A Sandwich Hybridization Assay Employing Enzyme Amplification for termination of Specific Ribosomal RNA from Unpurified Cell Lysates

ANALYTICAL BIOCHEMISTRY ARTICLE NO.

259, 258 –264 (1998)

AB982643

A Sandwich Hybridization Assay Employing Enzyme Amplification for Determination of Specific Ribosomal RNA from Unpurified Cell Lysates Ben Wicks,*,†,‡ David B. Cook,† Mike R. Barer,* Anthony G. O’Donnell,‡ and Colin H. Self† *Department of Microbiology, †Department of Clinical Biochemistry, and ‡Department of Agricultural and Environmental Science, University of Newcastle upon Tyne, Tyne NE2 4HH, United Kingdom

Received November 17, 1997

We have employed the power of the cyclic NADbased enzyme amplification system to the determination of 16S rRNA. This generally applicable system employs two oligonucleotide probes, one of which is captured on a microtiter well surface and the other labeled with alkaline phosphatase. The detection of very low levels of hybridization of the capture probe is then achieved by the means of the ultrasensitive enzyme-amplified assay system, resulting in a highly sensitive, convenient, and rapid technology which can be directly employed on unpurified samples. We have been able to demonstrate the detection of 20 amol (107 molecules) of pure rRNA, and specific signals from as few as 2000 bacterial cells have also been demonstrated. The total procedural time can be short—5 to 18 h— depending on the dynamic range and sensitivity required. RNA target in the range of 1012–108 molecules can be assayed within 5 h. Extending the substrate incubation time enables between 1011 and 107 molecules to be determined within 18 h. The system has great potential use with respect to studying the distribution and physiological states of cellular organisms. © 1998 Academic Press

There is a very great need for rapid, ultrasensitive, and convenient methods of determining RNA. The detection and quantitation of ribosomal RNA provides opportunities to study the physiology and specific distribution of any cellular organism. Levels of rRNA indicate the immediate protein synthetic capacity and growth state of microorganisms (1), while the determinative information inherent to rRNA molecules confers potential for highly specific recognition of the cell or organelle origins of detected molecules (2). 258

Present methods of rRNA analysis are generally time consuming, labor intensive, and relatively insensitive. Accurate quantification has required large (mg) amounts to be highly purified before OD260 nm measurements, orcinol determinations (3), or ethidium-bromide fluorimetry (4) can be performed. Although they are more sensitive, radiolabeling and quantitation after differential gradient centrifugation (5) are equally labor intensive and require that cells of the target organism can be propagated in a controlled environment. Current determinative methods such as dot-blot, Northern and cDNA hybridization, and RT-PCR1 also require extensive purification of the original RNA and are not easily or accurately made quantitative (6). These methods generally provide either physiological or determinative information, not both. Only with the recent development of fluorescence in situ hybridization directed at bacterial rRNA (7) have the possibilities for combined physiological and determinative studies been realized. These methods are presently limited to samples which are suitable for microscopy and require extended observation, often combined with digital image analysis, before quantitative results can be obtained. Here we describe a generally applicable rRNA assay directed, in this work, at the 16S subunit of bacterial ribosomes, which is quantitative, specific, and applicable to the rapid analysis of large numbers of samples. Unlike previously described RNA assay systems, all the components of this system can be obtained commercially, no highly specialized equipment is required, and the assay can be used and adapted without significant molecular biological experience. As no RNA pu1

Abbreviations used: RT, reverse transcription; DEPC, diethylpyrocarbonate; PBS, phosphate-buffered saline; TBSSC, Tris-buffered standard saline citrate; RVC, ribonucleoside vanadyl complex. 0003-2697/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

SANDWICH HYBRIDIZATION ENZYME-AMPLIFIED SPECIFIC rRNA DETECTION

259

subsequently cycled in a redox reaction between NAD1 and NADH1 using alcohol dehydrogenase and diaphorase (Fig. 1C). Ethanol provides electrons for the cyclic reaction and these are passed on via NADH1 to piodonitrotetrazolium violet, reducing it to strongly colored formazan with each turn of the cycle. Particular convenience comes from the fact that the end product measured is colored. Thus a simple conventional microplate reader may be used to read the results which can be expressed numerically. Some increase in sensitivity can be achieved by replacing the tetrazolium substrate with a fluorogenic substrate, although the dynamic range of this alternative form of the system is compromised by the need to minimize the concentration of the fluorogenic substrate (11). MATERIALS AND METHODS

FIG. 1. Assay format. (A) Hybridization of RNA, biotin-labeled capture probe, and alkaline phosphatase-labeled detector probe in solution. (B) Capture of probe/target hybrids onto avidin-coated solid phase for enzyme assay. (C) Schematic diagram of cyclic enzyme amplification reaction. Abbreviations: AP, alkaline phosphatase; B, biotin; ADH, alcohol dehydrogenase; INT, p-iodonitrotetrazolium violet.

rification or separation is required, samples from a wide variety of sources can be used and possible target loss during such procedures is eliminated. The assay involves sandwich hybridization with two oligonucleotide probes, a biotinylated capture probe and an enzyme-labeled detector probe (Fig. 1). Solution, rather than solid-phase, hybridization was chosen because solution-phase kinetics are more favorable, with reactions reaching equilibrium in minutes rather than hours (8). A major further advantage of this approach is that crude sample can be directly employed without any prior purification. Material not bound to the capture probe following hybridization is washed away, eliminating the need for target purification (9). Ultrasensitivity and rapidity were achieved by adoption of an enzyme detection system which uses a cyclic amplification step to enhance the detection of alkaline phosphatase using NADP as substrate (10). In this system alkaline phosphatase catalyzes the dephosphorylation of NADP to NAD1; the NAD1 generated is

Bacterial strains and growth. Bacterial strains used were Escherichia coli (ATCC 11303), Vibrio vulnificus (strain C1748T), and Vibrio sp. S14 (CCUG 15956). Organisms were grown aerobically with shaking in LB15 broth at 37°C (E. coli, V. vulnificus) or 30°C (Vibrio sp. S14). Colony-forming units were determined by drop plate counts on LB media solidified with 1.5% (w/v) agar. Materials and solutions. All glassware was baked at 180°C for 4 h and solutions were made up in diethylpyrocarbonate (DEPC)-treated Milli-Q water (Millipore) to inactivate RNase. DEPC, 0.1% (v/v), was added and incubated at 37°C for 1 h followed by autoclaving at 121°C for 15 min. Chemicals were obtained from Sigma Chemical Co. (Poole, Dorset, UK) unless otherwise stated. Phosphate-buffered saline (PBS) was 0.15 M NaCl, 0.03 M Na2HPO4, 0.09 M NaH2PO4, pH 7.6. Hybridization buffer consisted of 0.9 M NaCl, 20 mM Tris–HCl (pH 7.6), and 0.05% (w/v) N-laurylsarcosine. Tris-buffered standard saline citrate (TBSSC) was 0.15 M NaCl, 15 mM sodium citrate in 20 mM Tris–HCl, pH 7.6. Wash buffer was TBSSC 0.1% (v/v) Tween 20. Phosphate buffer was 20 mM NaH2PO4, pH 7.2. TE buffer was 10 mM Tris-HCl plus 1 mM EDTA, pH 8.0. RNA standard was E. coli 16S 1 23S rRNA fraction from Boehringer Mannheim (Lewes, East Sussex, UK) at 4 mg/ml. Oligonucleotide probes. All probes were synthesized by R&D Systems (Abindgon, Oxfordshire, UK). The general capture probe B3 was designed to target the 59 end of the 16S rRNA molecule of the g subdivision of the domain Bacteria (59-GCC AGC GTT CAA TCT GAG CCA TGA TCA AAC TCT TCA ATT T-39) and was a 40-mer labeled at the 39 end with biotin. It is complementary to a region corresponding to positions 1 to 40 in the E. coli 16S rRNA. The V. vulnificus-specific capture probe VULNIF (59-CCG

260

WICKS ET AL.

AGA AAC AAG TTT CTC TGT CGT-39) is a 59 biotinlabeled 25-mer complementary to a region corresponding to positions 68 to 94 in the E. coli 16S rRNA. The detector probe AP-1 was a bacterial-domain-specific probe (59-GCT GCC TCC CGT AGG AGT-39) described previously (12) conjugated to calf intestinal alkaline phosphatase at its 59 terminus. Crude cell lysates. Cells were removed directly from growth medium and mixed with 8 vol of lysis buffer (TE buffer, pH 8.0, containing 0.2 mg/ml lysozyme and 20 mM RVCs) and incubated at 37°C for 10 min. Then 1 vol of 10% (w/v) N-laurylsarcosine was added and the cells were vortexed briefly, heated to 99°C for 1 min, and chilled on ice. The lysate was added directly into the hybridization mixture. RNA purification. Exponential phase cells were harvested at 13,000 rpm for 5 min in a benchtop MSE microcentrifuge. The total RNA was extracted using the RNEASY kit (Qiagen) according to the manufacturer’s instructions. Ribonucleoside vanadyl complex solution was then added to a final concentration of 20 mM and samples were stored frozen at 220°C. RNA integrity was checked by electrophoresis on 1.8% TAE (0.04 M Tris–acetate, pH 8.0, 1 mM EDTA) agarose gel. RNA concentration was determined by measuring OD260nm. Preparation of coated microwells. Strips of eight flat-bottom Maxisorp microwells (Nunc) were incubated overnight at 4°C with 150 ml of Neutravidin (Pierce) 10 mg ml21 in TBSSC. Wells were washed five times with 250 ml of wash buffer, dried at 37°C for 1 h, and stored desiccated in the dark until required. Hybridization, capture, and washing. The RNA extracts or whole-cell lysates were added to 1011 molecules (200 fmol) of both the enzyme-labeled AP-1 detector probe and the biotinylated capture probe, 23 Denhardt’s reagent, sonicated salmon testes DNA at 0.1 mg/ml, and hybridization buffer to 10 ml. The mixture was hybridized in 0.7 ml Eppendorf tubes at 65°C for 2 h and then cooled on ice. The hybridization mixture was diluted to 100 ml with wash buffer and added to Neutravidin-coated microwells which were sealed and incubated for 80 min at 21°C. The wells were then washed six times with 150 ml wash buffer to remove any unbound detector probe. Amplified enzyme assay. To detect the bound alkaline phosphatase activity, 100 ml of substrate solution was added to wells which were sealed and incubated for 0.5–16 h at 21°C. Substrate solution was 0.1 mM NADP (NAD1-free grade; Boehringer Mannheim), 50 mM diethanolamine, pH 9.5, 0.7 M ethanol, 1 mM MgCl2, 1 mM ZnSO4. Following substrate incubation 100 ml of amplifier solution was added and the change

FIG. 2. Assay sensitivity using dilutions of E. coli 16 1 23S rRNA standards. SB is substrate blank. Following hybridization, capture, and washing, the substrate was incubated overnight at 21°C; amplifier was then added. The change in OD490 nm was monitored for 10 min. Error bars represent 95% confidence interval of six replicates.

in OD490nm, expressed as milli-absorbance units per minute (mAU/min), was measured over 10 min in kinetic mode on a Bio-Tek EL312LE microplate reader (Bio-Tek Instruments, Winooski, VT). Amplifier solution was 1 mM iodonitrotetrazolium violet, 0.05% (w/v) Triton X-100 in phosphate buffer to which was added approximately 300 unit ml21 yeast alcohol dehydrogenase (Boehringer Mannheim) and 1.2 unit ml21 pig heart diaphorase dehydrogenase (Boehringer Mannheim). On receipt from the supplier both enzymes were dialyzed against four changes of 20 mM sodium phosphate buffer, pH 7.2, at 4°C for 48 h to remove contaminating NADH/NAD1. The optimum ratio of ADH to diaphorase was determined for each new batch of enzymes as described previously (13). RESULTS

Each stage of the procedure was analyzed and optimized separately before the present level of sensitivity was obtained. Hybridization temperature, capture time, and pH of wash buffer were all critical factors. Assay sensitivity using pure RNA. The sensitivity of the assay was determined using the capture probe B.3 and doubling dilutions of purified E. coli RNA standards as target (Fig. 2). Assay detection limits were defined as the minimum target required to generate a mean signal significantly greater than the negative control (t test) (14). The assay response was linear and 107 molecules of RNA (20 amol) were detected above background (t test ,0.05). The negative control containing no target RNA was not significantly greater than the background signal of the substrate blank, indicating that the enzyme conjugate was binding specifically. Ultrasensitive enzyme as-

SANDWICH HYBRIDIZATION ENZYME-AMPLIFIED SPECIFIC rRNA DETECTION

say requires extremely low levels of background, but with careful optimization of the capture and washing conditions it was possible to eliminate most nonspecific conjugate binding. RNA detection in whole-cell lysates. To establish that the assay was able to detect RNA from biological samples E. coli cells were lysed using lysozyme/detergent and the crude lysate was applied directly into the assay (Fig. 3). RVCs were included in the lysis buffer to inhibit ribonuclease degradation of the target RNA. Signals were compared with a ‘‘no-cells’’ control rather than simply a ‘‘no-target’’ blank, although it was found that the components of the lysis procedure did not generate any increase in background (data not shown). Lysates from 500 cfu gave average signals higher than the no-cell control, while 2000 cells gave a signal significantly greater than the background of the no-cell control (t test ,0.05). Significance between the sample and the background values was determined using t test. Detection of phylogenetically specific RNA. In early experiments the broad specificity of the assay was confirmed by inclusion of an excess of yeast RNA. Such material never gave signals above the background (data not shown). The potential to differentiate RNA from more closely related organisms was assessed at the species level within the genus Vibrio by comparing the results obtained using a V. vulnificus-specific capture probe to detect RNA from a member of the same genus (Vibrio sp. S.14). Approximately 1 fmol (6 3 108 molecules) of RNA from E. coli, V. vulnificus, and Vibrio sp. S.14 was assayed using either the nonspecific capture probe B.3 or the

261

FIG. 4. Specificity determined using a nonspecific capture probe (B.3) and a capture probe specific for V. vulnificus (VULNIF). Target is 1 fmol of purified RNA from E. coli, V. vulnificus, and Vibrio sp. S.14. Error bars represent 95% confidence interval of six replicates.

V. vulnificus-specific capture probe VULNIF (Fig. 4). The nonspecific assay detected RNA from all three taxa with approximately equal efficiency, showing that for these organisms variations in some parts of the target sequence did not affect the binding of probes directed at conserved regions. When the V. vulnificus-specific capture probe VULNIF was used only the V. vulnificus RNA was detected; neither the E. coli nor the Vibrio sp. S.14 RNA generated signal significantly above background. The signal achieved using the V. vulnificus probe was, however, about 50% less than that with the B.3 probe. No attempt was made to investigate reasons for this. DISCUSSION

FIG. 3. Assay of crude whole E. coli cell lysates using the nonspecific capture probe B.3. Doubling dilutions of exponential-phase E. coli were lysed and 1-ml aliquots containing 250 – 8000 colony-forming units were added directly into the assay. Control was PBS containing no cells which was run through the lysis protocol. Substrate was incubated overnight at 21°C. The change in OD490 nm was monitored for 10 min. Error bars represent 95% confidence interval of six replicates.

Our goal in the present work was to develop a highly sensitive quantitative assay for RNA which could be conveniently used. Two critical parameters were speed and the ability to use the system on unpurified material. The full assay system is based on the hybridization of target RNA with a capture probe and an alkaline phosphatase detector probe subsequently determined colorimetrically by means of enzyme amplification. The enzyme amplification system was capable of detecting 106 molecules of phosphatase-labeled oligonucleotide (data not shown). This was translated into a full system capable of detecting less than 20 amol of 16S RNA. This is equivalent to 107 RNA target molecules and is comparable to the sensitivity achieved by others employing more complex assay formats, including luminometric detection (15). The number of ribosomes within a bacterial cell is closely linked with its physiological state (1, 16, 17). Consequently an accurate measure of rRNA content

262

WICKS ET AL.

can be used as an indicator of cellular activity or growth state (7). An actively growing bacterial cell contains 104–105 ribosomes (18). If cells can be lysed and their rRNA released with high efficiency, it should be possible to detect 102–103 cells with the assay described. The limit of detection actually found was 2 3 103 colony-forming units. This level of sensitivity indicates that the efficiency of cell lysis and target release was reasonably high. In separate experiments this was confirmed as .95% by measuring release of radioactivity from cells grown in the presence of [14C]-uridine (data not shown). As few as 2000 actively growing cells were detectable; however, in order to assay rRNA in slow-growing or stationary cells, the number of target cells must be increased to compensate for the decrease in rRNA content. RNase-mediated target degradation was prevented in these crude lysates by the inclusion of 20 mM RVCs in the lysis buffer (19). Assay sensitivity was 1000-fold greater than by the method of Mabilat et al. (20), who used an rRNA-targeted alkaline-phosphatase-labeled probe in conjunction with a simple fluorogenic substrate to detect bacterial cells, but used a capture probe already coated to a plastic solid phase and subsequent hybridization at 37°C rather than 65°C in liquid phase. Though this may have resulted in less inactivation of the alkaline phosphatase marker enzyme, they used the unamplified fluorogenic substrate 4-methylumbelliferone phosphate, which is less sensitive than the enzyme amplification detection method. The assay was also 10- to 1000-fold more sensitive than DNA-targeted bacterial detection systems using direct or indirect enzyme detection (21, 22). The ability to differentiate between rRNA from different bacterial species was clearly demonstrated. There were nine mismatches in the 25-base VULNIF target region in Vibrio sp. S.14, a relatively substantial difference between the probe and the nonspecific target tested. Where distinction between targets with higher levels of homology is required, it would be desirable to determine the minimum probe–target difference that could be discriminated. It should be noted that species-level discrimination was achieved on the basis of the capture probe specificity alone; the specificity could be increased further by incorporating a specific detector probe into the system. The VULNIF probe was tested against only a small number of organisms; validation of its specificity would require more extensive testing. A difference in signal strength between the two capture probes B.3 and VULNIF was noted when the V. vulnificus RNA was assayed. This may have been due to differences in the hybridization efficiency and dissociation temperatures of the two probes. Alternatively the capture efficiency of the two probes may have var-

ied because of their different hybridization positions within the folded 16S rRNA molecule. The binding site for the nonspecies-specific probe B.3 was at the 59 end of the 16S rRNA molecule, making the biotin label easily accessible to the neutravidin-coated solid phase. In contrast, the binding site for the VULNIF capture probe was located 100 bases downstream of the 59 end, which could have made capture onto the solid phase sterically less favorable. The enzyme-labeled detector probe was specific for all of the domain Bacteria, but could be made specific for lower levels of taxonomic rank, such as a particular genus or species. The capacity to quantify specific rRNA without prior purification is an important feature of this assay. Previous investigations have generally used purified extracts for both quantitative and phylogenetic studies. Such purifications are never 100% efficient and this limits quantitative work to situations in which rRNA can be radioactively labeled so that loss of target molecules can be accurately estimated and yields determined (5). While this conventional approach probably remains the method of choice for in vitro studies in which large cell populations can readily be labeled and prepared for analysis, it cannot be applied to natural samples or experimental systems in which independent analysis of several different organisms is required. The assay described circumvents these problems by its capacity to analyze crude cell extracts, its sensitivity, and its phylogenetic specificity. Several components of bacterial lysates could cause background signal in the assay. However, any possible problems with the ultrasensitive enzyme cycling system due to the use of crude samples are avoided by extensive washing of the plates after sandwich formation to remove any potentially interfering substances. There are few limitations to establishing assays based on the principles outlined here and the necessary resources are within the grasp of most laboratories with a microplate reader. The database of published small subunit rRNA sequences of both prokaryotes and eukaryotes is freely available online (23), facilitating the design and checking of candidate probe sequences. Commercial oligonucleotide synthesis with a variety of labels is widely available so that in-house synthesis or cloning and labeling of DNA fragments for use as probes is not necessary (24, 25). Although we prepared all the enzyme amplification reagents ourselves they can be bought in kit form (Dako Ltd., High Wycombe, UK). The range of applications could include ribosomal RNA analysis of bacterial, archeal, and eukaryotic cells, as well as measurement of the mitochondrial or chloroplast RNA content of eukaryotic populations. Furthermore, the approach does not require that the organism to be studied should be growing or even culturable (26, 27).

SANDWICH HYBRIDIZATION ENZYME-AMPLIFIED SPECIFIC rRNA DETECTION

We have described the assay as both rapid and ultrasensitive. Although maximal sensitivity was achieved within 18 h, an extremely useful level of sensitivity is achievable within 5 h. Extending the substrate incubation time illustrates one of the several ways of further enhancing the assay sensitivity. The technique is not in competition with RT-PCRbased techniques for determining RNA. Rather it provides a practical and easily applicable method for quantifying any specific rRNA in samples without purification. Actively growing cells containing maximal rRNA levels were used to demonstrate the potential whole-cell sensitivity of the assay. To measure the rRNA content of cells under different physiological conditions (e.g., starvation), the number of target organisms can simply be increased to compensate for the lower rRNA levels. The assay has successfully been used to monitor the rRNA levels of bacterial cells through exponential growth into starvation. Depending on the target used a range of either purified rRNA targets or intact cells can be incorporated into each assay to provide a standard curve against which results can be compared. The assay system is fast. Overall, the assay enables substantial sample throughout (50 to 100 samples can be processed by a single operator at one time) with assay times down to 5 h. The predominant influence on procedure time is the primary substrate incubation step; alteration of this affects the dynamic range and sensitivity of the assay. After 30 min incubation the assay can quantify between 1011 and 108 molecules of 16S rRNA. Extending the incubation to 12 h enables between 1011 and 107 molecules of target to be quantified. The cyclic enzyme amplification system used here has been widely applied in immunoassays to detect hormones and microbial antigens (28 –30). Although King and Ball used this approach to detect digoxigenin-labeled PCR products in a capture hybridization system (31), we have found no reports of cyclic enzyme amplification being used in the detection of RNA. POTENTIAL IMPROVEMENTS

While the present assay has met our requirements in terms of speed and sensitivity, it serves mainly to indicate the potential in such enzyme-amplified approaches to specific RNA assay. There are a number of ways in which the system can be further optimized. For example, the activity of the alkaline phosphatase label is reduced during conjugation to the oligonucleotide from an initial activity of 2500 to only 70 units mg21. A more efficient conjugation process could increase the activity of the enzyme and correspondingly the assay sensitivity by as much as 35-fold. Second, the 2-h hy-

263

bridization at 65°C reduced the activity of the enzyme label by a further 40% (data not shown). This loss of activity could be overcome by replacing the calf intestinal alkaline phosphatase with a thermally stable form of the enzyme (32). Third, advances in the sensitivity achievable by enzyme amplification to zeptomole (10221 mol) or even subzeptomole levels for alkaline phosphatase have been made (33), resulting in assays capable of detecting under 250 molecules (34). It appears that by further increasing the performance of the enzyme amplification detection method even greater sensitivity may be achieved. In conclusion, we have developed a simple and robust 16S rRNA assay which can rapidly detect less than 104 bacterial cells. The assay requires no specialized training or equipment, can process large numbers of samples simultaneously, and generates quantitative data on the amount of specific rRNA present in a sample. The generic nature of the assay makes it potentially applicable to the detection and quantitation of any cells or ribosome-containing organelles of interest. With the further improvements noted above our next goal of increasing the sensitivity of the assay to that required to detect single cells should be readily achievable. ACKNOWLEDGMENT We thank the Natural Environment Research Council for supporting B.W. and making this work possible.

REFERENCES 1. Bremer, H., and Dennis, P. P. (1987) in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology (Neidhardt, F. C., Ingraham, J. L., Low, K. B., Magasanik, B., Schaechter, M., and Umbarger, H. E., Eds.), Vol. 2, pp. 1527– 1542, Am. Soc. Microbiol., Washington, DC. 2. Giovannoni, S. J., Delong, E. F., Olsen, G. J., and Pace, N. R. (1988) J. Bacteriol. 170, 720 –726. 3. Gerhardt, P., Murray, R. G. E., Wood, W. A., and Kreig, N. R. (1994) Methods for General and Molecular Bacteriology, Am. Soc. Microbiol., Washington, DC. 4. Thoresen, S. S., Clayton, J. R., Dortch, Q., and Ahmed, S. I. (1983) J. Plankton Res. 5, 253–261. 5. Fla¨rdh, K., Cohen, P. S., and Kjelleberg, S. (1992) J. Bacteriol. 174, 6780 – 6788. 6. Roy, G., Roy, R., and Mitra, S. (1996) Anal. Biochem. 246, 45–51. 7. Delong, E. F., Wickham, G. S., and Pace, N. R. (1989) Science 245, 1360 –1363. 8. Thompson, J., and Gillespie, D. (1987) Anal. Biochem. 163, 281– 291. 9. Ranki, M., Palva, A., Virtanen, M., Laaksonen, M., and So¨derland, H. (1983) Gene 21, 77– 85. 10. Self, C. H. (1985) J. Immunol. Methods 76, 389 –393. 11. Cook, D. B., and Self, C. H. (1993) Clin. Chem. 39, 965–971. 12. Stahl, D. A., and Amann, R. (1991) in Nucleic Acid Techniques in Bacterial Systematics (Stackebrandt, E., and Goodfellow, M., Eds.), pp. 205–248. Wiley, Chichester, England.

264

WICKS ET AL.

13. Johannsson, A., and Bates, D. L. (1988) in ELISA and Other Solid Phase Immunoassays (Kemeny, D. M., and Challacombe, S. J., Eds.), pp. 85–106, Wiley, London. 14. Ekins, E. P. (1997) in Principles and Practice of Immunoassay (Price, C. P., and Newman D. J., Eds.) 2nd ed. pp. 173–207, Macmillan Reference, London. 15. Ishii, J. K., and Ghosh, S. S. (1993) Bioconjugate Chem. 4, 34 – 41. 16. Davis, B. D., Luger, S. M., and Tai, P. C. (1986) J. Bacteriol. 166, 439 – 445. 17. Rosset, R., Julian, J., and Moriier, R. (1966) J. Mol. Biol. 18, 308 –320. 18. Neidhardt, F. C., Ingraham, J. I., and Schaechter, M. (1990) Physiology of the Bacterial Cell: A Molecular Approach, Sinauer, Sunderland, MA. 19. Berger, S. L. (1987) Methods Enzymol. 152, 227–234. 20. Mabilat, C., Bukwald, S., Machabert, N., Desvarennes, S., Kurfurst, R., and Cros, P. (1994) Int. J. Food Microbiol. 28, 333–340. 21. Kolberg, J. A., Besemer, D. J., Stempien, M. M., and Urdea, M. S. (1989) Mol. Cell. Probes 3, 59 –72. 22. Lamoureux, M., Fliss, I., Messier, S., Blais, B. W., Holley, R. A., and Simard, R. E. (1996) J. Appl. Bacteriol. 81, 626 – 634. 23. Larsen, N., Olsen, G. J., Maidak, B. L., McCaughey, M. J., Overbeek, R., Macke, T. J., Marsh, T. L., and Woese, C. R. (1993) Nucleic Acids Res. 21, 3021–3023.

24. Yehle, C. O., Patterson, W. L., Bogulslawski, S. J., Albarella, J. P., Yip, K. F., and Carrico, R. J. (1987) Mol. Cell. Probes 1, 177–193. 25. Miller, C. A., Patterson, W. L., Johnson, P. K., Swartzell, C. T., Wogoman, F., Albarella, J. P., and Carrico, R. J. (1988) J. Clin. Microbiol. 26, 1271–1276. 26. Wayne, L. G., Brenner, D. J., Colwell, R. R., Grimont, P. A. D., Kandler, O., Krichevsky, M. I., Moore, L. H., Moore, W. E. C., Murray, R. G. E., Stackebrandt, E., Star, M. P., and Truper, H. G. (1987) Int. J. Syst. Bacteriol. 37, 463– 464. 27. Whiteley, A. S., O’Donnell, A. G., MacNaughton, S. J., and Barer, M. R. (1996) Appl. Environ. Microbiol. 62, 1873–1879. 28. Ball, J. K., Beards, G. M., and Desselberger, U. (1992) J. Virol. Methods 37, 149 –154. 29. Dhahir, F. J., Cook, D. B., and Self, C. H. (1992) Clin. Chem. 38, 227–232. 30. Vercabrera, L., Handzel, V., and Laszlo, A. (1994) J. Immunol. Methods 177, 69 –77. 31. King, J. A., and Ball, J. K. (1993) J. Virol. Methods 44, 67–76. 32. Hartog, A. T., and Daniel, R. M. (1992) Int. J. Biochem. 24, 1657–1660. 33. Bates, D. L. (1995) Int. Labmate 20, 11–13. 34. Bates, D. (1996) Mol. Biol. Cell 7, 3885–3885.