Quantitative Detection of Messenger RNA by Solution Hybridization and Enzyme Immunoassay*

A novel nucleic acid detection technique is described for the quantitative measurement of eukaryotic mRNA in biological samples. The procedure involves two steps: a hybridization reaction in solution with a biotinylated cDNA probe, and a conventional enzyme im- munoassay that uses a monoclonal antibody for DNA-RNA hybrids to detect the specific mRNA*cDNA complexes. The method has comparable sensitivity to 32P-based methods and yields results that are quanti- tative and highly reproducible. Furthermore, the test can be performed using unfractionated cytoplasm without the need for extraction with organic solvents. This technique provides a rapid and quantitative method for studying changes in cellular mRNA levels, and it is suitable for testing large numbers of samples. acid hybridization assays are powerful for the transcriptional activity of genes. Messenger RNA (mRNA)

A novel nucleic acid detection technique is described for the quantitative measurement of eukaryotic mRNA in biological samples.
The procedure involves two steps: a hybridization reaction in solution with a biotinylated cDNA probe, and a conventional enzyme immunoassay that uses a monoclonal antibody for DNA-RNA hybrids to detect the specific mRNA*cDNA complexes.
The method has comparable sensitivity to 32P-based methods and yields results that are quantitative and highly reproducible. Furthermore, the test can be performed using unfractionated cytoplasm without the need for extraction with organic solvents. This technique provides a rapid and quantitative method for studying changes in cellular mRNA levels, and it is suitable for testing large numbers of samples.
Nucleic acid hybridization assays are powerful tools for analyzing the transcriptional activity of genes. Messenger RNA (mRNA) has most frequently been detected by the labor-intensive technique of mixed phase hybridization assay (1, 2). The utility of this method is limited by the fact that nucleic acids bound to nitrocellulose filters have unfavorable reassociation kinetics, and that nonspecific binding of probe can interfere with detection of small amounts of specific mRNA (3,4). It has been difficult to standardize assays from experiment to experiment because the preparation of filters and their performance characteristics are highly variable. A further disadvantage is that standard assays rely on the use of radiolabeled probes that have a short functional half-life and must therefore be regenerated regularly.
We have devised a method for measuring individual species of cytoplasmic mRNA which overcomes many of these problems. With this method, mRNA is rapidly recovered from unfractionated cytoplasm by detergent treatment and proteolysis. Specific mRNA is hybridized to biotin-labeled cDNA probes in the liquid phase, and labeled DNA. RNA hybrids are detected by enzyme immunoassay.

The One-step
Acid Guanidinium Isothiocyanute/Pherwl-Chloroform Extraction Method of Chomczynski and Sacchi (Rapid Guunidinium Method (13))-Briefly, cells were pelleted and resuspended in 1 ml of 4 M guanidinium thiocyanate, 25 mM sodium citrate, (pH 7.0), 0.5% Sarkosvl. 0.1 M 2-mercaDtoethano1 (solution D). This mixture was mixed with 0.1 ml of 2M sodium ace&e (pH 4), i.0 ml of watersaturated phenol (pH 5.0), and 0.2 ml of chloroform/isoamyl alcohol (2&l), vortexed for 10 s, and cooled on ice for 15 min. After centrifugation at 4 "C for 20 min at 10,000 x g, the aqueous phase was retrieved, mixed with an equal volume of isopropyl alcohol, and incubated at -20 "C for 1 h. After another centrifugation at 10,000 X g, the RNA pellet was resuspended in 0.3 ml of solution D and precipitated with 2 volumes of 95% cold ethanol and sodium acetate at a final concentration of 0.3 M for 1 h at -70 "C. After a final centrifugation at 10,000 x g, the RNA pellet was resuspended in diethyl pyrocarbonate (DEPC)-treated water with 0.5% sodium dodecyl sulfate (SDS) as a weak RNase inhibitor. Concentrations of total RNA were measured by UV spectrophotometry.

A Single-step
Cytoplasmic RNA Isolation Method That Was Modified from the Procedure of Gough (14) and That of White and Bancroft (15)-Cell pellets were resuspended in 50 ~1 of 10 mM Tris, (pH 7.4), 1 mM EDTA, and 0.5% Nonidet P-40, vortexed for 10 s, incubated for 10 min on ice to lyse the cell membrane without disrupting the nuclear membrane, and centrifuged at 15,000 x g for 4 min to pellet nuclei. The supernatant was mixed with 50 ~1 of 2 x proteinase K buffer containing 0.2 M Tris, (pH 7.5), 25 mM EDTA, 0.3 M NaCl, 2% SDS, and 400 rg/ml proteinase K. The sample was incubated for 30 min at 55 "C and tested immediately in the hybridization assays.
Solution Hybridization RNA preparations, as described above, were mixed with 2.5-10 ng of a biotinylated DNA probe in a total of 200 ~1 of hybridization buffer containing 2 X SSC (1 X SSC is 150 mM NaCl plus 15 mM sodium citrate (pH 7.0)), 10 mM HEPES buffer (pH 7.4), 1 mM EDTA, and 0.3% SDS. The mixture of nucleic acids was heated to 100 "C for 10 min, quick cooled on ice, and hybridized overnight in a water bath at 78 "C. After completion of the hybridization reaction, samples were cooled to room temperature, and 40 pl of 10% Triton X-100 was added to form micelles with SDS.

Enzyme
Immunoassay for Detection of DNA. RNA Hybrids Fifty-p1 aliquots of the hybridization reaction mixture were tested in triplicate wells of black microtiter plates (MicroFLUOR B 96 Umicroplate 011-010-7201 from Dynatech Laboratories). Plates were prepared by incubation overnight at 4 "C with 50 ccl/well streptavidin (or anti-hiotin antibody) at a concentration of 1 rg/ml in 0.06 M carbonate buffer (pH 9.4). Before use, plates were washed six times with 10 mM sodium phosphate buffer (pH 7.2), 0.15 M NaCl, and 0.05% Tween 20 (PBST). Following addition of samples, plates were incubated at 37 "C for 2 h and then washed with PBST. Fifty ~1 of a solution containing Fab' fragments of a monoclonal antibody to DNA.RNA hybrids conjugated to @-D-galactosidase (gift of Robert Carrico, Miles Laboratories, Elkhardt, IN) at a concentration of 0.025 pg/ml in PBST with 0.5% gelatin and 0.5% mouse serum was added to each well, and plates were incubated for 1 h at 37 "C. Plates were washed with PBST, and then 50 &well substrate solution containing 0.1 mM 4-methylumhelliferyl P-D-galactoside (Sigma) in PBS with 1 mM MgClz and 50 pg/ml bovine serum albumin was added. After a 2-5-h incubation at room temperature, the amount of fluorescent methylumbelliferone that was generated by the enzymatic degradation of substrate was measured in a Dynatech MicroFLUOR microtiter plate fluorometer (detection wavelength of 365 nm, emission wavelength of 450 nm).

Statistical Annlysis
Semiquantitative results can be obtained by simple inspection of graphs of log fluorescence for various RNA dilutions. Quantitative comparisons of any two samples can be done by a parallel line assay method (16). This provides both a ratio of mRNA content and confidence limits for that ratio. have shown that the temperature optimum for perfectly matched targets is 78 "C and that probe concentrations in the range of 20-1000 rig/ml give optimal results (17). Biotin-labeled DNA. RNA hybrids are captured on microtiter plates coated with streptavidin or anti-biotin antibody. (Triton X-100 is added to samples before this step to form micelles with the SDS and prevent desorption of protein from the solid phase.) Following removal of unbound nucleic acids by washing, the amount of biotin-labeled duplexes bound to the solid phase is detected by reaction with an enzyme-labeled monoclonal antibody directed at DNA. RNA hybrids. The preparation and binding properties of the monoclonal antibody have been described previously (18)(19)(20). Of note, the reactivity of the antibody exhibits little dependence on the base composition of the nucleic acid hybrid, and the antibody has a very low affinity for either single-or double-stranded DNA. Bound DNA. RNA hybrids are measured by the addition of a fluorogenic substrate. For concentrations of RNA between 10 and 3000 pg/ml, the amount of fluorescence generated by the enzymatic reaction is proportional to the amount of captured hybrid. Signals are dependent on probe length, but DNA-RNA hybrids as short as 25 base pairs can still be recognized by the monoclonal antibody.

Comparison of Solution
Hybridization/Enzyme Immunoassay Method to Isotopic Slot Blot for Detection of mRNA-The performance of the method for detection of eukaryotic mRNA was first evaluated in an assay for IL-l@ mRNA in LPSstimulated THP-1 cells. The solution phase nonisotopic method was compared with a standard slot blot method (21) using a 32P-labeled probe (specific activity, 2 x lo4 cpm/ng). THP-1 cells were passaged for 24 h to provide a stable baseline and then stimulated for 2 h with LPS. Total RNA was extracted by the rapid guanidinium method. One hundred pg of extracted RNA was treated with 23 units of RNase-free DNase I for 15 min at room temperature and then boiled for 3 min. A titration curve of the extracted RNA was prepared in DEPC-treated water and subjected to analysis by either solution phase hybridization followed by enzyme immunoassay (EIA) detection, or slot blot assay with autoradiogram (Fig. 1). The nonisotopic assay could detect an IL-l message in as little as 50 ng of total RNA with a signal of 16 + 6 fluorescence units (fu). The background reactivity generated by the probe was 3 & 1 fu. The standard filter assay with an isotopic probe and overnight autoradiogram had a comparable detection limit. Specificity of the Assay for DNA. RNA Hybrids-To prove that the EIA was detecting DNA.RNA hybrids, a separate experiment was performed with human peripheral blood adherent cells as the source of IL-l/3 mRNA. Total cellular RNA, before and after treatment with Tl RNase, was hybridized with the biotin-labeled IL-lb probe, and the reaction products were tested by EIA (Table I). A total of 4 pg of RNA, extracted by the rapid guanidinium method, was incubated with 1000 units of Tl RNase for 45 min at 37 "C. The nucleic acids were boiled for 10 min and tested in the assay for IL-lfi mRNA. Stimulated cells gave a reactivity of 2,180 + 22 fu which, after digestion with Tl RNase, was reduced to 51 f 6 fu. Controls to demonstrate the specificity of the assay in-  pGEM-3 plasmid without an insert (38 t 6 fu). The elimination of reactivity by Tl RNase treatment as well as the low level of reactivity of the labeled probe in the absence of specific target sequences illustrate the specificity of the monoclonal antibody for probe-specific DNA. RNA hybrids. IL-l@ mRNA Quantitation in RNA Extracted from Cells by Three Different Methods-THP-1 cells were incubated overnight in 2% fetal bovine serum. Equal numbers (2 x 10") of cells were stimulated the next day with LPS (10 pg/ml) and harvested at various times after stimulation.
RNA was extracted by one of three methods: (a) purification of RNA by guanidinum isothiocyanate/cesium chloride centrifugation; (b) extraction of RNA by the one-step acid guanidinium isothiocyanate/phenol-chloroform method; or (c) recovery of cytoplasmic RNA by a single-step method. In all cases, the RNA equivalent of 1.0 x 10" cells was mixed with biotinylated IL-18 probe and hybridized in solution. DNA. RNA hybrids were detected by EIA (Table II). Compared with either of the other methods, the rapid cytosolic extraction method gave higher signals at all measured time points. The baseline (control) fluorescence detected at t = 0 with cytosol-extracted material was 2.6 times higher than the fluorescent signal generated by cesium chloride-purified RNA. Likewise, the peak fluorescence measured at 120 min was 2.0 x the value for purified RNA. However, when normalized to the baseline fluorescent signal at t = 0, all three RNA extraction methods yielded equivalent ratios of stimulated to unstimulated cells (Fig. 2).
Use of the Solution Hybridization/Enzyme Immunoassay for Quantitative Analysis of mRNA-The EL4 murine thymoma cell line expresses IL-2 mRNA within 2-4 h after stimulation with IL-1 plus phorbol 12-myristate 13.acetate. RNA was recovered from 2 x 10" stimulated cells by the single-step cytoplasmic RNA isolation procedure. Aliquots of this solution were then diluted 2-fold (solution A/2) or 4-fold (solution A/4) with a buffer containing equal parts of Tris/EDTA with 0.25% Nonidet P-40 and 2 x proteinase K buffer (without proteinase K). Serial 2-fold dilutions of each solution were prepared in DEPC/water with SDS, and then each sample was hybridized with biotinylated IL-Z cDNA at a final concentration of 12.5 rig/ml. Labeled hybrids were detected by EIA (Fig. 3). The peak signal in stimulated cells (2.0 x 106) was 939 + 87 fu, compared with 27 f 1 fu from an equal number of unstimulated cells and 3 f 1 fu from probe reacted with buffer alone. Standard deviations for mean fluorescent signals of three replicates were 5 + 2.6% of the mean. The assay was linear over 2 log,, dilutions.
Semiquantitative results, obtained by direct inspection of the graph, demonstrate that each of the dilution curves is parallel to the others and shifted by a factor of approximately 2. Quantitative results were obtained by analyzing the fluorescence data for undiluted mRNA, 1:2 mRNA, and 1:4 mRNA by the parallel line method (16). Six-point calculations were made, selecting fluorescence values between 100 and 500 for each dilution (Fig. 3). This range was selected because the lines appeared to be steepest there and because there was a suggestion of nonlinearity with fluorescence values over 500. The comparisons made were undiluted to 1:2, undiluted to 1:4, and 1:2 to 1:4. The ratios for these comparisons should have been 2.0,4.0, and 2.0, respectively. The calculated values were 2.26 with confidence limits of 2.12 and 2.41; 4.77 with confidence limits of 4.44 and 5.12; and 2.08 with confidence limits of 1.94 and 2.22. These values are close to but not identical with the theoretical ones, which confirms the visual impression that concentrated samples give slightly lower fluorescence values than would be expected. The analyses of variance showed no serious deviations from parallelism and no evidence of line curvature. The g values were 0.004, which confirmed assay validity.
A novel nucleic acid detection technique is described for the measurement of eukaryotic messenger RNA in biological samples. The procedure involves two steps: a hybridization reaction in solution with a biotinylated DNA probe, and a conventional enzyme immunoassay to detect specific, labeled DNA-RNA hybrids. The method is practical and yields results that are highly reproducible and objective. The nonisotopic detection method avoids the biohazards, inconvenience, and cost associated with use of radioactivity while achieving a sensitivity comparable to 32P-based methods. It should be noted that although a fluorogenic substrate was used in these experiments, equal sensitivity can be achieved with an alka-line phosphatase-labeled antibody and a calorimetric substrate (22). A second major advantage of the assay over techniques that use autoradiographic detection is the ability to quantitate results easily and accurately. The assays for Il-l@ and IL-2 mRNA were generally linear over two log,, dilutions, and samples containing equivalent amounts of mRNA gave similar fluorescence values. Another advantage of this method is that samples can be tested without the need to extract RNA with organic solvents. However, when maximum sensitivity is required, extracted samples may be required to reduce background reactivity of negative controls. The methods described in this report can in principle be applied to measure the expression of any gene whose sequence is known. The minimal requirement for the assay is a sequence of 25-50 bases although a probe of 250-350 bases is preferred for maximal sensitivity.
When used in combination with a rapid method for recovery of cytoplasmic RNA, the solution hybridization/enzyme immunoassay technique offers a rapid and quantitative means of studying changes in cellular mRNA levels.