Anal. Chem. 2002, 74, 1302-1306
Bioimprinting of Polymers and Sol-Gel Phases. Selective Detection of Yeasts with Imprinted Polymers F. L. Dickert* and O. Hayden
Institute of Analytical Chemistry, University of Vienna, Wa¨hringerstrasse 38, A-1090 Vienna, Austria
Coated quartz crystal microbalances were modified with a surface-imprinting process using whole yeast cells. These molded polymer and sol-gel surfaces show honeycomb-like structures as shown by atomic force microscopy. Reinclusion of cells allows a selective on-line monitoring of these microorganism concentrations in water over 5 orders of magnitude. The sensitivity to cells holds up in growth media up to 21 g/L. Even cell fragments can be detected in flowing conditions. The highly robust polymers on the sensor devices are suitable for biotechnological applications. Molecularly imprinted polymers (MIPs) are of growing interest for their potential applications as artificial enzymes (plastizymes),1,2, advanced materials for solid-phase extraction,3,4 and thin coatings for sensor devices.5-7 In most cases, the templates used for the synthesis of imprinted polymers are subnanometer molecules, such as solvents,8 toxic components, or drugs.9,10 Few reported works dealed with biopolymers11,12 or bacteria13 as templates for MIPs. Recently, templating processes with living yeast cells were reported for the preparation of ordered and porous sol-gels.14 Here, we describe a versatile procedure to form regular patterns of molded polymer and sol-gel surfaces with yeasts as templates, which results in packed receptor sites directly on the coated transducer surfaces. This efficient surface-imprinting * Corresponding author. E-mail:
[email protected]. (1) Wulff, G.; Gross, T.; Scho¨nfeld, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 1962-1964. (2) Toorisaka, E.; Yoshida, M.; Uezu, K.; Goto, M.; Furusaki, S. Chem. Lett. 1999, 387-388. (3) Sellergren, B. Trends Anal. Chem. 1999, 18, 164-174. (4) Vlatakis, G.; Andersson, L. I.; Mu ¨ ller, R.; Mosbach K. Nature 1993, 361, 645-647. (5) Yano, K.; Karube, I. Trends Anal. Chem. 1999, 18, 199-204. (6) Dickert, F. L.; Hayden, O. Adv. Mater. 2000, 12, 311-313. (7) Panasyuk, T. L.; Mirsky, V. M.; Piletsky, S. A.; Wolfbeis, O. S. Anal. Chem. 1999, 71, 4609-4613. (8) Dickert, F. L.; Hayden, O. Trends Anal. Chem. 1999, 18, 192-199. (9) Nicholls, I. A.; Andersson, L. I.; Mosbach, K.; Ekberg, B. Trends Biotechnol. 1995, 13, 47-51. (10) Liang, C.; Peng, H.; Zhou, A.; Nie, L.; Yao, S. Anal. Chim. Acta 2000, 415, 135-141. (11) Shi, H. Q.; Tsai, W. B.; Garrison, M. D.; Ferrari, S.; Ratner, B. D. Nature 1999, 398, 593-597. (12) Slade, C. J. J. Mol. Catal. B 2000, 9, 97-105. (13) Alexander, C.; Vulfson, E. N. Adv. Mater. 1997, 9, 751-755. (14) Chia, S.; Urano, J.; Tamanoi, F.; Dunn, B.; Zink, J. I. J. Am. Chem. Soc. 2000, 122, 6488-6489.
1302 Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
process with living cells was combined with mass-sensitive devices.15 EXPERIMENTAL SECTION Surface Imprinting. Soft lithographic techniques comparable to microcontact printing as proposed by Whitesides et al.16,17 were used for the molding of polymer surfaces (Figure 1). The flat cell lawns on the stamps are prepared on microscope slides, which are coated with the microorganism and pressed against a nonadhesive Teflon surface. The quartz crystal microbalance (QCM) devices are spin- or drop-coated with adequately diluted and prepolymerized solutions of the selected polymer. The freshly prepared cell stamps are pressed on the coated transducer shortly after the solvent of the polymer solution is evaporated. Constant pressing force on the sandwich, stamp-template-sensor layertransducer, provides good contact between the polymer and the cell walls. In the case of polyurethanes, the curing process of the polymer layer is normally done overnight under ambient conditions (Figures 2 and 3). The stamp and the cells are easily removed with hot water, leaving behind a honeycomb-like structure of the molding process on the QCM electrodes (Figure 4). Mass-Sensitive Devices. We are using AT-cut 10-MHz quartz blanks (diameter 10 mm) for a screen-printing process of binary or even ternary electrode structures on a single piezocrystal (cf. Figure 1). After printing the gold paste on the piezo substrate, the electrode structure is burned in to yield highly robust electrodes to withstand mechanical stress and inhibit electrode liftoffs by creeping effects of aggressive media. Compared to commercial QCMs with vapor-deposited gold electrodes, our QCMs are mechanically more robust. The dual-electrode structure allows excellent compensation of viscosity and temperature effects on a single piezocrystal. In this work, we coated the whole QCM, facing the aqueous phase with the polymer, and imprinted one electrode area. The reference electrode with nonimprinted polymer is used to eliminate unspecific adhesion effects during the differential measurement. Single-cell detection was performed with shear wave surface acoustic wave (SAW) devices, designed from a LiTaO3 substrate and gold electrodes, operating at a frequency of 428 MHz.18 (15) Dickert, F. L.; Hayden, O.; Halikias, K. P. Analyst 2001, 126, 766-771. (16) Xia, Y.; Whitesides, G. M. Angew. Chem., Int. Ed. Engl. 1998, 37, 550575. (17) Kane, R. S.; Takayama, S.; Ostuni, E:; Ingber, D. E.; Whitesides, G. M. Biomaterials 1999, 20, 2363-2376. 10.1021/ac010642k CCC: $22.00
© 2002 American Chemical Society Published on Web 02/07/2002
Figure 1. Surface-imprinting process of precoated quartz crystal microbalances with dual-electrode structures. Nonimprinted and surface-imprinted polymer on a single piezocrystal allows excellent differential measurements as well as direct comparison studies of imprinting effects and microbial adhesion.
Figure 3. Magnified part of a polymer-coated gold electrode/quartz surface area. The lighter imprinted areas represent the edge of the electrode area to the quartz substrate. The bar has a length of 25 µm.
Figure 4. Sol-gel layer from titananiumIV ethylate imprinted with S. cerevisiae. The coatings are extremely robust and scratch resistant. Figure 2. Tapping mode AFM picture showing the ongoing imprinting process on a coated QCM electrode. The polymerizing material is halfway incorporating single S. cerevisiae cells and forms packed and highly ordered biomimetic receptors.
Measurements. The dual QCMs were horizontally placed on the bottom of a 250-µL flow cell with the liquid overflowing the transducer. The samples as well as the flow cell were thermostated. This setup allows measurements without interferences from air bubbles. The self-made oscillator circuits are designed for applications even under highly damped conditions. The (18) Dickert, F. L.; Tortschanoff, M.; Bulst, W. E.; Fischerauer, G. Anal. Chem. 1999, 71, 4559-4563.
frequencies are measured with a Hewlett-Packard 53131A frequency counter and transferred to a PC via a HP-IB bus. Atomic force microscopy (AFM) measurements were done with a DI Nanoscope IIIa. Polymer Coating Materials. The sensor coatings were highly cross-linked materials. Pre-polyaddition of the polyurethanes was performed in tetrahydrofuran (THF) with functionality ratios of isocyanato groups:hydroxy or amino groups of 1:2. Reactive 4,4′diisocyanatodiphenylmethane (30% triisocyanato, technical grade) was used for the polyurethanes. Linker groups were bisphenol A and 4,4′-diaminodiphenyl ether. The cross-linker rates of phloroglucinol, triethanolamine, or melamine were varied. The polymAnalytical Chemistry, Vol. 74, No. 6, March 15, 2002
1303
erization continued under stirring conditions at 70 °C until the gel point was reached. The polymerizing solutions were diluted in THF and subsequently used for the coating process. The curing of the polyurethane layer was done under ambient conditions. As mentioned before, stamps and cells were removed by washing in hot water. The polyurethane layers show excellent robustness even in ultrasonic baths. Sol-Gel Coatings. We tried different sol-gel materials for their imprinting capabilities. Surface-imprinted tetraethoxysilane (TEOS), binary mixtures of TEOS/methyltriethoxy- or diethyltriethoxysilane and titaniumIV ethylate (TE) as well as TE/(3aminopropyl)methyldiethoxysilane composites showed good imprinted surfaces comparable with the results gathered with polyurethane. The TE layers were prepared from a 0.1 M aqueous ethanol solution with hydrochloric acid as catalyst with poly(ethylene gylcol) as plasticizer. The TEOS layers were synthesized again from various concentration ratios of aqueous ethanol solutions. However, the best surface-imprinted layers are formed with TE sol-gels. We allowed the polymerization of a prenetwork for 30 min before using the TE/TE composite solutions for coating and printing (Figure 4). Chemicals and Microorganisms. The reagents were used as received from Merck and Fluka. We used lyophilized bacteria and compressed and active dried yeasts. Escherichia coli strain W (ATCC9637) was a product of Sigma. L. oenus as well as the yeast strains were provided from Uvaferm. Saccharomyces cerevisiae from Anker was in a compressed form. Yeasts and bacteria are resuspended in a 1/15 M KH2PO4/Na2HPO4 buffer of pH 6. Cell concentrations were determined using a Neubauer improved erythrometer. The freshly prepared cell suspensions for the measurements showed no significant agglomeration of microorganisms (controls performed with light microscopy).
Figure 5. Sensor effect of a S. cerevisiae imprinted and a nonimprinted polyurethane. (a) Polyurethane cross-linked with phloroglucinol: the sensor difference between imprinted and nonimprinted polymer illustrates the effectiveness of cell adhesion via surface imprinting. Five orders of cell concentrations can be monitored in situ and in flowing conditions. (b) Polyurethane layer without phloroglucinol as cross-linker: all measurements were done in phosphate buffer pH 6, 25 °C and at a flow rate of 5 mL/min.
RESULTS AND DISCUSSION Surface Imprinting Results. Since the QCMs have a diameter of 1 cm, we are able to perform AFM testing (topography, elasticity of the coating, surface properties before/after usage in the flow cell) of the surface-imprinted materials on the coated transducer. The yeast-imprinted sensor coatings show a regular honeycomb-like surface. Section analyses with AFM showed that the depth of the imprinted sites was ∼1 µm. Due to the versatile printing technique, we are able to generate imprinted surfaces at least up to several square centimeters. Figure 4 shows a magnification of a coated electrode on a dual QCM. A S. cerevisiae-imprinted surface has ∼240 000 grooves on a 2-mm-diameter gold electrode. The yeast and bacteria strains used have a size distribution resulting from the cultivation process. For example the S. cerevisiae-imprinted polyurethane shows striking similarities in the cell size distribution of a yeast fermentation which is gathered with electrical flow impedance measurements.19 Figure 2 illustrates single yeast cells surrounded by the polymerizing material. Measurement of Cell Concentrations. Sensor responses are only observed on sensitive electrodes coated with imprinted films. No unspecific adhesion of microorganisms is detected on nonimprinted polymers under flowing conditions. Using differential measurements between sensitive (imprinted polymer) and reference (nonimprinted) electrode, we were able to monitor cell
concentrations in the range from 104 to 109cells/mL in flowing conditions of 10 mL/min. Figure 5 shows a sensor response to different concentrations of S. cerevisiae. The adhered yeast cells are simply washed away with a high flow rate pulse of several hundred milliliters per minutes for a few seconds. When we suppose an average weight of 30 pg/cell of S. cerevisiae and a mass resolution of 1 ng/Hz for a 10-MHz device, we are detecting roughly 3000-4000 yeast cells/1 kHz mass effect or about 1-2% of our imprinted sites are constantly filled with yeast under flowing conditions and 1-kHz sensor effect. Additionally, we were successful in detecting a single yeast cell-MIP interaction by using dual 428-MHz shear transverse wave resonators. These acoustic wave devices can be applied for liquidphase measurements and the increased resonance frequency appreciably improves the sensitivity of the transducer. As can be seen in Figure 6, we compared the SAW frequency responses of 5-µm quartz microspheres with those of the yeast cells. The adhesion of the microspheres to the SAW surface was optically controlled and the effect in Figure 6A could be interpreted as the adhesion of one microsphere. Thus, the sensor response in Figure 6B might be interpreted as the effect of one cell, taking into consideration the different specific weights of quartz and the cells. Preliminary results with bacteria-imprinted polymers show roughly 5 times smaller sensitivities with cell concentrations comparable with the presented yeast results (data not shown). The bacteria desorb within a time scale similar to the adhesion process (response time similar to yeast adhesions).20 Physical Influences. The imprint process maximizes the interaction surface between enriched cell and molded layer material. This geometrical fit is the basis to high and specific affinity interactions. We thoroughly determined the dependence
(19) Homepage of MCS Diagnostics, http://www.mcsdiagnostics.nl.
(20) Hayden, O.; Dickert, F. L. Adv. Mater. 2001, 13, 1480-1483.
1304 Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
Figure 6. Sensor effect of a silica microsphere with an approximate 5-µm-diameter (A) and a yeast cell (B) adhered on the delay area of a 428-MHz SAW.
Figure 7. Sensor effects with varying pH values.
of the imprint effect on various physical parameters. Yeasts are well known for their sedimentation behavior.21 Therefore we varied the angle of the coated QCM relative to a horizontal position with the coated side facing upward in the flow cell. We could see that the sensor response gradually decreased the more the QCM was rotated into a vertical position. It should be noted that the aqueous solutions flows are highly turbulent through a flow cell volume of 250 µL (flow rates used are in the range of 5-10 mL/min) and again virtually no yeast adhesion occurs on the nonimprinted coatings. The most interesting effect was the independence of the sensor effect from the flow rate. We varied the flow rate in the range between 0.8 and 10 mL/min and did not observe a pronounced influence of the sensor effect. Chemical Influences. We varied the pH value around physiological conditions and the buffer concentration with a phosphate buffer. The highest sensor effects to S. cerevisiae in the case of polyurethane layers with hydroxy functionalities was around pH 6 (Figure 7). The increase of the buffer concentration did not influence the sensor effects. Yeast and bacterial surfaces normally possess a net negative surface charge at physiological pH values.22 It has been reported that cell surface charge influences the binding of microorganisms to charged polymers.23 We chose pH 6 to gain high yeast sensitivities, since we were mainly interested in detecting S. cerevisiae. Interestingly, we found an optimum phloroglucinol concentration for maximum imprint effect of the cross-linked polyurethane, which is similar to our results gathered with MIPs (21) Bos, R.; van der Mei, H. C.; Busscher, H. J. FEMS Microbiol. Rev. 1999, 23, 179-230. (22) Mozes, N.; Leonard, A. J.; Rouxhet, P. G. Biochim. Biophys. Acta 1988, 945, 324-334. (23) van Loosdrecht, M. C. M.; Lyklema, J.; Norde, W.; Schraa, G.; Zehnder, A. J. B. Appl. Environ. Microbiol. 1987, 53, 1898-1901.
Figure 8. Effect of added growth medium to a buffered yeast solution (pH 6). In the presence of yeast extract (a) the sensor response is blocked by adhered cell surface residues. The same growth media concentrations without the yeast extract (b) do not interefere with the sensor response.
Figure 9. Sensor response pattern of three different yeast strains as analyte and template, respectively, on same polyurethane layers (pH 6, 25 °C, flow rate of 10 ml/min).
for polyurethane-coated optochemical devices.24 The layers have good imprinting properties with increasing cross-linker rates. Anyhow, too high amounts of cross-linker lead to brittle polyurethane layers. For the same layer composition, we tested the effectiveness of our sensor coatings in growth medium. Increasing concentrations of yeast extract, peptone hydrolysate and dextrose (YPD) medium up to 21 g/L resulted in decreasing sensor response for S. cerevisiae (Figure 8a). In other words, cell surface fragments from the yeast extract inhibited a proper interaction between yeast and imprinted surface. The sensor effect was not (24) Dickert, F. L.; Thierer, S. Adv. Mater. 1996, 8, 987-988.
Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
1305
influenced by the medium when we removed the yeast extract from the medium and repeated the experiment (Figure 8b). Selectivity of the Imprinted Polymers. Polyurethane layers with hydroxy components were imprinted with yeast of same genera S. cerevisiae, Saccharomyces bayanus, and Saccharomyces diastaticus. The sensor response to equal concentrations of the yeast strains is shown in Figure 9. S. cerevisiae and S. bayanus have pronounced template properties and show the best adhesion on their respective imprinted polymer. S. diastaticus has less templating properties but gives the best sensor response on the S. bayanus-imprinted layer. The imprinting effect can be seen particularly between S. bayanus and S. cerevisiae. The templating effect of both yeasts leads to increased adhesion on the MIPs (template yeast ) analyte yeast). The sensor effect is reduced in the case where, for example, S. cerevisiae (analyte) is exposed to the S. bayanus-imprinted MIP and vice versa, although the yeasts are from the same genera. Sol-Gel Layers. The titanium oxide sol-gel layer in Figure 4 illustrates the result of surface imprinting, which can also be done with this type of sensor material. Compared to polymer sensor coatings, the sol-gel layers are even more robust. In the case of the titanium oxide surface of the QCM, the coating is scratch resistant and can also be deemed as mechanical protection for the electrode structure. The surface structure of the imprinted sol-gel layers is comparable to the MIPs. However, only minor sensor responses to, for example, S. cerevisiae were observed so far (
