Anal. Chem. 1994,66, 341-344
A Novel Immunosensor for Herpes Viruses Bernd Konlg' and Mlchael Gratzel Institut de Chimie Physique 11, Ecole Polflechnique FWerale de Lausanne, CH 10 15 Lausanne, Switzerland
We have developed a reusable piezoelectric immunosensor for the detection of the following human herpes viruses: herpes simplex viruses types 1 and 2, varicella-zoster virus, EpsteinBarr virus, and cytomegalovirus. Synthetic peptides, representing different surface antigens of the five viruses, were used to generate virus-specific monoclonal antibodies. Apply an antibody layer via protein A immobilization onto a 10-MHz AT-cut crystal resulted in 5 X 104-1 X lo9 viruses on the electrode surface in a linear frequency change and a long-term stability of 8 weeks when the modified crystal was stored dry over silica gel blue at room temperature. Under these conditions, the coated crystal can be used 18 times without detectable loss of activity. The name of the family of herpes viruses derives from the prototype, the herpes simplex virus. Due to their homogeneous structure, it is not possible to distinguish thedifferent members of this family in the electron microscope. Nearly 100 different types of herpes viruses are known,' and of the 7 known human herpes viruses, 6 are extremely prevalent:2 (i) herpes simplex viruses types 1 and 2 (HSV-1 and HSV-2), (ii) varicellazoster virus (VZV), (iii) cytomegalovirus (CMV), (iv) EpsteinBarr (EBV), and (v) human herpes virus no. 6 (HHV-6). Herpes viruses are large, enveloped viruses with a diameter of 120-200 nme3 The frequency of reactivation varies significantly from one herpes virus to an~ther.l.~JFollowing retrograde axonal transport of the virus from the infected body surface, HSV persists in the neuron and establishes both lytic and latent infections. The infections of HSV- 1 are known as herpes labialis, herpes corneae, and stomatitis aphthosa, and herpes encephalitis;' those of HSV-2 are known as herpes gentialis. In contrast, VZV rarely reactivates during the live of an infected host. The transmission of this virus occurs primarily in an epidemic fashion via respiratory tract inoculation of aerosolized particles shed by infected individuals during the incubation period of the primary infection, ~ a r i c e l l a .The ~ reactivated VZV infection, known as zoster, is contagious, but the spread is limited mainly to one of direct contact. The clinically evident recurrences of EBV, CMV, and HIV-6are limited to patients with impaired cellular immunity. However, asymptomatic shedding of EBV and CMV from the oropharynx occurs f r e q ~ e n t l y .This ~ is the source of infection for susceptible hosts. It has not yet been documented that HHV-6 reactivates in healthy humans; however, a recent (1) Burkhardt, F. Mikrobiologiscbe Diagnostik; Thieme Verlag: Stuttgart, 1992; pp 417429. (2) Frenkel, N.; Schrimer, E.C.; Wyatt, L. S.;Katsafanas, G.; Roffman, E. Proc. Narl. Acad. Sci. U.S.A. 1990,87, 748. (3) Davison, A. J.; Scott, J. E. J. Gen. Virol. 1986,67,1759. (4) Corey, L.; Spear, P. G. N . Engl. J. Med. 1986,314, 686. ( 5 ) Croen, K. D.; Straus, S.E. Annu. Rev. Microbiol. 1991, 45, 265.
0003-2700/94/0366-0341$04.50/0
0 1994 American Chemical Society
report has shown that human herpes virus no. 6 can infect natural killer cells.7 Because of the uniform structure of the herpes viruses it is a difficult and time-consuming process today to determine the type and the number of the herpes viruses. An alternative to the established methods is the use of antibody-based biosensors. These sensors detect the antigen concentration either by direct competitive and displacement reactions similar to immunoassays or by direct changes in transducer output.8 An example for the latter type of system is the piezoelectric (PZ) crystal detector. With this system, assays both in the gas phase and in solution are possible. PZ devices consist of an oscillating quartz containing an adsorbent (antibody) on its surface that selectively interacts with the analyte. Adsorption of the analyte (antigen) increases the mass of the crystal and decreases proportionally the resonance frequency of oscillation. In 1885, Raleigh postulated the theoretical foundation of piezoelectricity. However, the use of PZ devices was realized only after Sauerbrey derived the frequency-to-mass relationship as described by the following e q ~ a t i o n : ~
where AF is the change in resonance frequency of the coated crystal, F is the resonance frequency of the crystal, AM is the mass increase, and A is the area of the coated crystal. The detection limit of this technique is about g. In 1964, King reported the use of a P Z crystal as an analytical detector.1° For the following years, the work with these detectors has focused mainly on the use of inorganic and organic coatings for gaseous environmental pollutants (for a recent review, see ref 11). Since recent work in cell biology makes biologically active molecules, such as antibodies, available in large amounts, it was possible to develop a new class of PZ biosensors-the piezoimmunosensor. The general philosophy behind biosensor technology is realtime output, high analyte sensitivity and specificity, simplicity of use, and cost e f f e c t i ~ e n e s s . ~Therefore, ~-~~ biosensors are being developed for a wide range of applications such as in the food industry, environmental monitoring and processing, (6) Dulbccco, R.; Ginsberg, H.S. Virology; J. B. Lippincott Co.: Philadelphia,
PA, 1988; pp 161-177. (7) Lusso, P.; Malnati, M. S.;Garzino-Demo, A.; Crowley, R. W.; Long, E. 0.; Gallo, R. C. Nature 1993, 362,458. (8) Karube, I.; Suzuki, M.Biosensors 1986,2, 343. (9) Sauerbrey, 0. Z.Z. Pbys. 1959,155, 206. (10) King, W. H. Anal. Cbem. 1964,36,1735. (1 1) Guilbault, G. G.; Jordan, J. CRC Crit. Rev. Anal. Cbem. 1988,19, 1. (12) Lowe, C . R. Biosensors 1985,1, 3. (13) Turner, A. P. F. Sens. Actuators 1989,17, 433. (14) Taylor, R. F. Bioinstrumentation: Research. Developments, and Applications; Buttersworth Boston, MA, 1990; pp 355-392.
AnaMicalChemistty, Vol. 66, No. 3,February 7, 1994 341
biotechnology, and at present mainly clinical diagnostics.15J6 The piezoimmunosensor system offers many of the properties desired for ideal immunosensors. We will report in this paper the development of a piezoimmunosensor for the detection of five different human herpes viruses. In contrast to the well-established techniques, this method represents a rapid, sensitive, and reliable alternative in clinical diagnostics of viruses, which will speed up the therapy of herpesvirus infections. The strong complexation between the modified gold electrode and virus protein-specific antibodies was exploited to construct a reusable immunosensor. The sensor was applied to purified or enriched virus suspensions as well as to viruses in human specimens. The sensitivity, specificity, reusability, and long-term stability of the system are also reported.
EXPERIMENTAL SECTION Reagents. Goat anti-mouse 1gG was from Sigma (Buchs, Switzerland), and poly(ethy1enimine) from Aldrich (Buchs, Switzerland). Fluoren-9-ylmethoxycarbonyl(Fmoc) amino acids were from Bachem (Heidelberg, Germany), trifluoroacetic acid was from Baker (Gross-Gerau, Germany), and liquid chromatographic (LC) solvents were from Merck (Darmstadt, Germany). All other reagents and solvents were analytical reagent grade or better (Fluka, Buchs, Switzerland). Apparatus. The PZ crystals (Universal Sensors, New Orleans, LA) used in this work were AT-cut with a basic resonance frequency of 10 MHz. The crystal consists of a 15 X 0.2 mm quartz wafer, which is placed between 7.5-mm gold electrodes, mounted in a ceramic holder with a plug. An identical (uncoated) crystal was used as reference to correct for temperature and humidity fluctuations. The crystal frequency was monitored with a P Z 105 immunobiosensor detector (Universal Sensors). Measurement Procedures. Mo, the frequency difference between the coated crystal and the reference crystal, was determined first (AF, = frequency of the uncoated reference crystal Cf,) minus frequency of the coated crystal with fr >fm). The coated crystal was subsequently dipped for 60 min at room temperature into a stirred 1-mL virus suspension in phosphate-buffered saline (pH 7) or human specimens with a single herpes virus in different concentrations (from 5 X lo4 to 5 X lo9 virus particles). The specimens were treated as described elsewhere. After rinsing with phosphate-buffered saline and air-drying, the coated crystal was placed in the test chamber and a new frequency difference was measured ( M I ) . was A shift in the frequency difference (AF = MI - Mo) related to the amount of cells adsorbed onto the crystal. Safety Considerations. Viruses are highly pathogenic organisms. It is therefore strongly recommended that work be done within a laminar flow hood. All instruments should be sterilized carefully after use. Immobilization Procedure.16 Three different methods of immobilization were employed, as follows. (1) Immobilization via a Thin Poly(ethy1enimine) (PEI) Layer. A methanol solution (5 mL) containing 4% (v/v) poly(ethy1enimine) was spread at room temperature on the
urn),
(15) Kbsslinger, C.; Drost, S.; Aberl, F.; Wolf, H.; Koch,S.; Woias, P. Biosens. Bioelecrron. 1992, 7, 391. (16) Kdnig, B.; GrBtzel, M. Anal. Chim. Acra 1993, 276, 329.
342
AnalflicalChemistry, Vol. 66,No. 3, February 1, 1994
electrode of the crystal for 20 s. After air-drying the crystal was washed with methanol. The crystal was then immersed in an aqueous 0.4 M glutaraldehyde solution (pH 7) for 30 min and washed with water. Subsequently, 5 p L of a 3 mg/ mL antibody solution was placed on each side of the crystal for 30 min. Unreacted aldehyde groups were blocked with a solution of 0.1 M glycine in 20 mM phosphate-buffered saline (pH 7). The crystal was subsequently rinsed with phosphate-buffered saline and distilled water and dried in air. (2) Immobilization via a Thin Silane Layer. The crystal was dipped in a 35 mM solution of (yaminopropy1)triethoxysilane (APTES) in benzene for 1 h at room temperature. After drying at 70 OC, the crystal was washed several times with acetone and incubated for 1 h in 0.6 M glutaraldehyde (pH 7). Following washing with distilled water, 5 p L of a 3 mg/mL antibody solution was spread on both sides of the crystal. After drying for 30 min, the crystal was washed with phosphate-buffered saline and distilled water and subsequently air-dried. ( 3 ) Immobilization via Protein A-Gold. The P Z crystal was dipped for 45 min in 1.2 N NaOH, then washed with distilled water, and placed for 20 min in 1.2 N HC1. The crystal was washed with distilled water and ethanol and dried. Then 5 pL of a protein A solution (2 mg/mL, pH 7) was added to the electrodes on both sides of the crystal. After drying, the crystal was rinsed with distilled water and 5 pL of a 3 mg/mL antibody solution was spread over the electrode surface, dried, washed with phosphate-buffered saline, then rinsed with distilled water, and dried again. Virus Stock Cultures. The virus stock cultures were maintained at standard conditions as described elsewhere.' For the experiments, the viruses were isolated from the cell homogenate by ~entrifugationl~ and, where possible, prepared in highly pure form. The number of viruses was determined according to standard procedures.17 Note. When working with viruses, care should be taken not to inhale infectious aerosoled particles and to avoid contact with these pathogens. Immunization. Synthetic peptides corresponding to special virus matrix proteins were used as antigens: (i) HSV-lgC, (ii) HSV-2gC, (iii) CMV 65-kDa matrix protein, (iv) EBVvirus capsid antigen (VCA), and (v) VZV-major capsid protein. Peptide synthesis, immunization, and production of monoclonal antibodies was done as described el~ewhere.~*J~
RESULTS AND DISCUSSION We have developed a reusable piezoelectric immunosensor for the detection of five human herpes viruses: herpes simplex types 1 and 2, varicella-zoster virus, cytomegalovirus, and Epstein-Barr virus. From the three established immobilization procedures (poly(ethylenimine), (7-aminopropy1)triethoxysilane, and protein A) to fix the antivirus antibodies to the gold electrodes, the best results in sensitivity, reusability, and stability were obtained with the protein A method (data not (17) Bonin, 0. Quantirariu-uirologische Merhodik; Thicme Vcrlag: Stuttgart. Germany, 1973. (18) Atherton, E.; Shcppard, R. C. The Peptides: AMlysis, Synthesis, Biology; Academic Press: San Dicgo, CA, 1987; Vol. 9, Chapter 1. (19) Campbell, A. M. MonocloMI Anribody and Immunosensor Technology; Elsevier: Amsterdam, 1991; Chapters 4-1 1 .
Table I. Frequency Changer due to Blndlng (5 X l o 7 Vlrur Partkhr) of Purlfled Herpes Vlrurer to the AntCVlrur Antlbody Modlfled Crystal Surface’
800
-
AF (Hz) immobilizationmethod poly(ethy1enimine) (yaminopropy1)triethoxysilane protein A a
herpes simplex type 1
EpsteinBarr Wrue
cytomegalovuus
523 f 12 517 f 11
386 f 17 379 f 12
502 f 14 514 f 15
590f8
432f7
586f9
e
600-
c
su
400-
2
200 -
ProteinA immobilizationprocedure. Average of 10experiments
f SEM.
io4
io5
io8
lo7
io8
io9
ioio
Number 01 cella
shown), which is in good agreement with earlier results.iOJ When the electrode was precoated with a thin layer of protein A and subsequently with the corresponding antibody, the crystal was stable for 8 weeks when stored dry over silica gel blueat rmm temperature (datanot shown) andcould be reused 18 times without detectable loss of activity (data not shown). Additionally,it provided the most reproducibleresults although reproducibility was not a problem, even with the other immobilization procedures (Table I). As shown in Figure 1, a plot of frequency change versus number of pure virus particles is linear from 5 X 104 to 1 X lo9cells for all of the human herpes viruses tested so far, with correlation coefficients of 0.999 and 1, and is saturated at 5 X lo9 cells. From the dimensions of the different human herpesviruses, it is possible to roughly estimate the theoretical maximum number of bound cells for our system (about 1.5 X lo9 cells can be bound to an area of 4.4 X lo7 pm2; mean diameter of the virus particles 160 nm). Comparison of this with the number of bound cells showed that over 90% of the electrode surface is occupied with cells. The lowest frequency change is seen with herpes simplex virus type 2 and EpsteinBarr virus, whereas herpes simplex type 1, cytomegalovirus, and varicella-zoster virus showed a 30%increased signal. We believe that this is due to our virus preparation, because it is known that with some herpes viruses, e.g., varicella-zoster v i r u ~it, ~is not possible to generate adequate titers of cell-free stocks. Therefore some cell material might adhere at the electrode surface. To test the specificity of the piezoimmunosensor, we immobilized different antiherpes virus antibodies via protein A on the gold electrode and compared the frequency shift of the pure virus particles with the virus in complex human specimen. As seen in Figure 2, with the cytomegalovirus both samples gave nearly identical results, making this technique feasible for the determination of different herpes viruses in human specimens. Similar results were obtained with herpes simplex viruses types 1 and 2 and varicella-zoster virus (data not shown). These results also support the choice of the immobilizationprocedure, because even with human specimens containing a lot of other cell material like bacteria, cell compartments, and body fluids no unspecific binding to the modified crystal surface is observed. To further characterize the specificity as well as the selectivity of our system, a mixture of herpes simplex viruses (20) Kdnig, B.; Gritzcl, M. Clin. Biochem. Anal. 1993,26, 1567. (21) Plomcr, M.; Guilbault, G. G.; Hock,B. Enzyme Microb. Technol. 1992,14, 30.
Flgure 1. Gomparison of frequency changes with increasing numbers of herpessimplex vlrw type 1 (e),herpes simplextype 2 (0), varicellaEpstein-Barr virus (A),and, cytomegabvirus(A).The zoster vkus 0, plot shows that the frequency response is linear for all viruses from 5 X io4 to 1 X io0 cells with correlation coefficients from 0.999 to 1. A saturation is seen with more than 1 X loo viruaes due to the complete occupancy of the electrode surface. Each point repregents the average of five experiments with an incubation time of 60 mhr (protein A immobllization method). The number of vlruses was determined according to standard procedures.17
800 -
r c
600-
E 400-
fP
I:
200 V ‘
io4
io5 io8 io7
io8
io9
iolo
Number 01 celb
Flgure 2. Frequency changes with increasing numbers of pure cytomegalovirus (0)and cytomegalovirus in human specimens (e). Both samples are linear in the range of 5 X 104-1 X I O 0 celis with correlation coefficients of 0.999 and 1, respectively. Each point representsthe average of five experiments with an incubationtime of 60 min (protein A immobilizationmethod). The number of viruses was determined according to standard procedures.l7
types 1 and 2, cytomegalovirus, Epstein-Barr virus, and varicella-zoster virus was applied to the anti-EBV-VCA antibody-modified electrode. Figure 3 shows the result of this experiment. As can be seen, even with a complex mixture of structurally similar viruses it is possible both to detect selectively a specific virus and to determine its number in the specimen, a condition of use in clinical diagnostics. Other important factors affecting the sensitivity and reusability of the piezoimmunosensor are the time of incubation of the antibody-coated crystal with the antigen (virus), the amount of antibody used for the coating process, and the method of regeneration. Because of the different behavior of the different antivirus antibodies produced, the incubation time also differs by a factor of 3. Figure 4 shows that a 30min incubation gives optimal sensitivity for herpes simplex virus type 2. However, other viruses need shorter (17 min) or longer (60 min) times. To have similar conditons, we therefore used in all experiments a 60-min incubation time. Coating the crystal with 3 mg/mL antibody solution was considered optimal. Up to 3 mg/mL, the sensitivity of the Analytcal Chemistry, Vol. 66,No. 3,February 1, 1994
943
800
1 3
io4 io5
io6 io7 Number
Of
io8 iog
io1O
cella
Flgm3. Frequency responsewith I n c r d n g numbers of pure EpsteinBarr virus (0)and Epstein-Barr virus in a suspension with HSV-1, HSV-2, CMV, and VZV (e). The frequency change is linear in the range of 5 X io4-1 X I O 0 virus particles wlth correlation coefficients of 0.999 and 1, respectively. This experiment demonstrates the selectivity of the pieroimmunosensor to detect a special herpes virus (EBV)lna mlxtwe with other humanherpesviruseswithout interference. Each point representsthe averageof five experknsntswith an incubation time of 60 min (protein A immobilization method). The number of viruses was determined according to standard procedures.17
modified crystal increased with an increase in the antibody concentration used for the immobilization procedure. Above this concentration, the frequency change was no longer dependent upon the antibody concentration used (data not shown). From the four different methods of regeneration (0.2 M glycine hydrochloride, 0.2 M ethanolamine, 8 M urea, and peptide competition), we prefer to elute the bound antigen (virus) with antigen-specificsynthetic peptides in a competition process, because this method gave the best results in terms of stability of the modified crystal surface and its reusability, as has also been demonstrated earlier,22because this procedure prevents contact of the antibodies and the modified gold electrode with harsh chemicals. With this improved regeneration method, the crystal can be used 18 times without detectable loss of activity (data not shown). The above results show that it is possible to detect human herpes viruses with a piezoimmunosensor. This technique is
Tlma (mlnutea)
Ftgurr 4. Resonance frequency shift (AF)of the human herpessimplex virus type 2 anti-HSV-2gCantlbody reaction(protein A ImmobHization method) as a function of time for a cell concentration of 1 X 100 cells/mL. After 30 min, the antlgen-antlbody reactlon Is nearly complete. Each polnt represents the average of five experlments.
cheap, rapid (about 90 min), sensitive (detection limit for human herpes viruses is about 5 X lo3 cells), and selective. Compared with the established laboratory techiques, such as ELISA, immunofluorescence microscopy, and polymerase chain reaction (PCR), it has some advantages. ELISA tests are simple, can often be done automatically, and are therefore used routineously in clinical diagnostics, but they are slow (hours). Immunofluorescence microscopy also makes use of the highly specific antigen-antibody reaction, but needs labeled antibodies as well as expensive equipment and, in addition, is time consuming. Today, the last alternative, the polymerase chain reaction, is limited to specialized laboratories, and in the case of human herpes viruses this technique is not applicable, because the human herpes virus no. 6 and the cytomegalovirus have some DNA homologies, making it impossible to distinguish between these two types of viruses. The piezoimmunosensor, however, is a simple and reliable system and seems to be well-suited for virological diagnostics. Received for revlew June 4,
1993.
Accepted September
8, 1993.' ~~~~~
(22) K(lnig, B.;GrBtzel, M. Anal. Chim. Acta 1993, 280, 37.
344
Analytical Chemktty, Vol. 66,No. 3, February 1, 1994
Abstract published in Advance ACS Absfracfs.November 1, 1993.
