sulfate, vitamin B12, and thiamine.

11. A cycling biochip sensor of Claim 7, wherein the attached protein is an anion binding protein capable of forming a reversible heat-labile complex with a ligand selected from among citrate, phosphate, glycerol-3-phosphate, and sulfate.

12. A cycling biochip sensor of Claim 7, wherein the protein comprises phosphate-binding protein from Escherichia coli.

13. A cycling biochip sensor of Claim 1, wherein the semiconductor substratum is selected from among Si, Ge, Ga, As, SiC, silicon on sapphire, metal, and metal/metal oxide.

14. A cycling biochip sensor of Claim 13, wherein at least a portion of the surface of the semiconductor substratum is oxidized.

15. A cycling biochip sensor of Claim 13, wherein at least a portion of the surface of the semiconductor substratum is covered with a metal, metal oxide, or metal/metal oxide.

16. A cycling biochip sensor of Claim 1, wherein the means for attaching the protein to the semiconductor substratum comprises a linking compound covalently bonding the protein to the semiconductor substratum.

17. A cycling biochip sensor of Claim 14, wherein the means for attaching the protein to the semiconductor substratum comprises a linking compound covalently bonding the protein to the semiconductor substratum.

18. . A cycling biochip sensor of Claim 17, wherein the means for attaching the protein to the semiconductor substratum comprises:

R₁-NH-(CH₂)_n-Si-R₂ wherein:

R₁ is the linking compound, n is an integer from about 2 to about 7, and

R₂ is the oxidized substratum.

19. A cycling biochip sensor of Claim 16, wherein the linking compound is selected from the group consisting of dimethyl adipimidate, dimethyl-3,3'-dithiobispropionimldate, dimethyl pimelimidate, dimethyl suberimidate, bis-[2-(sulfosuccinimidooxycarbonyloxy)ethyl]sulfone, disuccinimydylsuberate, disuccinimidyltartrate, dithiobis(succinimidylpropionate) 3,3'-dithiobis(sulfosuccinimidylpropionate), ethylene glycolbis-(succinimidylsuccinate), bis-[2-sulfosuccinimidooxyearbonyloxy)ethyl]sulfonate and the heterobifunctional cross-linking compound is selected from those having functional groups independently selected from N-hydroxysuccinimidyl, maleimidyl, pyridyl, N-hydroxy-sulfosuccinimidyl, alkyl α -keto halides, benzyl halides, alkyl or aryl α -haloamide, aryl isothiocyanate and azidophenyl.

20. A cycling biochip sensor of Claim 1, wherein the means for attaching the protein to the semiconductor substratum comprises one or more of the group consisting of biotin, avidin, and streptavidin.

21. A cycling biochip sensor of Claim 1, further comprising means for heating the sensor to a temperature sufficient to dissociate the protein-ligand complex.

22. A kit for detecting a ligand in aqueous solution, comprising a cycling biochip sensor comprising a semiconductor substratum having immobilized thereon a protein capable of forming a reversible heat-labile binding complex with the ligand, the protein capable of denaturing upon heating to release the ligand and renaturing upon cooling to regenerate the biochip sensor.

23. A kit of Claim 22, further comprising a sterile package containing the cycling biochip sensor.