Description

Field of the Invention

[0001] The present invention relates to the field of sensors, particularly biosensor chips for detecting an analyte in a sample. The sensor chip comprises a substrate layer and one or more organic conjugates bound to the substrate in form of a self- assembled monolayer (SAM). The present invention relates as well to a novel conjugate used to form a self-assembled monolayer (SAM). The present invention relates also to methods for regenerating the biosensor chip, thus enabling the repeated use the biosensor chips.

Background Art

[0002] In the past decades, the biological and medical fields have discovered the great advantages in the use of biosensors and biochips capable of characterizing and quantifying (bio)molecules. Considerable attention has been drawn to functionalization of surfaces by forming ordered organic films of few nm to several hundred-nm thickness. Self-assembled monolayers (SAMs) are ordered molecular assemblies formed by the adsorption of amphiphilic, surfactant-type molecules on surfaces. SAMs provide one simple route to functionalize surfaces by organic molecules containing free anchor groups. The monolayer produced by self-assembly allows tremendous flexibility with respect to several applications depending upon their terminal functionality

(hydrophilic or hydrophobic control) or by varying the chain length (distance control).

[0003] The interaction of biotin and avidin or streptavidin has been exploited for use in many protein and nucleic acid detection and purification methods. Avidin-biotin binding is the strongest known non-covalent interaction between a protein and ligand. The bond formation between biotin and avidin is very rapid, and once formed, is unaffected by extremes in pH, temperature, organic solvents and other denaturing agents. These features of avidin make detecting or purifying biotin-labeled proteins or other molecules particularly useful for a number of biomedical applications. However, a disadvantage of the strong binding is that it is essentially irreversible under physiological conditions.

[0004] The known strategies for reversible immobilization of biotinylated probe molecules fall into two categories: (1 ) (strept)avidin remains permanently bound on the solid surface, or (2) (strept)avidin is removed from the surface together with the biotinylated sensor molecules. [0005] So far, strategy 1 has only been established in affinity chromatography but not in biosensing.

[0006] Several versions of strategy 2 have successfully been implemented in biosensing: The widely used carboxymethyl-dextran chip (CM5) has been derivatized with single-stranded DNA (ssDNA). A conjugate of streptavidin with complementary ssDNA is immobilized by double strand formation (BIAcore application note "Biotin CAPture Kit", GE data file 28-9577-47 AA). When desired, the DNA double helix is rapidly broken with diluted HCI and fresh streptavidin is immobilized. The method is inapplicable in presence of DNA-binding proteins and can only be used for the biosensors of one company.

[0007] The second example for strategy 2 is illustrated by Fig. 1 : an acid-sensitive mutant of avidin serves as bridge between a biotinylated chip and biotinylated sensor molecules. When desired, the avidin mutant M96H is rapidly dissociated into four denatured subunits by the combination of citric acid and SDS (Pollheimer et al., 2013). The problem of nonspecific protein and DNA adsorption towards avidin M96H was solved by additional mutations which lowered the isoelectric pH (pi) to 7 (Taskinen et al., 2014) but these caused increased sensitivity to acid, so that the "standard regeneration" step (pH 2.5) for antibody-antigen interaction (Fig. 1AB) caused already some undesired loss of biotinylated antibody (Pollheimer et al., 2013; Zauner et al., 2015). In conclusion, acid-sensitive avidin mutants with pl=7 do not provide for sufficiently stable immobilization of biotinylated sensor molecules, as needed in routine biosensing.

[0008] In the third version of strategy 2, desthiobiotin-functionalized sensing surfaces were used for reversible binding of streptavidin. Dissociation of streptavidin

(±biotinylated molecule on top of streptavidin) was induced by free biotin (0.1 - 2 mM). Unfortunately, the kinetics of streptavidin dissociation was slow and the degree of chip regeneration was between 70 and 90% (Knoll et al, 2000; Yoon et al. 2001 ) which is unacceptable for biosensing applications. Similarly, the recovery of (strept)avidin from desthiobiotin-agarose was between 70 and 90% (Hirsch et al., 2002), in spite of prolonged washing with free biotin (50 mM, 2x) and subsequently with HCI (50 mM). Moreover, the removal of streptavidin as such from the desthiobiotin surface is a necessary but insufficient criterion for chip regeneration: Statistically biotinylated proteins (such as antibodies) can crosslink adjacent streptavidin molecules on the chip surface, preventing chip regeneration, even if the interaction between streptavidin and the chip is abolished (Pollheimer et al., 2013; compare also Fig. 8C and 8E).

Theoretically, the latter problem can be solved by using only desthiobiotinylated sensor molecules. In practice, this is impossible because countless sensing molecules are commercially available in the biotinylated form but essentially none with a desthiobiotin label.

[0009] Currently, after each sample injection, the bound analyte molecule must be removed to allow for the next measurement with another sample ("standard

regeneration" in Fig. 1 ). However, it is not always possible to quantitatively dissociate the analytes without denaturing the immobilized sensor molecules (Goode et al., 2015). In such a situation it would be necessary to use a new chip for each individual measurement, which is unrealistic because of high material and labor cost. As exemplified in Fig. 1 , the favorable alternative is switchable immobilization of biotinylated sensor molecules. Hereby the biotinylated probe molecules must remain stably bound under the typical measurement conditions, yet easily removed and replaced by new ones on a time scale of minutes, if desired.

Summary of invention

[0010] It is the objective of the present invention to provide a biosensor comprising a self-assembled monolayer formed by a novel compound and a method for

regenerating the biosensor to enable repeated use.

[001 1] The objective is solved by the subject of the present invention.

[0012] According to the invention there are provided biosensor chips, biosensing methods and comparable heterogeneous assay formats in which biotinylated sensor molecules (e.g., antibodies) are stably immobilized on (strept)avidin-functionalized surfaces and cognate analyte molecules (e.g., antigens) are reversibly bound to the sensor molecules. The purpose is either detection and quantification of the analyte molecules in a liquid sample, or biological interaction analysis between sensor molecules and analyte molecules (Fig. 1AB).

[0013] According to the invention there is provided a conjugate of Formula I,

A— (CH2) _n — X— L— Y— D

wherein

A is a moiety which provides for stable anchoring to a solid surface;

n is 0 or an integer of 1 to 22, X is selected from the group consisting of ether, thioether, ester, amide, urethane, urea, hydrazone, oxime, acetal bond, and a triazole product formed from azide with alkyne,

L denotes a hydrophilic polymer,

Y is selected from the group consisting of ether, thioether, ester, amide, urethane, urea, hydrazone, oxime, acetal bond, and a triazole product formed from azide with alkyne, and

D denotes desthiobiotin or a derivative, and

the total linear chain of— (CH2)n— X— L— Y— comprises at least 27 backbone atoms.

[0014] According to the invention there is provided a sensor chip comprising a substrate and a self-assembled monolayer comprising a compound of Formula I.

[0015] According to the invention there is provided a method for detecting an analyte molecule in a sample, comprising the steps of:

(a) coating the sensor chip with a sensor molecule,

(b) contacting said sensor molecule with a sample whereupon a complementary analyte molecule binds to the sensor molecule; and

(c) detecting the bound analyte molecule.

[0016] According to the invention there is provided a method of regenerating a sensor chip, comprising the steps of:

(a) washing the sensor with a chaotropic agent,

(b) incubating the sensor with a proteolytic enzyme, and

(c) washing the sensor with an ionic surfactant.

Brief description of drawings

[0017] Fig. 1 is a schematic representation of the regeneration of a sensor chip.

[0018] Fig. 2A depicts various SAMs. "db" stands for desthiobiotin.

[0019] Fig. 2B depicts examples of suitable hydrophilic polymers.

[0020] Fig. 3 is a schematic representation of the binding of streptavidin to SAMs with different lengths of hydrophobic segments (tilted) and hydrophilic segments (vertical).

[0021] Fig. 4 is a graph reporting the denaturation temperature of horse heart metmyoglobin in EPPS buffer at various concentrations of biotin or urea.

[0022] Fig. 5 is a sensorgram measured by surface plasmon resonance (SPR), showing imperfect chip regeneration by pulses of 200 mM biotin (pH 8.0). The mixed SAM contained 20% component 9 and 80% component 8, as illustrated in Fig. 1 B. [0023] Fig. 6 is an SPR sensorgram showing perfect chip regeneration by sequential injection of guanidinium thiocyanate (GTC), pepsin, and SDS (two cycles). The mixed SAM was the same as in Fig. 5.

[0024] Fig. 7 is a schematic representation (panel A-D) and the corresponding SPR sensorgram (panel E) showing easy removal of streptavidin plus mono-biotinylated bovine serum albumin (BSA) by GTC without the need for a protease. The mixed SAM was the same as in Fig. 5.

[0025] Fig. 8 is a schematic representation (panel A-D) and the corresponding SPR sensorgram (panel E) showing the need for protease action if streptavidin is

crosslinked by a statistically biotinylated protein. The mixed SAM was the same as in Fig. 5.

[0026] Fig. 9 is a measured SPR sensorgram demonstrating much slower chip regeneration by all kinds of reagents (GTC, pepsin, biotin) if the mixed SAM (80% component 8) contains 20% component 10 instead of 20% component 9, i.e., if the tether between the SAM surface and desthiobiotin is much longer (Fig. 3D) than the minimal requirement (Fig. 3B).

[0027] Fig. 10 is a measured SPR sensorgram demonstrating very unstable binding of streptavidin if the fraction of component 9 is lowered to 1 % (as compared to 20% in Fig. 5-8), i.e., if most streptavidin molecules are bound to one desthiobiotin residue only (see Fig. 3C).

[0028] Fig. 1 1 is a measured SPR sensorgram proving absence of protein adsorption both before and after coating of the mixed SAM with streptavidin. The mixed SAM was the same as in Fig. 5.

[0029] Fig. 12 is a segment from a measured SPR sensorgram. Before t = 1800 s, streptavidin and biotinylated DNA probe had been injected in flow cell 2 (FC2) and then only streptavidin in FC1. The mixed SAM was the same as in Fig. 5.

[0030] Fig. 13 is a segment from a measured SPR sensorgram. Before t = 3600 s, streptavidin had been injected in both flow cells, biotin-BSA in FC2, and then biotin- p rote in G in FC1. The mixed SAM was the same as in Fig. 5.

[0031] Fig. 14 is a graph showing the data evaluation from Fig. 13 by the "double referencing method" (solid lines, obtained by subtraction of FC2 from FC1 and subtraction of the sample buffer difference curve from all other difference curves) as well as fitting of the data by the "bivalent analyte model" which assumes binding of one lgG2kappa by two adjacent molecules of biotin-protein G (dotted curves).

Description of embodiments

[0032] The present invention was guided by our hypothesis about the reason for slow and incomplete removal of streptavidin from desthiobiotin-surfaces. Streptavidin was not only retained by biospecific binding to immobile desthiobiotin residues but also by physisorption, the synergy between the two being the likely reason for slow

displacement by biotin. This would explain the slow and incomplete removal of streptavidin from the desthiobiotin-surfaces of Knoll et al., 2000, and Yoon et al., 2001. The situation in the study Knoll et al., 2000 is illustrated in Fig. 3A. Here the high fraction of component 1 was known to cause significant protein adsorption (Prime et al., 1991 and 1993). In the study of Yoon et al., 2001 , the highly cationic polymer PAMAM was the probable reason for the slow desorption of streptavidin from this sticky surface.

[0033] In the present invention, the intention was to invert the nonspecific effect of the surface from attraction to repulsion of streptavidin. The repulsion should be strong enough to help in the displacement of streptavidin by free biotin, while at the same time the repulsion should be small enough to allow for stable binding of streptavidin to the surface-linked desthiobiotin residues in absence of free biotin.

[0034] One way for creating a protein-repelling surface is formation of a self- assembled monolayer (SAM) with high protein resistance. SAMs which carry a dense brush of oligo(ethylene glycol) chains (OEG) are disclosed in Prime et al., 1991 and 1993, preferably with four ethylene oxide units or longer (Hahn et al. 2007, see molecule 4 in Fig. 2A). The same high protein resistance is achieved if the SAM components lack long alkyl chains and consists only of PEG chains with terminal thiols (Nileback et al., 201 1 , see molecules 6 and 7 in Fig. 2A). Interestingly, analogous SAMs with high protein resistance can be formed on H-terminated silicon, on acidic metal oxides, or on silanizable surfaces (glass, semiconductor oxide, metal oxide), respectively, if the thiol groups in 4, 6, and in homologous molecules are replaced by terminal vinyl groups (Booking et al., 2005; Yam et al., 2004), terminal

phosph(on)ate/catechol groups (Gnauck et al., 2007; Dalsin et al., 2005a and 2005b; Kotsokechagia et al., 2012), or terminal chloro/alkoxy-silane groups (Boozer et al., 2003) - or if dense PEG brushes are formed on pre-activated surfaces (Piehler et al. 2000). Dense bottle-brush brushes, and dense coatings with multi-arm PEGs (Cha et al., 2004) and cross-linked hydrogels (Groll et al., 2005) serve the same purpose, whereby the PEG chains can be replaced by poly(2-methyl/ethyl-2-oxazolines) (POX, Fig. 2B) (Zhang, 2000) or poly(hydroxyethyl methacrylate) (pHEMA, Fig. 2B) (Konradi et al., 2007). Bolduc et al. (2010) discovered that SAMs formed from the hydrophilic pentapeptide HHHDD (carrying a 3-mercaptopropionyl group on its N-terminus, see Fig. 2B) are highly protein-resistant, even in serum. High protein resistance was also found for SAMs consisting of peptoids prepared from N-substituted glycine (Fig. 2B, n=0) (Lau et al., 2012, R = methyl; Lau et al. 2015, R = 2-methoxyethyl) or from N- substituted beta-alanine (Fig. 2B, n=1 , R = methyl) (Lin et al., 201 1 ). All peptide and peptoidic SAMs were hydrophilic, with water contact angles <40°. For simplicity, all hydrophilic polymers in Fig. 2B were drawn with a thiol for anchoring to gold surfaces, although some of the SAMs had been published with other anchoring groups on other surfaces.

[0035] Mixed biotin SAMs (as depicted in Fig. 3A - 3D) are easily prepared on gold surfaces (Knoll et al., 2000; Jung et al., 2000; Pollheimer et al., 2013), due to the fact that alkanethiol derivatives can be packed at a very high lateral density (Harder et al. 1998). In contrast, few studies are known about comparable mixed SAMs on glass, semiconductor oxides, metal oxides, and H-terminated silicon or diamond, especially not about mixed biotin SAMs with molecular dimensions as in Fig. 3A - 3D. A major reason is that the anchoring groups on glass and metal oxides are larger than the tiny thiol groups which serve as anchors on gold surfaces. For compensation of lower anchor site density, longer hydrophilic polymer chains are typically grafted to the surface to create a protein-resistant monolayer on glass and related surfaces.

[0036] Unfortunately, the biotinylation of the long polymer chains on such brush-like surfaces bears the risk that all four binding sites of streptavidin are occupied by surface-linked biotin groups (Fig. 3E), rendering such a surface useless for subsequent binding of biotinylated probe molecules. For example, this phenomenon was observed when applying streptavidin to the biotinylated d extra n chip from XanTec Bioanalytics, even in case of the so-called two-dimensional d extra n chip where the d extra n chains were supposed to be short and linked to the surface via many attachment sites

(unpublished observation of the author). [0037] Several studies have shown that the problem of streptavidin occlusion (Fig. 3E) can be overcome if the biotinylated polymer chains are packed at such high lateral density that streptavidin can only bind to the surface of the polymer brush (Fig. 3F), reserving two binding sites for subsequent binding of biotinylated probes. A reliable method to this end is activation of glass with neat epoxysiiane, followed by solvent-less reaction with liquid PEG diamine (Piehler et al., 2000). In spite of their great length (MW-2000), the PEG chains are packed at such a great density that streptavidin can only bind to the surface of the biotinylated PEG brush (Biswas et al., 2014). The unusual reaction conditions with neat epoxysiiane, followed by solvent-less PEG diamine were chosen because pre-existing methods had proven unreliable to achieve comparably high lateral densities of the grafting sites.

[0038] In parallel, layers with very high packing densities of linear hydrophilic polymer chains were also achieved by surface attachment of brushes consisting of a polylysine backbone with many pendant linear PEG chains (Huang, et al., 2002). Again, streptavidin could bind with two out of its four binding sites to this surface (Fig. 3G). Extremely high packing densities were also achieved with bottle-brush brushes carrying linear POX side chains (Zhang, 2000) (Fig. 3G). As an alternative to comb- or brush-like copolymers, the linear hydrophilic polymer chains can also be connected to a single core in a star-like manner (Fig. 3H). Indeed, surfaces coated with such multi- arm PEGs turned out to be highly protein resistant (Cha et al., 2004; Groll et al., 2005). Biotinylation of the termini of the hydrophilic PEG chains resulted in only bivalent binding of streptavidin, as needed for further binding of biotinylated probe molecules (Heyes et al., 2007).

[0039] Biotin-SAMs on gold (Fig. 3A - 3D) must not carry biotin residues on more than 30% of the oligoethylene glycol (OEG) chains, for otherwise streptavidin is prevented from binding (Jung et al., 2000). No such limitation was found on surfaces coated with long PEG chains (Biswas et al., 2014; Heyes et al., 2007). Obviously the surface covered by one long PEG chain (e.g., MW-2000, n~50) is much larger than in SAMs with OEG chains, resulting in a significantly lower biotin density (Fig 3F - 3H). Moreover, the long, flexible polymer chains can easily adapt to the surface and the binding site geometry of bound streptavidin (Fig. 3F - 3H).

[0040] When functionalized with desthiobiotin, the surfaces sketched in Fig. 3F - 3H are anticipated to show a similar functional behavior as the optimized SAM in Fig. 3B: Streptavidin can bind bivalently, thereby compensating for the lower binding strength of desthiobiotin as compared to biotin. At the same time, the limited extensibility of individual polymer chains (Kienberger et al. 2000) beyond the interface of the densely packed layer pulls bound streptavidin towards the protein-repelling polymer surface. On desthiobiotin-SAMs (Fig. 3B), this antagonism was found essential to ensure stable binding of streptavidin under normal measuring conditions and rapid removal upon rinsing with 200 mM biotin and/or 6 M GTC.

[0041] In conclusion, the principles of regenerative chip preparation can be adapted from gold surfaces to glass and oxide surfaces if the thin SAM is replaced by densely packed layers of longer linear polymers which form such dense brushes that streptavidin can only bind two polymer-linked desthiobiotins.

[0042] The branching sites depicted in Fig. 3G and Fig. 3H do not concern the linear polymer chains as such but their attachment to a linear or circular core structure which is then immobilized on the solid surface (Groll et al., 2005; Heyes et al., 2007; Huang et al., 2002) or prepared on the surface (Zhang, 2000). Branching of the core structure serves to force the linear chains into a maximally packed state, where the linear polymer chains are arranged in a quasi-parallel fashion and where the biotinylated ends of the hydrophilic chains can reach the layer surface, as needed for binding of streptavidin. Attachment of the linear polymer chains to a core polymer (Fig. 3G) or a circular core (Fig. 3H) is only an option preferred by some workgroups to achieve high packing density but it is not intrinsically necessary (Fig. 3F), as shown by Piehler et al. (2000) and Biswas et al. (2014).

[0043] In a first aspect, a conjugate is provided having the structural Formula I,

A— (CH ₂ )n— X— L— Y— D

wherein

A is a moiety which provides for stable anchoring to a solid surface;

n is 0 or an integer of 1 to 22,

X is selected from the group consisting of ether, thioether, ester, amide, urethane, urea, hydrazone, oxime, and acetal bond, as well as the triazole product formed from azide with alkyne,

L denotes a hydrophilic polymer (including hydrophilic peptides and peptoids), Y is selected from the group consisting of an ether, thioether, ester, amide, urethane, urea, hydrazone, oxime, acetal bond, as well as the triazole product formed from azide with alkyne, and

D denotes desthiobiotin or a derivative thereof, and

the total linear chain of— (Ch n— X— L— Y— comprises at least 27 unbranched backbone atoms, with the proviso that D is not biotin.

[0044] As used herein, the term "unbranched" refers to a linear and straight backbone atom chain without any branches. However, the backbone atoms may optionally bear a substituent or a sidechain. In contrast, in a branched polymer a substituent, e.g., a hydrogen atom on a monomer subunit is replaced by another covalently bonded chain of that polymer. Examples of suitable polymers are depicted in Fig. 2B. Not suitable for the present inventions are polymers with densely branched structure and a large number of end groups such as dendrimers. In some embodiments of the invention the linear unbranched polymer chains may be attached to a core polymer or a circular core.

[0045] In one embodiment of the first aspect the total linear unbranched chain of — (CH ₂ )n— X— L— Y— comprises 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42 ,43, 44, 45, 50, 75, 100, 125, 150, 175 or up to 200 backbone atoms.

[0046] In one embodiment of the first aspect the total linear unbranched chain of — (CH ₂ )n— X— L— Y— comprises 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42 ,43, 44, 45, 50, 75, 100, 125, 150, 175 or up to 200 backbone atoms.

[0047] In one embodiment of the first aspect the total linear chain of— (CH2)n— X— L— Y— comprises 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42 or 43 backbone atoms.

[0048] In one embodiment of the first aspect the total linear chain of— (CH2)n— X— L— Y— comprises 39, 40, 41 or 42 backbone atoms.

[0049] In one embodiment of the first aspect the total linear chain of— (CH2)n— X— L— Y— comprises 41 backbone atoms.

[0050] In one embodiment of the first aspect the total linear unbranched chain of — (CH ₂ )n— X— L— Y— comprises 100, 125, 150, 175 or up to 200 backbone atoms.

[0051] In one embodiment of the first aspect A represents an organosulfur group, an organosilicon group, a fatty acid, a hydroxamic acid, a phosphonate or phosphate group, a catechol moiety (1 ,2-hydroxybenzene or 1 ,2,3-trihydroxybenzene), or a terminal vinyl group. [0052] The term "hydrophilic polymer" refers to any suitable polymer which preferably avoids non-specific adsorption of proteins and nucleic acids. Suitable hydrophilic polymer are for example, polyethylene glycol (PEG), polyoxazolines, polyethylene oxide) (PEO), polyvinyl pyrrolidone) (PVP), poly(methacrylic acid) (PMA), poly(acrylic acid) (PAA), poly(hydroxyethyl methacrylate) (pHEMA), polyvinyl alcohol) (PVA), peptides, peptoids, cellulose derivatives or a natural polymer such as d extra n.

[0053] In one embodiment of the first aspect the hydrophilic polymer represents poly(ethylene glycol), polyacrylamide, polyvinyl alcohol), hydroxylethylcellulose (HEC), poly(/V-hydroxyethyl acrylamide) (PHEA), hydroxylpropyl methylcellulose (HPMC), poly(2-hydroxyethyl methacrylate) (pHEMA), polyvinyl pyrrolidone), poly(acrylic acid), dextran, hyaluronic acid, poly (2-methacryloyloxyethyl phosphorylcholine) (PMPC), or poly(2-methyl/ethyl-2-oxazoline) (POX), hydrophilic peptides and hydrophilic peptoids.

[0054] As used herein the term "peptide" refers to a peptide which comprises at least 3 to 10 amino acids which may be the same or different. Preferably the peptide comprises 4, 5, 6, 7 or 8 amino acids.

[0055] In one embodiment of the first aspect L represents polyethylene glycol.

[0056] In one embodiment of the first aspect the polyethylene glycol comprises at least 3, 4, 5, 6, 7, 8, 9, 10, 1 1 or 12 ethylene oxide moieties.

[0057] In one embodiment of the first aspect the polyethylene glycol comprises 5, 6, 7, 8 or 9 ethylene oxide moieties.

[0058] In one embodiment of the first aspect the polyethylene glycol comprises 6, 7 or 8 ethylene oxide moieties.

[0059] In one embodiment of the first aspect X denotes an amide group.

[0060] In one embodiment of the first aspect Y denotes an amide group.

[0061] The term "derivative of desthiobiotin" refers to a compound which mimics the binding properties of desthiobiotin as closely as possible, being homologous to desthiobiotin by differing from desthiobiotin only by insertion of one or two Ch groups in an existing C-C or N-H bond, or by insertion/deletion of up to 4 Ch groups in the linear hydrocarbon chain, or by exchange of the terminal carboxyl group by an amino alcohol or thiol group. The term derivative of desthiobiotin does not encompass biotin.

[0062] In one embodiment of the first aspect the desthiobiotin derivative is of

Formula II, Formula II

wherein R ¹ , R ² and R ³ are independently from one another H, or a substituted or C1-3 alkyl, C2-3-alkenyl, C2-3-alkynyl, and

R ⁴ is -(CH ₂ ) _P -R ^a , wherein

p is an integer from 1 to 8, and

R ^a is selected from the group consisting of -COOH, -CO-NH-NH2 (hydrazide), -CHO (aldehyde), -NH-NH2, -O-NH2, -NH ₂ , -NH-CH3, -OH, -SH, or -N ₃ or alkyne (as used in click chemistry).

[0063] In a further embodiment, the desthiobiotin derivative is of Formula II, wherein R ¹ and R ² are H, R ³ is -CH ₃ , and R ⁴ is -(CH ₂ )5-COOCH ₃ .

[0064] In a further embodiment, the desthiobiotin derivative is of Formula II, wherein

R ² is H, R ¹ and R ³ are -CH ₃ , and R ⁴ is -(CH ₂ ) ₅ -COOH.

[0065] In a further embodiment, the desthiobiotin derivative is of Formula II, wherein R ¹ is -CH2CH3, R ² is H, R ³ is -CH ₃ , andR ⁴ is -(CH ₂ ) ₅ -COOH.

[0066] In a further embodiment the conjugate is

[0067] Most surfaces adsorb protein. Thus, the main challenge in controlling the interaction of proteins with surfaces lies in finding surfaces that resist nonspecific adsorption of proteins. Inert surfaces provide the background necessary for spatially restricting protein adsorption or for preparing surfaces that bind only specific proteins and are also used in the construction of biosensors. Thus, protein-non-adsorbing surfaces are important in the field of biosensing, especially in label-free methods where nonspecific adsorption cannot be discriminated from specific binding of anaiyte.

[0068] Streptavidin was found to be stably bound on protein-resistant SAMs where a fraction of the OEG chains carries biotin residues, provided that the biotin-terminated alkanethiol (BAT) is longer than the matrix alkanethiol (MAT) by at least 7 backbone atoms (Jung et al., 2000, see molecules 4 and 5 in Fig. 2A). In case of wild-type streptavidin, the protein-repelling effect of the OEG layer was not very evident.

Streptavidin mutants, however, were rapidly displaced by free biotin if the lateral density of biotin was low (BAT/MAT < 1/99). Interestingly, stable binding of these mutants on the mixed SAM was restored by using a higher lateral density of biotin (20- 30%). These experiments showed that reduced affinity of streptavidin mutants on protein-resistant SAMs can be compensated by the enhanced stability of bivalent binding on adjacent biotin residues.

[0069] In the above example, the affinity of streptavidin for a biotin-functionalized OEG-terminated SAM was lowered by mutations in streptavidin. This situation is analogous to the studies of Pollheimer et al., 2013 and Taskinen et al., 2014 where the affinity of avidin for a biotin-SAM was lowered by mutations of avidin. In the present invention, an opposite strategy for affinity reduction was chosen: the biotin residues on the OEG-terminated SAM were replaced by desthiobiotin, while sticking to wild-type streptavidin. To the best of our knowledge, no protein-resistant SAM or other protein- resistant surface has been functionalized with desthiobiotin before this study.

[0070] Our hypothesis on the fine-tuning of streptavidin affinity for a desthiobiotin- SAM is illustrated in Fig. 3. The alkyl chains of the SAM are depicted with the correct 60° tilt relative to the surface and the hyd rated OEG chains with their orthogonal orientation (Harder et al., 1998). As mentioned above, Fig. 3A corresponds to a mixed SAM formed from molecules 1 and 2 (Fig. 2A). The affinity is lower than on the analogous biotin-SAM (Knoll et al., 2000) but the linker segment between SAM and desthiobiotin is sufficiently long (12 atoms, indicated by Δ12 in Fig. 2A) to abolish steric strain and the dodecanol matrix is adsorptive towards proteins (Prime et al., 1991 and 1993), resulting in high overall stability of streptavidin binding and slow displacement by biotin. Fig. 3B applies to the mixed SAM formed from components 8 and 9 (Fig. 2A) at a molar ratio of 80/20. This mixed SAM is strictly analogous to our previously published biotin-SAM (Pollheimer et al., 2013), except that biotin has been replaced by desthiobiotin. Here, the linker between desthiobiotin and the SAM surface is shorter (Δ9, Fig. 2A), thus streptavidin is pulled towards the protein-repelling OEG layer. In spite of these adverse factors, it was anticipated that bivalent binding would ensure stable binding of streptavidin in absence of biotin. At 1 % surface density of

desthiobiotin, the same SAM was anticipated to exhibit monovalent binding of streptavidin (Fig. 3C), faster displacement of streptavidin by free biotin, and even spontaneous loss of streptavidin in absence of free biotin. In contrast, much slower displacement of streptavidin by free biotin was expected in case of a mixed SAM formed from components 8 and 10 (Fig. 2A, Δ21 ) at a molar ratio of 80/20, because here the desthiobiotin component is much longer than the matrix component, thereby reducing steric repulsion of streptavidin from the SAM surface (Fig. 3D) The shortest OEG derivative positively tested for high protein resistance is structure 3 in Fig. 2A (Prime et al., 1993). Probably a protein-resistant SAM can also be formed by the hypothetical structure 3b which lacks one methylene group in the alkyl chain of structure 3 and has a linker length of 20 atoms. When applying the rule of Jung et al. (2000), according to which the biotin component in a mixed SAM should be at least 7 atoms longer, then the corresponding desthiobiotin component should have a linker length of n = 27 atoms (structure 3c in Fig. 2A). This linker length n = 27 appears to be the lower limit for desthiobiotin components if they should be useful for the purpose of this invention.

[0071] In the below examples it is shown that all aspects of the hypothesis in Fig. 3 were confirmed. Moreover, the mixed SAM with 80% of component 8 (MAT) and 20% of component 9 (desthiobiotin-terminated alkanethiol, DBAT) turned out to be a preferred embodiment which provides for completely stable binding of streptavidin under all typical conditions of measurement and of "standard regeneration" (Fig. 1AB), while at the same time at least 99% of streptavidin plus the biotinylated sensor molecules are removed and replaced within few minutes when desired (Fig. 1 CD). This success, however, was only possible by using guanidinium thiocyanate, rather than free biotin, for displacement of streptavidin, and by proteolytic cleavage of the cross- linked layers formed from streptavidin and statistically biotinylated proteins (Fig. 8).

[0072] In a further aspect, the present invention encompasses a sensor chip comprising a substrate and a self-assembled monolayer comprising a conjugate of Formula I as described above.

[0073] In a further aspect, the present invention encompasses a sensor chip, wherein the substrate is selected from metal, semiconductor, metal oxides, semiconductor oxides, glass, metal and semiconductor surfaces primed with coordinating transition metal ions, hydrogen-terminated silicon, diamond, or plastic. [0074] In a further aspect, the present invention encompasses a sensor chip, wherein the substrate is selected from the group consisting of gold, silver, copper, lead, mercury, AI2O3, ΤΊΟ2, Nb2Os, Ta2Os, ITO, iron oxides, SnO2, S1O2, AgO, CuO.

[0075] In a further aspect, the present invention encompasses a sensor chip, wherein the substrate is gold.

[0076] In a further aspect, the present invention encompasses a sensor chip, wherein the self-assembled monolayer is a mixed self-assembled monolayer comprising a matrix component and a conjugate of Formula I as described above.

[0077] A further aspect of the invention relates to the sensor chip, wherein the matrix component (MAT) is a conjugate of Formula III,

A— (CH ₂ )n— X— L,

wherein A, n, X, and L are as defined above.

[0078] In one embodiment the total linear chain of— (Ch jn— X— L comprises at least 20 backbone atoms.

[0079] In one embodiment the total linear chain of— (Ch n— X— L comprises 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30,0 31 , 32, 33, 34, 35, 36, or 37 backbone atoms.

[0080] In one embodiment the total linear chain of— (CH2)n— X— L comprises 25, 26, 27, 28, 29, 30,0 31 , 32, 33, 34, or 35 backbone atoms.

[0081] In one embodiment the total linear chain of— (Ch n— X— L comprises 32 backbone atoms.

[0082] In one embodiment of the invention A is a sulfur-containing group, preferably a thiol group.

[0083] In one embodiment of the invention X is an amide group.

[0084] In one embodiment of the invention the matrix component is selected from the group consisting of

[0085] A further aspect of the invention relates to the sensor chip as described above, wherein the matrix component is at least 4, 5, 6, 7, 8, 9, or 10 backbone atoms shorter compared to the compound of Formula I. [0086] A further aspect of the invention relates to the sensor chip, wherein in said mixed self-assembled monolayer the ratio of compound of general Formula I to matrix component is in the range of about 3 to 50%, preferably in the range of about 5 und 40%, more preferred in the range of about 10 und 20%.

[0087] In one embodiment of the invention the mixed SAM is formed from

components 8 and 9 or 8 and 10 of Fig. 2A at a molar ratio of 80/20.

[0088] A further aspect of the invention relates to the sensor chip, wherein the self- assembled layer is functionalized by biospecific binding of a biomolecule.

[0089] In one embodiment of the invention the biosensor is functionalized for biospecific binding of biotin.

[0090] A further aspect of the invention relates to a sensor chip, wherein the functionalizing molecule is avidin, streptavidin, bradavidin, rhizavidin, or avidin-related proteins (AVRs) as well as mutants and/or hybrids and/or fusion proteins of said biotin- binding proteins.

[0091] A further embodiment of the invention relates to a method for detecting an analyte molecule in a sample, comprising the steps of:

(a) functionalizing the sensor chip with a biomolecule,

(b) contacting the sensor chip with a sample whereupon the sensor molecule binds to the biomolecule,

(b) contacting said sensor chip with a complementary analyte molecule which specifically binds to the sensor molecule;

(c) detecting the bound analyte molecule.

[0092] In one embodiment of the invention, the biomolecule is streptavidin or avidin.

[0093] A sensor molecule according to the invention comprises a biotin moiety and a target moiety. Examples of target moieties include synthetic molecules, nucleotides, nucleic acids, aptamers, peptide nucleic acids, peptides, proteins, enzymes, and antibodies. The term "antibody" as used herein encompassed monoclonal antibodies (mAbs), immunoglobulin (Ig) or immunoglobulin class G (IgG), heavy-chain antibodies (HcAb's), or fragments thereof such as fragment-antigen binding (Fab), Fd, single- chain variable fragment (scFv), or engineered variants thereof such as for example Fv dimers (diabodies), Fv trimers (triabodies), Fv tetramers, or minibodies, single-domain antibodies like VH or VHH or V-NAR, and darpins. [0094] The analyte molecule is complementary to the target moiety. Complementary according to the invention refers to molecules which specifically can bind to the target moiety. Thus, for an antibody as target moiety, the corresponding antigen is used as analyte molecule.

[0095] A further embodiment of the invention relates to the method for detecting an analyte molecule, wherein the sensor molecule comprises an antibody and the analyte molecule is an antigen.

[0096] In one embodiment of the invention the sensor molecule comprises an antigen and the analyte molecule is an antibody.

[0097] A further embodiment of the invention relates to the method for detecting an analyte molecule, wherein the self-assembled monolayer is coated with a biomolecule such as for example, avidin or streptavidin and the sensor molecule is labelled with biotin, desthiobiotin, a peptide (e.g. Streptag), or an aptamer. An aptamer according to the invention refers to an oligonucleotide or peptide that binds a specific target molecule. More specifically, aptamers can be classified as DNA, RNA or XNA aptamers usually consisting of short strands of oligonucleotides and peptide aptamers. Peptide aptamers consist of a short variable peptide domain, attached at both ends to a protein scaffold.

[0098] The analyte molecule may be tagged or labeled for detection such that a detectable signal is produced. Typically, the detectable signal is produced by any of the tags or labels known in the art. The terms "tag" or "label", as used herein, refers to any substance attachable to an analyte molecule, in which the substance is detectable by a detection method. Non-limiting examples of labels applicable to this invention include but are not limited to luminescent molecules, chemiluminescent molecules, fluorochromes, fluorescent quenching agents, colored molecules, radioisotopes, scintillants, massive labels (for detection via mass changes), biotin, avidin,

streptavidin, protein A, protein G, antibodies or fragments thereof, Grb2, polyhistidine, Ni ²⁺ complexes, Flag tags, myc tags, heavy metals, enzymes, alkaline phosphatase, peroxidase, luciferase, electron donors/acceptors (e.g., methylviologen or methylene blue), acridinium esters, labels for electrochemiluminescence (e.g., rubidium

complexes, quantum dots), and colorimetric substrates.

[0099] In a further aspect of the invention a method for regenerating a sensor chip is encompassed. The method comprises the steps of: (a) washing the sensor chip with a chaotropic agent,

(b) incubating the sensor chip with a proteolytic enzyme, and

(c) washing the sensor chip with an ionic surfactant.

[00100] A further aspect of the invention encompasses the method as described above, wherein the chaotropic agent is urea, guanidinium chloride, or guanidinium thiocyanate. The chaotropic agent may be used in combination with HCI or SDS.

Preferably the chaotropic agent is guanidinium thiocyanate.

[00101] The chaotropic agent is used in the range of 6 to 8 M, preferably 6 M.

[00102] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein the proteolytic enzyme is pepsin.

[00103] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein pepsin is used at low pH, preferably about pH 2 and in an amount of about 0.5 - 5 mg/ml, preferably 1 - 4 mg/ml, more preferred of about 2 mg/ml.

[00104] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein the ionic surfactant is an anionic surfactant, preferably sodium dodecyl sulfate.

[00105] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein the sensor chip is pretreated with 100 - 500 mM biotin, preferably with 150 - 300 mM, more preferred with about 200 mM.

[00106] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein at least 95%, preferably 97%, more preferably at least 98% of the self-assembled monolayer is regenerated. Regenerated according to the inventions means that the fraction of the non-covalently bound complex comprising the

biomolecule, the sensor molecule and the analyte molecule is removed by the regeneration procedure and the sensor chip can be reused for further cycles of sensor molecule immobilization and analyte detection. Thus, the self-assembled monolayer of the sensor chip remains functional.

[00107] A further aspect of the invention encompasses the method for regenerating a sensor chip, wherein said sensor chip is reused and again regenerated. In one embodiment of the invention, the sensor chip is regenerated at least 3 times, at least 10 times or at least 20 times. Due to the effective method for regenerating the sensor chip is still usable after several cycles of regeneration. Examples

The Examples which follow are set forth to aid in the understanding of the invention but are not intended to, and should not be construed to limit the scope of the invention in any way. The Examples do not include detailed descriptions of conventional methods, e.g., preparing of suitable substrates and/or self-assembled monolayers. Such methods are well known to those of ordinary skill in the art.

Example 1 - Synthesis of components 9 and 10

[00108] Components 9 and 10 (Fig. 2A) were synthesized by reacting the N- hydroxysuccinimide ester of 16-bromohexadecanoic acid with the appropriate NH2- PEG-NH-Boc (Polypure, Oslo, Norway) in chloroform and N,N-diisopropyl-N- ethylamine (DIEA).The product was purified by washing the chloroform solution with phosphoric acid, sodium carbonate, saturated NaCI, and by chromatography on Sephadex LH-20 (in chloroform). The Boc group was removed with trifluoroacetic acid (TFA) in dichloromethane and the N-hydroxysuccinimide ester of desthiobiotin was coupled to the deprotected amino group in DMF and DIEA. Excess of desthiobiotin NHS-ester was hydrolyzed in pyridine/water (9/1 ) at 30-35°C for 2 h to allow for easier removal of NHS and desthiobiotin by washing a chloroform solution of the crude product with 0.1 M phosphoric acid, with 10% sodium carbonate, and with saturated NaCI solution. The bromine atom was replaced by thioacetate in DMF/potassium thioacetate, the product was purified by chloroform/water extraction, and the acetylthio groups were hydrolyzed with K-tert-butoxide in methanol under strict exclusion of oxygen up to the moment of acidification with TFA. The reaction mixture was diluted with chloroform and washed with aqueous NaCI solution (-2.5 M), yielding the pure product in the chloroform layer. The reaction conditions and purification methods (extraction, gel filtration on Sephadex LH20 in chloroform) were closely similar as in the synthesis of the analogous biotin component (Pollheimer et al., 2013).

Example 2 - Preparation of mixed desthiobiotin-SA s on gold

[00109] The matrix component 8 (Fig. 2A, synthesized as described in Pollheimer et al., 2013) and the desthiobiotin components (9 or 10) were mixed from chloroform stock solutions at the desired molar ratio (80/20, except for Example No. 6 where the molar ratio was 99/1 ), dried, redissolved in THF, and treated with zinc/acetic acid, as described for the analogous mixed biotin-SAM components (Pollheimer et al., 2013). This step ensured a statistical distribution of the symmetric and mixed disulfides of 8 and 9 (or 8 and 10) after reoxidation in air and greatly improved the functional properties of the mixed SAM (Pollheimer et al, 2013). Pretreatment of gold and SAM formation was performed as described for the analogous biotin-SAM (Pollheimer et al., 2013).

Example 3 - Test for sensor chip regeneration by high concentrations of free biotin

[001 10] A 400 mM stock solution of free biotin was prepared by addition of Tris base to a final pH of 8.0. Mixed SAMs with 20% desthiobiotin component 9 were prepared as described in Example No. 2. Streptavidin was bound in both flow cells of the SPR biosensor (BIAcore X), and biotinylated goat IgG (biotin-lgG) was immobilized on top of streptavidin in flow cell 2 (FC2). Free biotin was injected at 0.1 mM, 1 mM, 10 mM concentration, as well as at 100 mM, 200 mM, and 400 mM concentration.

[001 1 1] It was found that >100 mM free biotin was required to remove most or all chip-bound protein molecules on a time scale of few minutes (data not shown, comparable data are found in Fig. 5). It was speculated that such high biotin

concentrations might have a two-fold effect: (i) the well-known competition for the biotin-binding sites of streptavidin which are occupied by the desthiobiotin residues of the chip and by the biotin residues of biotin-lgG, and (ii) a hitherto unreported activity of deprotonated biotin as a very strong denaturant. The latter seemed likely because of the amphiphilic nature of deprotonated biotin, as well as because of the urea segment of biotin. This hypothesis was tested by measuring the denaturation temperature (Tm) of an unrelated protein (horse heart metmyoglobin) at different concentrations of free biotin (in EPPS buffer, pH 8.0) and comparing it with the effect of normal urea (Fig. 4). Metmyoglobin was chosen because it allows for easy monitoring of denaturation in a UV-vis spectrophotometer: Denaturation causes release of the heme group, whereupon the molar extinction coefficient of the Soret band (ε ₄ 09 = 17100 M- cnrr ¹ ) drops by a factor of about 6 (Puett et al., 1973). The data in Fig. 4 confirmed that biotin is a much stronger denaturant than urea. According to the fit lines, 200 mM biotin causes the same change of the melting temperature (from 72 to 67°C) as 1.4 M urea. The two-fold role of concentrated biotin (as a competitor and as a general destabilizing agent) explains why biotin is able to not only remove streptavidin from the desthiobiotin surface but also the network formed between streptavidin and statistically biotinylated proteins (see data in Fig. 5 and sketch in Fig. 8C). [001 12] The relatively best results were obtained with 3-5 min injections of 200 mM biotin. Longer injection times or higher biotin concentrations seemed to cause less removal of protein. The likely explanation for this unexpected observation is that biotin inserts between the OEG chains of the mixed SAM, causing an increased resonance angle in SPR. As exemplified in Fig. 5, the bound material (protein and/or inserted biotin) can largely be removed by SDS (0.5%). However, the original baseline was never perfectly restored unless the sensor chip was washed with 6 M guanidinium thiocyanate (GTC, Fig. 5).

Example 4 - Test for sensor chip regeneration by strong denaturant and proteolytic digestion

[001 13] As shown in Fig. 6, perfect removal of streptavidin and biotinylated protein was achieved when the sensor chip was treated with two cycles of the injection sequence GTC (6 M), pepsin (2 mg/ml in 100 mM glycine/HCI, pH 2.5) and SDS (0.5%). This order was more efficient than injecting pepsin before GTC. Obviously the denaturation of the biotinylated proteins (protein G in FC1 , solid line, IgG in FC2, dashed line) by GTC greatly increases their sensitivity to pepsin. When tested under comparable conditions, the inclusion of 4 mM tris(2-carboxyethyl)phosphine (TCEP) appeared to enhance the effect of 6 M GTC. Nevertheless, TCEP -free GTC was much preferred because TCEP is sensitive to oxidation by air, thus (immediately before the injection) GTC must be mixed with acidic TCEP hydrochloride and with the appropriate amount of a base, and this step cannot be performed by standard autosamplers.

Example 5 - Mechanistic analysis of the effects of denaturant and protease

[001 14] The mechanistic roles of GTC and pepsin were revealed by comparison of Fig. 7 with Fig. 8. In both figures, flow cell 1 was loaded with streptavidin (FC1 , solid trace). Here, GTC was always able to remove all bound streptavidin before pepsin and SDS were injected. The difference between Fig. 7 and 8 concerns the type of biotinylated protein which was bound on top of streptavidin in flow cell 2 (FC2, dashed trace). In Fig. 7, we used "mono-biotin-BSA" which carried a single biotin label per protein (see sketch in Fig. 7C). This protein had been prepared by labeling the single free cysteine of BSA with maleimide-PEG2-biotin (Pierce). In Fig. 8, however, we used statistically biotinylated BSA which formed a cross-linked network on top of streptavidin (see sketch in Fig. 8C). Interestingly, GTC was able to completely remove the double layer of streptavidin and mono-biotin-BSA even before injection of pepsin (dashed trace in Fig. 7E). This was explained by the fact that mono-biotin-BSA cannot crosslink adjacent streptavidin molecules (see sketch in Fig. 7D). In contrast, GTC had almost no effect on the double layer of streptavidin and statistically biotinylated BSA but these proteins were easily removed in the subsequent injection of pepsin (dashed trace in Fig. 8E). Taking into account the well-known resistance of streptavidin to all kinds of proteases, the latter finding demonstrates that pepsin acts by cleavage of the biotinylated protein (see sketch in Fig. 8D).

[001 15] The data in Fig. 7 and 8 indicate that GTC does not dissociate streptavidin into its four subunits, for otherwise no pepsin would be required to decompose the cross-linked protein double layer in Fig. 8. The data also show that GTC can only extract desthiobiotin residues but not ordinary biotin residues from streptavidin, for otherwise the network with statistically biotinylated BSA would also have been dissociated by GTC alone.

Example 6 - Mixed SAM with a longer linker between the SAM surface and

desthiobiotin

[001 16] In this study, we prepared the two kinds of mixed SAMs. The matrix component was always compound 8 but the desthiobiotin component was either compound 9 or compound 10 (see Fig. 2A). The purpose was to show that the shorter desthiobiotin component 9 pulls bivalently bound streptavidin very close to the protein- repelling SAM surface (Fig. 3B), causing acceleration of streptavidin removal by free biotin or by GTC. The longer desthiobiotin component 10, however, would allow for more relaxed bivalent binding of streptavidin (Fig. 3D), where repulsion between the OEG layer and streptavidin should be absent and the removal of streptavidin by GTC or free biotin should take more time.

[001 17] These expectations proved correct, as shown by the comparison of Fig. 8 (short component 9) with Fig. 9 (long component 10). The first injection of GTC removed 100% of bound streptavidin in Fig. 8 but only 92% in Fig. 9 (see solid traces). Even more pronounced was the difference with respect to the cross-linked double layer of streptavidin and statistically biotinylated BSA (dashed traces). Here the first injection of pepsin removed almost 93% of the proteins in Fig. 8 but only 68% in Fig. 9. As can be seen from Fig. 9, many additional injections of the different regeneration reagents were required until the desthiobiotin-SAM with the longer tether was fully regenerated by removal of all bound proteins, in contrast to all SAMs with shorter tethers (Fig. 5 and 6). These findings confirmed our concept that the short tether causes steric strain and repulsion from the SAM when streptavidin is bivalently bound via short tethers (Fig. 3B), while longer tethers cause little or no strain and repulsion (Fig. 3D).

[001 18] In conclusion, fast and complete regeneration of a desthiobiotin sensor chip is only possible if the SAM is protein-resistant, if the tether is short enough to cause steric strain and repulsive interaction between the SAM and streptavidin, and if the proper reagents are chosen and injected in the right order.

Example 7 - Comparison of bivalent and monovalent binding of streptavidin on the

SAM

[001 19] In the above examples, steric strain and protein repulsion (as in Fig. 3B) were shown to be of great help when streptavidin was to be removed by GTC or biotin. Fortunately, this partial destabilization of streptavidin binding was only apparent when applying GTC or biotin but no in their absence (see Fig. 5 - 9). The following example demonstrates that the reason was bivalent binding of streptavidin on mixed SAMs with 20% desthiobiotin density. We prepared a mixed SAM with 1 % of the shorter component 9 and 99% of the matrix component 8. The functional consequences are shown in Fig. 10. Now a large fraction of the initially bound streptavidin molecules spontaneously dissociate within few minutes (solid and dashed trace). Removal of the remaining streptavidin (± biotin-BSA on top) by GTC is also much easier than on mixed SAMs with 20% desthiobiotin component 9 (compare Fig. 8). By analogy to the biotin- SAMs of Jung et al. (2000), the reason for such weak affinity must be monovalent binding, as illustrated in Fig. 3C.

[00120] In conclusion, the mixed SAM of components 8 and 9 in an 80/20 ratio (Fig. 3B) provides for an optimal situation where strain and repulsion of streptavidin facilitate its removal by GTC, pepsin, and biotin. At the same time, bivalent binding is able to ensure stable binding of streptavidin in absence of these reagents, in spite of strain and protein repulsion.

Example 8 - Protein resistance of a chip-bound streptavidin monolayer

[00121] In the above examples, the protein resistance of the SAM itself was seen as a helpful parameter which enhances the efficiency of GTC and biotin with respect to streptavidin desorption. In biosensing applications it is necessary that high protein resistance is also observed after immobilization of streptavidin. Fig. 1 1 shows that this is perfectly true when SAMs are formed with the 80/20 mixture of components 8 and 9. The test protocol was the same as in Pollheimer et al. (2013) and Taskinen et al.

(2014). Only negligible amounts of lysozyme, BSA, and IgG were bound when injecting these proteins at high concentration (1 mg/ml).

Example 9 - Absence of DNA adsorption and high functionality of immobilized DNA

[00122] Fig. 12 demonstrates that nonspecific DNA adsorption to sensor chip-bound streptavidin is below the detection limit. The figure shows only the essential part of the whole experiment which followed the same protocol as in Taskinen et al. (2014). A streptavidin monolayer was formed in FC2 and functionalized with biotin-labeled DNA probe (a 30 mer with the same sequence as in Taskinen et al. (2014)). Then a streptavidin monolayer was formed in FC1 and biotin-BSA was run over both flow cells (the first injection plotted in Fig. 12). After injection of unlabeled BSA, the same DNA strand as before was injected, except that it lacked the biotin label. No binding was detected in any of the two flow cells, proving absence of nonspecific adsorption.

Injection of the complementary DNA strand also gave no adsorption in FC1 which contained only biotin-BSA. In FC2, however, pronounced hybridization to the

complementary, previously immobilized biotin-DNA was observed, indicating good accessibility of immobilized biotin-DNA for analyte binding.

Example 10 - Biological Interaction Analysis of biotin-protein G and goat IgG

[00123] Immobilized biotin-protein G and soluble human lgG2i constitute a good test system for Biological Interaction Analysis in a biosensor (Pollheimer et al., 2013;

Zauner et al., 2015). lgG2K was found to act as bivalent analyte which simultaneously binds to two adjacent biotin-protein G molecules on the chip surface. The same experiment was here performed on the usual mixed desthiobiotin-SAM (80/20 mixture of components 8 and 9), streptavidin was bound in both flow cells, FC2 was blocked with biotin-BSA, and FC1 was functionalized with biotin-protein G (outside the plot range of Fig. 13). Then, a dummy injection of glycine (100 mM, pH 2.7) was applied (the first injection shown in Fig. 13) and found to have no effect in any flow cell. This shows the favorable stability of the wild-type streptavidin monolayer on the

desthiobiotin-SAM in Fig. 13, as compared to avidin mutant M96H on the biotin-SAM in Pollheimer et al. (2013). Then, different concentrations of lgG2i were injected in sample buffer (1 μΜ BSA in the running buffer) and always removed with glycine (pH 2.7). Double referencing was performed as described (Pollheimer et al., 2013), resulting in an overlay of the corrected binding curves (solid traces in Fig. 14). These curves could well be fitted by a single set of kinetic constants when using the bivalent analyte model (dotted traces in Fig. 14), while the same discrepancy between measured and calculated binding curves as in Poiiheimer et al. (2013) was observed when using the simple Langmuir model which assumes 1 : 1 binding between protein G and lgG2ic. The kinetic constants and equilibrium constants were in good agreement with those obtained before (Poiiheimer et al., 2013).

[00124] The enhanced stability of the wild-type streptavidin monolayer at pH 2.7 was the first advantage over avidin mutant monolayers (Poiiheimer et al. 2013; Taskinen et al. 2014; Zauner et al., 2015). The second advantage was lower nonspecific binding of protein and DNA on streptavidin (exemplified in Fig. 1 1 and 12) as compared to

"neutralized avidin M96H" or "switchavidin" (Taskinen et al., 2014, Zauner et al., 2015), and the third was a higher capacity for binding of biotinylated sensor molecules and analytes, typically by about 50%.

Literature References

BIAcore application note: "Biotin CAPture kit", GE data file 28-9577-47 AA.

Biswas et al. (2014) Soft Matter 10, 2341 -2345.

Booking et al. (2005) Langmuir 21 , 10522-10529.

Bolduc et al. (2010) Anal. Chem. 82, 3699-3706

Boozer et al. (2003) Sensors and Actuators B: 90, 22-30.

Cha et al. (2004) Proteomics 4, 10965-10976.

Dalsin et al. (2005a) Langmuir 21 , 640-646.

Dalsin et al. (2005b) Materials Today 8, 38-46.

Ding et al. (1999) Bioconjugate Chem. 10, 395-400.

Ding et al. (2001 ) Nature 41 1 , 59-62.

Gnauck et al. (2007) Langmuir 23, 377-381.

Goode et al. (2015) Langmuir 31 , 6267-6276.

Groll et al. (2005) Langmuir 21 , 1991 -1999.

Harder et al. (1998) J. Phys. Chem. B 102, 426-436.

Hahn et al. (2007) Chemical Monthly 138, 245-252.

Heyes et al. (2007) Mol. Biosyst. 3, 419-430.

Hirsch et al. (2002) Anal. Biochem. 308, 343-357.

Holmberg et al. (2005) Electrophoresis 26, 501 -510. Huang et al. (2002) Langmuir 18, 220-230.

Jung et al. (2000) Langmuir 16, 9421 -9432.

Kienberger et al. (2000) Single Molecules 1 , 123-128.

Knoll et al. (2000) Coll. Surf. A 161 , 1 15-137.

Konradi et al (2006) Langmuir 22, 8571 -8575.

Kotsokechagia et al. (2012) Langmuir 28, 1 1490-1501.

Kulkarni et al. (2004) Bioconjugate Chem. 15, 747-753.

Lau, K. H. A. et al. (2012) Langmuir 28, 16099-16107.

Lau, K. H. A. et al. (2015) Adv. Mater. Interfaces 2, 1400225.

Lin, S. et al. (201 1 ) Biomacromolecules 12, 2573-2582.

Morag et al. (1996a) Biochem. J. 316, 193-199.

Morag et al. (1996b) Anal. Biochem. 243, 257-263.

Muller et al. (1993) Science 262, 1706-1708.

Nednoor et al. (2005) Chem. Mater. 17, 3595-3599.

Nileback et al. (201 1 ) Biosens. Bioelectron. 28, 407-413.

Nordlund et al. (2003) FEBS Lett 555, 449-454.

Piehler et al. (2000) Biosens. Bioelectron. 15. 473-481.

Poliheimer et al. (2013) Bioconjugate. Chem. 24, 1656-1668.

Prime et al. (1991 ) Science 252, 1 164-1 167.

Prime et al. (1993) J. Am. Chem. Soc. 1 15, 10714-10721.

Puett et al. (1973) J. Biol. Chem. 248, 4623-4634.

Taskinen et al. (2014) Bioconjugate. Chem. 25, 2233-2243.

Yam et al. (2004) Chem. Commun. 2510-251 1.

Yoon et al. (2001 ) Langmuir 17, 1234-1239.

Zauner, Dominik (2015) Sensors and Actuators B 229, 646-654.

Zhang, Ning (2000) Ph.D. thesis at the Technical University of Munich, Germany.