biosensors Article

A Flow SPR Immunosensor Based on a Sandwich Direct Method Mauro Tomassetti 1, *, Giorgia Conta 1 , Luigi Campanella 1 , Gabriele Favero 2 , Gabriella Sanzò 2 , Franco Mazzei 2 and Riccarda Antiochia 2 1 2

*

Department of Chemistry, Sapienza University of Rome, P.le Aldo Moro 5, 00185 Rome, Italy; [email protected] (G.C.); [email protected] (L.C.) Department of Chemistry and Drug Technologies, Sapienza University of Rome, P.le Aldo Moro 5, 00185 Rome, Italy; [email protected] (G.F.); [email protected] (G.S.); [email protected] (F.M.); [email protected] (R.A.) Correspondence: [email protected]; Tel.: +39-6-4991-3722

Academic Editor: Nicole Jaffrezic-Renault Received: 2 March 2016; Accepted: 4 May 2016; Published: 13 May 2016

Abstract: In this study, we report the development of an SPR (Surface Plasmon Resonance) immunosensor for the detection of ampicillin, operating under flow conditions. SPR sensors based on both direct (with the immobilization of the antibody) and competitive (with the immobilization of the antigen) methods did not allow the detection of ampicillin. Therefore, a sandwich-based sensor was developed which showed a good linear response towards ampicillin between 10´3 and 10´1 M, a measurement time of ď20 min and a high selectivity both towards β-lactam antibiotics and antibiotics of different classes. Keywords: ampicillin; flow immunosensor SPR; sandwich method

1. Introduction β-lactams antibiotics are a class of antibiotics with a structure containing a β-lactam ring, i.e., a four-term cyclic amide. The action mechanisms of most of these antibiotics interfere with the synthesis of the cell wall of the bacteria foreign to the human body [1]. Therefore, the β-lactam antibiotics generally act as bactericides. However, some bacteria are capable of producing particular enzymes, called β-lactamases, that are able to inactivate the same antibiotics, making them completely ineffective [1]. Among the antibiotics of the penicillin group, ampicillin shows a broader spectrum of activity than penicillin G. The cephalosporins, such as ceftriaxone (which will also be considered in this study), do not belong to the group of penicillins, but they belong to the same class of β-lactam antibiotics. Therefore, they have the same action mechanism of penicillins, although they show a wider antibacterial spectrum and a greater resistance to β-lactamase. Of course, several strategies have been developed for the analytical determination of these antibiotics, for example chromatographic [2–5], mass-spectrometry [6–9], or microbial screening methods [10,11], but also methods based on sensors, biosensors and immunosensors [12–19]. Our research group developed different sensors for β-lactam antibiotics detection, initially ISES (Ion Selective Electrodes) [20] and most recently amperometric immunosensors [21]. Recent research carried out in our laboratories have been aimed at developing a traditional highly sensitive amperometric immunosensor [21] for penicillin G and other β-lactam antibiotics based on the competitive method, obtaining results that support the validity of this approach [21]. The LOD (low detection limit) was of the order of 10´10 M, but the analysis time was very long (about 1 h). This type of immunosensor also exhibited a poor selectivity towards other β-lactam antibiotics but a good selectivity towards antibiotics which do not belong to this class: it has therefore been used for

Biosensors 2016, 6, 22; doi:10.3390/bios6020022

www.mdpi.com/journal/biosensors

Biosensors 2016, 6, 22

2 of 13

the determination of the “pool” of antibiotics of the β-lactam class in common real matrices such as river waste water which can be more or less polluted by these types of antibiotics. In the present research, we examined the possibility of developing an immunosensor for the direct determination of a β-lactam antibiotic, specifically ampicillin, by using the flow SPR technique, especially in order to shorten the analysis time. The SPR immunosensors are analytical devices capable of detecting, in a selective manner, chemical or biological species, which do not require sample pretreatments and the use of competitive methods. However, the first results showed that it was not possible to realize an SPR immunosensor using the simplest and most common direct method, as the SPR signal obtained when, for instance, the penicillin G molecule that formed the immunocomplex with the antibody immobilized onto the SPR optical surface, was too low, probably because of the insufficiently high molecular weight of this antibiotic. In addition, the attempt to use a competitive method by immobilizing the antigen (instead of the antibody) and making it compete with the antigen to be determined in order to bind he free antibody in solution, was not successful, as described in this work. Moreover, this competitive method would not show any potential advantage compared to the traditional competitive method with amperometric detection in terms of measurement times, which would have been about the same for both methods. In fact, once the antibody complex is formed, its hydrolytic cleavage, in order to regenerate the optical surface of the SPR (that is the gold plate functionalized with the immobilized and not complexed antigen) to make a subsequent SPR measurement, resulted in being extremely difficult, as described in the “Results” section of the present paper. The research has therefore moved in the direction of the development of an immunosensor for ampicillin detection that is able to operate with a sandwich geometry (1st immobilized antigen molecule—antibody—2nd antigen molecule) that allowed for decreasing the analysis time and simplifying the measurement method. 2. Materials and Methods 2.1. Materials Sodium dihydrogen phosphate NaH2 PO4 (ě99.0%), sodium dibasic phosphate Na2 HPO4 (ě99.0%), EtOH (96%), 11-mercaptoundecanoic acid (MUA) (95%), 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS) (98%), glycine, ethanolamine (ě99%), ceftriaxone and erythromycin were supplied by Sigma-Aldrich (St. Louis, MO, USA). The SPR plates, composed of a layer of Au of a thickness of 50 nm on a glassy support, were supplied by XanTec bioanalytics GmbH (Dusseldorf, Germany). The oil with a refractive index of 1.6100 ˘ 0.0002 was supplied by Cargille Laboratories (Cedar Grove, NJ, USA). Sodium Ampicillin was provided by Farmitalia—Carlo Erba (Milan, Italy); anti-Ampicillin was provided by Sigma (St. Louis, MO, USA); penicillin G was produced by Fluka Analyticals (St. Louis, MO, USA); fosfomycin was produced by Crinos S.P.A. (Villa Guardia, Como, Italy). All solutions were prepared with deionized water (R = 18.2 mΩ ˆ cm at 25 ˝ C; TOC (Total Organic Carbon) < 10 g¨ mL´1 ) obtained by a Millipore Milli-Q system (Millipore, Molsheim, France). 2.2. Apparatus SPR flow measurements were realized by using BioSuplar 400T apparatus (Analytical µ-Systems—Department Of Mivitec GmbH, Sinzing, Germany), with a laser diode with low power (630–670 nm) as the light source. This instrument allows for the qualitative analysis of molecules on the basis of the mass of the sandwich antibody formed on the plate, which produces a variation in the resonance angle. This signal is then amplified and transduced. The output signal is successively processed as a function of time in the form of a sensorgram.

Biosensors 2016, 6, 22 Biosensors 2016, 6, 22

3 of 13 3 of 12

3. Results Results In this this section, section,we we described carried out during the experimental described all all the the stepssteps carried out during the experimental work in work order in to order to develop a sandwich immunosensor for ampicillin detection and to construct the relative develop a sandwich immunosensor for ampicillin detection and to construct the relative calibration calibration curve. curve. 3.1. SPR Immunosensor 3.1. SPR Immunosensor and and Flow Flow Measurements Measurements An An SPR SPR immunosensor immunosensor is is an an analytical analytical device device capable capable of of selectively selectively detecting detecting chemical chemical or or biological species without the need of sample pretreatment. It is composed of three main elements: biological species without the need of sample pretreatment. It is composed of three main elements: a abioreceptor bioreceptor biological component), a transducer, finally a system for signal processing. (or(or biological component), a transducer, andand finally a system for signal processing. The The bioreceptor is immobilized on the surface of the optical sensor and is responsible for the variation bioreceptor is immobilized on the surface of the optical sensor and is responsible for the variation of of the resonance angle, the characteristic parameter of this technique, related to the species to be the resonance angle, the characteristic parameter of this technique, related to the species to be determined. fact, we we basically basically measure determined. In In fact, measure the the variation variation of of the the resonance resonance angle, angle, due due to to the the interaction interaction between thethe bioreceptor, in our complex, immobilized on a suitably between the theanalyte analyteand and bioreceptor, in case our the caseantibody the antibody complex, immobilized on a functionalized gold metal film. The “optical” signal generated is transduced into an electric signal, suitably functionalized gold metal film. The “optical” signal generated is transduced into an electric then amplified and processed in the form ofform a sensorgram (resonance angle vs.angle time). signal, then amplified and processed in the of a sensorgram (resonance vs. time). The SPR flow device (Figure 1) has a prism as the main element, installed on rotating plate, plate, The SPR flow device (Figure 1) has a prism as the main element, installed on aa rotating automatically controlled by a computer. The different samples are introduced into a proper cell automatically controlled by a computer. The different samples are introduced into a proper cell by by means pump. The sensor consists of a of gold sheet,sheet, 50 nm50 thick, a glassy means of ofaaperistaltic peristaltic pump. The sensor consists a gold nm placed thick, on placed on asupport. glassy Once functionalized by forming a so-called SAM (Self-Assembled Monolayers), the plate is plate placed support. Once functionalized by forming a so-called SAM (Self-Assembled Monolayers), the is on the prism surface using an oil with a refractive index equal to that of the prism and the support placed on the prism surface using an oil with a refractive index equal to that of the prism and the itself, in order optical continuity. Finally, theFinally, plate isthe fixed with system fora mechanical support itself, to in ensure order to ensure optical continuity. plate is afixed with system for anchorage. After positioning the SAM on the prism, the assembly of the SPR flow device is completed mechanical anchorage. After positioning the SAM on the prism, the assembly of the SPR flow device with a cell suitable for flow analysis. When the software for data processing is ready, it is possible is completed with a cell suitable for flow analysis. When the software for data processing is ready,to it start the analysis. To observe the SPR phenomenon, the polarized light, emitted by a diode laser at low is possible to start the analysis. To observe the SPR phenomenon, the polarized light, emitted by a power, is reflected from the wafer on from a detector and on monitored asand a function of the diode laser at low power, is reflected the wafer a detector monitored as resonance a function angle, of the thanks to the prism rotation [22,23]. Successively, the SPR angle change caused by the formation of the resonance angle, thanks to the prism rotation [22,23]. Successively, the SPR angle change caused by sandwich was recorded, in the form of a sensorgram, as a function of time. The registered variation the formation of the sandwich was recorded, in the form of a sensorgram, as a function of time. The of resonance angle isof proportional to the isamount of complexed analyte. of The results showed on The the registered variation resonance angle proportional to the amount complexed analyte. sensorgram haveonseveral phases (Figure corresponding to the presence of different in the results showed the sensorgram have2)several phases (Figure 2) corresponding to solutions the presence of ˝ flow cell. The SPR used unit is the Unit of Resonance (r.u.), which corresponds to a change of 0.0001 different solutions in the flow cell. The SPR used unit is the Unit of Resonance (r.u.), which of the resonance corresponds to aangle. change of 0.0001° of the resonance angle.

Figure Plasmon Resonance Resonance(SPR) (SPR)operating operatingininflow flowmode. mode. Figure1.1.Measurement Measurementinstrument instrument used used for for Surface Surface Plasmon

Biosensors 2016, 2016, 6, 226, 22 Biosensors

4 of 124 of 13 4 of 12

Biosensors 2016, 6, 22

Resonance Units (r.u.) Resonance Units (r.u.)

1000 1000 FR FR

900 900 800 800 700 700 600 600

FD FD

FA FA LB LB

a a 4

2 10 4 2 10

4

2,5 10 4 2,5 10

4

3 10 4 3 10 Time (s) Time (s)

4

3,5 10 4 3,5 10

4

4 10 4 4 10

Figure 2. Example of sensorgram obtained with a sandwich format. For each measurement, after Figure 2. the Example of baseline sensorgram obtained withphosphate sandwich format. For each measurement, checking original (LB) by flowing buffer 50 mM, pHeach 7.2, the analyte after that after Figure 2. Example of sensorgram obtained with aa sandwich format. For measurement, checking the original baseline (LB) by flowing phosphate buffer 50 mM, pH 7.2, the analyte that flowed into the system upon the steady state has been reached (FA); then, the dissociation step (FD) checking the original baseline (LB) by flowing phosphate buffer 50 mM, pH 7.2, the analyte that flowed flowed into the system upon the steady state hassmall beenbold reached (FA); then, the dissociation step (FD) started by flowing again phosphate buffer (the (a) the indicates the Δ r.u, relative the by into the system upon the steady state has been reached (FA);line then, dissociation step (FD) to started started by flowing again phosphate buffer (the small bold line (a) indicates the Δ r.u, relative towas the formation of the immunocomplex between analyte and bioreceptor); the regeneration step (FR) flowing again phosphate buffer (the between small bold line (a) indicates the ∆ r.u, relative tostep the(FR) formation formation immunocomplex analyte bioreceptor); was of performedof bythe glycine-HCl solution pH 2.4 inally theand phosphate bufferthe to regeneration obtain the baseline. the immunocomplex between analyte and bioreceptor); the regeneration step (FR) was performed by performed by glycine-HCl solution pH 2.4 inally the phosphate buffer to obtain the baseline. glycine-HCl solution pH 2.4 inally the phosphate buffer to obtain the baseline. 3.2. Self Assembled Monolayer (SAM) 3.2. Self Assembled Monolayer (SAM) SAMs areMonolayer highly ordered structures [24] that form spontaneously. A typical SAM is generally 3.2. Self Assembled (SAM) SAMs are orderedcompounds structures [24] spontaneously. A typical SAM is generally constituted by highly organosulfur suchthat as form the 11-mercaptoundecanoic acid (MUA), which constituted by organosulfur compounds such as the 11-mercaptoundecanoic acid (MUA), SAMs are binds highly structures [24] that form spontaneously. typical SAMthe iswhich generally covalently to ordered the gold surface of the plate. The preparation is carriedA out by dipping plate covalently to the2gold surface of the plate. preparation is carried by dipping the plate in a solution of MUA mMcompounds in ethanol for about 18 h. A SAM structure is out therefore obtained, with constituted by binds organosulfur such asThe the 11-mercaptoundecanoic acid (MUA), which in solution ofgroups MUA of 2 mM in ethanol for form aboutAu-S 18 h.covalent A SAM structure is therefore obtained, with thea mercapto the MUA able bonds, while the groups, covalently binds to the gold surface of thetoplate. The preparation is carried outcarboxylic by dipping the plate the mercapto groups of the MUA able 3, to form Au-S covalent bonds, binding while the carboxylic groups, which remain free, 2asmM shown in Figure available subsequent with the antibody or with in a solution of MUA in ethanol for are about 18 h. for A SAM structure is therefore obtained, which remain free, as shown in Figure 3, are available for subsequent binding with the antibody or the antigen. After 18 h, the plate with SAM is taken away from the solution, washed with ethanol the mercapto groups of h, thetheMUA able to form Au-Saway covalent bonds, whilewashed the carboxylic groups, the After 18 with SAM is taken fromof the solution, with andantigen. let to dry. The prism isplate rinsed with ethanol, then a drop the oil is placed onto theethanol prism whichand remain as shown inisFigure are ethanol, available for subsequent binding with the antibody or the let and tofree, dry. prismSAM-modified rinsed 3, with surface the The prepared gold plate.then a drop of the oil is placed onto the prism antigen. After 18 h, the plate with SAM is taken away from the solution, washed with ethanol and let surface and the prepared SAM-modified gold The stabilization of the SPR signal is thenplate. carried out by flowing 50 mM phosphate buffer into to dry. The prism is rinsed with ethanol, then a drop of the is placed50 onto the prism and surface and the The stabilization of the SPR signal is then carried out by flowing mM phosphate buffer into the cell at pH = 7.2 for about 1 h. This process is crucial as itoil rehydrates the MUA surface creates the cell at pH = 7.2 for about 1 h. This process is crucial as it rehydrates the MUA surface and creates prepared SAM-modified plate. a stable baseline on thegold sensorgram. a stable baseline on the sensorgram.

Figure 3. Schematic representation of SAM (Self Assembled Monolayer) with free –COOH and Figure 3. Schematic representation of SAM (Self Assembled Monolayer) with free –COOH and covalent Figure 3. bonds Schematic representation of SAM (Self Assembled Monolayer) with free –COOH and covalent Au-S. bondscovalent Au-S. bonds Au-S.

The stabilization of the SPR signal is then carried out by flowing 50 mM phosphate buffer into the cell at pH = 7.2 for about 1 h. This process is crucial as it rehydrates the MUA surface and creates a stable baseline on the sensorgram.

Biosensors 2016, 6, 22

5 of 13

3.3. Choice of the Measurement Format Biosensors 2016, 6, 22

5 of 12

As mentioned in the Introduction and successively described in the Results and Discussion section, 3.3.found Choicethat, of theafter Measurement Format it was the immobilization of the antigen (ampicillin) on SAM and the association of its antibody (anti-ampicillin), a strong bond is successively formed, not described easy to break through As mentioned in the Introduction and in the Resultsthe andhydrolysis Discussionwith the classic normally used for this purpose [21–23]. This suggestedon theSAM development section,eluent it wassolutions found that, after the immobilization of the antigen (ampicillin) and the of association of its antibody (anti-ampicillin), a strong is formed, not easy to break the of a sandwich geometry for the measurement instead of abond competitive format type. After through the injection hydrolysis with the classic eluent solutions normally used for this purpose [21–23]. This suggested Gly-HCl at pH 2.4, the antibody complex, immobilized on the plate surface, and consisting of the first the development of a sandwich geometryantigen for the measurement instead of a competitive type.with immobilized antigen—antibody—second is in fact “regenerated” by breakingformat the bond After the injection of suggests Gly-HCl at pHthe 2.4, the antibody complex, immobilized on the and platethe surface, the second antigen. This that bond between the second antigen molecule antibody and consisting of the first immobilized antigen—antibody—second antigen is inmolecule fact “regenerated” is certainly much weaker than that between the antibody and the first antigen immobilized by breaking the bond with the second antigen. This suggests that the bond between the second on the plate, which however is not easily broken by the treatment with the same solvent (Gly-HCl at antigen molecule and the antibody is certainly much weaker than that between the antibody and the pH 2.4). first antigen molecule immobilized on the plate, which however is not easily broken by the treatment with the same solvent (Gly-HCl at pH 2.4).

3.4. Immobilization of Ampicillin

3.4. of Ampicillin TheImmobilization immobilization procedure was studied by monitoring the sensorgram reported in Figure 4.

The firstThe operation was theprocedure SAM stabilization, occurred the cell byreported the phosphate immobilization was studiedwhich by monitoring theinsensorgram in Figurebuffer 4. flowing 50 mM at pH = 7.2 for about 1 h. This allowed for recording a first base line (LB line, Figure The first operation was the SAM stabilization, which occurred in the cell by the phosphate buffer 4). Successively, 1:1 mixture of for 0.5 about M EDC M NHS allowedatofirst flow in line the (LB cell line, for about flowing 50amM at pH = 7.2 1 h.and This0.1 allowed forisrecording base Figure15 4).min in order to activate themixture carboxyl groups of the (A line, Successively, a 1:1 of 0.5 M EDC andSAM, 0.1 Mwith NHSaisconsequent allowed to raising flow in of thethe cellsignal for about 15 min in order the carboxyl groups of the SAM,buffer, with apH consequent raising of the signal Figure 4). The plateto is activate then washed with 50 mM phosphate = 7.2, with a resulting lowering (Asignal line, Figure The plate is thenthe washed with 50 mM phosphate buffer, 7.2, with resulting of the due to4).separation from surface of the excess of reagents (B pH line,= Figure 4).a At this point, ´ 2 lowering of the signal due to separation from the surface of the excess of reagents (B line, Figure a solution of ampicillin 10 M is allowed to flow in the cell for about 40 min (FI line, Figure 4) in4). order At this a solution ampicillin 10−2 M is allowed flow ingroups the cell of forthe about 40 and min the (FI line, to form thepoint, covalent bondsof(amidic bonds) between the to carboxy SAM antigen Figure 4) in order to form the covalent bonds (amidic bonds) between the carboxy groups of the (Figure 5). Finally, the cell was washed with phosphate buffer (C line, Figure 4). In order to avoid SAM and the antigen (Figure 5). Finally, the cell was washed with phosphate buffer (C line, Figure non-specific interactions with the activated –COOH groups, the latter are blocked by ethanolamine 4). In order to avoid non-specific interactions with the activated –COOH groups, the latter are flowing 1 mM (Figure 5) for about 15 min (D line, Figure 4. The cell was then washed with phosphate blocked by ethanolamine flowing 1 mM (Figure 5) for about 15 min (D line, Figure 4. The cell was buffer (Ewashed line, Figure then with 4). phosphate buffer (E line, Figure 4). As shown in 4, ampicillin waswas effectively immobilized onon the functionalized As shownFigure in Figure 4, ampicillin effectively immobilized the functionalizedsurface. surface. In In fact, the base linebase (C),line after the stepthe ofstep ampicillin immobilization (FI),(FI), is located at r.u. values greater fact, the (C), after of ampicillin immobilization is located at r.u. values greaterthan the base (B) before ampicillin immobilization. than line the base line (B)the before the ampicillin immobilization. 1000

A 900

r.u.

800

D

700

FI

600

C

B

LB

E

500 0

1 10

4

2 10

4

3 10

4

4 10

4

5 10

4

6 10

4

7 10

4

Time (s)

Figure 4. Immobilization of ampicillin: (LB) original baseline generated by flowing phosphate buffer;

Figure 4. Immobilization of ampicillin: (LB) original baseline generated by flowing phosphate buffer; (A) activation with EDC/NHS; (B) washing with phosphate buffer; (FI) immobilization of ampicillin (A) activation with EDC/NHS; (B) washing with phosphate buffer; (FI) immobilization of ampicillin 10−2 M; (C) washing with phosphate buffer; (D) inactivation of –COOH groups which may have not ´2 10 M; (C) washing with phosphate buffer; (D) inactivation of –COOH groups which may have not reacted with ethanolamine; (E) washing with phosphate buffer. reacted with ethanolamine; (E) washing with phosphate buffer.

Biosensors 2016, 6, 22 Biosensors 2016, 6, 22

6 of 13 6 of 12

−2 M (R) and inactivation of Figure 5. Schematic representation representation of of SAM SAM association associationwith withampicillin ampicillin10 10´2 M (R) and inactivation –COOH groups groups which which may may have have not not reacted reacted by byethanolamine ethanolamine(R’). (R’). –COOH

3.5. 3.5. Association Association of of Anti-Ampicillin Anti-Ampicillin The now ready ready for for the the formation formation of of the the antibody antibody complex. complex. Subsequent Subsequent injections injections of The surface surface is is now of different solutions of antibody at increasing concentrations have been performed. These different solutions of antibody at increasing concentrations have been performed. These have have been been ´3−3M. produced fromaastock stocksolution solutionwith witha aconcentration concentration 1 10 × 10 M.The Thefirst firststep stepwas wastoto flow mM produced from of of 1ˆ flow 5050 mM of of phosphate buffer at pH = 7.2 to obtain the base line (LB line, Figure 6). The first antibody solution, phosphate buffer at pH = 7.2 to obtain the base line (LB line, Figure 6). The first antibody solution, at a −10 M, was introduced into the cell observing the increase of the relative at a concentration of10 6 ´×1010M, concentration of 6 ˆ was introduced into the cell observing the increase of the relative signal signal until a plateau is reached (A, 6) Figure 6) the dueformation to the formation of the antibody At thisa until a plateau is reached (A, Figure due to of the antibody complex.complex. At this point, point, a washing procedure with phosphate buffer was carried out. Then, a solution of Gly-HCl, washing procedure with phosphate buffer was carried out. Then, a solution of Gly-HCl, pH = 2.4 was pH = 2.4 was into should, the cell,inwhich in theory, dissociate the nti-ampicillin complex between introduced intointroduced the cell, which theory,should, dissociate the complex between and nti-ampicillin and ampicillin (FR line, Figure 6) through a hydrolysis reaction with the consequent ampicillin (FR line, Figure 6) through a hydrolysis reaction with the consequent regeneration of the regeneration of the activated where only the immobilized antigen present. Theto signal was activated surface, where only surface, the immobilized antigen is present. The signaliswas expected return to expected return tolevel, the initial a second addition of the antibody. the initialto reference beforereference carryinglevel, out a before secondcarrying additionout of the antibody. On the contrary, it is On contrary, is easy to that see the on the sensorgram the complexand between anti-ampicillin easythe to see on the it sensorgram complex betweenthat anti-ampicillin ampicillin immobilizedand on ampicillin immobilized on the surface after is practically not dissociated afterpH the treatment with the gold surface is practically notgold dissociated the treatment with Gly-HCl, = 2.4. The surface Gly-HCl, = 2.4. Thehence surface being regenerated, it ismeasurement impossible towith carrythe out a second not being pH regenerated, it isnot impossible to carry outhence a second same plate, measurement with the same plate, so that it would also be impossible to build, for example, a so that it would also be impossible to build, for example, a calibration curve using a “competitive” calibration curve “competitive” methodwas withtherefore the samecontinued plate. The with experiment was additions therefore method with the using same aplate. The experiment successive ´ 10 continued with successive additions of further solutions of antibody at increasing concentrations, of further solutions of antibody at increasing concentrations, 8 ˆ 10 M (B line, Figure 6) and −10 −9 M (C line, Figure 6), until, ´ 3 M, × 10 6) and × 10 concentration at least 18 ˆ 10´9 M M (B (C line, line, Figure Figure 6), until,1 for concentration above at least 10 for almost all theabove immobilized −3 10 M, almost all the antigenwith molecules had practically reacted with the immobilized antibody, by antigen molecules hadimmobilized practically reacted the antibody, by forming the antibody forming the antibody immobilized complex. The Δ r.u. values obtained were plotted against the complex. The ∆ r.u. values obtained were plotted against the concentration of anti-ampicillin, obtaining concentration of anti-ampicillin, obtaining an antibody saturation curve, as shown in Figure 7. The an antibody saturation curve, as shown in Figure 7. The obtained “saturation curve” allowed for obtained “saturation curve” allowed for evaluating the concentration ofsaturating antibody solution of evaluating the concentration of antibody solution capable of virtually with thecapable antibody virtually saturating with the antibody the plate already activated with the antigen, by formation of the plate already activated with the antigen, by formation of the antibody complex, immobilized on ´3 M, underis the antibody immobilized the plate itself. concentration atoperating least higher than the plate itself.complex, This concentration is on at least higher thanThis 1 ˆ 10 the suitable −3 1conditions. × 10 M, under the operating suitable conditions. It should be noted that the instrument was It should be noted that the instrument was able to detect the antibody complex whenable the to detect the antibody complex whenorders the increase of itslower, concentration was several of increase of its concentration was several of magnitude since anti-ampicillin is aorders molecule magnitude since anti-ampicillin a molecule with quite a high molecular weight (150 kDa). with quite alower, high molecular weight (150iskDa).

Biosensors 2016, 6, 22 Biosensors 2016, 6, 22

7 of 13 7 of 12

Figure 6. Antibody complexation on the SAM coated with the antygen: (A) 6 × 10−10 M; (B) 8 × 10−10 M; Figure 6. Antibody complexation on the SAM coated with the antygen: (A) 6 ˆ 10´10 M; (C) 1 × ´10 10−9 M, anti-ampicillin flowing successive solutions; (FR) regeneration step by flowing (B) 8 ˆ 10 M; (C) 1 ˆ 10´9 M, anti-ampicillin flowing successive solutions; (FR) regeneration Gly-HCl at pH 2.4. step by flowing Gly-HCl at pH 2.4.

Figure 7. “Saturation curve” with anti-ampicillin, obtained by flowing antibody solutions: 6 × 10−10 Figure.7 −10 −9 M, 1 × 10−3 M). M, 1 × 10curve” M, 8 7. × 10 Figure “Saturation with anti-ampicillin, obtained by flowing antibody solutions: 6 ˆ 10´10 M,

8 ˆ 10´10 M, 1 ˆ 10´9 M, 1 ˆ 10´3 M).

3.6. Formation of the Sandwich and Construction of a Calibration Curve for Ampicillin

3.6. Formation of the Sandwich and Construction of aparagraph Calibrationconfirmed Curve for the Ampicillin The experiments described in the previous difficulty with applying a competitive method and therefore suggested the choice to use the sandwich geometry to obtain a The experiments described in the previous paragraph confirmed the difficulty with applying calibration curve for ampicillin. Therefore, after completely saturating the plate with a 0.1 M a competitive method and therefore suggested the choice to use the sandwich geometry to obtain antibody solution and washing it with buffer, as described in the previous paragraph, the experience a calibration curve for ampicillin. Therefore, after completely saturating the plate with a 0.1 M started by recording the initial baseline when 50 mM phosphate buffer at pH = 7.2 was introduced in antibody solution and washing it with buffer, as described in the previous paragraph, the experience the flow cell containing the plate activated with the antibody complex immobilized on its surface. started by recording the initial baseline when 50 mM phosphate buffer at pH = 7.2 was introduced Successively, several injections of solutions at increasing ampicillin concentration, obtained by in the flowthe cellsame containing the plate with been the antibody complex immobilized onthe its curve surface. diluting stock solution (1 activated × 10−1 M), have carried out. The following parts of Successively, several injections of solutions at increasing ampicillin concentration, obtained by diluting were obtained: the flow of the first ampicillin solution at a concentration of 1 × 10−3 M on the theantibody same stock solution complex (1 ˆ 10´1caused M), have been carriedofout. The following partssignal of theiscurve were immobilized a small increase the signal; once a stable reached, ´3 M on the antibody obtained: the flow of the first ampicillin solution at a concentration of 1 ˆ 10 phosphate buffer was again injected. A small decrease of the signal was observed, but at the end of immobilized complex caused a smallsignal increase of the signal; to once stable signal is reached, phosphate the decrease, a positive difference (r.u.) compared thea original base line was registered, buffer was againthe injected. A small decrease the signal was observed, but at the end of the decrease, which proved sandwich formation. Theofsubsequent injection of a solution of Gly-HCl at pH 2.4 a positive signal (r.u.) of compared toLastly, the original base line registered, which proved caused adifference characteristic increase the signal. by flowing againwas phosphate buffer 50 mM at pHthe = 7.2, theformation. signal returned to a value of r.u. equal that of the initial original line (Figure 2). This sandwich The subsequent injection of atosolution of Gly-HCl at pHbase 2.4 caused a characteristic increase of the signal. Lastly, by flowing again phosphate buffer 50 mM at pH = 7.2, the signal returned

Biosensors 2016, 6, 22

8 of 13

to a value of r.u. equal to that of the initial original base line (Figure 2). This showed that the sandwich Biosensors 2016, 6, 22 8 of 12 2016, 6, 22 8 of 12 hadBiosensors broken, but the antibody complex remained fixed on the plate. The same steps were repeated ´1 M. An increasing value for the at increasing ampicillin concentrations 5 ˆantibody 10´3 M to 1 ˆ 10remained showedthat that thesandwich sandwich hadbroken, broken,from butthe the complex fixedon onthe theplate. plate.The The showed the had but antibody complex remained fixed −3 and −1 M. difference between the signal at observed after the washing with phosphate buffer the initial original same steps were repeated increasing ampicillin concentrations from 5 × 10 M to 1 × 10 An same steps were repeated at increasing ampicillin concentrations from 5 × 10−3 M to 1 × 10−1 M. An increasing value for the difference between the signal observed after the washing with phosphate base line (example of Figure 8) has been recorded, which demonstrated how the interaction between increasing value for the difference between the signal observed after the washing with phosphate buffer andthe themolecule initialoriginal original base line(example (example ofestablished Figure8) 8)has has beentime recorded, which demonstrated thebuffer 2nd antigen andbase the line antibody wasof each giving rise demonstrated to the sandwich and initial Figure been recorded, which ´1 Mwas how the interaction between the 2nd antigen molecule and the antibody established eachtime time is geometry. A typical example of a recorded measurement for 1 ˆ 10 ampicillin concentration how the interaction between the 2nd antigen molecule and the antibody was established each −1 M giving rise to the sandwich geometry. A typical example of a recorded measurement for 1 × 10 −1 reported Figure It was clear, however, that example this observation was measurement to be supported a “blank” giving in rise to the 8. sandwich geometry. A typical of a recorded for 1by × 10 M ampicillin concentration reported inFigure Figure8.8.ItItwas was clear,however, however, that thisobservation observation wasto to ∆ test,ampicillin which is concentration described in the following paragraph. In conclusion, in the described experience, the isisreported in clear, that this was be supported by a “blank” test, which is described in the following paragraph. In conclusion, in the be supported by a “blank” which is described the following paragraph. conclusion, the of in r.u. values related to each test, antigen addition haveinbeen determined and theIn difference, ininterms describedexperience, experience,the theΔΔin inr.u. r.u.values valuesrelated relatedto toeach eachantigen antigenaddition additionhave havebeen beendetermined determinedand and described resonance units, between the final r.u. values (after washing with phosphate buffer), obtained with the the difference, in terms of resonance units, between the final r.u. values (after washing with the difference, in termsantigen of resonance units, between r.u. values (after washing association of the second with the antibody and the the final r.u. values relative to the initial with original phosphatebuffer), buffer),obtained obtainedwith withthe theassociation associationof ofthe thesecond secondantigen antigenwith withthe theantibody antibodyand andthe ther.u. r.u. phosphate baseline, was always positive and growing. The so obtained ∆ in r.u. values were finally plotted values relative to the initial original baseline, was always positive and growing. The so obtained Δ in relative to the initial ampicillin original baseline, was always positive and growing. The so obtained Δ in as avalues function of the relative concentration (see calibration line, Figure 9). For ampicillin r.u. values were finally plotted as a function of the relative ampicillin concentration (see calibration r.u. values were finally plotted as a function of the relative ampicillin concentration (see calibration concentrations lower 10´3 M, it is likely that the than sandwich thethe sensitivity the line, Figure Figure 9). 9). For Forthan ampicillin concentrations lower 10−3−3 M, M,isititformed, likelybut that sandwichof line, ampicillin concentrations lower than 10 isis likely that the sandwich isis method does not enough it. not seem enough to highlight it. formed, but theseem sensitivity of to thehighlight method does formed, but the sensitivity of the method does not seem enough to highlight it. 1300 1300

E E

1200 1200

r.u. r.u.

1100 1100 1000 1000 900 900 800 800 700 700 600 600 0 0

5000 5000

1 104 1 10

4

4

1,5 104 1,5 10

Time (s) Time (s)

with sandwich format. Δ r.u (bold line) is related Figure 8. Example of ampicillin detection 1 × 10´1 M Figure 8. Example of of ampicillin withsandwich sandwich format. ∆ r.u (bold is related MMwith format. Δ r.u (bold line)line) is related Figure 8. Example ampicillindetection detection11ˆ× 10−1 −1 M, before and after the association of ampicillin. to the concentration (E) 1 × 10 ´1 −1 to the concentration 10 M, thethe association of ampicillin. to the concentration (E)(E)1 1ˆ×10 M,before beforeand andafter after association of ampicillin. −1

Figure 9. Comparison between blank curve and calibration curve of ampicillin.

Figure Comparisonbetween between blank blank curve curve of of ampicillin. Figure 9. 9.Comparison curveand andcalibration calibration curve ampicillin.

Biosensors 2016, 6, 22

9 of 13

3.7. “Blank” Curve for Ampicillin In order to make a “blank” curve, the usual SAM was prepared by dipping the gold plate for about 18 h in a solution of 2 mM 11-mercaptoundecanoic acid and ethanol. After the fixed time, the SAM was washed with ethanol to remove the excess reagent. Then, the plate was left to dry. The assembly was completed by placing on the prism a drop of oil and then the prepared plate with the SAM. Once obtained, the base line through the flow of phosphate buffer (50 mM at pH = 7.2) for about 1 h, the –COOH groups have been activated with a solution of EDC 0.5 M and NHS 0.1 M at a 1:1 ratio (freshly prepared) for about 15 min, then, the plate was washed with phosphate buffer and finally a solution of ampicillin 10´1 M was injected and allowed to flow for about 40 min. The plate was then washed again with 50 mM phosphate buffer at pH = 7.2. In order to be sure that all the sites on the gold plate had been saturated, a solution of 1 mM ethanolamine was injected to block any carboxyl groups that had not reacted. At this point, after removing the excess ethanolamine by washing with phosphate buffer, the following ampicillin solutions at increasing concentrations have been injected: (a) 10´3 M; (b) 5 ˆ 10´3 M; (c) 5 ˆ 10´2 M; (d) 7 ˆ 10´2 M. The ∆ in r.u. values, determined after each addition, were plotted as a function of increasing concentrations of ampicillin. The “blank” curve thus obtained and the calibration curve described in the previous section were both reported in a single graph (Figure 9) in which each point represents the average of three separate determinations. The equations of the straight lines obtained and the main analytical data are summarized in Table 1. The fact that the two lines show very different slopes (the slope of the “blank” curve is very low) indicates that the sandwich geometry was actually formed and allowed the construction of the calibration curve, while the immobilization of the antigen alone on the SAM, as in the case of the “blank” curve, gave a signal which is almost zero and barely visible only at the highest antigen concentrations. Table 1. Analytical data using an SPR device of both calibration and blank curves. Equation of the Regression Straight Line

Linearity Range (M)

Calibration straight line (Y = ∆ r.u., X = M )

y = 7.89 (˘ 1.63) + 530.8 (˘ 29.1)x

1 ˆ 10´3 –1 ˆ 10´1

“Blank” curve (Y = ∆ r.u., X = M)

y = 0.91 (˘ 0.15) + 185.3 (˘ 3.4)x

1 ˆ 10´3 –7 ˆ 10´2

Pooled SD (%)

LOD (M)

0.9940

5.1

10´3

0.9993

4.4

10´3

R2

3.8. Selectivity Tests In theory, the response of each immunosensor should be very selective because of the specificity of the antigen-antibody reaction. However, this selectivity should be experimentally verified due to the cross-reactivity, a situation in which it is possible that side reactions occur at the immunosensor due to a chemical compound very similar to the antigen. Consequently, the selectivity was assessed considering both compounds with a structure similar to ampicillin, such as other β-lactam antibiotics (penicillin G and ceftriaxone) and other antibiotics with very different structures (fosfomycin and erythromycin), as possible interferents. The first selectivity test involved penicillin G (Figure 10), an antibiotic with the same β-lactam ring of ampicillin. Different concentrations of penicillin G have been tested, similar to those used to construct the calibration curve of ampicillin, and injected in succession in the cell as reported in Section 3.6, obtaining the sensorgram shown in Figure 10. It is possible to note that the instrument seems at first to detect the presence of Penicillin G in the cell but also that, after washing with 50 mM phosphate buffer at pH = 7.2, the signal returned to the original base line. This can be ascribed to the fact that the sandwich in the case of penicillin G may not form, or at least it is not detected by the SPR transducer, probably because of the low molecular weight of penicillin G. In other selectivity tests (Figure 11), possible interferences of both β-lactam, such as ceftriaxone, or cephalothin, and non

resulted in being almost zero (line A of Figure 11). Similar experiments have been carried out with resulted in being almost zero (line A of Figure 11). Similar experiments have been carried out with erythromycin, another non β-lactam antibiotic. Its injection resulted in a decreasing signal (line B of erythromycin, another non β-lactam antibiotic. Its injection resulted in a decreasing signal (line B of Figure 11), as if this addition would cause a partial separation of the antibody complex. However, Figure 11), as if this addition would cause a partial separation of the antibody complex. However, this assumption does not seem fully explanatory, as both fosfomycin and erythromycin showed Δ in this assumption does not seem fully explanatory, as both fosfomycin and erythromycin showed Δ in r.u. values equal to zero after the subsequent injection of 50 mM phosphate buffer at pH = 7.2 with a Biosensors 2016, 6, 22 to zero after the subsequent injection of 50 mM phosphate buffer at pH = 7.2 with 10 of 13 r.u. values equal a return of the signal to the original baseline on the sensorgram. These results showing that the latter return of the signal to the original baseline on the sensorgram. These results showing that the latter two antibiotics do not create any interference could perhaps be predictable, as they are non β-lactam two antibiotics do not create any interference could perhaps be predictable, as they are non β-lactam antibiotics. Based onsuch this as evidence, it is or possible to conclude the immunosensor developed for β-lactam antibiotics, fosfomycin, erythromicin, were that evaluated by using the same procedure antibiotics. Based on this evidence, it is possible to conclude that the immunosensor developed for ampicillin results employed detection for penicillin G. in being very selective towards this antibiotic. ampicillin detection results in being very selective towards this antibiotic.

Figure 10. Selectivity: penicillin by Figure Selectivity: complexation complexationof penicillinG by using using sandwich sandwich immunosensor immunosensor format. format. Figure 10. 10. Selectivity: complexation ofof penicillin GG by using sandwich format. −4 M; (B) −3 M; (C) −3 M; (D) −2immunosensor −2 M. Penicillin G solutions: (A) 5 × 10 1 × 10 5 × 10 1 × 10 M;´2(E) 5(E) × 10 ´4 ´3 ´3 ´2 Penicillin G solutions: (A) 5 ˆ 10 M; (B) 1 ˆ 10 M; (C) 5 ˆ 10 M; (D) 1 ˆ 10 M; 5 ˆ 10 −4 −3 −3 −2 −2 Penicillin G solutions: (A) 5 × 10 M; (B) 1 × 10 M; (C) 5 × 10 M; (D) 1 × 10 M; (E) 5 × 10 M. M. 850 850

r.u.r.u.

800 800

750 750

700 700

A A

C C B B

650 650

600 600 0 0

4

1 10 4 1 10

4

2 10 4 2 10

4

3 10 4 3 10 Time (s) Time (s)

4

4 10 4 4 10

4

5 10 4 5 10

4

6 10 4 6 10

Figure 11. Selectivity test: (A) Fosfomycin 1 × 10−3 M; (B) Erythromicin 5 × 10−3 M; (C) Ceftriaxone 4 × −3 ´3 (C) Figure 11. Selectivity Selectivity test: test: (A) × 10 (B) Erythromicin 10−3 Ceftriaxone 4× Figure (A) Fosfomycin Fosfomycin 11 ˆ 10´3M;M; (B) Erythromicin5 5× ˆ 10M; M; (C) Ceftriaxone −3 M. 11. 10 −3 ´3 M. M. 410ˆ 10

4. Conclusions 4. Conclusions Also in the case of ceftriaxone, a cephalosporin, which is a β-lactam antibiotic but does not belong The purpose of this experimental research was the construction of an immunosensor for the purpose this experimental research was the construction of an immunosensor for 11). the to theThe same class ofofampicillin, the formation of the sandwich was not detected (line C of Figure direct determination of ampicillin, using the SPR technique operating in flow conditions. In order to direct of ampicillin, the SPR technique In order in to In the determination case of fosfomycin, the ∆ inusing r.u. before and after theoperating injectioninofflow this conditions. antibiotic, resulted being almost zero (line A of Figure 11). Similar experiments have been carried out with erythromycin, another non β-lactam antibiotic. Its injection resulted in a decreasing signal (line B of Figure 11), as if this addition would cause a partial separation of the antibody complex. However, this assumption does not seem fully explanatory, as both fosfomycin and erythromycin showed ∆ in r.u. values equal to zero after the subsequent injection of 50 mM phosphate buffer at pH = 7.2 with a return of the signal to the original baseline on the sensorgram. These results showing that the latter two antibiotics do not create any interference could perhaps be predictable, as they are non β-lactam antibiotics. Based on

Biosensors 2016, 6, 22

11 of 13

this evidence, it is possible to conclude that the immunosensor developed for ampicillin detection results in being very selective towards this antibiotic. 4. Conclusions The purpose of this experimental research was the construction of an immunosensor for the direct determination of ampicillin, using the SPR technique operating in flow conditions. In order to reach the aim of the work, the results obtained suggested the development of a sandwich geometry (first molecule of immobilized antigen—antibody—second antigen molecule) as the best format. The formation of the sandwich was confirmed by comparing the slope value of the calibration curve of ampicillin with the slope of the “blank” curve, which resulted being very low. Furthermore, from the data resulting from the selectivity tests, it was found that none of the other tested antibiotics, both those belonging to very different classes of antibiotics and those structurally similar to the class of β-lactam, resulted in being potential interferents. This is a remarkable result, as it highlights the great selectivity of the SPR immunosensor for ampicillin in the concentration range from 10´3 M to 10´1 M. A disadvantage of the developed SPR immunosensor, compared to traditional “competitive” immunosensors (non-SPR), is that traditional immunosensors are generally more sensitive devices [21] compared to the SPR immunsensor described in this research, even if they have often proven to be much less selective compared to the SPR immunosensor. However, for certain applications such as the control of commercial pharmaceutical specialties, a high sensitivity is not mandatory. Moreover, good selectivity means that the method could be used in the case of specialties containing different active molecules in association. Another important positive aspect to be emphasized is the analysis time. In fact, a direct determination of ampicillin in a much shorter time (maximum 20 min) compared to the time required by the traditional competitive methods (which generally require more than 1 h [21] to complete the measure) is possible, thanks to the formation of the sandwich. Acknowledgments: This work was funded by the Sapienza University of Rome, “Protezione dell’Ambiente e dei Beni Culturali (CIABC)” center and by “Istituto per lo studio dei Materiali Nanostrutturati (ISMN)” of CNR (Consiglio Nazionale delle Ricerche). Author Contributions: M.T. conceived and designed the experiments; G.C. performed the experiments and prepared the figures; G.S. analyzed the data; L.C. and F.M. contributed reagents/materials/analysis tools; M.T. and R.A. wrote the paper; G.F. gave suggestions on the structure of the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations The following abbreviations are used in this manuscript: SPR LOD SAM r.u. MUA Gly EDC NHS

Surface Plasmon Resonance Low detection limit Self Assembled Monolayer Unit of Resonance 11-mercaptoundecanoic Acid Glycine 1-ethyl-3-(3-dimethyl aminopropyl) Carbodiimide N-hydroxysuccinimide

References 1.

Murray, P.R.; Rosenthal, K.S.; Pfaller, M.A. Microbiologia Medica, 6th ed.; Edra Masson: Milan, Italy, 2015; pp. 167–169 and 187–190.

Biosensors 2016, 6, 22

2.

3.

4.

5.

6. 7.

8.

9.

10. 11. 12. 13. 14.

15. 16.

17. 18.

19.

20. 21.

12 of 13

Benito-Peña, E.; Partal-Rodera, A.I.; Léon-González, M.E.; Moreno-Bondi, M.C. Evaluation of mixed mode solid phase extraction cartridges for the preconcentration of beta-lactam antibiotics in wastewater using liquid chromatography with UV-DAD detection. Anal. Chim. Acta 2006, 556, 415–422. [CrossRef] Marazuela, M.D.; Bogialli, S. A review of novel strategies of sample preparation for the determination of antibacterial residues in foodstuffs using liquid chromatography-based analytical methods. Anal. Chim. Acta 2009, 645, 5–17. [CrossRef] [PubMed] Briscoe, S.E.; McWhinney, B.C.; Lipman, J.; Roberts, J.A.; Ungerer, J.P.J. A method for determining the free (unbound) concentration of ten beta-lactam antibiotics in human plasma using high performance liquid chromatography with ultraviolet detection. J. Chromatogr. B 2012, 907, 178–184. [CrossRef] [PubMed] Samanidou, V.F.; Evaggelopoulou, E.N.; Papadoyannis, I.N. Development of a validated HPLC method for the determination of four penicillin antibiotics in pharmaceuticals and human biological fluids. J. Sep. Sci. 2006, 2, 1550–1560. [CrossRef] Jin, H.; Kumar, A.P.; Paik, D.H.; Ha, K.C.; Yoo, K.C.; Lee, Y.I. Trace analysis of tetracycline antibiotics in human urine using UPLC–QToF mass spectrometry. Microchem. J. 2010, 94, 139–147. [CrossRef] Heller, D.N.; Smith, M.L.; Chiesa, O.A. LC/MS/MS measurement of penicillin G in bovine plasma, urine, and biopsy samples taken from kidneys of standing animals. J. Chromatogr. B 2006, 830, 91–99. [CrossRef] [PubMed] Becker, M.; Zittlau, E.; Petz, M. Residue analysis of 15 penicillins and cephalosporins in bovine muscle, kidney and milk by liquid chromatography–tandem mass spectrometry. Anal. Chim. Acta 2004, 520, 19–32. [CrossRef] Ohmori, T.; Suzuki, A.; Niwa, T.; Ushikoshi, H.; Shirai, K.; Yoshida, S.; Ogura, S.; Itoh, Y. Simultaneous determination of eight β-lactam antibiotics in human serum by liquid chromatography–tandem mass spectrometry. J. Chromatogr. B 2011, 879, 1038–1042. [CrossRef] [PubMed] Pikkemaat, M.G.; Rapallini, M.L.; Dijk, S.O.; Elferink, J.W. Comparison of three microbial screening methods for antibiotics using routine monitoring samples. Anal. Chim. Acta 2009, 637, 298–304. [CrossRef] [PubMed] Musser, M.B.; Anderson, K.L.; Rushing, J.E.; Moats, W.A. Potential for milk containing penicillin G or amoxicillin to cause residues in calves. J. Dairy Sci. 2001, 84, 126–133. [CrossRef] Brand, U.; Reinhardt, B.; Rüther, F.; Scheper, T.; Schügerl, K. Bio-field-effect transistors as detectors in flow-injection analysis. Anal. Chim. Acta 1990, 238, 201–210. [CrossRef] Goldfinch, M.J.; Lowe, C.R. Solid-phase optoelectronic sensors for biochemical analysis. Anal. Biochem. 1984, 138, 430–436. [CrossRef] Adrian, J.; Pasche, S.; Voirin, G.; Adrian, J.; Pinacho, D.G.; Font, H.; Sánchez-Baeza, F.; Marco, J.M.P.; Diserens, J.M.; Granier, B. Wavelength-interrogated optical biosensor for multi-analyte screening of sulfonamide, fluoroquinolone, β-lactam and tetracycline antibiotics in milk. Trends Anal. Chem. 2009, 28, 769–777. [CrossRef] Thavarungkul, P.; Dawan, S.; Kanatharana, P.; Asawatreratanakul, P. Detecting penicillin G in milk with impedimetric label-free immunosensor. Biosens. Bioelectron. 2007, 23, 688–694. [CrossRef] [PubMed] Benito-Peña, E.; Moreno-Bondi, M.C.; Orellana, G.; Maquieira, A.; Van Amerongen, A. Development of a novel and automated fluorescent immunoassay for the analysis of β-lactam antibiotics. J. Agric. Food Chem. 2005, 53, 6635–6642. [CrossRef] [PubMed] Park, E.K.; Jung, W.C.; Lee, H.J. Application of a solid-phase fluorescence immunoassay to determine amoxicillin residues in fish tissue. Acta Vet. Hung. 2010, 58, 83–89. [CrossRef] [PubMed] Gamella, M.; Campuzano, S.; Conzuelo, F.; Esteban-Torres, M.; De Las Rivas, A.B.; Reviejo, A.J.; Muñoz, R.; Pingarrón, J.M. An amperometric affinity penicillin-binding protein magnetosensor for the detection of β-lactam antibiotics in milk. Analyst 2013, 138, 2013–2022. [CrossRef] [PubMed] Huth, S.P.; Warholic, P.S.; Devou, J.M.; Chaney, L.K.; Clark, G.H. Parallux™ beta-lactam: A capillary-based fluorescent immunoassay for the determination of penicillin-G, ampicillin, amoxicillin, cloxacillin, cephapirin, and ceftiofur in bovine milk. J. AOAC Int. 2002, 85, 355–364. [PubMed] Campanella, L.; Tomassetti, M.; Sbrilli, R. Benzylpenicillinate liquid membrane ion-selective electrode: Preparation and application to a real matrix (Drug). Ann. Chim. 1986, 26, 483–497. Merola, G.; Martini, E.; Tomassetti, M.; Campanella, L. New immunosensor for β-lactam antibiotics determination in river waste waters. Sens. Actuators B Chem. 2014, 199, 301–313. [CrossRef]

Biosensors 2016, 6, 22

22.

23.

24.

13 of 13

Tomassetti, M.; Martini, E.; Campanella, L.; Favero, G.; Sanzò, G.; Mazzei, F. Lactoferrin determination using flow or batch immunosensor surface plasmon resonance: Comparison with amperometric and screen-printed immunosensor methods. Sens. Actuators B Chem. 2013, 179, 215–225. [CrossRef] Tomassetti, M.; Martini, E.; Campanella, L.; Favero, G.; Sanzó, G.; Mazzei, F. A new surface plasmon resonance immunosensor for Triazine pesticide determination in bovine milk: A comparison with conventional amperometric and screen-printed immunodevices. Sensors 2015, 15, 10255–10270. [CrossRef] [PubMed] Love, J.C.; Estroff, L.A.; Kriebel, J.K.; Nuzzo, R.G.; Whitesides, G.M. Self-Assembled Monolayers of thiolates on metals as a form of nanotechnology. Chem. Rev. 2005, 105, 1103–1169. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).