Extraction or Adsorption? Voltammetric Assessment of Protamine

Apr 30, 2015 - Cation-exchange extraction of polypeptide protamine from water into an ionophore-based polymeric membrane has been hypothesized as the ...
0 downloads 0 Views 903KB Size
Download PDF
Article pubs.acs.org/ac

Extraction or Adsorption? Voltammetric Assessment of Protamine Transfer at Ionophore-Based Polymeric Membranes Mohammed B. Garada, Benjamin Kabagambe, and Shigeru Amemiya* Department of Chemistry, University of Pittsburgh, 219 Parkman Avenue, Pittsburgh, Pennsylvania 15260, United States S Supporting Information *

ABSTRACT: Cation-exchange extraction of polypeptide protamine from water into an ionophore-based polymeric membrane has been hypothesized as the origin of a potentiometric sensor response to this important heparin antidote. Here, we apply ion-transfer voltammetry not only to confirm protamine extraction into ionophore-doped polymeric membranes but also to reveal protamine adsorption at the membrane/water interface. Protamine adsorption is thermodynamically more favorable than protamine extraction as shown by cyclic voltammetry at plasticized poly(vinyl chloride) membranes containing dinonylnaphthalenesulfonate as a protamine-selective ionophore. Reversible adsorption of protamine at low concentrations down to 0.038 μg/mL is demonstrated by stripping voltammetry. Adsorptive preconcentration of protamine at the membrane/water interface is quantitatively modeled by using the Frumkin adsorption isotherm. We apply this model to ensure that stripping voltammograms are based on desorption of all protamine molecules that are transferred across the interface during a preconcentration step. In comparison to adsorption, voltammetric extraction of protamine requires ∼0.2 V more negative potentials, where a potentiometric super-Nernstian response to protamine is also observed. This agreement confirms that the potentiometric protamine response is based on protamine extraction. The voltammetrically reversible protamine extraction results in an apparently irreversible potentiometric response to protamine because back-extraction of protamine from the membrane extremely slows down at the mixed potential based on cation-exchange extraction of protamine. Significantly, this study demonstrates the advantages of ion-transfer voltammetry over potentiometry to quantitatively and mechanistically assess protamine transfer at ionophore-based polymeric membranes as foundation for reversible, selective, and sensitive detection of protamine.

T

potential response10 based on a nonequilibrium cationexchange reaction between aqueous protamine and a membranous co-ion, i.e., sodium ions.6 Inherently, the superNernstian response is not selective for protamine and is sensitive to both protamine and a co-ion. In addition, a potentiometric response to protamine is somehow irreversible, which has been ascribed to irreversible extraction of protamine into an ionophore-based polymeric membrane.6−9,11−18 Recently, we applied ion-transfer voltammetry as a powerful electroanalytical method to study protamine transfer across the interface between two immiscible electrolyte solutions (ITIES). Advantageously, a voltammetric response to protamine was obtained by reversibly and selectively transferring only protamine across the ITIES despite the presence of highconcentration aqueous electrolytes19−25 in contrast to irreversible and nonselective potentiometry.6−9 Specifically, we obtained cyclic voltammograms (CVs) of protamine transfer

he biomedical importance of polypeptide protamine as a heparin antidote1 has driven extensive development of potentiometric protamine sensors for more than 2 decades.2,3 Originally, a potentiometric protamine-sensitive electrode was developed by Meyerhoff and co-workers4 immediately after their pioneering work on the heparin counterpart.5 Their protamine sensor initially employed an ionophore-free polymeric membrane based on a lipophilic cation exchanger, tetrakis(4-chlorophenyl)borate, to yield a super-Nernstian response to protamine with a charge of ∼+20 in contrast to a small Nernstian response with little sensing utility.6 A clinical application of a potentiometric protamine-sensitive electrode was demonstrated by replacing the cation exchanger with a protamine-selective charged ionophore, dinonylnaphthalenesulfonate (DNNS), for end-point detection of heparin neutralization with protamine in undiluted whole blood samples.7,8 Moreover, a detection limit was lowered from 0.5 to 0.02 μg/ mL by rotating a DNNS-based membrane electrode, which enhanced mass transport of protamine from an aqueous sample to the membrane/water interface.9 Dependence of a detection limit on mass transport of protamine confirmed that the resultant super-Nernstian response is due to the mixed© XXXX American Chemical Society

Received: February 15, 2015 Accepted: April 30, 2015

A

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

■

by externally polarizing a micropipet-supported interface26,27 between water and the organic solution of lipophilic electrolytes. Without an ionophore, protamine was extracted into the polar nitrobenzene phase but was only adsorbed at the interface between water and the nonpolar organic phase, i.e., 1,2dichloroethane and 1,6-dichlorohexane.19,20 Samec and coworkers also observed voltammetric adsorption and desorption of protamine at macroscopic 1,2-dichloroethane/water interfaces without a protamine-selective ionophore.21 By contrast, we demonstrated protamine extraction into the 1,2-dichloroethane phase doped with DNNS by employing micropipet CV and amperometry.22,23 The CVs and chronoamperograms of protamine transfer were quantitatively analyzed to reveal that DNNS forms a 20:1 complex with protamine to facilitate extraction of the electrically neutral and lipophilic complex.22 Protamine extraction into the 1,2-dichloroethane phase across the macroscopic ITIES was also observed voltammetrically by Osakai and co-workers, who used not only DNNS24 but also bis(2-ethylhexyl) sulfosuccinate25 as protamine ionophores. In this work, we apply ion-transfer voltammetry for the first time to obtain a greater mechanistic understanding of protamine transfer at ionophore-based polymeric membranes. Interestingly, we observe not only protamine extraction into the membranes but also protamine adsorption at the membrane/ water interface. Our observation of both extraction and adsorption of protamine at the same membrane/water interface is unique because protamine was either extracted or adsorbed at the ITIES in previous voltammetric studies.19−25 In addition, we find that protamine adsorption is thermodynamically much more favorable than protamine extraction. Nevertheless, protamine extraction dominantly contributes to a superNernstian response to protamine, which has been hypothesized in previous potentiometric studies6−9,11−18 and is confirmed experimentally in this study by employing the mixed-potential model.10,28,29 Moreover, we reveal that a mixed-potential response based on thermodynamically reversible protamine extraction is kinetically irreversible because back-extraction of protamine from a membrane slows down dramatically at the mixed potential. Experimentally, we employ polarizable membrane/water interfaces to voltammetrically study protamine transfer, which contrasts to classical potentiometry at nonpolarizable membrane/water interfaces.6−9 Specifically, a plasticized poly(vinyl chloride) (PVC) membrane is doped with tetradodecylammonium (TDDA) salts of DNNS and tetrakis(pentafluorophenyl)borate (TFAB) as an ionophore and a supporting electrolyte, respectively, to eliminate cation-exchange sites from the membrane. The DNNS-doped PVC membrane is supported by a gold electrode coated with a lipophilic analogue of poly(3,4-ethylenedioxythiophene) (PEDOT) as a voltammetric cation-to-electron transducer.30−33 Adsorption and desorption of protamine at a high concentration of 28 μg/mL are readily observed by cyclic voltammetry. Adsorptive stripping voltammograms of protamine at lower concentrations down to 0.038 μg/mL are measured and quantitatively analyzed to demonstrate that DNNS-facilitated protamine adsorption follows the Frumkin isotherm.21 In addition, protamine is voltammetrically extracted at more negative potentials, where super-Nernstian responses to protamine are observed when cation-exchange sites are added to the DNNS-doped plasticized PVC membrane used for voltammetry.

THEORY Here, we develop a model based on adsorptive preconcentration of low-concentration protamine at a polymeric membrane doped with an excess amount of ionophore for stripping voltammetry. We assume that equilibrium adsorption of protamine at the membrane/water interface is maintained to follow the Frumkin isotherm, i.e.,21 βc w(0, t ) =

⎛ Γ(t ) ⎞ Γ(t ) exp⎜ −g ⎟ Γmax − Γ(t ) ⎝ Γmax ⎠

(1)

with ⎛ ΔG 0 ⎞ ads ⎟ β = exp⎜ − ⎝ RT ⎠

(2)

where cw(0,t) is the concentration of protamine at the aqueous side of the interface at a preconcentration time, t, Γ(t) is the corresponding total surface excess concentration of a protamine−ionophore complex, Γmax is the saturated surface concentration of the complex, g represents the strength of interactions among complexes at the interface, and ΔG0ads is the potential-dependent standard Gibbs energy of interfacial complexation between protamine and ionophores. Consequently, the current carried by protamine during a preconcentration step, il(t), is limited by mass transport of lowconcentration protamine from the aqueous solution to the interface, which is accelerated by rotating an electrode to yield il(t ) = 0.62z iFADw 2/3ω1/2ν−1/6[c0 − c w(0, t )]

(3)

where zi is the charge transferred by an protamine molecule across the interface, Dw and cw are the diffusion coefficient and concentration of protamine in the aqueous phase, respectively, ω is the rotation speed, and ν is the viscosity of the aqueous electrolyte solution. Noticeably, the preconcentration potential across the membrane/water interface is negative enough to completely deplete protamine at the aqueous side of the interface at t = 0, i.e., cw(0,0) = 0, to yield the Levich equation from eq 3 as given by il(0) = il = 0.62z iFADw 2/3ω1/2ν−1/6c0

(4)

Since protamine is only adsorbed at the preconcentration potential, il(t) is related to Γ(t) as t

Γ(t ) =

∫0 il(τ ) dτ z iFA

=

Q a(t ) z iFA

(5)

where Qa(t) is the total charge based on protamine adsorption at the interface. Overall, combination of eq 1 with eqs 3 and 5 yields ⎛ dQ a(t ) ⎤ Q a(t ) Q a(t ) ⎞ 1⎡ ⎥= ⎢i l − ⎟ exp⎜ −g il ⎣ dt ⎦ c0β[z iFA Γmax − Q a(t )] ⎝ z iFA Γmax ⎠

(6)

Equation 6 is solved numerically by using Mathematica 8 (Wolfram Research, Champaign, IL, USA) to plot Qa(t) against t with Qa(0) = 0 as the initial condition.

■

EXPERIMENTAL SECTION Chemicals. Protamine sulfate (Grade III from herring, histone-free), tetradodecylammonium (TDDA) bromide (≥99.0%), PVC (high molecular weight, selectophore), 2B

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry nitrophenyl octyl ether (oNPOE, selectophore, ≥99.0%), bis(ethylhexyl)sebacate (DOS, selectophore, ≥97.0%), and tetrabutylammonium (TBA) chloride (≥97.0%) were purchased from Aldrich (Milwaukee, WI, USA). LiCl (≥98.5%) was purchased from Fisher Scientific (Fair Lawn, NJ, USA). Tetramethylammonium (TMA) hydroxide (25% in H2O, Trace Select Ultra, Aldrich) and hydrochloric acid (≥30%, Trace Select, Aldrich) were used as high-purity supporting electrolytes for stripping voltammetry and potentiometry. The potassium salt of TFAB (Boulder Scientific Co., Mead, CO, USA) was used to yield TDDATFAB as organic supporting electrolytes34 and also to serve as a cation exchanger for potentiometry. A TDDA salt of DNNS was prepared by metathesis of TDDA bromide and dinonylnaphthalene sulfonic acid (Nacure 1052, King Industries, Norwalk, CT, USA) as reported elsewhere.22 All sample solutions were prepared by using ultrapure water (18.2 MΩ·cm and TOC of 3 ppb) from the Milli-Q Advantage A10 system equipped with Q-Gard T1 pak and Quantum TEX cartridge (EMD Millipore, Billerica, MA, USA).33 The Milli-Q system was fed with the water (15.0 MΩ·cm) purified from tap water by using Elix 3 Advantage (EMD Millipore). Sample solutions were prepared by using polypropylene volumetric flasks (VITLAB GmbH, Grossostheim, Germany) and poured into polypropylene beakers (VITLAB GmbH) for electrochemical measurement. To prevent airborne contamination during storage, the flasks were filled with Milli-Q water and the beakers were immersed in Milli-Q water filled in polypropylene wide-mouth jars (Thermo Scientific, Marietta, OH, USA). Electrode Modification. A 5 mm diameter gold disk attached to a rotating disk electrode tip (Pine Research Instrumentation, Raleigh, NC, USA) was modified with an oxidatively doped film of tetradecyl- or decyl-substituted poly(3,4-ethylenedioxythiophene) (PEDOT-Cn) and then with an oNPOE/PVC or DOS/PVC membrane. A relatively thick PEDOT film was electrodeposited on a gold electrode as reported elsewhere31 to prevent significant polarization of the conducting polymer film for voltammetric ion-to-electron transduction. A relatively thick PVC membrane was drop cast on the PEDOT-Cn-modified gold electrode to ensure the semiinfinite diffusion of an extracted analyte ion in the membrane.31 Specifically, a PVC membrane was drop cast from a solution containing 4 mg of PVC, 1.07 mg of TDDADNNS, 16 mg of oNPOE or DOS, and 2.2 mg of TDDATFAB in 1.0 mL of THF. For potentiometry, 0.134 mg of KTFAB was added to the THF solution to serve as a cation-exchange site. A total of 30 μL of the THF solution was drop cast to the PEDOT-Cncoated electrode in the form of 6 μL aliquots added every 3 min. Electrochemical Measurement. An electrochemical workstation (CHI 900A or CHI 600A, CH Instruments) was used for voltammetric measurement. A Pt-wire counter electrode was employed in the following three-electrode cells

Each type of electrochemical measurement required additional setups. A piece of Teflon tube34 was put on a PVC/ PEDOT-Cn-modified gold electrode for cyclic voltammetry with cell 1 to define a disk-shaped membrane/water interface with a diameter of 1.5 mm. A PVC/PEDOT-Cn-modified gold electrode was rotated during cyclic voltammetry and stripping voltammetry with cell 2 by using a modulated speed rotator (Pine Research Instrumentation). For stripping voltammetry, the rotator and electrochemical cell were placed in an Ar-filled polyethylene glovebag (AtmosBag, Aldrich), which was protected from airborne contaminants inside a class 100 vertical laminar flow hood (model AC632LFC, AirClean Systems, Raleigh, NC, USA).33 In addition, Milli-Q water was collected inside of the hood, while sample solutions were prepared inside of the bag itself. An as-prepared electrode was contaminated during preparation and was cleaned in the background Milli-Q water solution of supporting electrolytes by repeating stripping voltammetric measurements until no contaminant response was detected. Cell 2 was also used for potentiometry with a 16-channel potentiometer (EMF 16, Lawson Lab, Malvern, PA, USA) in its “Differential” setting, where a platinum wire was connected from the potentiometer and immersed into the sample solution. An electrode was rotated to quickly replace potassium ion in the membrane with TMA+ in a protamine-free background solution and also to achieve steady-state transport of protamine during potentiometry.

■

RESULTS AND DISCUSSION CV of Protamine Adsorption. We employed ion-transfer CV to find that protamine is adsorbed at the DNNS-doped membrane/water interface much more favorably than extracted into the membrane. Specifically, protamine adsorption occurs at more positive potentials by ∼0.2 V than protamine extraction, which will be discussed later. Background-subtracted CVs based on adsorption and desorption of 28 μg/mL protamine (Figure 1) were obtained when relatively positive potentials were applied to a DNNS-doped PVC membrane coated on the gold electrode. The gold electrode was modified also with an intermediate layer of alkyl-substituted PEDOT as a voltammetric ion-to-electron transducer.30−33 Protamine adsorption was dominant at these positive potentials when a DNNS-doped PVC membrane was plasticized with either oNPOE or DOS (red lines in part A or B, respectively, of Figure 1). The negative potential sweep resulted in protamine transfer across the membrane/water interface. The transferred protamine molecules were adsorbed at the interface and were desorbed from the interface on the positive potential sweep, which yielded a sharper bell-shaped wave without a diffusional tail. Adsorption and exhaustive desorption of protamine were also confirmed by integrating the background-subtracted CVs. The resultant charge based on protamine transfer across the DOS/ PVC membrane (dashed line in Figure 1B) increased and then decreased to zero at the end of the potential cycle. This result confirms that protamine was only adsorbed at the interface and was not extracted into the ∼14 μm thick membrane,31 from which extracted ions cannot be exhaustively stripped during CV as demonstrated for TBA+ transfer (Supporting Information Figure S-1). Moreover, the charge under the voltammogram at the oNPOE/PVC membrane (Figure 1A) returned to ∼5% of the maximum value. This residual charge is due to the protamine molecules that are still adsorbed at the interface when a potential cycle was completed at the positive limit of

Ag|AgCl|x M protamine sulfate in 10 mM LiCl|PVC|PEDOT‐Cn|Au

(cell 1) Ag|AgCl|y M protamine sulfate in 10 mM TMACl|PVC|PEDOT‐Cn|Au

(cell 2)

The concentrations of protamine sulfate are given in Results and Discussion. The current carried by protamine from the aqueous phase to the membrane phase was defined to be positive. All electrochemical experiments were performed at 22 ± 3 °C. C

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

interface between water and the DNNS-doped oNPOE/PVC membrane. This observation excludes the possibility that protamine−DNNS complexes are adsorbed at the interface owing to a low solubility of the complexes in the membrane only at high protamine concentrations. Specifically, protamine was potentiostatically adsorbed at different preconcentration times, t, and then stripped from the membrane/water interface by linear-sweep voltammetry. A membrane-coated electrode was rotated at 2000 rpm to achieve steady-state transport of protamine between the aqueous solution and the membrane/ water interface as assumed in our theoretical model. The resultant stripping voltammograms of 0.75 μg/mL protamine (Figure 2A) showed clear desorption peaks. The peak current saturated quickly at t = 120 s while the whole voltammograms broadened and shifted to negative potentials at t = 240 and 480 s (inset of Figure 2A). The saturation of the peak current corresponds to the saturation of the membrane/water interface with protamine. Much longer preconcentration times were

Figure 1. Background-subtracted CVs (red lines) and their integrations (dashed lines) with 28 μg/mL protamine at DNNSdoped PVC membranes with (A) oNPOE and (B) DOS as plasticizers (cell 1). The potential was applied to the gold electrode, swept at 0.1 ′ . V/s, and defined against E0TBA

the potential window. The residual charge is not due to protamine extraction, which requires much more negative potentials as discussed later. Protamine transfer at the DOS/PVC membrane required less negative potentials than that at the oNPOE/PVC membrane (Figure 1). The averages of positive and negative peak potentials with DOS- and oNPOE-plasticized membranes were 0.253 and 0.128 V, respectively. This result indicates the formation of more stable protamine−DNNS complexes at the less polar DOS/PVC membrane, where stronger electrostatic interactions are exerted between the sulfonate group of an ionophore molecule and the guanidinium group of a protamine molecule.22 Noticeably, the applied potential was defined ′ against the formal potential of TBA+ transfer, E0TBA (Supporting 31 Information Figure S-1). Advantageously, our voltammetric electrodes were reusable for repetitive CV measurements of protamine transfer as well as the following CV measurement of TBA+ transfer. The reusability of our electrodes contrasts to the irreversible responses of potentiometric protamine-sensitive electrodes.6−9,11−18 Adsorptive Stripping Voltammetry. We employed stripping voltammetry to demonstrate that protamine at low concentrations of 0.75 and 0.038 μg/mL is adsorbed at the

Figure 2. Stripping voltammograms of (A) 0.75 and (B) 0.038 μg/mL protamine at rotating electrodes (2000 rpm) with DNNS-doped oNPOE/PVC membranes at different preconcentration times (cell 2). The potential was applied to the gold electrode, swept at 0.1 V/s, and ′ . Preconcentration potentials were (A) −0.063 defined against E0TBA and (B) −0.087 V.

D

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry required for saturation of stripping voltammograms with 0.038 μg/mL protamine (Figure 2B). Significantly, adsorptive stripping voltammetry of protamine is analytically advantageous in terms of reversibility, sensitivity, and selectivity. Stripping voltammograms were obtained repeatedly at different preconcentration times in contrast to irreversible potentiometry.6−9,11−18 A protamine concentration of 0.038 μg/mL is comparable to a detection limit of 0.02 μg/ mL for a potentiometric protamine-sensitive electrode under the rotating-electrode configuration.9 A protamine concentration of 0.038 μg/mL is equivalent to 7.5 nM protamine sulfate with a molecular weight of 5,000 and is 106 times lower than the background concentration of relatively hydrophobic TMA+, thereby indicating high protamine selectivity. A detection limit of 0.038 μg/mL protamine, however, is compromised by the limited capacity of the membrane/water interface for protamine adsorption (see later discussion). In fact, low-nanomolar and picomolar detection limits have been demonstrated for stripping voltammetry based on extraction of monovalent and divalent ions into the solid-supported thin polymeric membrane.30,32,33,35,36 Moreover, an even lower detection limit is expected for exhaustive stripping voltammetry based on extraction of protamine because of its much higher charge.33 Frumkin Isotherm for Protamine Adsorption. Stripping voltammograms of protamine adsorption were successfully analyzed by using the Frumkin isotherm21 to confirm that all protamine molecules that are transferred across the membrane/ water interface during a preconcentration step are desorbed from the interface. This result confirms dominant adsorption of protamine at the preconcentration potentials, which were positive enough to minimize protamine extraction. Specifically, the charge based on protamine desorption from the interface, Qd(t), was determined experimentally and compared with the theoretical charge based on protamine adsorption during a preconcentration step, Qa(t), as obtained by solving eq 6 based on the Frumkin adsorption isotherm. The charge under a stripping voltammogram at each preconcentration time, Qex(t), is a sum of Qd(t) and the charge due to background processes, Qbg, which mainly originates from the charging of the membrane/water interface. A plot of Qd(t) (=Qex(t) − Qbg) versus t was compared with a theoretical plot of Qa(t) versus t (Figure 3). The best fits of Qd(t) plots with Qa(t) plots yielded the parameters listed in Table 1 in addition to zi = +20 and A = 0.196 cm2. A charge of +20 carried by each protamine molecule is consistent with its positive charges based on arginine groups.37 Moreover, zi = +20 was obtained from the il values (Table 1) by using eq 4 with ν = 0.010 cm2/s and Dw = 1.0 × 10−6 cm2/s.20 These results indicate each protamine molecule carries +20 charges across the membrane/water interface during preconcentration and stripping steps. Analysis of stripping voltammograms by using the Frumkin isotherm provided quantitative insights into protamine adsorption at the membrane/water interface. The ΔG0ads and Γmax values (Table 1) are similar to those determined for protamine adsorption at the interface between water and the 1,2-dichloroethane solution of tetrakis(4-chlorophenyl)borate as protamine binding sites.21 The ΔG0ads and Γmax values are higher at 0.038 μg/mL than at 0.75 μg/mL, which is ascribed to the application of a more negative potential to the membrane/ water interface during preconcentration of 0.038 μg/mL protamine. Importantly, a relatively small difference, ΔEp, of ∼25 mV in the preconcentration potentials resulted in a

Figure 3. Charges based on protamine desorption (closed circles) as obtained from the stripping voltammograms of (A) 0.75 and (B) 0.038 μg/mL protamine in Figure 2A,B, respectively. Solid lines indicate the charges based on protamine adsorption at different preconcentration times as calculated by solving eq 6 with the parameters listed in Table 1

Table 1. Parameters from Figure 3A,B c0 (μg/mL) c0 (nM)a il (nA) ΔG0ads (kJ/mol) Γmax (nmol/cm2) g Qbg (μC) a

0.75 150 126 52 3.7 3.4 5.28

0.038 7.5 6.26 58 9.0 3.4 2.10

Molecular weight of 5,000 was used for protamine sulfate.

substantial difference of 6 kJ/mol in the ΔG0ads values owing to the high charge of protamine, where a difference in ΔG0ads values (=zi,appFΔEp) is only 2.5 kJ/mol for an adsorbed ion with an apparent charge, zi,app, of +1. In addition, a positive g value of 3.4 (Table 1) indicates attractive interactions among DNNS− protamine complexes at the interface. By contrast, negative g values in a range between −1.3 and −16.9 were obtained for protamine adsorption at the 1,2-dichloroethane/water interface, which indicates repulsive interactions among the protamine complexes of tetrakis(4-chlorophenyl)borate.21 CV of Protamine Extraction. Ion-transfer CV of protamine at DNNS-doped PVC membranes demonstrated protamine extraction at much more negative potential than E

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

peak current based on protamine adsorption was weakly dependent on the rotation speed (Supporting Information Figure S-2), which indicates that a sufficiently large amount of protamine was initially adsorbed at the aqueous side of the membrane/water interface in the presence of a high concentration of 28 mg/mL protamine. Importantly, protamine is extracted reversibly as demonstrated by the peak current response based on back-extraction of protamine from the membrane upon the positive potential sweep (Figure 4). The charge under a peak-shaped reverse response based on backextraction around 0 V is much smaller than the charge under the coupled peak-shaped or sigmoidal forward response around −0.1 V. This result confirms that protamine was extracted into the membrane on the forward potential sweep and was not exhaustively stripped from the membrane on the reverse potential sweep owing to the diffusional loss of extracted protamine into the bulk of the relatively thick membrane. Similarly, protamine was extracted into DNNS-doped DOS/ PVC membranes as discussed in Supporting Information (Figures S-3 and 4). Potentiometry Based on Protamine Extraction. We performed potentiometry with our solid-supported membrane electrodes to demonstrate that a super-Nernstian response to protamine is based on protamine extraction as hypothesized in previous potentiometric studies.6−9,11−18 In our potentiometric experiments, the potassium salt of TFAB was added as a cation exchanger to an oNPOE/PVC membrane with the same composition as that used for voltammetry. The PVC membrane was conditioned and studied potentiometrically in an aqueous solution of TMA+ (cell 2), which was used to voltammetrically study protamine extraction (see earlier discussion). The resultant membrane composition was different from the composition of the DNNS-doped membranes employed for traditional potentiometric protamine-sensitive electrodes, which contained only a DNNS salt of sodium ion both as an ionophore and a cation exchanger.6−9 Nevertheless, we were able to obtain such super-Nernstian responses to protamine as expected for the traditional potentiometric electrodes (Figure 5). When a concentrated protamine solution was added to a background solution of 10 mM TMACl to yield 0.75 μg/mL protamine, the membrane potential increased rapidly and reached a steady-state value of ∼45 mV with respect to the background value (Figure 5A). A smaller and slower potentiometric response was obtained for 0.038 μg/mL protamine (a potential change of ∼9 mV for 25 min in Figure 5B). These potentiometric responses to protamine were superNernstian and were much larger than expected for a Nernstian response to protamine with a charge of approximately +20 (∼3 mV/decade). Noticeably, a low concentration of 0.038 μg/mL protamine was detectable by potentiometry with the solidsupported electrode, which was rotated at 2000 rpm to enhance protamine transport from the aqueous solution to the membrane/water interface.9 We employed the mixed potential model10,28,29 to confirm that a super-Nernstian response to protamine is based on protamine extraction. In our potentiometric experiments, an equilibrium potential, Eeq, was achieved before protamine was added to the aqueous phase because TMA+ was initially present in both membrane and aqueous phases; i.e.,

protamine adsorption. Protamine extraction was clearly observed by using TMA+ as an aqueous supporting electrolyte (cell 2) to maximize the negative side of the potential window. Extraction of more lipophilic TMA+ is thermodynamically less favorable than extraction of Li+ (cell 1) because Li+ binds more strongly to DNNS− in the membrane. A small current response based on protamine extraction into a DNNS-doped oNPOE/ PVC membrane was observed at approximately −0.1 V, which is ∼0.2 V more negative than a peak current response based on protamine adsorption (Figure 4A). Protamine extraction was

Figure 4. CVs of 28 μg/mL protamine at DNNS-doped oNPOE/PVC membranes at different rotation speeds in cell 2. The potential was applied to the gold electrode, swept at 0.1 V/s, and defined against ′ . E0TBA

coupled with mass transport of protamine from the aqueous solution to the membrane/water interface. The current response based on protamine extraction was enhanced by rotating the membrane-modified electrode faster (Figure 4B− D). The resultant limiting current was proportional to the square root of the rotation speed, ω, as expected from the Levich equation (eq 4). The experimental slope of 0.29 ± 0.1 μA/(rad)1/2 was consistent with a value of 0.28 μA/(rad)1/2 as calculated from eq 4 with zi = +20, A = 0.196 cm2 (5 mm diameter disk), Dw = 1.0 × 10−6 cm2/s, ν = 0.01 cm2/s, and c0 = 5.6 μM (a molecular weight of 5,000 for protamine sulfate). This result indicates that each protamine molecule carries ∼+20 charges across the membrane/water interface. By contrast, the

TMA+(w) ⇌ TMA+(m)

(7)

Under the equilibrium condition, the net current carried by TMA+ across the interface was zero as shown by a schematic F

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

to yield the zero net current at Emix. The mixed potential was based on cation-exchange extraction as given by P20 +(w) + 20DNNS−(m) + 20TMA+(m) → DNNS20P(m) + 20TMA+(w)

(8)

20+

where P is protamine and DNNS20P is a protamine−DNNS complex. Importantly, protamine was extracted into the membrane to yield a mixed potential of up to 50 mV with respect to Eeq (Figure 5). In fact, this mixed potential was close to the negative limit of the voltammetric potential window based on TMA+ transfer and was negative enough to voltammetrically drive protamine extraction to a mass-transport limit (Figure 4). Importantly, back-extraction of protamine is extremely slow at the mixed potential, thereby yielding an irreversible potentiometric response to protamine.6−9,11−18 When the membrane that contains extracted protamine molecules is exposed to a protamine-free TMA+ solution, the reverse reaction of eq 8 is thermodynamically favored. This cationexchange reaction, however, is kinetically prevented by extremely slow back-extraction of protamine at Emix (red curve in Figure 6). By contrast, reversible protamine extraction was confirmed by voltammetry, where a sufficiently positive potential can be applied externally to the membrane/water interface to drive the back-extraction of protamine (Figure 4).

■

CONCLUSIONS In this study, we took advantage of powerful ion-transfer voltammetry to mechanistically assess protamine extraction into ionophore-doped polymeric membranes as well as protamine adsorption at membrane/water interfaces. Interestingly, protamine−DNNS complexes are formed at interfaces and then extracted into the membranes when additional potentials of approximately −0.2 V are applied to the interfaces. This twostep mechanism for DNNS-facilitated extraction of protamine polycation contrasts to the single-step mechanism for ionophore-facilitated extraction of small monovalent and divalent cations.31 Significantly, this study supports that protamine extraction serves as the response mechanism of potentiometric protamine-selective electrodes.6−9,11−18 Moreover, we employed the mixed potential model10,28,29 to reveal that thermodynamically reversible protamine extraction results in a kinetically irreversible potentiometric response to protamine at the mixed potential. Ion-transfer voltammetry will be useful also to study both thermodynamics and kinetics of protamine transfer at nonpolarizable membrane/water interfaces22 as employed for potentiometric protamine-sensitive electrodes.6−9 This study establishes both extraction and adsorption of protamine as promising response mechanisms for voltammetric protamine sensors, thereby widening the scope of reversible, selective, and sensitive detection of this important heparin antidote. Protamine adsorption was useful for exhaustive stripping voltammetry to detect down to ∼0.038 μg/mL protamine, which is nearly as low as the lowest detection limit achieved by potentiometry.9 This detection limit, however, was significantly compromised by the limited capacity of the membrane/water interface for protamine adsorption. In fact, a low-picomolar detection limit is expected for stripping voltammetry based on extraction of highly charged protamine into a polymeric membrane with a high preconcentration capacity33 when the viscous membrane is thin enough to

Figure 5. Potentiometric responses to (A) 0.75 and (B) 0.038 μg/mL protamine at oNPOE/PVC membranes doped with the TDDA salts of DNNS and TFAB as a protamine ionophore and a cation exchanger, respectively, in cell 2. The electrodes were rotated at 2000 rpm.

voltammogram of TMA+ transfer (black curve in Figure 6). When protamine was added to the TMA+ solution, the phase

Figure 6. Schematic voltammograms when protamine is initially present only in the aqueous phase (blue curve) or the membrane phase (red curve). The black curve schematically represents an iontransfer voltammogram of TMA+ in both bulk aqueous and membrane phases.

boundary potential shifted to the mixed potential, Emix, such that the negative current based on back-extraction of TMA+ from the membrane is balanced by the positive current based on protamine extraction from the aqueous phase (blue curve) G

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

(24) Osakai, T.; Yuguchi, Y.; Gohara, E.; Katano, H. Langmuir 2010, 26, 11530. (25) Shinshi, M.; Sugihara, T.; Osakai, T.; Goto, M. Langmuir 2006, 22, 5937. (26) Girault, H. H. Electrochemistry at Liquid−Liquid Interfaces. In Electroanalytical Chemistry; Bard, A. J., Zoski, C. G., Eds.; Taylor & Francis: Boca Raton, FL, USA, 2010; Vol. 23, p 1. (27) Amemiya, S.; Wang, Y.; Mirkin, M. V. Nanoelectrochemistry at the Liquid/Liquid Interfaces. In Specialist Periodical Reports in Electrochemistry; Compton, R. G., Wadhawan, J. D., Eds.; RSC: London, 2013; Vol. 12, p 1. (28) Kakiuchi, T.; Senda, M. Bull. Chem. Soc. Jpn. 1984, 57, 1801. (29) Langmaier, J.; Samcová, E.; Samec, Z. Anal. Chem. 2007, 79, 2892. (30) Kim, Y.; Rodgers, P. J.; Ishimatsu, R.; Amemiya, S. Anal. Chem. 2009, 81, 7262. (31) Ishimatsu, R.; Izadyar, A.; Kabagambe, B.; Kim, Y.; Kim, J.; Amemiya, S. J. Am. Chem. Soc. 2011, 133, 16300. (32) Kabagambe, B.; Izadyar, A.; Amemiya, S. Anal. Chem. 2012, 84, 7979. (33) Kabagambe, B.; Garada, M. B.; Ishimatsu, R.; Amemiya, S. Anal. Chem. 2014, 86, 7939. (34) Guo, J.; Amemiya, S. Anal. Chem. 2006, 78, 6893. (35) Kim, Y.; Amemiya, S. Anal. Chem. 2008, 80, 6056. (36) Garada, M. B.; Kabagambe, B.; Kim, Y.; Amemiya, S. Anal. Chem. 2014, 11230. (37) Ando, T.; Yamasaki, M.; Suzuki, K. Protamines: Isolation, Characterization, Structure and Function; Springer-Verlag: New York, 1973. (38) Guo, J.; Yuan, Y.; Amemiya, S. Anal. Chem. 2005, 77, 5711.

exhaustively strip protamine. Intrinsically high sensitivity will be required for voltammetric detection of protamine in blood samples, where interfacial adsorption of blood proteins38 will interfere protamine adsorption and, subsequently, protamine extraction.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b00644.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: 412-624-8611. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (Grant CHE-1213452). We thank Steve Knight, King Industries, for providing us with samples of DNNS. We also thank Prof. Mark E. Meyerhoff, Department of Chemistry, University of Michigan, and Prof. Eric Bakker, Department of Chemistry, University of Geneva, for helpful discussions as well as sharing unpublished results.

■

REFERENCES

(1) Carr, J. A.; Silverman, N. J. Cardiovasc. Surg. 1999, 40, 659. (2) Fu, B.; Bakker, E.; Yun, J. H.; Wang, E.; Yang, V. C.; Meyerhoff, M. E. Electroanalysis 1995, 7, 823. (3) Meyerhoff, M. E.; Fu, B.; Bakker, E.; Yun, J.-H.; Yang, V. C. Anal. Chem. 1996, 68, 168A. (4) Yun, J. H.; Meyerhoff, M. E.; Yang, V. C. Anal. Biochem. 1995, 224, 212. (5) Ma, S. C.; Yang, V. C.; Meyerhoff, M. E. Anal. Chem. 1992, 64, 694. (6) Fu, B.; Bakker, E.; Yun, J. H.; Yang, V. C.; Meyerhoff, M. E. Anal. Chem. 1994, 66, 2250. (7) Ramamurthy, N.; Baliga, N.; Wahr, J. A.; Schaller, U.; Yang, V. C.; Meyerhoff, M. E. Clin. Chem. 1998, 44, 606. (8) Ramamurthy, N.; Baliga, N.; Wakefield, T. W.; Andrews, P. C.; Yang, V. C.; Meyerhoff, M. E. Anal. Biochem. 1999, 266, 116. (9) Ye, Q. S.; Meyerhoff, M. E. Anal. Chem. 2001, 73, 332. (10) Amemiya, S. Potentiometric Ion-Selective Electrodes. In Handbook of Electrochemistry; Zoski, C. G., Ed.; Elsevier: New York, 2007; p 261. (11) Shvarev, A.; Bakker, E. J. Am. Chem. Soc. 2003, 125, 11192. (12) Shvarev, A.; Bakker, E. Anal. Chem. 2005, 77, 5221. (13) Fordyce, K.; Shvarev, A. Anal. Chem. 2008, 80, 827. (14) Xu, Y.; Shvarev, A.; Makarychev-Mikhailov, S.; Bakker, E. Anal. Biochem. 2008, 374, 366. (15) Gemene, K. L.; Meyerhoff, M. E. Anal. Biochem. 2011, 416, 67. (16) Crespo, G. A.; Afshar, M. G.; Bakker, E. Angew. Chem., Int. Ed. 2012, 51, 12575. (17) Crespo, G. A.; Ghahraman Afshar, M.; Dorokhin, D.; Bakker, E. Anal. Chem. 2014, 86, 1357. (18) Bell-Vlasov, A. K.; Zajda, J.; Eldourghamy, A.; Malinowska, E.; Meyerhoff, M. E. Anal. Chem. 2014, 86, 4041. (19) Amemiya, S.; Yang, X.; Wazenegger, T. L. J. Am. Chem. Soc. 2003, 125, 11832. (20) Yuan, Y.; Wang, L.; Amemiya, S. Anal. Chem. 2004, 76, 5570. (21) Trojanek, A.; Langmaier, J.; Samcova, E.; Samec, Z. J. Electroanal. Chem. 2007, 603, 235. (22) Yuan, Y.; Amemiya, S. Anal. Chem. 2004, 76, 6877. (23) Rodgers, P. J.; Amemiya, S. Anal. Chem. 2007, 79, 9276. H

DOI: 10.1021/acs.analchem.5b00644 Anal. Chem. XXXX, XXX, XXX−XXX