SPR investigations of the formation of intermediate layer of the immunosensor bioselective element based on the recombinant Staphylococcal protein A

A. E. Rachkov1, A. O. Bakhmachuk1, 2, O. B. Gorbatiuk1, M. J. Matsishin1, 3, R. V. Khristosenko4, Iu. V. Ushenin4, A. P. Soldatkin1, 3 1 Institute of Molecular Biology and Genetics, NAS of Ukraine, 150, Akademika Zabolotnoho Str., Kyiv, Ukraine, 03680 2 Educational and Scientifi c Center «Institute of Biology», Taras Shevchenko National University of Kyiv, 64/13, Volodymyrska Str., Kyiv, Ukraine, 01601 3 Institute of High Technologies, Taras Shevchenko National University of Kyiv, 2, korp.5, Pr. Akademika Hlushkova, Kyiv, Ukraine, 03022 4 V. Ye. Lashkaryov Institute of Semiconductor Physics, NAS of Ukraine, 41, Prospect Nauki, Kyiv, Ukraine, 03028 oleksandr_rachkov@yahoo.com


Introduction
The development of approaches and tools of analytical biotechnology can help signifi cantly to upgrade the quality of human life by improving the methods of biomedical diagnostics, molecular engineering, environmental monitoring, food analysis, drug discov-ery, etc.As compared with existing standard analytical approaches, the biosensor methods of analysis have a number of advantages: they provide easy, fast, accurate, highly sensitive, specifi c, and cheap procedure of the measurement.Besides, the real-time measurements are possible, while only a minimal probe pretreatment is necessary.A biosensor is a self-con-tained device consisting of two functional parts.A bioselective element is in direct contact with a physical transducer, which transforms the information from biorecognition domain into an electrical or an optical signal [1].
The use of the variety of immune components is very attractive in this area of research.The immunosensors seem to be promising and helpful technique, offering a high specifi city through the use of immune molecules, simple operation, uncomplicated sample preparation and high sensitivity [2].The effective immobilization of either an antibody or an antigen on the sensing surface with preserving their functional activity is an important step in achieving a highly reliable immunosensor.This immobilization step affects the sensitivity and specifi city of the sensor.However, during the immobilization of antibodies their antigen-binding activity is usually much lower in comparison with the same antibody activity in a free state due to the low density of the antigenbinding sites.The main reasons for this are believed to be a random orientation of antibodies on the sensor surface and a steric hindrance caused by the infl uence of the surface of a solid substrate.
Although physical adsorption does not require the use of several materials and complex reactions, it also leads to the serious drawbacks, such as the protein denaturation, very low stability and random orientation of the proteins.So they experience conformational changes that make their functional sites inactive or inaccessible, leading to the complicated interaction with the analyte [3,4].The covalent immobilization by coupling IgG to a chemically activated matrix ensures much more reliable attachment to a sensor surface.However, this method has the disadvantage that is a heterogeneous immobilization and, therefore, some of the attachment sites interfere with the protein functional regions.
Regardless of physical adsorption or covalent binding, the immunoglobulin molecules are arranged on the substrate surface in a disordered manner, part of the active sites is covered due to the direct contact with the substrate, resulting in the loss of the recognition combining ability of the target molecule [5].A specifi c orientation (or an oriented immobilization) is a promising alternative to physical adsorption or covalent binding.It can be accomplished by using an intermediate molecules directly attached to the sensor surface.One of the best candidates for the role of the intermediate molecule is immunoglobulin-binding Staphylococcal protein A (SPA).SPA selectively binds the Fc-domain of antibody leaving the Fabregion available to detect an antigen.The SPA molecule structure includes a signaling sequence [6], a IgG-binding region consisting of fi ve highly homologous domains, and a C-terminal anchoring part, which attaches the protein to the bacterial cell wall [7,8].SPA molecules are highly resistant to the denaturing factors: they are thermostable, resistant to a wide range of pH (1-12), and are not destroyed by trypsin cleavage [9].The fi rst use of SPA for the immunosensing applications was reported by Muramatsu et al. [10].They designed a piezoelectric biosensor for the human IgG detection by SPA covalently immobilized via 3-aminopropyltriethoxysilane and glutaraldehyde.Prusak-Sochaczewski and Luong proposed an idea of using SPA as an intermediate protein for immobilization of the antibodies for the human serum albumin detection.In that case, SPA was physically adsorbed onto the gold electrodes under conditions close to the isoelectric point of the protein A (pH 5.5) [11].
However, even if the antibodies are not randomly connected to the surface coated with SPA, the protein A itself meets the same problem of the random immobilization [12].An alternative method for immobilizing proteins is to use a genetic engineering approach, in which a specifi c attachment site is selectively introduced into a non-essential part of the recombinant protein.It was shown that a single cysteine residue introduced into the C-terminal part of the recombinant IgG-binding molecules can be used for their immobilization on a thiol-containing solid matrix [13].It was also shown that the recombinant protein A with a specially introduced cysteine residue increases the reliability of protein immobilization via thiol adsorption on the gold sensor surface [12].In our previous work [14] the original plasmid pET24-SPA-6HisCys was constructed, the recombinant protein A contained all fi ve IgG-binding doma-ins, 6His-tag and C-terminal cysteine residue (SPA-Cys) was obtained and its immobilization on a gold sensor surface was demonstrated.
The optical biosensors based on the surface plasmon resonance (SPR) allow direct registration in real time a wide range of intermolecular interactions without any labels [15].Unlike the well-known, but extremely expensive bench-top SPR spectrometers «Biacore», SPR spectrometers «Plasmon» developed at V. Ye.Lashkaryov Institute of Semiconductor Physics of National Academy of Sciences of Ukraine are small, rather simple in operation, and much cheaper devices [16,17].Therefore, the aim of this work was to investigate the formation of intermediate layer of the immunosensor bioselective element based on the recombinant protein A from Staphylococcus aureus with cysteine residue (SPA-Cys), and its interactions with human IgG using the SPR spectrometer «Plasmon».
Human IgG was purifi ed by affi nity chromatography as described in [14].The genetically engineered fusion protein containing fi ve IgG-binding domains of SPA and the bacterial alkaline phosphatase with enhanced catalytic properties (SPA-BAPmut) was obtained as described in [18].

Synthesis and purifi cation of SPA-Cys
The recombinant Staphylococcal protein A with cyste ine residue (SPA-Cys) was synthesized and purifi ed as described earlier [14].Briefl y, the DNA seque nces encoding the IgG-binding region of SPA, His-tag and cysteine were genetically fused and expressed in E. coli in the soluble form using a modifi ed auto-induction protocol.The purifi cation of the target protein was performed by immobilized-metal affi nity chromatography in the native conditions.SPA-Cys was obtained in a functionally active state with purity ~95 %.

Competitive ELISA
The wells of the immunological plate were coated overnight at 4 °C with 100 μL 4 μg/ml IgG in carbonate buffer, pH 9, and then washed with PBS containing 0.1 % Tween 20 (PBST).The surface of the wells was blocked to prevent nonspecifi c sorption with 150 μl 2 mg/ml of skim milk powder solution in PBS.After 1 hour incubation at 37 °C the plates were washed with PBST.The mixtures of 50 μl 4 μg/ml fusion protein containing the IgG-binding domains of SPA and the bacterial alkaline phosphatase with enhanced catalytic properties (SPA-BAPmut) with 50 μl PBS or with 50 μl SPA-Cys solutions of different concentrations were added to the wells.After 1 hour incubation at 37 °C the plates were washed with PBST and phosphatase buffer (0.1 M Tris-HCl, pH 9.5; 0.14 M NaCl, 0.015 M MgSO 4 ), and 100 μl of substrate solution (p-nitrophenyl phosphate in phosphatase buffer) was added.After incubation for 30 min at room temperature, color development was stopped with 1 M NaOH (50 μL per well) and the absorbance was measured at 405 nm using the micro plate reader «Titertek Multiskan MCC/340» (Germany).

SPR spectrometric analysis of protein-protein interactions
The glass plates with a thin layer of gold were cleaned and mounted on the spectrometer prism as described in [19].SPR analysis was performed by using a measuring fl ow-cell of the spectrometer «Plasmon-4m» and the peristaltic pump «Ismatec» (the pump speed of ~40 μl/min).At fi rst the measuring fl ow-cell was thoroughly washed by working buffer solution (PBS) to stabilize the SPR signal.Then a sample (usually 120 μl) was injected and incubated with the pump switched off for 30 min for the protein immobilization on the sensor surface, and for 10 min for the interactions of immobilized components with their molecules-partners.After that the measuring fl ow-cell was washed by PBS again until there was a stable SPR signal.To distinguish an actual sen sor response caused by the interactions between a sample and the sensor surface or preliminary immobilized components from the signal caused by the random fl uctuations of medium refractive index, it is necessary to wash the fl ow-cell before and after each sample by the same buffer solution, and only then to determine a value of the SPR response.

SPA-Сys/IgG binding constant determination
To analyze the binding SPA-Cys and human IgG the classic Langmuir model of heterogeneous binding was used.In general, the interactions between immobilized SPA-Cys and IgG is described by where [SPA] is the number of available IgG-binding sites of the immobilized SPA-Cys, [IgG] is immunoglobulin concentration in the solution, [AG] is the number of complexes formed between the immobilized SPA-Cys and IgG.The process is characterized by equilibrium dissociation constant K d .Langmuir equation of the curve that describes this process is written as follows: Its linearized form, called the Skatchard equation, permits to get simply and clearly a value of the equilibrium constants and a value of the maximum number of SPA-Cys/IgG complexes: In order to obtain the characteristic parameters one can use the nonlinear least squares for the basic Langmuir equation or the least squares method for the Skatchard equation.

Results and Discussion
Before applying the immune components for the biosensor analysis, their activity was tested using the competitive format of ELISA.The SPA-Cys solutions of different concentrations and a fi xed concentration of the fusion protein containing IgG-binding domains of SPA and the bacterial alkaline phosphatase with enhanced catalytic properties (SPA-BAPmut) were added to the immobilized human IgG.If SPA-Cys actively binds to the immobilized IgG and competes with SPA-BAPmut, the results of ELISA will be in an inverse proportion to the concentration of SPA-Cys.The results shown in Fig. 1 confi rmed our expectations: the absorbance values while adding SPA-Cys and SPA-BAPmut are signifi cantly lower than the control values (only SPA-BAPmut) and are in the inverse proportion to the concentration of SPA-Cys.We can conclude that SPA-Cys interacts specifi cally with the immobilized IgG and demonstrates a suffi ciently high activity.
In the previous work [14], a possibility of the successful immobilization of the recombinant SPA-Cys on the gold sensor surface of SPR spectrometer «Plasmon-4m» was demonstrated after the injection of a sample of the purifi ed SPA-Cys into the measuring fl ow-cell and its incubation there.In the present work, the dependence of the immobilization level on the concentration of SPA-Cys is investigated.A linear part of this dependence was observed in the range from 0 to 0.5 μM SPA-Cys while 2 μM SPA-Cys showed a close-to-saturation level of immobilization (Fig. 2).
According to the conversion factor of the SPR response into the value of surface density of the immobilized protein [20], this value when using 2 μM SPA-Cys was 1.1 ± 0.2 ng/mm 2 .Given the molecular weight of SPA-Cys (34.5 kDa), we can calculate that in average approximately 51 nm 2 of the sensor surface falls on one molecule of the immobilized SPA-Cys.It means that a lot of free space (the potential sites for the nonspecifi c adsorption during the next stages of the sensor work) is left.In this regard, the need to block effi ciently the nonspecifi c adsorption on these free sites becomes even more obvious.
An attempt to apply BSA as a blocking agent did not give a satisfactory result: the saturation of the sensor response was achieved after 3 consecutive injections of 0.2 mg/ml BSA into the measuring fl ow-cell, and its growth (~ 0.08 angular degrees) corresponds to ~ 0.8 ng/mm 2 only.Even taking into consideration a bigger size of BSA molecules, the average value 41 nm 2 of the sensor surface per each immobilized protein molecules (SPA-Cys and BSA) indicates a quite large area of the free sensor surface.

AG AG AG
Unlike BSA, the fi rst injection of 0.2 mg/ml skim milk powder into the measuring fl ow-cell caused a very big sensor response.Washing the measuring cell by PBS led to a moderate signal reduction (refl ecting a relatively small portion of weakly adsorbed milk proteins compared with strongly immobilized ones).Each following injections gave gradually decreasing increment of sensor response, and four consecutive injections of 0.2 mg/ml skim milk powder were necessary to approach to the saturation of the sensor response (~0.35 angular degrees).It corresponds to protein surface density ~3.5 ± 0.4 ng/mm 2 .In average, approximately 12 nm 2 of the sensor surface falls on one molecule of the immobilized SPA-Cys and milk proteins.These values are in good agreement with parameters of the protein monolayer.
To check, whether the immobilized molecules of SPA-Cys retain their immunoglobulin-binding activity, human IgG was injected into the measuring fl ow-cell, the sensor surface which was previously treated by 1 μM SPA-Cys and free sites of the chip were blocked by milk proteins.The sensogram 1 in Fig. 3 shows that three consecutive injections of 10, 20 and 40 μg/ml IgG cause a signifi cant sensor response that assumes its active interactions with the immobilized SPA-Cys.
However, how much this result refl ects the interactions only between IgG and the immobilized SPA-Cys, and to what extent the nonspecifi c interactions with milk proteins or with bare portions of the gold surface are involved?To answer this question, the same IgG samples were injected into the measuring fl ow-cell, the sensor surface of which was treated by only milk proteins (without SPA-Cys).As seen in the sensogram 2 in Fig. 3 the IgG injections in the absence of SPA-Cys cause almost no sensor response.Thus, it is clear that IgG does interact with SPA-Cys, and does not with milk proteins.Furthermore, the sensogram 2 in Fig. 3 demonstrates a high effi ciency of the applied blocking procedure.Otherwise, IgG (in the absence of the immobilized SPA-Cys) would cause a noticeable sensor response, as it could be adsorbed on the unblocked sites of the gold surface.
For a further use of the prepared bioselective element, an effi cient regeneration procedure should be applied.In fact, to this end, researchers use various solvents or reagents that change the pH and ionic strength of the solution, acting on a charge of interacting molecules and thus their tertiary structure [21].In this case, after treating the sensor surface with a solution of 40 mM sodium citrate buffer (pH 2.5) [22] the level of sensor signals almost backs to the values that preceded the IgG injections (Fig. 3).It shows a quite effective disruption of the links between the immobilized SPA-Cys and IgG and the removal of the latter.The subsequent injections of new IgG samples showed that such regeneration procedure has not es-sentially affected the level of immunoglobulin-binding activity of the immobilized SPA-Cys.Thus, the re-use of the immunosensor bioselective element formed on the basis of SPA-Cys is possible.Does a level of the immobilized SPA-Cys infl uence on a value of the sensor response at IgG injections?To answer this question we conducted the following experiment with four bioselective elements obtained at different SPA-Cys concentrations.As shown in Fig. 4 the values of the SPR response are directly dependent on the surface density of immobilized SPA-Cys and on the IgG concentration, at least in the range from 10 to 40 μg/ml.
To investigate the effect of storage time of the immunosensor bioselective element, formed on the basis of SPA-Cys, on its properties, the IgG samples of the same concentration (10 μg/ml) were injected within some period of time into the measuring fl ow-cell, the sensor surface of which was previously treated by 1 μM SPA-Cys and was blocked by milk proteins.A gradual decrease in the sensor response was observed, however, the fi rst ten days the immunosensor bioselective element kept a relatively high immunoglobulin-binding activity.Note that this experiment was conducted at a rather high ambient temperature (25-27 °C) without the use of any preservatives.So, there is some opportunity for improving the achieved level.
The injection of the IgG samples of various concentrations (from 0.5 to 100 μg/ml or from 3 to 670 nM) into the measuring fl ow-cell with the sensor previously treated by 1 μM SPA-Cys (surface density ~1 ng/mm 2 ) allowed us to get the dependence of response on the IgG concentration.In turn, the data presented in the coordinates mol/cm 2 of the bound IgG vs. mol/l of the free IgG (Fig. 5), using the Langmuir model yielded the values of equilibrium dissociation constant for IgG/SPA-Cys interaction 8.5 ± 0.7× 10 -8 M (K a = 1.2 ± 0.1 × 10 7 M -1 ) and the value [AG] max = 1.5 ± 0.1 × 10 -12 mol/cm 2 .The use of Ska tchard plot provided very close data.
The equilibrium binding constant obtained for IgG binding to the immobilized SPA-Cys indicates their quite strong interaction and its value is consistent with the values for IgG binding to the immobilized SPA determined by others researchers (K a = 1.8-4.8× 10 7 M -1 ) [23][24][25].In conclusions, when studying the process of immobilization of the recombinant protein A from Staphylococcus aureus with the C-terminal cysteine residue (SPA-Cys) on a gold surface of the SPR spectrometer the direct dependence on the concentration of SPA-Cys in the range from 0.2 to 2 μM was revealed.The effi ciency of blocking the nonspecifi c adsorption sites on the sensor surface with milk proteins and lack of IgG interaction with them as well as the direct dependence of the sensor response on the IgG concentration and surface density of the immobilized SPA-Cys were shown.The determined equilibrium binding constant for IgG binding to the immobilized SPA-Cys indicates their quite strong interaction and its value is consistent with the literature data.
The successful immobilization of SPA-Cys on a gold surface of the SPR spectrometer while preserving its high immunoglobulin-binding activity, selectivity and stability of the sensor response confi rms the effi ciency of SPA-Cys as intermediate component for the creation of the immunosensor bioselective elements.
The bioselective elements based on SPA-Cys can be used as a «universal platform» for the oriented immobilization of antibodies against almost unlimited variety of antigens, as well as for an express screening of immunological status e.g., when testing immunodefi ciency.

Fig. 2 .
Fig. 2. The dependence of the sensor response (index of immobilization level) on the injected SPA-Cys concentration