Scribd Logo
Upload

Welcome to Scribd!

  • Bio Recognition

    Uploaded by

    Lucas Eduardo
    22 views7 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    22 views7 pages

    Bio Recognition

    Uploaded by

    Lucas Eduardo

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 7
    Biosensors & Bioelectronics 14 (1999) 405-411 Specific bio-recognition reactions observed with an integrated Mach-Zehnder interferometer M. Weisser *!, G. Tovar *, S. Mittler-Neher *, W. Knoll **, F. Brosinger >, H. Freimuth *, M. Lacher*, W. Ehrfeld > * Max-Planck-Insitu fir Polymerforschung, Ackermannseg 10, D-$5128 Mainz, Germany © MM Institut fr Mikrotechnik Mainz, GmbH, Carl-Zeiss-Straple 18-20, 55129 Mainz, Germany Received 6 April 1998; accepted 12 November 1998 Abstract ‘The combination of an integrated Mach-Zehnder-interferometer (iMZI) at the bottom of a fluidic microchannel system with supramolecular interfacial binding layers optimized for biosensing purposes is described. The model system used is based on the highly specific interaction of streptavidin to its ‘ligand’ biotin: a single monolayer of a correspondingly derivatized silane-com- pound is formed by a self-assembly procedure on top of the channel rib guiding the light through the channels. Injection of a streptavidin solution which leads to the formation of a protein monolayer of d= 2.8 nm in effective thickness results in a phase shift of the sample light relative to the reference channel of A®=6n, in good agreement with the theoretical sensitivity of A@jAd,=2.9 x/nm for a protein layer (n= 1.45) calculated for the device. © 1999 Elsevier Science S.A. All rights reserved. Keywords: Integrated optics; Mach-Zehnder interferometer; Biorecognition reaction; Self-assembled monolayers; Biotin—Streptavidin 1. Introduction The vision of the ‘laboratory on a chip’ has stirred up a considerable interest in microsystem technology and its associated enabling technologies (Widmer et al., 1996). In addition to the design, construction, and performance optimization of the active key components like microreactors, pumps and mixers, the correspond ing research and development thereby has focused on elements required for integrated diagnostics and process control (Ehrfeld, 1996). Numerous designs and examples have been presented in the literature for the on-line monitoring of process parameters and the characterization of reaction prod- ucts, based on electrical, fluorescence optical and den- sitometric detection principles (Puers, 1996). * Corresponding author, Tel +49-6131-379160; fax: + 49-6131- 379360. E-mail address: knoll@mpip-mainz.mpg.de (W. Knoll) "Present address: Laboratory for Surface Chemistry & Technol ogy, University of Stuttgart, Nobelstrape 12, 70569 Stuttgart, Ger- many, “Present address: IBM Deulschland Speichersysteme GmbH, Hechtsheimer Strap 2, $5131 Mainz, Germany. Surprisingly, little attention has been paid so far to the continuous monitoring of process parameters by con- cepts based on integrated optics (Heidemann et al. 1993; Ingenhoff et al., 1993; Gauglitz and Ingenhoff, 1994; Zappe et al., 1994; Brosinger et al., 1997) despite the obvious compatibility of the elements of micro- fluidics and opto-boards (Kunz ct al., 1994; Schmitt et al., 1997). Here, we present first results of our efforts to com- bine the analytical capabilities of an integrated Mach- Zehnder interferometer (iMZI) to detect and quantify the specific binding of a receptor protein from solution to its surface-bound ligands with a miniaturized micro- fluid cell fabricated from _ silicon/siliconoxynitride (Brosinger et al., 1997). 2, Experimental 2.1. The Mach-Zehnder interferometer The fabrication and some performance parameters of the employed integrated iMZI have been described in (0956-5663/99)$ - see front matter © 1999 Elsevier Science S.A. All rights reserved. PII: $0956-5663(98)00124.9 406 M, Weisser et al. / Biosensors & Bivelectronics 14 (1999) 405-411 detail in a previous paper (Brosinger et al., 1997). Some of the characteristic features and dimensions are sum- marized in Fig. 1. The two branches of the iMZI had an effective length Lyeq,= 12 mm (Fig. 1(a)). The SiO, buffer layer (Fig. 1(b)) on top of the Si substrate had a thickness of ~ 2.5 um. The active guiding structure was made from siliconoxynitride with a layer thickness of 350 nm (d+h) in Fig. 1(b)) and a refractive index n= 1.55. Beam confinement was achieved by etching a rib into this layer, 2 um wide and ~ 55 nm in height, A. In order to preclude unwanted environmental inftu- ences, the whole chip was covered with 2 wm SiO,, The (a) (b) fluid channel =~ Ww, sensitive area waveguiding layer h with etched rib’ Ee £ ibe es buffer layer substrate —{ (©) Plexiglass cover Parafiim ‘500 jim thick. LSJ Chip imzi Fig. 1. (a) Waveguide chip lay-out with integrated MZI; (b) eross-sec~ tion of the guiding structure with integrated uid channel; (c) sche ‘matic of the final assembly consisting of the chip, a parafilm seal with two separate fluid channels, and a Plexiglas cover with separate inlet and outlet for the two mierocuvette channels, position of the liquid cell is defined by reactive ion etching of a window in the SiO, cover (Fig. 1(b)). The structure was completed by pressing a poly- (methylmethacrylate) (PMMA) cover slip to the chip with a parafilm mask (500 jum thick) as a seal in-be- tween separating two independent liquid channels in contact to the two optical branches of the iMZI. This mask was structured such as to define two separate channels for the liquid flow (Fig. 1(€)). Four holes in the Plexiglas cover allowed for the individual perfusion of the liquid channels in contact with the two branches of the iMZI. The TE-polarized chopped light of a 5 mW HeNe Laser (4 = 632.8 nm) was endfire-coupled to the inter- ferometer, and the out-coupled light monitored with lock-in technique. 2.2. Functionalization of the iMZI sensing surface In order to act as a specific (bio-) sensor, the interfer- ometer had to be functionalized at its surface by a thin coating that exposes selected binding sites to the ana~ lyte in the aqueous phase passing through the fluid channels. This surface layer has to fulfil two require- ments; (I) the binding sites, e.g. ligands, antibodies, or DNA capture probes have to be organized at the interface such as to allow for maximum binding of the analyte (receptors, antigens, or complement DNA strands); and (2) the chemical nature and the molecular architecture of the surface layer should minimize the nonspecific binding (NSB) of all components in the solution passing through the cell, ‘As a model reaction, we selected to study the binding of streptavidin, a tetrameric protein of molecular mass 60 kDa, to its ‘ligand’ biotin, which is easily coupled to other (linker-) molecules through its carboxy-endgroup not involved in the recognition and binding reaction (Wilchek & Bayer, 1990). In order to ensure a molecu- larly controlled architecture of the interfacial layer, a self-assembly strategy was chosen that had been suc- cessfully employed for the functionalization of Au sur faces (Spinke et al., 1993a,b): long chain derivatives of dimethyl-monomethoxy-silanes, carrying as a func- tional endgroup either a -COOH-moiety or the biotin group, are co-adsorbed from binary solutions (Tovar, 1995). The structure formula of the two compounds used in this study (kindly provided by Boehringer Mannheim) are given in Fig. 2a). The binding capacity of the mixed monolayers towards streptavidin molecules were first characterized by surface plasmon spectroscopy (SPS) (Knoll, 1997). For this purpose, the ‘Ag substrates used for the reflectivity measurements in the Kretschmann configuration were pre-coated by evaporating SiO,, thus mimicking the glassy surface of the iMZI. Details of this preparation, e.g. the reactive -M, Weisser et al. / Biosensors & Bioelectronics 14 (1999) 405-411 407 06 08 1.0 Fig, 2. (2) Structure formula of the two end-functionalized silane-derivatives used in the assembly of a binary mixed SAMS (b) effective thicknesses measured by surface plasmon spectroscopy after injection and binding of streptavidin (~~) for binary mixed SAMs of different biotin content, biotin prior to injection deposition of SiO,, its activation in a HO containing RF-plasma or by KOH treatment in order to generate the silanol groups required for the silan-coupling reac- tion, and the monolayer formation and annealing have been described in detail elsewhere (Tovar, 1995) This preparation strategy allows for the build-up of a stable monomolecular layer, functionally optimized for the specific binding of a single layer of streptavidin while simultaneously NSB is minimized (Fig. 2(b)). The latter could be checked by blocking all binding sites of streptavidin by excess free biotin before injection into the liquid cell. The effective thickness values given in Fig. 2(b) are calculated from the corresponding angular shifts of the surface plasmon resonance induced by the respective layer formation using a Fresnel formalism ‘The data labelled NSB (—@-) refer to measurements of the adsorption of streptavidin, the binding sites of which were blocked by free and assuming a refractive index of n,=1.45 for the protein layer. The obtained maximum effective thick- ness of the protein monolayer of 2.8 nm has to be compared to the protein’s dimension of ~4.5 x 4.5 x 5.2 nm?, and hence corresponds to a surface coverage of ~ 45-50%, It should also be noted that the shift of the resonance angle induced by 2 2.8 nm thick strep- tavidin layer is equivalent to a bulk refractive index change (probed by the evanescent surface plasmon field) relative to pure buffer (0.5 M NaCl) of An=3.0 10-°. Based on these results, we choose a binary mixture with a mole fraction of the biotinylated com- pound of Xpioxin = 0.3 for the surface functionalization of the opto-chip. The whole device was kept in the 408 M. Weisser etal + branch Biosensors & Bioelectronics H,O—> 05 M Nact 14 (1999) 405-411 H,O—> 0.5 MNacl : 2" branch ar) : | | (b) | ahaa cE tisec tisec 10] 10 | 4 sol se z4 a4 2| q| | q o Tbr ae Fa Tr sec tisec Fig. 3. Sensitivity and symmetry test of the iMZI: (a) gives the output signal obtained while replacing the water of one of the liquid channels by 0.5 M NaCl solution; the slight drift of the signal for r> 150 s is presumably due to some leakage between the two liquid channels owing to insufficient isolation by the parafilm. This effect is absent in the case of symmetric liquids (cf (b)) and hence does not constitute a real instability (of the device; (b) shows the full reversibility of this 11.7 phase shift by also exchanging the water in the second channel by 0.5 M NaCl methanolic silane solution for 18 h, carefully rinsed with pure methanol, dried in a stream of nitrogen, and annealed for 2 h at 120°C and 10~* mbar. 3. Results 3.1. Monitoring bulk refractive index changes After mounting the silanized micro-cell and filling the fluid channels with water (milli Q quality), the two branches were sequentially perfused with 0.5 M NaCl solution propelled by a peristaltic pump which serves three purposes: (1) the procedure was a stability test of the surface functionalization; (2) the recording of the output intensity modulated by the asymmetric replace- ment (in one channel only) of the water against the NaCl-solution with its concomitant refractive index change is a sensitivity test and calibration of the sensor performance; and (3) the replacement of the liquid in the second branch allowed for a symmetry test of the device response function, ‘The result of this sequence of operations in displayed. in Fig. 3. Fig. 3(a) shows the output intensity modula- tion caused by the exchange of the milliQ water by 0.5 M NaCI solution. The intensity variation corresponds to a total phase change of 11.7, which is fully reversed upon the exchange of the liquid in the second branch (Fig. 3(b)). The time course of these intensity oscilla- tions is given by the rate of solution exchange which in ‘turn is determined, in our case, by the pumping effi- ciency of the peristaltic pump. 3.2. Monitoring the binding of streptavidin to the biotin-functionalized self-assembled monolayer As can be seen from Fig. 4(a), the injection of a 10~* M streptavidin solution (in 0.5 M NaCl) into the first liquid channel causes an intensity modulation corre- sponding to ~6z. Injection of the identical strep- tavidin solution into the second channel reverses the oscillations and even overcompensates the phase shift of the first branch by x (Fig. 4(b)). This originates from the fact that both fluid channels are not completely identically functionalized with the same density of bi- otin sites. As a consequence, the exposure of the two slightly different sensor channels leads to a slightly M, Weisser et al. / Biosensors & Bivelectronics 14 (1999) 405-411 408 a 10) ~ z4 Pa + Streptaviain + Streptavidin +05 MNacl i 4% branch 2° branch 41% branch 7 () () (©) —____. ol \ fut 7 705 3004000580 wae To TO eo T8058 sec tse Use ‘al pow} ci a4 a, | 75 1 a 720 a 75 7 = 780 ti sec sec Fig. 4, Specifie binding of streptavidin to the biotin-functionalized silane-SAM on the waveguide structure: (a) gives the output signal after the injection of a 10 M streptavidin solution (in 0.5 M NaCl) to the fist channel only. The resulting 6x phase shift is overcompensated by 2 if streptavidin is also injected into the second channel (b). Rinsing the first channel by pure solution (0.5 M NaCl) causes a further phase shift AO 1.2n within ~ 1500 s (). different optical response owing to a different density of bound streptavidin. In the following we hence use an average value of the phase change of Ad = 6,57. If the first channel is rinsed with pure 0.5 M NaCl solution (a procedure which is known to also remove the fraction of bound streptavidin which is only phy- sisorbed). a phase shift of 1.2 is observed (Fig. 4(0)) 4, Discussion The presented results have shown that the combina- tion of an iMZI with an integrated microftuidic han- dling system allows for the sensitive monitoring of even minute refractive index changes occurring either in the bulk of the liquid solution, An,, probed by the evanes- cent field of the guided optical mode, or by a thin layer of material, Ad,, built up by adsorption directly onto the channel waveguide surface (or, of course, both). The basic mode of operation of an (integrated) iMZI is the following: An optical phase (velocity) change A®, in one of the branches of the interferometer (Lukosz and Tiefenthaler, 1988) 2 am" Od, where: / is the wavelength of the light, Vis the effective refractive index of the structure describing the propaga- tion behavior of the guided mode, and @N/én, and @N/éd; depending on optical channel parameters like cross-section and refractive indices, is transformed into an output power signal oscillating by interference ef ied) reference chan- A© = Lesos 2 (an +Waa) a fects with the power of an (unmo nel according to P ea + cos A®) @ with P, being the light power coupled into the device. In the case of a bulk refractive index change of the liquid in the micro-cuvette induced by replacing pure HO by 0.5 M NaCl solution, the obtained phase shift equalled At = 11.72. With the refractomettically deter- mined index change between water (71 = 1.3318) and 0.5 M NaCl (n= 1.3369) of An, = 5.1-10-%, one obtains an experimental sensitivity of A@/An, = 2.3- 10-7 which hhas to be compared to the theoretical sensitivity of ‘A@/An, = 2.1 10°x based on an approximate expres- sion for 6N/én, given for a planar waveguide structure (Lukosz and Tiefenthaler, 1988). 410 M, Weisser et al. / Biosensors & Bivelectronics 14 (1999) 408-411 The concept of introducing a selective surface func- tionality for specific biorecognition and binding by the co-assembling of corresponding end-functionalized lane molecules developed for large-area planar sub- strates also proved to be ideally suited for the modifica- tion of the channel waveguide surfaces for bio-sensing purposes with the iMZI. These binary mixed self-assem- bly monolayers (SAMs) provide an interfacial architec- ture which can be optimized for specific binding reactions while simultaneously suppressing NSB (Spinke et al., 1993a,b). The sensitivity of a given channel waveguide for small thickness variations induced by a binding reaction to its surface is determined by 2N/éd;. By inserting the approximation given for planar waveguides (Heide- mann et al., 1993; Lukosz and Tiefenthaler, 1988) into Eq, (1), one obtains a theoretical sensitivity of the iMZI for monitoring biorecognition reactions of A®/Ad,= 2.9 minm for n, = 1.45. Again, the experimentally deter- mined phase change upon binding of (a monolayer of) streptavidin to the biotin-sites on the channel, A®~ 6.57, compared well to this expectation if we assume the formation of a protein monolayer of 2.8 nm thick- ness as in the case of the SPS experiment (cf. Fig. 2 (b)). ‘Another way to compare the two experiments is based on the following consideration: The formation of a streptavidin monolayer with n= 1.45 and d= 2.8 nm at the surface of the planar functionalized substrate in the SPS set-up results in a shift of the resonance angle corresponding to a bulk refractive index change of An, =3-10->. With the theoretical A@/An, = 2.1 - 10°, this should result in a phase shift of A®= 6.3x, which is in excellent agreement with the observation. Accepting a lower bound for the unambiguous moni toring of binding events of A® = 1/20, one thus derives at an experimentally determined lower limit of the sensitivity of the device of ~ 1/100 of a streptavidin monolayer, corresponding to Adin = 0.02 nm which is equivalent to a mass coverage of ~ 2 ng/cm’. This sensitivity is at least one order of magnitude better than obtainable with surface plasmon spec- troscopy (Spinke et al., 1993a). One has to keep in mind, however, that this is possible only because of the relatively long sensing path, integrating all the subtle phase changes of the propagating mode. The propaga- tion (= integration) length in the SPS experiment of Fig. 2 was in the order L, +30 yum only, and could be further reduced to the few jum range (e.g. by using Au substrates (Aust et al., 1994)). L,, on the other hand, determines the minimum spot size of a highly integrated multi-element detector array operated in a surface plas- mon microscopic mode for massive parallel (bio-) sens ing (Zizlsperger and Knoll, 1998). ‘The observed phase change of A@

    You might also like