Anal. Chem. 2005, 77, 2632-2636
Electrochemical Method for the Detection of Lipase Activity Gintaras Valincius,*,† Ilja Ignatjev,† Gediminas Niaura,† Maryte3 Kazˇeme3 kaite3 ,† Zita Talaikyte3 ,† Valdemaras Razumas,† and Allan Svendsen‡
Institute of Biochemistry, Mokslininku 12, LT-08662 Vilnius, Lithuania, and Novozymes A/S, Smørmosevej 25, DK-2880 Bagsvaerd, Denmark
A novel electrochemical technique for the general assay of lipase activity is described. The method utilizes a solidsupported lipase substrate, which is formed by dripping and drying a small amount of an ethanol solution of 9-(5′ferrocenylpentanoyloxy)nonyl disulfide (FPONDS) onto gold modified by a hexanethiol self-assembled monolayer. The redox ferrocene group of FPONDS generates the electrochemical signal, the intensity of which is proportional to the number of FPONDS molecules at the interface. Electrochemical and surface-enhanced infrared absorption spectroscopic data, as well as control experiments with an engineered, deactivated mutant enzyme, demonstrate that the wild-type lipase from Thermomyces lanuginosus is capable of cleaving the ester bonds of FPONDS molecules via an enzymatic hydrolysis mechanism, which includes the adsorption of the lipase onto the substrate surface. The hydrolysis liberates the ferrocene groups from the interface triggering a decay of the electrochemical redox signal. The rate of the electrochemical signal decrease is proportional to the lipase activity/concentration. These data suggest a general method for the direct measure of enzymatic activity of lipases. The abundance and variety of lipolitic enzymes make them one of the most important participants of the chemistry of life. Triacylglycerol hydrolases (EC 3.1.1.3) that hydrolyze triacylglycerols at the oil/water interface have extensive applications in oleochemistry as detergent additives and digestive aids, as well as in the paper and food industries.1-3 Unlike other bond-cleaving enzymes, e.g., proteases or restriction enzymes, hydrolyses by lipases are carried out in heterogeneous multiphase systems. In many cases, the environment of the enzyme at the substrate/ liquid interface plays an important or even crucial role for the overall enzymatic activity of these proteins.1-4 Thus, the ability to monitor enzymatic activity of lipases under these conditions is of paramount importance. * Corresponding author. E-mail:
[email protected]. † Institute of Biochemistry. ‡ Novozymes A/S. (1) Schmid, R. D.; Verger, R. Angew. Chem., Int. Ed. 1998, 37, 1608-1633. (2) Bornscheuerand, U. T.; Kazlauskas, R. J. Hydrolases in Organic Synthesis: Regio- and Stereoselective Biotransformations; Wiley-VCH: Weinheim, 1999. (3) Houde, A.; Kademi, A.; Leblanc, D. Appl. Biochem. Biotechnol. 2004, 118, 155-170.
2632 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
In this article, we describe a new electrochemical technique, which allows the direct measurement of lipase enzymatic activity/ kinetics. Unlike other methods for studying lipase kinetics, e.g., pH stat, back-titration, colorimetry, or Langmuir-Blodgett film balance techniques,1-3 this technique utilizes solid-supported substrates, which, in principle, could be engineered in the form of microsensors for the detection of the activity and specificity of lipases. The technique is described using the Thermomyces lanuginosus lipase (TLL).5 However, we believe the electrochemical principle described here should be applicable to other bondcleaving enzymes (see, for example, ref 6 in which the fluorescencebased technique was used to monitor bond-cleaving events induced by proteases). EXPERIMENTAL SECTION The solid-supported substrates were made of polycrystalline gold electrodes7 initially covered by a hexanethiol (HT) selfassembled monolayer (SAM; 1 mM HSC6H13/95% ethanol, 3 h). Atop of this SAM, a layer of a ferrocene-terminated long-chain ester disulfide, 9-(5′-ferrocenylpentanoyloxy)nonyl disulfide (FPONDS, structure shown in Chart 1),7 is deposited via a simple drip-and-dry procedure. Specifically, the first step gives the chemically anchored HT-SAM, which shields the underlying metal surface from direct contact with the FPONDS and, later, with the protein solutions. Next, a 5-µL droplet of a 95% ethanol containing 25 µM FPONDS is placed atop the gold-supported HT-SAM, and the solvent is allowed to evaporate. The solution spread over the whole electrode plane, including the metal and the insulating resin parts. Commercially available gold electrodes (Bionalytical Systems, Inc., West Lafayette, IN) were used. The concentration and the volume of the ethanol solution is chosen such that, after evaporation, the effective surface concentration of the FPONDS molecules is close to the packing density of one monomolecular layer.8 The overlayer of deposited FPONDS molecules is stable in aqueous buffer solutions for prolonged periods of time.7 It is (4) Kobayashi, A.; Sato, Y.; Mizutani, F. Biosci. Biotechnol. Biochem. 2001, 65, 2392-2396. (5) Svendsen, A.; Clausen, I. G.; Patkar, S. A.; Borch, K.; Thellersen, M. In Methods in Enzymology; Rubin, B., Dennis, E. A., Eds.; Academic Press: New York, 1997; Vol. 284, Part A, pp 317-340. (6) Salisbury, C. M.; Maly, D. J.; Ellman, J. A. J. Am. Chem. Soc. 2002, 124, 14868-14870. (7) See Supporting Information for more details. (8) For SEIRA spectroscopy measurements, ∼3-fold surface densities of FPONDS were used. 10.1021/ac048230+ CCC: $30.25
© 2005 American Chemical Society Published on Web 03/19/2005
Chart 1. Structure of 9-(5′-Ferrocenylpentanoyloxy)nonyl Disulfide (FPONDS)
apparent that, the combination of these two steps, with SAM formation using molecules other than HT and with different FPONDS concentrations will enable one to vary the microenvironment of the FPONDS molecules and, thus, affect the activity and specificity of different lipases and other hydrolytic enzymes. For the surface-enhanced infrared absorption (SEIRA) spectroscopy experiments, the chemically deposited gold films were prepared as follows. Polished 0.3-0.4-mm-thick silicon (15 × 15 mm samples) wafers were immersed for 10 min in a 10% (NH4)2S2O8 solution in concentrated sulfuric acid. Then, the electrodes were washed with copious amounts of deionized water. The wafers were transferred to a 40% water solution of NH4F for 3 min, to dissolve the SiO2 layer from the surface, dried in a stream of argon, and then put onto the metallic plate termostated at 60 °C. For the gold plating, we used a solution described by Miyake et al.9 We premixed 60 µL of gold plating solution with 120 µL of 2% HF solution, and the whole liquid amount was allowed to spread onto the freshly NH4F-pretreated silicon wafer surface to obtain a gold film ∼1.2-1.5 cm2 in area. After 60-90 s, the process was stopped by intense rinsing with water. Then, the gold-coated wafer was rinsed with 95% ethanol and immediately transferred to a solution of HT (1 mM) for 3 h. Electrochemical measurements were carried out on a EG&G Versastat computerized potentiostat (Princeton Applied Research, Princeton, NJ). A three-electrode conventional cell (10 cm3) was thermostated at 25 ( 0.2 °C. A platinum plate (∼2 cm2) served as an auxiliary electrode, and a sodium chloride saturated calomel electrode was used as the reference (E ) 239 mV vs standard hydrogen electrode). The real surface area of the working electrode was estimated from the Au surface oxidation charge (390 ( 10 µC‚cm-2). SEIRA measurements were performed on Perkin-Elmer, model Spectrum GX FT-IR spectrometer equipped with DGTS detector. The spectral resolution was set at 4 cm-1, and all of the spectra were acquired by 500 scans. Buffers were made using ASC reagent grade sodium salts and sodium hydroxide purchased from Sigma-Aldrich Chemie GmbH. The buffer containing 0.1 M NaClO4, 0.01 M monobasic sodium phosphate, adjusted with NaOH to pH 7.0 was used throughout the work. The Millipore-Q purified (18.2 MΩ‚cm) water was used for the preparation of solutions. T. lanuginosus proteins (both native and mutant forms) were provided by Novozymes A/S (Bagsvaerd, Denmark) in the form of ∼3-6 mg/mL solutions with producer certified activities. RESULTS AND DISCUSSION The initial curve in Figure 1 shows the voltammetric redox response due to the presence of the ferrocene group in the (9) Miyake, H.; Ye, S.; Osawa, M. Electrochem. Commun. 2002, 4, 973-977.
Figure 1. Cyclic voltammetry curves of FPONDS-based substrate upon addition of TLL in the amount of 373 LU in 2 mL of buffer solution. One standard lipase activity unit 1 LU ) the amount of enzyme that hydrolyzes 1 µmol of glycerol tributyrate/min.5 The numbers next to the curves indicate the time (min) elapsed after the addition of TLL. All curves recorded at 100 mV/s scan rate, in the potential interval from 0 to 0.5 V vs SSCE. Buffer composition: 0.1 M NaClO4, 0.01 M phosphate buffer, pH 7.0, 25 °C. Initial surface concentration of FPONDS: 2.7 × 10-10 mol/cm2. Estimated surface area of the gold electrode: 0.042 cm2. Inset: decay of the normalized integrated peak current of the electrode in the quiescent solutions at various lipase activities: 0 (squares), 3.73 (circles), and 373 LU (triangles) in 2 mL of buffer solution.
FPONDS molecules on the surface. Without external polarization, the response was stable for at least 2 h in a quiescent buffer solution. However, moderate agitation of the solution with a magnetic stirrer bar (∼100 rpm) results in a slight decay of the initial signal at a rate of less than 0.2%/min. On the other hand, under continuous potential cycling, in the interval from 0 to 0.5 V (vs saturated sodium chloride calomel electrode; SSCE), the voltammetric signal starts to decay even in the quiescent solution. The rate of decay is ∼0.4%/cycle. Noteworthy, when calculating initial rates of interfacial enzymatic reactions, small spontaneous deterioration of the signal (usually below 2-3%) generally results in insignificant errors. Upon lipase addition (Figure 1), changes in the cyclic voltammetry curve are clearly seen, which can be assigned to the TLL activity. The peak current, which magnitude is proportional to the number of ferrocene groups in the surface layer, decreases upon lipase injection. As can be seen from the inset of Figure 1, the rate of decrease of the peak current integral is proportional to the amount of injected enzyme. This suggests that the rate of electrochemical signal decay can be used to determine the TLL activity. Although the decrease of the redox response is due to changes in the ferrocene groups at the surface, it is not clear from these data alone that the current decrease is due to the hydrolytic cleavage of the ester bond by TLL. There are other potential Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2633
Table 1. Integrated Intensities of Some SEIRA Bands as a Function of the Substrate Incubation Time in the Quiescent TLL-Containing Solutiona integrated intensity, au incubation time, h
γ(CH)Fc at 1105 cm-1
ν(CdO) at 1734 cm-1
ν(CH)Fc at 3090 cm-1
0 1 15
0.0170 (100) 0.00652 (38.4) 0.000564 (3.3)
0.3232 (100) 0.1360 (42.1) 0.00922 (2.9)
0.03234 (100) 0.01024 (31.7)
a Solution composition: 0.1 M NaClO , 0.01 M phosphate buffer, 4 pH 7.0, 25 °C. TLL amount: 746 LU in 4 mL of the buffer solution. Initial surface amount of FPONDS: ∼1.67 × 10-9 mol/cm2. Values in parentheses, percent.
Figure 2. SEIRA spectra of the FPONDS/HT/Au/Si substrate in the middle- and high-frequency regions measured after different times of incubation in the quiescent TLL-containing solution: (a) 0, (b) 1, and (c) 15 h. The background spectrum of the HT/Au/Si substrate was subtracted, A total of 500 scans at 4-cm-1 resolution were accumulated. Incubation solution: 4 mL of 0.1 M NaClO4, 0.01 M phosphate buffer of pH 7.0, containing 746 LU of TLL. Initial amount of FPONDS is ∼1.7 × 10-9 mol/cm2.
reasons for the disappearance of signal. For example, the protein might compete with the FPONDS molecules for surface adsorption sites. To show that the ester bond is specifically cleaved by the lipase, we recorded the SEIRA spectra of the substrate exposed to the TLL solution (Figure 2). The SEIRA signal was measured in a transmission mode using chemically deposited Au films on silicon.7,9 To increase the sensitivity, we first recorded the spectrum of the gold surface modified by HT-SAM. Then, the FPONDS overlayer was deposited, and the sum spectrum of the physisorbed FPONDS and HT-SAM was recorded. The difference of these two spectra, which is shown in Figure 2a, is the pure spectrum of FPONDS. In the middle frequency domain (Figure 2a, left graph), the narrow out-of-plane ferrocene ring vibrational band at 1105 cm-1,10 and strong ester carbonyl band at 1734 cm-1 are clearly seen. In the high-frequency range (Figure 2a, right graph), the CH stretching vibrational band of the ferrocene ring at 3090 cm-1 and a family of the symmetric and antisymmetric CH stretching vibrational bands of the alkyl chain at 2853 and 2926 cm-1, respectively, are evident.10 The peak positions of CH modes are sensitive to the order of the polymethylene chain.11-13 The frequencies in Figure 2 are higher by several reciprocal centimeters as compared with the well-ordered crystalline-like structures (2850 and 2917-2920 cm-1),11-13 indicating that the polymethylene chains of physisorbed FPONDS are disordered, containing a number of gauche defects. Upon addition of TLL, the SEIRA spectrum starts to change. Spectrum b in Figure 2 shows the SEIRA signal obtained after 1 h of incubation of the substrate in the unstirred enzyme solution. (10) Popenoe, D. D.; Deinhammer, R. S.; Porter, M. D. Langmuir. 1992, 8, 25212530. (11) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558-569. (12) Simon-Kutscher, J.; Gericke, A.; Hu ¨ hnerfuss, H. Langmuir 1996, 12, 10271034. (13) Love, J. C.; Wolfe, D. B.; Haasch, R.; Chabinyc, M. L.; Paul, K. E.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 2003, 125, 2597-2609.
2634 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
The first and the most important observation is that the decrease in the intensities of the ferrocene and carbonyl group bands occurs simultaneously (Table 1). After 1 h of TLL exposure, the integrated spectral intensities of these two bands decrease to ∼40% from the initial values. After 15 h of incubation in the lipase solution, both bands exhibit about the same residual integrated intensity level of ∼3%. On the other hand, incubation in the TLL solution for 1 h induces only slight perturbations on the intensities of CH vibrational bands (e.g., the integrated intensity of the 2853-cm-1 band decreases from 0.29 to 0.26), indicating that the main polymethylene chain of FPONDS remains on the surface. This supports the assumption that the removal of the ferrocene moieties from the surface occurs primarily via the cleavage of ester bond. Another important result is seen in Figure 2b. The amide I and amide II bands at 1652 and 1535 cm-1 (Figure 2b, left graph) as well as the NH stretching vibrational band at 3358 cm-1 (Figure 2b, right graph) appear upon exposing the substrate wafer to the lipase solution.14,15 These bands indicate protein anchoring to the surface, which is known to be a necessary step for the initiation of TLL enzymatic action.1,4,5 These spectral changes continue to increase with substrate/ TLL solution contact time. The ferrocene bands totally disappear (Figure 2c), concurrent with the reduction of the much stronger carbonyl band at 1734 cm-1 to barely above the spectral baseline level. The integral intensity of this spectral line is less than 3% of the initial value (Table 1). Thus, enzyme-induced cleavage of the ester bond proceeds until the substrate material is almost completely consumed. On the other hand, the protein bands (1535, 1652, and 3358 cm-1) continue to increase relative to the 1-h spectrum (Figure 2b). This suggests that the kinetics of the enzyme action might be determined by the protein adsorption/ anchoring rate in unstirred solutions. Even though Figure 2 presents clear evidence that the ester bond is cleaved upon injection of TLL, the SEIRA data alone do not yield direct prove that the bond rupture occurs through an enzymatic mechanism. To confirm the enzymatic nature of the ester bond hydrolysis, the following experiment was performed. It is known that the catalytic triad of amino acids Ser-His-Asp determines the hydrolytic activity of TLL with the Ser 146 acting as a nucleophile.16 Along with the wild-type (WT) TLL, we applied (14) Tamm, L. K.; Tatulian, S. A. Q. Rev. Biophys. 1997, 30, 365-429. (15) Neogi, P. Neogi, S. Stirling, C. J. M. J. Chem. Soc., Chem. Commun. 1993, 1134-1136. (16) Svendsen, A. Biochim. Biophys. Acta 2000, 1543, 223-238.
Consequently, it is possible to utilize the kinetic data of this process as an enzyme activity assay. In the range of standard lipase activity units (1 LU ) the amount of enzyme that hydrolyzes 1 µmol of glycerol tributyrate/min5) from 4 to 400 LU dissolved in 2 mL of the solution (0.1 M NaClO4, 0.01 M phosphate buffer, pH 7.0, 25 °C), using a 0.042-cm2 surface area working electrode, under modest agitation with the magnetic stirrer bar (∼100 rpm), the following linear relationship between the initial rate of integrated peak current decrease (Vini) and the TLL activity was observed:
Vini(nC/s) ) -0.024((0.001) × LU(µmol/min), R2 ) 0.9934
Figure 3. Normalized integrated voltammetric peak of the FPONDS/ HT/Au substrate vs time of the electrode incubation in the quiescent solutions containing WT TLL or S146A mutant TLL plus WT TLL. Concentrations of the WT TLL (triangles and open circles) and S146A mutant TLL (filled circles) in the solutions: 3.5 and 9.25 µg/mL, respectively. Activity of WT TLL: 373 LU in 2 mL of solution containing 0.1 M NaClO4, 0.01 M phosphate buffer of pH 7.0, 25 °C. Initial surface concentration of FPONDS is ∼2.6 × 10-10 mol/cm2, estimated surface area of the gold electrode: 0.042 cm2.
a mutant variant of the enzyme, in which Ser 146 in the catalytic triad was substituted with alanine (S146A mutant of TLL).17 Compared to the WT enzyme, the S146A protein exhibits ∼500 times lower hydrolytic activity.18 Thus, within the time frame used in the cyclic voltammetry measurement of the initial hydrolysis rate of FPONDS molecules (e10 min), there is virtually zero activity for this variant of TLL. Figure 3 compares the voltammetric responses in the presence of WT and mutant lipases. It is evident (Figure 3, filled circles) that the S146A mutant does not induce changes in the voltammetric signal. Within 60 min after the mutant injection, the integral of the peak current decreases by e4-5%, which is comparable to natural deterioration rate of the FPONDS layer (vide ultra). However, once the native lipase is injected, the integrated voltammetric signal immediately starts to decay (Figure 3, open circles). Figure 3 reveals one more interesting detail. The rate of signal decay is clearly dependent on whether the buffer contains an inactive mutant before the injection of WT enzyme. The electrochemical signal decays much slower (in Figure 3, 4.7 times) if the mutant enzyme is present in the system prior to the WT enzyme injection. We believe, this result is consistent with the idea that, being present in the solution, both active and inactive proteins compete for the surface. It is widely acknowledged that in order to cleave the ester bonds the lipase molecule must anchor itself to the hydrophobic substrate surface. Hence, when the mutant the S146A is present in the solution prior to injection of the WT enzyme, the inactive competitor occupies part of the substrate’s surface. Therefore, the reduced hydrolysis rate (Figure 3, compare open circles and triangles) testifies that the cleavage of the ester bond of FPONDS occurs via an enzymatic mechanism. (17) Peters, G.H.; Svendsen, A.; Langberg, H.; Vind, J.; Patkar, S. A.; Kinnunen, P. K. J. Colloids Surf., B 2002, 26, 125-134. (18) Svendsen, A. Unpublished data.
In principle, one would expect that Vini might depend on the physical state and location of the FPONDS molecules on the surface. It is known that SAMs of short-chain thiols, such as HT, are disordered, contain defects, and can be displaced by larger thiols/disulfides such as FPONDS. We have observed this. Indeed, displacement becomes noticeable when the HT-SAM is exposed to FPONDS solutions for long time periods (more than 3 h). However, we believe, on the basis of a number of observations, that the FPONDS molecules do not significantly penetrate into the HT-SAM, within the time scale of substrate fabrication (∼5 min) and are only in the form of an overlayer. First, the dripand-dry overlayer, though being sufficiently stable in water solutions, can be almost entirely removed by soaking the fabricated electrode for several minutes in ethanol (see Supporting Information, part 3), which is accompanied by the complete loss of the electrochemical signal. In contrast, our previous data19-21 with SAMs of 9-mercaptononyl-5′-ferrocenyl pentanoate (MNFcP), which is exactly half of FPONDS (see Chart 1), chemisorbed directly onto the Au, show no similar instability upon washing with ethanol. Second, in perchlorate electrolyte solutions, the full width of half-maximum (fwhm) of the cyclic voltammagram of the drip-and-dry FPONDS film is very narrow (e40 mV) while the MNFcP SAMs exhibits a much broader fwhm (g90 mV), indicating strong lateral attractive interactions in the FPONDS that is absent in the chemisorbed MNFcP SAMs. Third, in the presence of 0.1 M ClO4-, the midpoint potential of the current maximums of drip-and-dry FPONDS films is located at +0.30 V, while FPONDS molecules deposited via exchange mechanism (3 h) exhibit more negative values of about +0.23 V.7 This indicates a more hydrophilic environment22 of the ferrocene groups in the mixed SAMs. Taken together, these facts strongly suggest that the majority of the FPONDS molecules are in the state that we describe as physisorbed and that is distinctively different from the state of a classical SAM that might be formed via incubation of bare gold electrode in FPONDS solution or via exchange mechanism. Finally, it is worth mentioning that the choice of the FPONDS as a lipase target molecule was determined entirely by the stability (19) Kazˇeme˘ kaite˘ , M.; Bulovas, A.; Smirnovas, V.; Niaura, G.; Butkus, E.; Razumas, V. Tetrahedron Lett. 2001, 42, 7691-7694. (20) Kazakevicˇiene˘ , B.; Valincius, G.; Niaura, G.; Talaikyte˘ , Z.; Kazˇeme˘ kaite˘ , M.; Razumas, V. J. Phys. Chem. B 2003, 107, 6661-6663. (21) Valincius, G.; Niaura, G.; Kazakevicˇiene˘ , B.; Talaikyte˘ , Z.; Kazˇeme˘ kaite˘ , M.; Butkus, E.; Razumas, V. Langmuir 2004, 20, 6631-6638. (22) Rowe, G. K.; Creager, S. E. Langmuir 1991, 7, 2307-2312.
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2635
of a drip-and-dry layer. Similar molecules, like 9-mercaptononyl5′-ferrocenyl pentanoate, 9-bromononyl-5′-ferrocenylpentanoate and decyl-5-ferrocenylpenatnoate, were found to form less stable overlayers.7 We hypothesize that the stability of FPONDS dripand-dry layers is primarily determined by the size of this molecule. In addition, the extraordinarily narrow redox peaks of the substrates obtained by the drip-and-dry procedure suggest strong lateral attractive interactions between the FPONDS molecules, which may contribute to the stability of the layer. CONCLUSIONS We describe a novel electrochemical approach to evaluate the enzymatic activity of lipases. Specifically, we demonstrate that a solid-supported, presumably a physisorbed layer of 9-(5′-ferrocenylpentanoyloxy)nonyl disulfide, prepared via a simple drip-anddry technique, could be utilized as a substrate for the T. lanuginosus lipase. The kinetics of the interfacial enzymatic bond cleavage was followed electrochemically by the decrease of redox signal of the ferrocene reporter group, thus allowing direct monitoring of the enzymatic activity. It is conceivable that this
2636
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
approach, or simple modifications thereof, could be used for other bond-cleaving enzymes. ACKNOWLEDGMENT Financial support from the European Community under Contract NMP4-CT-2003-505211 is gratefully acknowledged. The authors are thankful to Dr. David Vanderah for the interesting and helpful discussions. SUPPORTING INFORMATION AVAILABLE (a) Description of the polycrystalline gold electrode preparation procedure, (b) details of the synthesis of FPONDS and other ferrocene terminated compounds, (c) experimental data supporting the hypothesis of the physisorbed nature of FPONDS molecules, and (d) drip-and-dry layer stability data. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review November 30, 2004. Accepted February 11, 2005. AC048230+
