A label free potentiometric sensor

8132019 A label free potentiometric sensor

httpslidepdfcomreaderfulla-label-free-potentiometric-sensor 17

A Label-Free Potentiometric Sensor Principle for the Detection of AntibodyminusAntigen InteractionsMahir S Ozdemir dagger Marcin Marczakdagger Hugo BohetsDagger K ristien Bonroysect Dirk Roymanssect Lieven Stuyversect

Koen Vanhouttedagger Marcin Pawlak∥ and Eric Bakker ∥

daggerPharmaceutical Development amp Manufacturing Sciences Janssen Research amp Development 2340 Beerse BelgiumDaggerOctens BVBA Sint-Michielskaai 34 2000 Antwerpen Belgiumsect Janssen Infectious Diseases minusDiagnostics BVBA 2340 Beerse Belgium∥Department of Inorganic and Analytical Chemistry University of Geneva Quai E-Ansermet 30 CH-1211 Geneva Switzerland

S Supporting Information

ABSTRACT We report here on a new potentiometric biosensing principle for thedetection of antibody minusantigen interactions at the sensing membrane surface

without the need to add a label or a reporter ion to the sample solution This isaccomplished by establishing a steady-state outward 1047298ux of a marker ion from themembrane into the contacting solution The immunobinding event at the sensingsurface retards the marker ion which results in its accumulation at the membranesurface and hence in a potential response The ion-selective membranes weresurface-modi1047297ed with an antibody against respiratory syncytial virus using click chemistry between biotin moleculesfunctionalized with a triple bond and an azide group on the modi1047297ed poly (vinyl chloride) group of the membrane The bioassay sensor was then built up with streptavidin and subsequent biotinylated antibody A quaternary ammonium ion served as themarker ion The observed potential was found to be modulated by the presence of respiratory syncytial virus bound on themembrane surface The sensing architecture was con1047297rmed with quartz crystal microbalance studies and stir eff ects con1047297rmedthe kinetic nature of the marker release from the membrane The sensitivity of the model sensor was compared to that of acommercially available point-of-care test with promising results

O ver the past few decades important advancements in the1047297eld of diagnostics have been achieved as a result of the

introduction of the enzyme-linked immune absorbent assay (ELISA) and polymerase chain reaction (PCR) based tests

which allowed large scale automation of serological testing andthe ampli1047297cation of nucleic acids respectively

Seventy percent of the in vitro diagnostic (IVD) market iscurrently shaped by tests based on these technologies forinfectious pathogens1 Despite their proven reliability thesetests are time-consuming and normally only performed in alaboratory environment In case of acute infections immediateavailability of the results is often not possible potentially resulting in a loss of precious time before a suitable therapy can

be startedIt has been shown that respiratory syncytial virus (RSV) cancause lower respiratory tract disease (LRD) in infants andpatients for example after hematopoietic cell transplantation(HCT) and result in substantial early mortality2minus6 Early disease detection and intervention even initiated at a time

when the viral load is at its highest can improve diseaseoutcome in previously healthy naturally infected children7

Indeed early information about the infecting agent obtainedfrom rapid diagnostic tests has been shown to signi1047297cantly alterthe management of the patientrsquos illness resulting in a reductionin diagnostic tests performed reduced antibiotic use more

accurate use of antivirals and better patient management ingeneral8

Unfortunately current point-of-care (POC) diagnostic testsdo not always meet the needs of patients and doctors since only qualitative results are obtained and the limit of detection isoften insufficient

Currently no commercial POC tests are based onpotentiometric readout Modern potentiometric sensors areestablished in the determination of inorganic ions in environ-mental (eg Fminus and NH+) and clinical applications (eg pHNa+ K + Ca2+ and Clminus)9 On the other hand electrochemical

biosensors for the label-free detection of antibody minusantigeninteraction have been researched extensively10 The detectionof bacteria by potentiometry has been reported11 and early

work by Rechnitz demonstrated potentiometry to be apromising tool to probe antibody minusantigen interactions12 Stillthe use of ion-selective electrodes (ISEs) as the transductionmethod of a label-free determination of immunobinding is in itsinfancy

In this study a potentiometric polymeric membraneelectrode was investigated as a possible POC diagnostic toolfor the direct detection of antibody minusantigen interactions Zero

Received February 18 2013 Accepted March 27 2013Published March 27 2013

Article

pubsacsorgac

copy 2013 American Chemical Society 4770 dxdoiorg101021ac400514u | Anal Chem 2013 85 4770minus4776



current ion 1047298uxes across ion-selective membranes are today well-understood and are known to often be responsible for thelower detection limit of ISEs Such 1047298uxes have been exploitedfor the design of steptrodes and switchtrodes where largepotential changes around a critical concentration can beobserved13 In the approach presented here such an ion 1047298ux is deliberately set up across the membrane electrode For thispurpose a marker ion is used to make the electrode responsiveto its steady-state concentration at the membrane surface Asimmunobinding occurs on the membrane surface the surface

binding layer imposes a resistance to mass transport of themarker ion This results in an increase in surface concentrationand hence of the measured potential This strategy does notrequire the addition of a marker or reporter ion to the samplesolution Such an apparently label-free approach is potentially

very attractive in view of practical handling of a biosensor in the1047297eld

The operating and performance characteristics of thedeveloped sensor were evaluated with RSV as a model systemCovalent membrane functionalization was achieved with arecently reported click chemistry approach14

EXPERIMENTAL SECTION

Materials High molecular weight poly vinyl chloride(PVC) potassium tetrakis(4-chlorophenyl) borate (KTpClPB)mesamoll streptavidin biotin tetrabutylammonium chloride(TBACl) bovine serum albumin (BSA) and tetrahydrofuran(THF) together with all reagents and solvents used in thesynthesis were purchased from Sigma Aldrich and used withoutfurther puri1047297cation To detect RSV we used Palivizumab(Synagis) as a monoclonal antibody recognizing the fusion (F)protein of the virus (mAb-RSV)15 In1047298uenza A (A2Aichi2682) was detected with CR8020 a monoclonal antibody (mAb-Flu)16 Monoclonal antibodies targeting HIV-1 gp-41(MH-SVM25 ATCC) (mAb-gp41) were all cultured in house

at Janssen Pharmaceutica NV All electrochemical measurements were performed with a

two-electrode con1047297guration using a Polyplast Pro RX (Hamilton) as a reference electrode and the polyethyleneglycol (PEG)-modi1047297ed PVC membrane as a working electrodeMeasurements were performed at ambient temperature with aConsort D130 using a home-built data acquisition software Allpotentiometry and QCM measurements were performed inphosphate buff er (Dulbeccorsquos Phosphate Buff ered SalineDPBS) All aqueous solutions were prepared by dissolvingthe appropriate salts in Milli-Q-puri1047297ed distilled water (DI)

Membrane Preparation and Electrode Construction A disk of 6 mm was cut from a microporous polypropylene1047297 ber support (25 μm thickness 55 porosity 0064 μm pore

size) The disk was glued to a PVC tube of a 6 mm diameterand 1 mm wall size by means of cyclohexanone The tube wasthen dried overnight This was followed by casting 35 μL of themembrane cocktail on top of the disk which contained 325(mm) PVC-N314 655 (mm) mesamoll 2 (mm)potassium tetrakis (4-chlorophenyl) borate (KTpClPB)dissolved in THF This membrane becomes cation responsiveaccording to the Hofmeister selectivity sequence with morelipophilic ions preferred over less lipophilic ones

This membrane composition was found to be the mostsuitable choice as a result of optimization After full evaporationof the solvent modi1047297cation of the surface was performed by so-called click chemistry as described previously14

As a result of click chemistry the azide-modi1047297ed PVC iscoupled with PEG molecules which are very hydrophilic andshield the sensor surface from any interfering particle availablein the sample When attached to the PVC surface alone PEGmolecules prevent nonspeci1047297c adsorption When modi1047297ed witha biotin molecule however resultant PEGminusBiotin linker(PEGminusB) can be used to attach the recognition elements onthe sensor surface by using the high affinity streptavidinminus biotincomplex With dependance on the type of biotinylated antibody selected a particular sensor can be made to be speci1047297c ornonspeci1047297c for a given target molecule The following twosurface modi1047297cation cocktails (ie PEG PEGminusB) were used inpreparing the electrodes

PEGCuSO4middot5H2O [46 mg (019 mmol)] and ascorbic acid[167 mg (095 mmol)] in 6 mL of water PEG derivative 5 mg(002 mmol)14 1047297rst dissolved in 025 mL THF then added tothe 6 mL H2O

PEGminus biotinCuSO4middot5H2O [46 mg (019 mmol)] andascorbic acid [167 mg (095 mmol)] in 6 mL of water

biotinminusPEG derivative 10 mg (002 mmol)14 1047297rst dissolved in025 mL THF and then added to the 6 mL H2O

All electrodes were stored in a humid environment over 24 h After this each electrode was rinsed with distilled water Theinside of the tube was 1047297lled with 500 μL of a 4 mgmL TBAClin DPBS The electrical connection was provided by means of asilver wire The electrodes were conditioned in DPBS for 48 hprior to measurement under constant stirring (300 rpm) tohydrate the membrane

Electrode Membrane Surface Modi1047297cation Thecurrent setup allows working in 4 independent cells In eachmeasurement cell one can place up to six electrodes and areference electrode in 5 mL DPBS under constant stirring at150 rpm After reaching a stable potential (ie potential driftless than 1 mV10 min) the electrodes were exposed tostreptavidin for 1 h under constant stirring at a 1047297nalconcentration of 25 μgmL

After each incubation step DPBS was refreshed Once astable baseline potential was established which usually took upto 30 min electrodes were incubated for 1 h with biotinylatedspeci1047297c or nonspeci1047297c antibodies with a 1047297nal concentration 25 μgmL Before exposing the electrodes to the antigen the buff er was refreshed

Sensor Working Mechanism The mechanism of action is based on the disturbance of the internal marker ion (TBACl)1047298ux as a result of target molecules binding on the sensorsurface As the membrane is very sensitive to the marker ionthat leaches out from the internal solution to the sample side

where the virus is present the binding of the virus to theantibody disturbs the 1047298ux of the marker ion changing its localconcentration in the vicinity of the sensor which then results in

a measurable potential change at the sensor Schematicrepresentation of the sensing principle is depicted in Figure 1

Quartz Crystal Microbalance Measurements QCMexperiments were carried out with a Q-Sense E4 (GothenburgSweden) instrument in a 1047298ow through cell to investigate

whether the click chemistry modi1047297cation antibody interactionand virus adherence were executed in accordance with theexpected electrode architecture In all QCM experiments goldchips uniformly spin-coated with the same PVC membranecocktail used for the potentiometric sensor Resultingmembrane layer thickness was monitored with ellipsometry and only those with desired thickness (around 100 nm) wereused for the QCM measurements

Analytical Chemistry Article

dxdoiorg101021ac400514u | Anal Chem 2013 85 4770minus47764771



Western Blot Analysis and ELISA The a

ffi

nities of themAb-RSV and mAb-gp41 (control) antibodies against thetarget antigen of interest (RSV) were tested with Western Blotanalysis The binding between the anti-RSV F protein antibody and RSV was tested by means of ELISA In these tests thesignal readout was monitored by varying the antibody concentration while keeping the viral concentration constantand vice versa

Potentiometric Measurements Two sets of potentio-metric measurements were performed measurements tosupport the working mechanism and the actual determinationof antibody minusantigen interactions

Measurements Supporting the Mechanism of Action Thefunctionality of the electrodes based on the out1047298ux of themarker ion was 1047297rst tested based on a potentiometric stir eff ect

As the elevation of the local concentration of TBACl in thesample side is precluded as a result of constant stirring whenthe stirring is absent the concentration is expected to increase

Additionally the mechanism of action (MOA) was testedexperimentally during electrode build-up since the addition of streptavidin and antibody binding on the sensor surface are alsoexpected to yield a signal

Experiments were conducted in which electrodes withoutsurface modi1047297cation were exposed to BSA to ascertain theMOA In these experiments ion strength of the marker ion waskept constant at the same concentration on both sides of theISE membrane BSA solutions were tested in the presence of interfering ions

Measurement of Virus Electrodes were exposed to four

consecutive virus spikes containing 103 PFU (plaque formingunits)mL each The time bet ween expositions was 10 min A POC test Binax NOW RSV17 was used to assess the sensitivity of the proposed sensor The Binax NOW RSV is a membrane-

based immunochromatographic technique designed to detectRSV fusion protein antigen in nasal washes and nasopharyngealswab specimens The test is based on anti-RSV antibodiesconjugated to visualizing particles and adsorbed onto anitrocellulose membrane to form a sample line Upon theaddition of the virus sample to the test strip and a 15 minincubation period the signal readout was done In allpotentiometric and parallel tests performed using Binax NOW the identical virus batches were used

THEORY

The sensing principle put forward here for the 1047297rst timeinvolves the continuous release of a label ion from an ion-selective membrane into the aqueous solution The system ishere understood by a steady-state concentration pro1047297le acrossthe membrane that is driven by the extraction of marker ion saltat the backside of the membrane A surface con1047297ned

immunoreaction is here understood to result in an intermediatediff usion layer between the membrane and the aqueousdiff usion layer The concentration pro1047297le across the threediff usion layers is schematically shown in Figure 2

The 1047298ux of marker ion j across the membrane is described by Fick rsquos 1047297rst law as

δ

δ = minus

minus J D

c c( ) (0) j

m jm j

m m jm

m (1)

where the phase labels are shown as m superscripts D jm is the

diff usion coefficient c jm(δ m) the molar concentration at the

backside of the membrane (position δ m see Figure 2) c jm(0) its

concentration at the sample side of the membrane and δ m isthe membrane thickness

The 1047298ux across the diff usion layer where immunoreactionoccurs is written in analogy as

δ

δ = minus

minus J D

c c(0) ( ) j

d jd j

d jd d

d (2)

where position 0 refers to the membrane surface and δ d to theend of this intermediate layer in contact with the samplesolution Finally the 1047298ux across the aqueous diff usion layer is

written similarly as

δ

δ = minus

minus J D

c c( ) (bulk) j j

jd

jaq aq aq aq

aq (3)

where δ d is the aqueous diff usion layer thickness at steady stateIt can be altered by the stirring rate of the solution At position

Figure 1 Schematic showing signal build-up as a result of thedisturbance of the ion 1047298ux set up across the membrane electrode A marker ion (TBACl) is used to make the electrode responsive to itssteady-state concentration at the membrane surface The internalmarker ion leaches out from the sample side reaching an (A)equilibrium and (B) antigenminusantibody binding occurs on themembrane surface resulting in an increase in surface concentrationand hence of the (C) measured potential which eventually (D) levelsoff

Figure 2 Schematic representation of the sensing principle and thesymbols used to describe the concentration changes at each position(layer thicknesses are not to scale) A concentration gradient of amarker ion across the sensing membrane results in its continuousrelease into the sample solution As a biorecognition event takes placethe concentration at the membrane surface is increased owing to the build-up of a diff usion barrier resulting in a potential increase Thedotted line in the binding layer indicates the concentration gradient in

the absence of a binding event





d we assume equilibrium and since both diff usion layers areaqueous we 1047297nd c j

aq (δ d) = c jd (δ d ) Furthermore ion-exchange

with other sample cations is excluded and we approximate theconcentration of marker ion at position 0 of the membrane

with the ion-exchanger concentration c jm(0) = cR

mThe observed boundary potential of the ion-selective

membrane is a function of the concentrations (strictlyactivities) across the aqueousminusmembrane interface and is

written here as

ϕ= Δ + E RT

z F

c

cln

(0)

(0)m

j j

jd

jmPB aq

0

(4)

where Δaq mϕ j

0 is the standard potential of ion transfer across thisinterface (a constant) and R T and F have their establishedmeanings The charge of the marker ion z j is here taken as 1

At steady-state all three 1047298uxes in eqs 1 2 and 3 are equalEliminating c j

d (d ) and J j solving the result for c jd (0) and

inserting it into eq 4 gives the potential response as

ϕδ

δ δ

δ

= Δ + + +

minus

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎧⎨⎩

⎧⎨⎩

⎫⎬⎭

⎫⎬⎭

E RT

z F c

D

D D

c c

ln (bulk)

( ) (0)

m j

j j

jm

j

d

jd

jm m

jm

PB aq 0 aq

m

aq

aq

(5)

In a label-free sensing approach one aims to avoid additionof the label to the sample solution Consequently we usec j

aq (bulk) = 0 and keep the diff usion coefficients and themembrane thickness and compositions constant to obtain fromeq 5

δ δ = + +

⎪ ⎪

⎪ ⎪⎧⎨⎩

⎫⎬⎭

E A RT

z F D Dln

j j

d

jdPB

aq

aq

(6)

where A is a constant A reduction of the apparent diff usion

coefficient in the surface-con1047297ned layer for example by asurface blocking event results in a potential increase Thischange is a direct function of the diff usion coefficient and theadditional diff usion layer thickness Figure 3 illustrates expectedpotential changes on the basis of eq 6 as a function of bindinglayer thickness and reduction in diff usion coefficient in thatlayer

Convective stirring of the solution is expected to give smallersignals (reduced value of δ aq ) In accordance with Figure 3 surface binding events on the scale of a few hundrednanometers are detectable with this approach if it induces animportant retardation of the marker ion

RESULTS AND DISCUSSION

Figure 4 shows the QCM data acquired from diff erent sensorsThe signal at the top panel was acquired from a gold crystalcoated with an antibody recognizing RSV F protein whereasthe middle was from a gold crystal modi1047297ed with an antibody recognizing HIV-1 gp41 protein (as a negative control) The

bottom signal was obtained from a crystal coated only withPEG serving as another control in the experiments Thisexperiment con1047297rms the build-up of the immunoreagents at theion-selective membrane surface The exposure of the

biotinylated membranes to streptavidin and the subsequentstep of binding the biotinylated antibodies were clearly

visualized by QCM PEG-modi1047297ed membranes (bottom)show no response to streptavidin as expected As can be

Figure 3 Calculated potential changes for the surface binding event atthe electrode surface that slows the diff usion of the marker ion(logarithmic diff usion coefficient on the x axis) according to eq 6Diff usion coefficient in the aqueous phase is taken as D j

aq = 10minus5 cm2

sminus1 Stirring the solution decreases the aqueous diff usion layer andhence the potential response

Figure 4 QCM data showing the change in resonance frequency occurring only for the (A) speci1047297c electrode (top) upon RSV exposure whereas the (B) middle (modi1047297ed with an antibody recognizing HIV-1 gp41 protein as a negative control) and the (C) bottom signal (coated only with PEG) show no response to RSV spikes (A and B) respond to streptavidin and antibody injections as aresult of streptavidin binding to the biotinylated surfaces (PEG-B) andsubsequent biotinylated antibody binding to streptavidin while (C)PEG electrode shows no response as expected





seen from Figure 4 a decrease in the resonance frequency uponexposure to 103 PFU RSV was observed only for the speci1047297celectrode (top) This indicates that the sensor surface wasmodi1047297ed as desired

The speci1047297city of the anti-RSV F antibody was tested with Western Blot analysis The data con1047297rmed a speci1047297c interactionof this antibody with the denatured F1 part of the RSV Fprotein (Figure 5 left) The binding between the anti-RSV F

antibody and the F protein in its native pretriggeredconformation was also tested by means of an ELISA In theseexperiments the signal readout was monitored by varying the

Ab concentration 01minus10 ( μgmL) while keeping the viralconcentration constant at 106 PFU and vice versa (ie primary

Ab of 10 μgmL RSV of 103minus

106 PFU) (Figure 5 right) Asseen in Figure 5 (right bottom) the signal increases when theantibody or the virus concentration are elevated indicating aninteraction between the two

The proposed response mechanism of the potentiometric biosensor principle was initially con1047297rmed based on a stir-eff ect As seen in Figure 6 A when the solution was stirred the signalread-out was low This is a consequence of the fact thatconstant stirring prevented the ion 1047298ux building up the localconcentration as predicted by eq 6 On the other hand whensample stirring was turned off the potential increased owing tothe increasing concentration of marker ion at the sensorsurface

The mechanism laid out in the theoretical part was

independently supported with BSA as a model analyte Inthis experiment two diff erent electrodes placed in the samemeasurement cell were exposed to BSA to test their responsesto any possible interaction with this molecule (Figure 6B) Inthis case the membrane consisted of a PVC which was notmodi1047297ed with any antibody The electrodes were 1047297rst exposedto a BSA 1047297ltrate (ie not containing any BSA) to con1047297rm thelack of potential increase in the absence of BSA and to rule outany impact of sample impurities Indeed no potential increase

was observed in response to the BSA 1047297ltrate (data not shown)This was followed by a stir eff ect as shown in Figure 6B to testelectrode response according to eq 5 Signals increased whenthe stirring was off whereas they all decreased as the stirring

was on A BSA spike of (50 mgmL) caused electrode signals torise The fact that BSA 1047297ltrate injections did not give rise to ameasurable potential change indicates that the observed signalchanges were brought about by the BSA binding on themembrane surface It is important to emphasize that noresponse was observed when the marker ion was already

present in the sample solution This is in accordance with theprinciple set forth above and indicates that a potential responseis indeed induced by retarding the marker ion and increasing itsaccumulation at the membrane surface Figure 7 (top) presents

Figure 5 Western Blot and ELISA tests showing the interaction between the anti-RSV F antibody and the F protein in its denaturedand native pretriggered conformation Dark-blue and light bluerepresent the responses with and without mAb-RSV primary antibodyrespectively

Figure 6 (A) Stir eff ect supporting the proposed mechanism of actionThe constant stirring prevented the ion 1047298ux from building up the localconcentration which was increased when the stirring was off (B)Clear signal responses following stir-eff ect recorded potentiometri-cally from a set of two electrodes in the same measurement cell as aresult of BSA binding to the sensor surface Potential of each electrodeindicated by red and blue respectively increases after BSA injection

Figure 7 Data acquired from (A) speci1047297c and (B C D) a group of 3types of nonspeci1047297c (ie control) electrodes The speci1047297c electrode was a PEG-B membrane modi1047297ed with mAb-RSV whereas (B) the1047297rst control was a PEG-B membrane modi1047297ed with mAb-gp41 Theother two sets were controls based on PEG-only membranes treated with mAb-RSV and mAb-gp41 recognition elements respectively





the experimental data on RSV detection using a set of (A)speci1047297c electrodes and three sets of control electrodes wereused [ie total number of 39 electrodes 16 of which werespeci1047297c (A)] As seen in this graph while (B C D) controlgroups show no response to the lowest virus concentration of 10 PFUmL the (A) speci1047297c electrodes respond to the virus A schematic of diff erent sensor architectures can be seen in Figure7 (bottom)

Among three control groups (B C and D) two controlgroups (C and D) showed no response to 10 and 100 PFUmLRSV injections It is also observed in Figure 7 that thediff erence in signal response between the speci1047297c (A) andcontrol electrodes (B C D) becomes larger as the virusconcentration increases (eg 1000 PFUmL) It should benoted that control group B was identical to the speci1047297c set A except for the monoclonal antibody used (ie mAb-gp41instead of mAb-RSV) so as to ascertain that diff erent signalresponses correspond to the electrode speci1047297city Additionallycontrol groups C and D were employed to monitor any interaction which might be due to mAbs binding on themembrane surface despite the PEG layers used

A comparison of the potentiometric signal readouts of a

speci1047297c and nonspeci1047297c electrode pair recorded within the samemeasurement cell is given in Figure 8 As shown in this 1047297gure

the signal amplitude of the speci1047297c electrode increases as theamount of viral particles in the cell is increased The potentialrecorded from the nonspeci1047297c electrode follows a steady

baseline except for the highest viral load This is probably dueto nonspeci1047297c interactions (ie background signal) as a resultof very high concentration of viral particles the measurementcell As these speci1047297c and nonspeci1047297c electrodes were exposed

to the same viral loading diff erent signal responses correspondto the electrode speci1047297city

To evaluate the sensitivity of our sensor we investigated the viral concentrations in batch mode with an orthogonaltechnique quantitative reverse-transcriptase polymerase chainreaction (qRT-PCR) The detection limit of the proposedsensor system was found to be approximately 3 logarithmicunits higher than that of the qRT-PCR (data not shown)

We used identical RSV batches in order to compare theresponse and sensitivity of the proposed sensor with thecommercial RSV test Table 1 shows the responses recordedfrom a pair of speci1047297c (ie PEG-B membranes modi1047297ed withmAb-RSV) as well as nonspeci1047297c electrodes (ie PEG-B

membranes modi1047297ed with mAb-gp41) to a set of RSV spikes of varying concentrations Plus and minus signs indicate thepresence and the absence of a signal response respectively Asshown in Table 1 only the electrode modi1047297ed with mAb-RSV responded to viral concentrations of 103 and 104 PFU It should

be emphasized that nonspeci1047297c electrodes and the Binax NOW test started responding to the RSV virus only at a viral load of 105 PFUmL

We conducted a limited number of preliminary experimentson the detection of In1047298uenza A (A2Aichi2682) to test theuniversality of the developed sensor by using CR8020 mAbsand mAb-gp41 (ie control) A design of experiment (DOE)

was conducted with a total of eight cells each containing threespeci1047297c and three nonspeci1047297c electrodes (ie 48 electrodes intotal) Electrodes were exposed to In1047298uenza A virus and to RSV (see Figure 2S of the Supporting Information) The majority of the electrodes were found to be clearly more responsive toIn1047298uenza A virus than to RSV although nonspeci1047297cinteractions were also observed especially at high virusconcentrations

CONCLUSIONS

A new potentiometric biosensor principle has been evaluated toexplore the concept of modulating the mass transport of amarker ion to detect antigenminusantibody interactions at the ion-selective membrane surface This was demonstrated by aninfectious disease model for RSV The proposed mechanism of action was evidenced by potentiometric experiments Theelectrode architecture was veri1047297ed with QCM measurements

Most electrochemical measurement systems require theaddition of an indicator ion to the sample solution (egCa2+)18 or a redox couple (eg amperometry) to recognize theantibody minusantigen interaction In this sensor principle themarker ion is delivered by the membrane in direction of theanalyte solution by diff usion through the membraneConsequently the evaluated sensing system renders label-free

detection of antibody minus

antigen interactions possible This can inturn make the sample preparation process less cumbersome forpoint-of-care applications

The sensor principle requires eff ective coverage of thesensing surface upon immunoreaction The area of the currentsensor is very large (a few millimeters in diameter) compared tothe 100 nm virus one wishes to detect Potentiometricmicroelectrodes with size ranges in the submicrometer rangehave been known for many years19 and should provide a muchmore favorable membrane to virus area than the systemsstudied here Miniaturization of the sensing system is expectedto increase the sensitivity of the sensor and hence lower thelimit of detection (LOD)

Figure 8 A comparison of the potentiometric signal readouts of an

electrode treated with mAb-RSV (solid line) and a control electrode with mAb-gp 41 (dashed line) recorded within the same measurementcell The signal amplitude of the speci1047297c electrode increases as the virus concentration in the cell is increased whereas that from thenonspeci1047297c electrode follows a steady baseline except for the highest viral load

Table 1 A Summary of Signal Responses Recorded from aPair of Electrodes to a Set of RSV Spikes Containing Varying

Amounts of Viral Particlesa

RSV (PFU) PEG-B Ab-RSV PEG-B Ab-gp41 Binax NOW

103 + minus minus

104 + minus minus

105 + + +

5 times 105 + + +aPlus and minus signs indicate the presence and absence of a signalrespectively Only the electrode treated with mAb-RSV responded to virus concentrations of 103 and 104 PFU





The delivery of the target molecules to the sensor surfaceachieved by stirring in the current study can be facilitated by using a micro1047298uidic channel which can transport the targetmolecule to the sensor surface more eff ectively

Despite the promising results presented in this worknonspeci1047297c interactions on the polymer surface remain apotential limitation of the current design Signals recorded fromPEG-only membranes treated with mAb-RSV (Figure 7)indicate a suboptimum shielding of the membrane surface

which needs to be overcome before the technique can beapplied in real application with more complex biologicalsamples

Although not investigated in our current study the Abisotype (eg use of IgG versus IgM) may have an impact onthe sensitivity of the sensor as a result of diff erent avidity of the

Abs Using multimeric antibodies as compared to theirmonomeric counterparts could improve sensitivity as moreantigen binding possibly results in more pronounced changes of the ion 1047298ux at the surface of the sensor Additionally the size of the antigen of interest can aff ect signal transfer as largermolecules can be expected to cause larger 1047298ux changes of themarker ion the membrane surface and hence yield larger

signalsThis early study aimed at exploring the concept of an

incorporated marker ion 1047298ux in potentiometric ion-selectivesensors as a new biosensor approach The data con1047297rm thefeasibility of detecting antibody minusantigen interactions by potentiometry To our knowledge this is the 1047297rst reporteduse of potentiometry on the basis of passive ion 1047298uxes inprobing antibody minusantigen interactions Such an apparently label-free approach may become an attractive platform forfuture progress in bioaffinity sensor research

ASSOCIATED CONTENT


Synthetic details and additional experiments This material isavailable free of charge via the Internet at httppubsacsorg

AUTHOR INFORMATION

Corresponding Author

E-mail ericbakkerunigech

Notes

The authors declare no competing 1047297nancial interest

ACKNOWLEDGMENTS

The authors thank the IWT (Innovatie door Wetenschap enTechnologie) for 1047297nancial support (WTO 080329)

REFERENCES(1) Kalorama Information Worldwide POC Diagnostic Test Markets(2) Martino R Porras R P Rabella N Williams J V Ra mila E

Margall N Labeaga R Crowe J E Jr Coll P Sierra J Biol Blood Marrow Transplant 2005 11 781minus796

(3) Nichols W G Gooley T Boeckh M Biol Blood MarrowTransplant 2001 7 11Sminus15S

(4) Khanna N Widmer A F Decker M Steffen I Halter JHeim D Weisser M Gratwohl A Fluckiger U Hirsch H H Clin

Infect Dis 2008 46 402minus412(5) McCarthy A J Kingman H M Kelly C Taylor G S Caul E

O Grier D Moppett J Foot A B Cornish J M Oakhill ASteward C G Pamphilon D H Marks D I Bone MarrowTransplant 1999 24 1315minus1322

(6) Erard V Chien J W Kim H W Nichols W G Flowers ME Martin P J Corey L Boeckh M J Infect Dis 2006 193 1619minus

1625(7) Saleeby C M Bush A J Harrison L M Aitken J A

DeVincenzo J P J Infect Dis 2011 204 996minus1002(8) Bonner A B Monroe K W Talley L I Klasner A E

Kimberlin D W Pediatrics 2003 112 363minus7(9) Bakker E Pretsch E Chemistry 2002 74 420A minus426A

(10) Vestergaard M Kagan K Tamiya E Sensors 2007 7 3442minus

3458(11) Zelada-Guillen G A Riu J Duzgun A Rius F X Angew

Chem Int Ed 2009 48 7334minus7337(12) DrsquoOrazio P Rechnitz G A Anal Chim Acta 1979 109 25minus

31(13) Vigassy T Morf W E Badertscher M Ceresa A De Rooij

N F Pretsch E Sens Actuators B 2001 76 477minus482(14) Pawlak M Mistlberger G Bakker E J Mater Chem 2012 22

12796minus12801(15) Johnson S Oliver C Prince G A Hemming V G Pfarr D

S Wang S C Dormitzer M OrsquoGrady J Koenig S Tamura J K Woods R Bansal G Couchenour D Tsao E Hall W C Young J F J Infect Dis 1997 176 1215minus1224

(16) Ekiert D C Friesen R H E Bhabha G Kwaks T Jongeneelen M Yu W Ophorst C Cox F Korse H J W M

Brandenburg B Vogels R Brakenhoff J P J Kompier R KoldijkM H Cornelissen L A H M Poon L L M Peiris M Koudstaal W Wilson I A Goudsmit J Science 2011 333 843minus850

(17) binaxreg NOW reg RSV Rapid Test Binax Inc Portland Maine1998 6

(18) Xu Y Bakker E Langmuir 2009 25 568minus573(19) Manz A Simon W Anal Chem 1987 59 74minus79





current ion 1047298uxes across ion-selective membranes are today well-understood and are known to often be responsible for thelower detection limit of ISEs Such 1047298uxes have been exploitedfor the design of steptrodes and switchtrodes where largepotential changes around a critical concentration can beobserved13 In the approach presented here such an ion 1047298ux is deliberately set up across the membrane electrode For thispurpose a marker ion is used to make the electrode responsiveto its steady-state concentration at the membrane surface Asimmunobinding occurs on the membrane surface the surface

binding layer imposes a resistance to mass transport of themarker ion This results in an increase in surface concentrationand hence of the measured potential This strategy does notrequire the addition of a marker or reporter ion to the samplesolution Such an apparently label-free approach is potentially

very attractive in view of practical handling of a biosensor in the1047297eld

The operating and performance characteristics of thedeveloped sensor were evaluated with RSV as a model systemCovalent membrane functionalization was achieved with arecently reported click chemistry approach14

EXPERIMENTAL SECTION

Materials High molecular weight poly vinyl chloride(PVC) potassium tetrakis(4-chlorophenyl) borate (KTpClPB)mesamoll streptavidin biotin tetrabutylammonium chloride(TBACl) bovine serum albumin (BSA) and tetrahydrofuran(THF) together with all reagents and solvents used in thesynthesis were purchased from Sigma Aldrich and used withoutfurther puri1047297cation To detect RSV we used Palivizumab(Synagis) as a monoclonal antibody recognizing the fusion (F)protein of the virus (mAb-RSV)15 In1047298uenza A (A2Aichi2682) was detected with CR8020 a monoclonal antibody (mAb-Flu)16 Monoclonal antibodies targeting HIV-1 gp-41(MH-SVM25 ATCC) (mAb-gp41) were all cultured in house

at Janssen Pharmaceutica NV All electrochemical measurements were performed with a

two-electrode con1047297guration using a Polyplast Pro RX (Hamilton) as a reference electrode and the polyethyleneglycol (PEG)-modi1047297ed PVC membrane as a working electrodeMeasurements were performed at ambient temperature with aConsort D130 using a home-built data acquisition software Allpotentiometry and QCM measurements were performed inphosphate buff er (Dulbeccorsquos Phosphate Buff ered SalineDPBS) All aqueous solutions were prepared by dissolvingthe appropriate salts in Milli-Q-puri1047297ed distilled water (DI)

Membrane Preparation and Electrode Construction A disk of 6 mm was cut from a microporous polypropylene1047297 ber support (25 μm thickness 55 porosity 0064 μm pore

size) The disk was glued to a PVC tube of a 6 mm diameterand 1 mm wall size by means of cyclohexanone The tube wasthen dried overnight This was followed by casting 35 μL of themembrane cocktail on top of the disk which contained 325(mm) PVC-N314 655 (mm) mesamoll 2 (mm)potassium tetrakis (4-chlorophenyl) borate (KTpClPB)dissolved in THF This membrane becomes cation responsiveaccording to the Hofmeister selectivity sequence with morelipophilic ions preferred over less lipophilic ones

This membrane composition was found to be the mostsuitable choice as a result of optimization After full evaporationof the solvent modi1047297cation of the surface was performed by so-called click chemistry as described previously14

As a result of click chemistry the azide-modi1047297ed PVC iscoupled with PEG molecules which are very hydrophilic andshield the sensor surface from any interfering particle availablein the sample When attached to the PVC surface alone PEGmolecules prevent nonspeci1047297c adsorption When modi1047297ed witha biotin molecule however resultant PEGminusBiotin linker(PEGminusB) can be used to attach the recognition elements onthe sensor surface by using the high affinity streptavidinminus biotincomplex With dependance on the type of biotinylated antibody selected a particular sensor can be made to be speci1047297c ornonspeci1047297c for a given target molecule The following twosurface modi1047297cation cocktails (ie PEG PEGminusB) were used inpreparing the electrodes

PEGCuSO4middot5H2O [46 mg (019 mmol)] and ascorbic acid[167 mg (095 mmol)] in 6 mL of water PEG derivative 5 mg(002 mmol)14 1047297rst dissolved in 025 mL THF then added tothe 6 mL H2O

PEGminus biotinCuSO4middot5H2O [46 mg (019 mmol)] andascorbic acid [167 mg (095 mmol)] in 6 mL of water

biotinminusPEG derivative 10 mg (002 mmol)14 1047297rst dissolved in025 mL THF and then added to the 6 mL H2O

All electrodes were stored in a humid environment over 24 h After this each electrode was rinsed with distilled water Theinside of the tube was 1047297lled with 500 μL of a 4 mgmL TBAClin DPBS The electrical connection was provided by means of asilver wire The electrodes were conditioned in DPBS for 48 hprior to measurement under constant stirring (300 rpm) tohydrate the membrane

Electrode Membrane Surface Modi1047297cation Thecurrent setup allows working in 4 independent cells In eachmeasurement cell one can place up to six electrodes and areference electrode in 5 mL DPBS under constant stirring at150 rpm After reaching a stable potential (ie potential driftless than 1 mV10 min) the electrodes were exposed tostreptavidin for 1 h under constant stirring at a 1047297nalconcentration of 25 μgmL

After each incubation step DPBS was refreshed Once astable baseline potential was established which usually took upto 30 min electrodes were incubated for 1 h with biotinylatedspeci1047297c or nonspeci1047297c antibodies with a 1047297nal concentration 25 μgmL Before exposing the electrodes to the antigen the buff er was refreshed

Sensor Working Mechanism The mechanism of action is based on the disturbance of the internal marker ion (TBACl)1047298ux as a result of target molecules binding on the sensorsurface As the membrane is very sensitive to the marker ionthat leaches out from the internal solution to the sample side

where the virus is present the binding of the virus to theantibody disturbs the 1047298ux of the marker ion changing its localconcentration in the vicinity of the sensor which then results in

a measurable potential change at the sensor Schematicrepresentation of the sensing principle is depicted in Figure 1

Quartz Crystal Microbalance Measurements QCMexperiments were carried out with a Q-Sense E4 (GothenburgSweden) instrument in a 1047298ow through cell to investigate

whether the click chemistry modi1047297cation antibody interactionand virus adherence were executed in accordance with theexpected electrode architecture In all QCM experiments goldchips uniformly spin-coated with the same PVC membranecocktail used for the potentiometric sensor Resultingmembrane layer thickness was monitored with ellipsometry and only those with desired thickness (around 100 nm) wereused for the QCM measurements






ffi










THEORY




δ

δ = minus

minus J D

c c( ) (0) j

m jm j

m m jm

m (1)






δ

δ = minus

minus J D

c c(0) ( ) j

d jd j

d jd d

d (2)



δ

δ = minus

minus J D

c c( ) (bulk) j j

jd

jaq aq aq aq

aq (3)















written here as

ϕ= Δ + E RT

z F

c

cln

(0)

(0)m

j j

jd

jmPB aq

0

(4)

where Δaq mϕ j





ϕδ

δ δ

δ

= Δ + + +

minus

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎧⎨⎩

⎧⎨⎩

⎫⎬⎭

⎫⎬⎭

E RT

z F c

D

D D

c c

ln (bulk)

( ) (0)

m j

j j

jm

j

d

jd

jm m

jm

PB aq 0 aq

m

aq

aq

(5)



δ δ = + +

⎪ ⎪

⎪ ⎪⎧⎨⎩

⎫⎬⎭

E A RT

z F D Dln

j j

d

jdPB

aq

aq

(6)










aq = 10minus5 cm2







































CONCLUSIONS











103 + minus minus

104 + minus minus

105 + + +













ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS




























ffi










THEORY




δ

δ = minus

minus J D

c c( ) (0) j

m jm j

m m jm

m (1)






δ

δ = minus

minus J D

c c(0) ( ) j

d jd j

d jd d

d (2)



δ

δ = minus

minus J D

c c( ) (bulk) j j

jd

jaq aq aq aq

aq (3)















written here as

ϕ= Δ + E RT

z F

c

cln

(0)

(0)m

j j

jd

jmPB aq

0

(4)

where Δaq mϕ j





ϕδ

δ δ

δ

= Δ + + +

minus

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎧⎨⎩

⎧⎨⎩

⎫⎬⎭

⎫⎬⎭

E RT

z F c

D

D D

c c

ln (bulk)

( ) (0)

m j

j j

jm

j

d

jd

jm m

jm

PB aq 0 aq

m

aq

aq

(5)



δ δ = + +

⎪ ⎪

⎪ ⎪⎧⎨⎩

⎫⎬⎭

E A RT

z F D Dln

j j

d

jdPB

aq

aq

(6)










aq = 10minus5 cm2







































CONCLUSIONS











103 + minus minus

104 + minus minus

105 + + +













ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS

































written here as

ϕ= Δ + E RT

z F

c

cln

(0)

(0)m

j j

jd

jmPB aq

0

(4)

where Δaq mϕ j





ϕδ

δ δ

δ

= Δ + + +

minus

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎪

⎧⎨⎩

⎧⎨⎩

⎫⎬⎭

⎫⎬⎭

E RT

z F c

D

D D

c c

ln (bulk)

( ) (0)

m j

j j

jm

j

d

jd

jm m

jm

PB aq 0 aq

m

aq

aq

(5)



δ δ = + +

⎪ ⎪

⎪ ⎪⎧⎨⎩

⎫⎬⎭

E A RT

z F D Dln

j j

d

jdPB

aq

aq

(6)










aq = 10minus5 cm2







































CONCLUSIONS











103 + minus minus

104 + minus minus

105 + + +













ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS



























































CONCLUSIONS











103 + minus minus

104 + minus minus

105 + + +













ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS








































CONCLUSIONS











103 + minus minus

104 + minus minus

105 + + +













ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS


































ASSOCIATED CONTENT



AUTHOR INFORMATION



Notes


ACKNOWLEDGMENTS

























Documents

A label free potentiometric sensor