Antibody-based immobilization of bioluminescent bacterial sensor cells

Talanta 55 (2001) 1029–1038

Antibody-based immobilization of bioluminescent bacterialsensor cells

J. Rajan Premkumar a, Ovadia Lev a,*, Robert S. Marks b, Boris Polyak b,Rachel Rosen a, Shimshon Belkin a

a Di�ision of En�ironmental Sciences, Fredy and Nadine Hermann Graduate School of Applied Science,The Hebrew Uni�ersity of Jerusalem, Jerusalem 91904, Israel

b Department of Biotechnology Engineering, The Institute for Applied Biosciences, Ben-Gurion Uni�ersity of the Nege�,PO Box 653, Beer-She�a 84105, Israel

Received 4 May 2001; received in revised form 20 August 2001; accepted 20 August 2001

Abstract

Whole-cell luminescent bioreporter sensors based on immobilized recombinant Escherichia coli are described andevaluated. The sensors were prepared by glutaraldehyde-anchoring of nonspecific anti-E. coli antibodies onaminosylilated gold or silica glass surfaces with subsequent attachment of the probe bacteria. We demonstrate thegenerality of the concept by attachment of several E. coli strains that express luciferase in response to differentphysiological stress conditions including heat shock, DNA damage (SOS), fatty acid availability, peroxide andoxidative stress. The sensors can be used either as single- or multiple-use disposable sensing elements or forcontinuous operation. We show compatibility with optical fiber technology. Storage stability of the sensors exceeded5 months with no measurable deterioration of the signal. Repeatability on exposure in successive days was �15%,as was sensor to sensor reproducibility. Sensitivity and detection limits of the immobilized cells were comparable tothat of non-immobilized bacteria. © 2001 Elsevier Science B.V. All rights reserved.

Keywords: Bioluminescence; Sensor; E. coli ; Antibody

www.elsevier.com/locate/talanta

1. Introduction

Genetically modified reporter microorganismscan be used as sensitive and inexpensive analyticaltools for laboratory and field toxicity assessment[1,2]. Such organisms can be ‘tailored’ to emit a

readily detectable signal in response to pre-deter-mined environmental conditions. This is normallyachieved by the selection of a gene induced inresponse to the desired stress, and the geneticfusion of its promoter to a gene or a group ofgenes coding for reporter proteins, the activity orpresence of which can be easily monitored. In E.coli and other microorganisms, the use of colori-metric, electrochemical, fluorescent or luminescentwhole-cell reporting has been successfully demon-strated [1,2]. In this paper, we make use of a set of

* Corresponding author. Tel.: +972-2-6585558; fax: +972-2-6586155.

E-mail address: [email protected] (R.S. Marks).

0039-9140/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.

PII: S 0039 -9140 (01 )00533 -1

mailto:[email protected]

J.R. Premkumar et al. / Talanta 55 (2001) 1029–10381030

genetically modified E. coli strains that emit adose-dependent luminescent signal in response tothe activation of several global defense circuits[3–5]. The specificity of the different strains wasdetermined by the choice of the gene promoter.

While the high sensitivity and applicability ofthese and similarly constructed bacterial sensorcells under controlled laboratory conditions hasbeen repeatedly demonstrated, less attention hasbeen devoted to their long-term storage and use inan immobilized form. The successful incorporationof such microorganisms in solid-state matrices is animportant step en-route to the conversion of thesebioassays into user-friendly sensing devices capableof continuous monitoring or multiple use detection.A few attempts to embed genetically engineeredprobe bacteria in solid matrices were reported,based on encapsulation in soft gels includingagarose [6], acrylamide [7], calcium and strontiumalginates [8], and agar [9]. Polyak and coworkers[10] reported that alginate-based matrices sufferfrom diffusion limitations due to the thick films butdid create viable biosensors [11]. Alginates are alsounstable in calcium-poor solutions, and deterioratein the presence of phosphate and other calciumchelators although this has not been shown to bethe case in field studies (Polyak and Marks, unpub-lished results). The low deformation resistance andthe biodegradability of most soft gels are additionalincentives for a search for an alternative encapsu-lating procedure. Deposition of a protective dialy-sis membrane was also reported [12] although thiscannot be considered a true encapsulation.

In this study we present a simple and generalapproach for the immobilization of recombinant E.coli on solid substrates and their use as disposableor multiple use bioluminescent sensors. The proce-dure is based on the immobilization of the engi-neered bacteria on antibody modified substrates.The technology for bonding antibodies on differentflat or porous substrates is very versatile and welldeveloped and it is often used as a first step in theconstruction of complex sensing elements. Thesuccessful combination of this generic approachwith reporting microorganisms paves the way fortheir incorporation on or in virtually any substrate.So, despite the fact that in this article we addressonly the immobilization of luminous bacteria onglass and gold substrates, it establishes a way thatcan be used for the integration of recombinantbacteria in other cell-based-biosensing-architec-tures as well. In this article, we demonstrate thevalidity of this generic approach for two types ofsurfaces: glass and gold-coated glass. We furtherdemonstrate compatibility of the technologywith different modes of optical fiber signal trans-duction.

2. Experimental

2.1. Preparation of the recombinant E. coli

The five E. coli strains used in this study arelisted in Table 1. Each contained a multi-copyplasmid in which a different gene promoter was

Table 1Bioluminescent bacterial strains and the inducers used in the present study

Strain Concentration ReferenceE. coli host strain Promoter Stress sensitivity Inducer used in(M)this studydesignation

RFM443 [3]TV1061 0.43agrpE EthanolHeat shock (generalstress)

[4]5.9×10−6NAbSOS (genotoxicity)recARFM443DPD2794RFM443 katG Oxidative (peroxides)DPD2511 H2O2 2.9×10−3 [13]

DPD2544 W3110 fabA Fatty acid availability [14]Phenol 5.2×10−4

(general stress)W3110 micF Oxidative (superoxide) MVc 3.9×10−3 [15]DPD2515

a 2% (v/v).b Nalidixic acid.c Methyl viologen

J.R. Premkumar et al. / Talanta 55 (2001) 1029–1038 1031

fused to the Vibrio fischeri luxCDABE genes. Theuse of the 5 lux genes under the same promoteralleviates the need for externally supplying along-chain aldehyde, the substrate for bacterialluciferase. Construction of the different plasmidshas been described in detail in the past in therespective references in Table 1.

2.2. Preparation of bacterial cultures

Prior to immobilization, the bacterial strainswere grown overnight with shaking at 37 °C inLuria–Bertani (LB) broth (NaCl 5 g l−1, Yeastextract 5 g l−1, Tryptone 10 g l−1) [16] containing100 �g ml−1 ampicilin to ensure plasmid mainte-nance. The stationary culture was diluted to ca.107 cells per ml with fresh LB medium and thenregrown under the same conditions. Growth wasmonitored using a Klett-Summerson colorimeter(Monostat Corporation) and the cells were har-vested at early exponential growth phase (20 Klettunits, corresponding to about 108 cells per ml aspreviously described [3,5,13].

2.3. Luminescence determination

The effects of different inducers on non-immo-bilized bacteria were measured by mixing early-exponential growth cultures aliquots (50 �l) withequal amounts of LB broth containing the appro-priate inducer concentration. The mixture wasplaced into duplicate wells of a 96-wells opaquewhite microtiter plate (Costar Europe, Bdhoeve-drop, The Netherlands). The microtiter plateswere incubated in a temperature controlled(26 °C) shaker, and the luminescence was periodi-cally measured by a microtiter plate luminometer(model Victor2, EG&G Wallac 1420, Turku, Fin-land). LB broth and bacterial samples without theinducer served as controls.

The luminescence response of the immobilizedbacteria was measured by a similar procedure,with the solid substrate (glass or gold), along withits attached bacteria placed into a 100 �l LBcontaining the appropriate inducer concentrationin duplicate microtiter plate wells. In all cases, thesignal is reported in the instrument’s arbitraryRelative Luminescence Units (RLU).

2.4. Optical fiber measurements

Multimode optical fibers, PUV 400 BM (Cer-maOptec, GmbH, Bonn), were used in these ex-periments. They present a pure core diameter of400 �m, with a refractive index of 1.4571 (at 633nm) and a cladding diameter of 440 �m, with arefractive index 1.4011 (at 633 nm). Their blacknylon jackets were stripped away from 1-cm longoptical fiber tips, which were then used for immo-bilization of the bioluminescent cells. The opticalfiber was placed in a 0.5 ml conical tube (Jonplast,Italy) within the test sample solution or in a 1-cmin diameter opaque flow cell (an opaque 0.5 cmplastic T junction). The optical fiber was fixed atthe epicenter of the light proof tube via a 100 �lpipette tip to avoid ambient light interference.The far end of the fiber was held by a fiber holder(FPH-DJ, Newport) and placed into an adjustablefiber mount (77837, Oriel). The photon countingsystem was comprised of a light-tight box and aHamamatsu HC135-01 photomultiplier tube sen-sor module, as was described elsewhere [10,11].The response is reported in the instrument’s rela-tive response units (photon counts per second).

2.5. Preparation of anti-E. coli modified glass slides

Glass plates (GP) (1 mm in thickness) werepurchased from Paul Marienfeld GmbH & Co.KG Laboratory Glassware Company (Germany).The plates were cut into squares (4×4 mm)which were cleaned successively by ethanol, soapsolution, piranha solution (H2SO4:H2O2 4:1), andfinally with NaOH (1 M). The clean plates weredried at 80 °C and then aminosylilated by immer-sion for 3 h in N [-3-(trimethoxysilyl)propyl]-ethylenediamine (EDAS) at 80 °C. Theamine-coated samples were rinsed with distilledwater and then immersed in 2% (v/v) glutaralde-hyde for 30 min. Excess glutaraldehyde was rinsedoff and the glass plates were immersed for 30 minin an aqueous antibody solution containing 13 gprotein l−1 (AB, Rabbit Anti-E. coli, DAKO,Denmark). The plates were rinsed again in dis-tilled water and immersed in the E. coli suspen-sion (ca. 2×108 cells per ml) for 30 min. The


coated glass plates were washed with LB medium(pH 7), and kept in microtiter plates (Dynatech,Chantilly, VA, USA) in the same solution. At-tachment of the bacteria to the optical fibers wascarried out in an identical manner. The bacteriawere applied only to the exposed upper 10 mm ofthe fiber.

2.6. Preparation of anti-E. coli modifiedgold-coated glass slides

Gold-coated glass slides (1 mm in thickness)(home made by electrodeposited gold on ITOslides) were cut in squares (4×4 mm) and dippedinto 0.1 M 2-aminoethanethiol hydrochloride for3 h. They were then rinsed with distilled waterand immersed in 2% glutaraldehyde for 30 min.Excess glutaraldehyde was rinsed off and the goldcoated glass slides were immersed in the aqueousantibody solution (13 g protein l−1) for 30 min.The plates were rinsed again in distilled water andimmersed in the E. coli suspension (ca. 2×108

cells per ml) for 30 min. The coated glass slideswere washed with LB medium (pH 7), and werekept in microtiter plates (Dynatech, Chantilly,VA, USA) in the same solution.

2.7. Chemicals

Glutaraldehyde, 2-Aminoethanethiol hydro-chloride, N [-3-(trimethoxysilyl)propyl]-ethylenedi-amine), nalidixic acid and phenol were purchasedfrom Aldrich. ACS grade hydrogen peroxide(Merck), ethanol (J.T. Baker), malathion (Ameri-can Cyanide Company), aldicarb (Union Carbide)and DDT (Union Carbide) were used.

3. Results

Microscopic observation of the immobilized E.coli cells attached to antibody modified glass andgold-coated slides revealed that the average sur-face density of the bacteria was ca. 105 cells percm2 corresponding to ca. 2% surface coverage. Ina series of attempts to increase the density of theattached bacteria, we have changed the durationof the silanization procedure, increased the bacte-

Fig. 1. Luminescence of TV1061 (E. coli ) glass sensors thatwere prepared by immersion of the antibody-modified glass inan E. coli culture (108 cells per ml) for the specified time. Thesensors were stored overnight in LB medium and then exposedto a 2% alcohol solution in LB medium for 1 h before thesignal was measured.

rial density during the deposition process, as wellas increased the E. coli deposition time on theantibody-coated substrates. None of these treat-ments increased significantly the surface coverage.Fig. 1 depicts an example of such a study. Weimmersed the antibody coated glass slides in thebacterial culture (108 cells per ml) for the specifiedamount of time, then exposed the slide to asolution of 2% alcohol in LB medium for 30 min,and finally recorded the induced luminescence.Fig. 1 shows that ca. 10 min immersion time ofthe antibody-coated glass in the bacterial suspen-sion was sufficient to attain maximal response.The maximum that appears in Fig. 1 for a 25 mindeposition probably reflects data scatter ratherthan true optimal operating conditions. There-fore, we used 30 min antibody binding duration inall subsequent studies. The rather low surfacecoverage may be attributed to incomplete cover-age of the solid surface by the antibodies or to thenonspecificity of the anti-E. coli polyclonal im-munoglobulins used in this study, which allow forcompetition of cell lysis products for antibodybinding sites.

3.1. The general immobilization concept

Fig. 2A–E compare the evolution of the biolu-minescence signal of the immobilized recombinantbacterial strains in Table 1 in response to their


model inducers. In all cases but one, no signal wasobtained in the absence of inducer (always thelowest line in the respective figures). The oneexception was strain DPD2511 (Fig. 2E), reacting

to the presence of hydrogen peroxide. This phe-nomenon was already described, and was shownto be caused by H2O2 formation in non-inocu-lated LB medium [13].

Fig. 2A depicts the stimulation of luminescencein strain TV1061, which reflects the activation ofthe bacterial heat shock system. It has previouslybeen shown [3,5] that this global circuit is inducedin response to a very broad variety of environ-mental insults and, therefore, induction of lu-minescence in this strain is an excellent indicationof general toxicity. The inducer used in this caseand during the rest of this study, ethanol, isroutinely employed as a general heat shock activa-tor. Fig. 2A shows the response of the two solidsensors (TV1061 deposited on glass (i) and goldsubstrates (ii)). The dynamic response of the twowas very similar and both were also similar to thephotoresponse of suspended (non-immobilized)bacteria curve (iii) in Fig. 2A. Fig. 2B–E similarlydepict the response of the other E. coli sensorstrains to their respective inducers using the glass-immobilized bacteria only. The responses of im-mobilized DPD2794 (a genotoxicity sensor) tonalidixic acid, DPD2515 (a sensor of superoxideradicals) to methyl viologen, DPD2544 (an addi-tional sensor of general toxic stress) to phenol andDPD2511 (a sensor of peroxide stress) to hydero-gen peroxide are all in line for those reported forthe non-immobilized cultures [4,13–15]. Sensitiv-ity and detection limits of the immobilized cellswere comparable to that of non-immobilizedbacteria.

Fig. 2.

Fig. 2. Time course of the evolution of luminescence ofdifferent E. coli reporters (as of Table 1) in response to theaddition the respective inducers. Luminescence response inblank LB medium is also shown (lowest curves in all figures).(A) (i) Glass-based, (ii) Gold-based TV1061 sensors and (iii)TV1061 suspension exposed to 0.434 M ethanol (right axis).(B) Glass-based DPD2794 sensors exposed to nalidixic acid (6�M). (C) Glass-based DPD2515 sensors exposed to methylviologen (3.9 mM). (D) Glass-based DPD2544 sensors exposedto phenol (0.5 mM). (E) Glass-based DPD2511 sensors ex-posed to hydrogen peroxide (2.9 mM).


3.2. Pesticide assays

It was demonstrated that heat shock gene ex-pression [17] is induced by numerous water con-taminants and thus E. coli harboring a fusion of aheat shock promoter to the lux genes (e.g. ofVibrio fischeri ) can serve as a general early warn-ing ecotoxicity sensor. In an attempt to examinethe applicability of strain TV1061 in its immobi-lized form, we have monitored its response toseveral pesticides. Three such examples are pre-sented in Fig. 3, following exposure to aldicarb,malathion, and DDT (0.1 mM each). In all cases,the increase in luminescence was apparent shortlyfollowing the addition of the pesticide, and reli-able signals could be obtained in 1 h. A similarresponse was obtained for non-immobilized bacte-ria (not shown). The sensitivity and detectionlimits of the immobilized cells were comparable tothat of non-immobilized bacteria.

3.3. Concentration-dependence of the luminescentresponse

Fig. 4 depicts the dose-dependency of the lu-minescent response of immobilized TV1061 cells(several identical sensors) to ethanol. Regardlessof the sampling time, maximal signal was ob-tained in the presence of 2% (0.43 M) of thisinducer. Higher concentrations imposed a toxiceffect on cell viability and on the bioluminescencereaction, causing a decrease in the overall signal.Similar response was observed for the suspendedbacteria as well.

3.4. Reproducibility, repeatability and stability

Several metrological characteristics of the heatshock based toxicity sensors were investigated.Fig. 5 demonstrates the time course of 4 differentglass-based sensors, prepared independently undersimilar conditions and exposed to 2% alcohol.The four sensors exhibited similar behavior,demonstrating good in-batch reproducibility. Fig.6 depicts typical multiple use of the same sensingelements in consecutive days (Fig. 6A) and weeks(Fig. 6B). The sensors were stored in LB mediumat 4 °C between tests. Each test began with rins-

ing of the sensing element with LB and introduc-ing it into 2% alcohol in the same medium. Thetime course of the luminescence signal wasrecorded daily or weekly. At the end of the cycle,the sensors were rinsed with fresh LB and storedin the same medium at 4 °C. The sensors showedgood repeatability even after multiple use (Fig.

Fig. 3. Toxicity test of Glass-based TV1061 sensors exposed to0.1 mM of (A) aldicarb; (B) malathion; and (C) DDT. Lu-minescence readings were taken 30 min after exposure of thesensing element to the tested pesticide solution. The lower linescorrespond to the response of sensors in LB medium withoutthe inducer.


Fig. 4. Concentration– luminescence curves for Glass-basedTV1061 sensors exposed to the specified concentration ofethanol in LB medium for (a) 130; (b) 220; (c) 280; (d) 330;and (e) 400 min.

placed the optical fiber in close proximity to a4×4 mm E. coli TV1061 glass- or gold-boundsensor and used the optical fiber as a wave-guide.In another approach, we immobilized the bacteriaonto the optical fiber tip according to the proce-dure delineated in the experimental section. In thelatter configuration, the optical fiber was usedboth as a sensing element and as a signaltransducer.

As a preliminary step towards the applicationof optical fibers as wave-guides, we have at-tempted to optimize the position of the fiber tip inrespect to the glass elements harboring the cells.Fig. 7 delineates the sensor response as a functionof the distance between the flat end of the opticalfiber and glass- and gold-based TV1061 sensors.In both cases, actual contact with the bacterialplain was sub-optimal. Elsewhere, the signal de-creased very gradually with the vertical (upward)distance, at approximately 1% mm−1. Only whenthe fiber reached the water surface and a meniscuswas formed between the fiber and the liquid (atca. 15 mm) the signal dropped rapidly andreached close to zero RLU when the fiber waspulled away from the solution. The gold-basedsensor showed a slightly higher response com-pared with the glass-based one, which can beattributed to increased collection efficiency due tolight reflection from the mirror like gold-coatedsurface. Otherwise, the signal–distance curve wasvery similar for the gold and glass based sensors.The signal was practically independent of thehorizontal location of the optical fiber as long asit was located above the surface of the sensor (notshown). In subsequent studies we located the opti-cal fiber at a distance of 2 mm from the center ofthe sensing element.

One of the most important advantages of solidsensing elements compared with suspended bacte-rial cultures is their compatibility with continuousmonitoring schemes. Fig. 8A and B demonstratethe ability to monitor toxicity (heat shock) stressunder continuous flow conditions. Fig. 8Ademonstrates the response of a coated opticalfiber that was exposed to 2% alcohol feed at thetimes indicated by the upward arrows in thefigure. The downward arrows indicate a switch toalcohol-free feed. The figure demonstrates that the

6C) and they were stable for several months (thesestudies are still continued).

3.5. Cell immobilization onto optical fibers

The use of optical fibers in conjunction withbacterial bioluminescent sensors was already re-ported for optical fibers coated by cells embeddedin alginates [10,11] and for unmodified opticalfiber in a cell suspension [18]. In the secondconfiguration, the optical fiber merely serves as awave-guide transducing the optical signal to thephotomultiplier [19,20]. In the present study wehave investigated variants of both approaches un-der continuous flow conditions. We have first

Fig. 5. In-batch repeatability of the luminescence time courseof four different Glass-based TV1061 sensors exposed to 2%ethanol in LB medium.


Fig. 6. Multiple use and stability studies of Glass-based TV1061 sensors exposed to 2% ethanol for the specified duration and thenstored in pure LB medium at 4 °C. (A). Luminescence time course after exposure of the sensors to alcohol at consecutive days. (B)Luminescence time course after exposure of the sensors to alcohol at consecutive weeks. The last frame corresponds to 20th week.(C) Stability curve denoting the luminescence after weekly consecutive exposure to 2% alcohol solution for 6 h (last point denotes20 weeks old sensor).

sensor can sense heat shock evolution (Fig. 8A)continuously. Fig. 8B demonstrates a continuousresponse of a wave-guide that was placed 2 mmfrom a 4×4 mm glass-based-TV1061 sensor in2% alcohol feed. The figure demonstrates that thesensor can sense heat shock evolution (Fig. 8B)continuously. The response of the cell modified

optical fiber sensor was very similar to the re-sponse of the combined glass sensors and opticalwave-guide. Fig. 8 demonstrates the technicalability to perform continuous monitoring in thetwo modes of operation. The difference betweenthe responses of the two sensors (Fig. 8A and B)to the second slug of 2% alcohol feed may stem


from different histories of the biological elements;it clearly illuminates the need for further researchon the standardization of the proposed technique.

4. Conclusions

The study described in this communication out-lines a possible configuration for the incorporationof genetically engineered microorganisms into afunctional biosensor. It is demonstrated that anti-body-attached E. coli can serve for the monitoringof toxicity, as well as other stress conditions, eitheras a disposable sensing element, as a multiple usesensing element, or in a continuous monitoringmode of operation. The assays are very simple, theresults are rapidly available, and the sensitivities areat least in the same order of magnitude as those ofsuspended bacteria. However, more researchshould be devoted to the standardization of thissensing technique; one of the principal thrustsshould be towards the minimization of signaldependence on sensor history.

Acknowledgements

This research was supported by grant number13139-1-98 of the Infrastructure program of theMinistry of Science and Culture, Israel; Instrumen-tation was partly funded by the Israel Science

Fig. 8. (A) Luminescence time course of TV1061 modifiedoptical fiber sensor under continuous flow of LB medium(flowrate, 1 ml min−1). (B) Luminescence time course of aGlass-based sensor with optical fiber signal transduction undercontinuous flow of LB medium (flowrate, 1 ml min−1). Up-ward arrows indicate initiation of the introduction of 2%alcohol in LB medium and downward arrows indicate pure LBmedium feed.

Foundation, founded by the Israel Academy ofSciences. We gratefully acknowledge strains pro-vided by T.K. Van Dyk and R.A. LaRossa, TheDuPont Company, USA.

References

[1] S. Daunert, G. Barrett, J.S. Feliciano, R.S. Shetty, S.Shrestha, W. Smith-Spencer, Chem. Rev. 100 (2000) 2705.

[2] S. Kohler, S. Belkin, R.D. Schmid, Fresenius J. Anal.Chem. 366 (2000) 769.

[3] T.K. Van Dyk, W.R. Majarian, K.B. Konstantinov,R.M. Young, P.S. Dhurjati, R.A. LaRossa, Appl. Envi-ron. Microbiol. 60 (1994) 1414.

[4] A.C. Vollmer, S. Belkin, D.R. Smulski, T.K. Van Dyk,R.A. LaRossa, Appl. Environ. Microbiol. 63 (1997) 2566.

[5] S. Belkin, D.R. Smulski, S. Dadon, A.C. Vollmer, T.K.Van Dyk, L.A. LaRossa, Water Res. 31 (1997) 3009.

[6] P. Corbisier, E. Thiry, L. Diels, Environ. Toxicol. WaterQual. 11 (1996) 171.

Fig. 7. Dependence of the measured luminescence on thedistance between the light collecting end of an optical fiber andTV1061 sensors exposed to 2% ethanol in LB medium. (a)Glass-based sensor; (b) Gold-based sensor.


[7] R.S. Shetty, S. Ramanathan, I.H. Badr, J.L. Wolford, S.Daunert, Anal. Chem. 71 (1999) 763.

[8] A. Heitzer, K. Malachowsky, J.E. Thonnard, P.R.Biewkowski, D.C. White, G.S. Sayler, Appl. Env. Micro-biol. 60 (1994) 1497.

[9] S. Sigmond, B. Kurdelova, M. Kurdel, Czechoslovak J.Phys. 49 (1999) 405.

[10] B. Polyak, E. Bassis, A Novodvorets, S. Belkin, R.S.Marks, Water Sci. Technol. 42 (2000) 305.

[11] B. Polyak, E. Bassis, A. Novodvorets, S. Belkin, R.S.Marks, Sensors Actuators B 74 (2001) 18.

[12] Y. Ikariyama, S. Nishiguchi, T. Koyama, E. Kobatake,M. Aizawa, Anal. Chem. 69 (1997) 2600.

[13] S. Belkin, D.R. Smulski, A.C. Vollmer, T.K. Van Dyk,R.A. LaRossa, Appl. Environ. Microbiol. 62 (1996) 2252.

[14] O. Bechor, D.R. Smulski, T.K. Van Dyk, R.A. La Rossa,S. Belkin, J. Biotechnol., (2001) In press.

[15] J.T. Oh, Y. Cajal, E.W. Skowronska, S. Belkin, T.K. VanDyk, M. Sasser, M.K. Jain, Biochem. Biophys. Acta 43(2000) 1463.

[16] J.H. Miller, Experiments in Molecular Genetics, ColdSpring Harbor Laboratory Press, Cold Spring Harbor,NY, 1972.

[17] T.K. Van Dyk, Appl. Environ. Microbiol. 58 (1992) 1414.[18] I. Yolcubal, J.J. Piatt, S.A. Pierce, M.L. Brusseau, R.M.

Maier, Anal. Chim. Acta 422 (2000) 121.[19] L. Tescionw, G. Belfort, Biotechnol. Bioeng. 42 (1993)

945.[20] S.M. Gautier, L.L. Blum, P.R. Coulet, Anal. Chem. Acta

243 (1991) 149.

Antibody-based immobilization of bioluminescent bacterial sensor cells

Documents

Transcript of Antibody-based immobilization of bioluminescent bacterial sensor cells