Methods and Kits for Performing Nucleic Acid Hybridization Assays

The present invention relates in general to methods and kits for performing nucleic acid hybridization assays and in particular to methods and kits for immobilizing a target nucleic acid on a solid support

One characteristic property of nucleic acid, which forms the heritable material of all living organisms, is its ability to form sequence-specific hydrogen bonds with a nucleic acid having a complementary sequence of nucleotides. This ability of nucleic acids to form sequence-specific hydrogen bonds (i.e., to hybridize) with complementary strands of nucleic acid is exploited in techniques generally called hybridization assays.

In a hybridization assay, a nucleic acid having a known sequence is used as a "probe" to search a sample for a "target" complementary sequence. Labelling of the hybrid formed by the probe and the target permits the detection and quantitation of target complementary sequence in the sample.

Because all strains of a particular microorganism share a genetic component in the form of nucleic acids susceptible to diagnosis by means of a hybridization assay, such hybridization assays are valuable research and medical tools. Detection of specific target nucleic acids enables accurate diagnosis of bacterial, fungal and viral disease states in humans, animals and plants. Additionally, the ability to probe for a specific nucleotide sequence is of potential use in the identification and diagnosis of human genetic disorders.

In one type of hybridization assay, called solution hybridization, a labelled probe nucleic acid is added to a solution of a sample to be searched for a target nucleic acid. In order to ensure that both the probe and a target are in a single-stranded state suitable for hybridization, the sample and probe are heated in order to break (denature) the hydrogen bonds which are found between complementary strands of a double-stranded probe or a double-stranded target, or which are found within secondary structure of a probe or target. Upon cooling, the reaction is reversed and double-stranded nucleic acid is allowed to form. The amount of double-stranded nucleic acid which forms may be determined by scintillation counting of the label on the probe after degradation of unhybridized single strands or after isolating double-stranded DNA by passing the hybridization solution over a hydroxyapatite column which selectively retains the double-stranded form. However, if either the probe or the target was introduced in double-stranded form, a reaction reforming (renaturing) double-stranded probe or a double-stranded target competes with the hybridization reaction between probe and target and thereby reduces the sensitivity of this technique.

In another approach to hybridization assays, the renaturation problem is circumvented by immobilizing denatured target nucleic acid on a support. Retention of a labelled probe on a support-bound target after passage of the support-bound target through a solution containing the probe permits detection and quantitation of the target by measurement of the amount of bound label. See, e.g., Falkow, et al., U.S. Pat. No. 4,358,535; and Shafritz, European Patent Application No. A1-0062286. Nevertheless, because the amount of labelled probe is far in excess of the amount of target present, non-specific binding of the labelled probe to the support may swamp the detectable signal from a target present in small amounts.

Still another approach to hybridization assays is called a "sandwich" hybridization. A two-step sandwich hybridization procedure involves the use of an immobilized target nucleic acid which is exposed in a first step to a first nucleic acid probe having a first portion complementary to the target and having a second portion which is not complementary to the target. In a second step, a second, labelled nucleic acid probe, which is complementary to the second portion of the first probe, is allowed to hybridize to the first probe, forming a "sandwich" comprising the first probe between the target and the second probe. Dunn, et al., Cell, 12: 23-36 (1977). The sandwich hybridization procedure is relatively easy to perform and is not seriously affected by protein or other biological contaminants. Ranki, et al., Gene, 21: 77-85 (1983). However, a two-step sandwich hybridization assay involves considerable delay associated with immobilization of the sample on a filter.

A one-step sandwich assay involves the use of a first nucleic acid probe immobilized on a filter. The first nucleic acid probe is complementary to a first portion of a target nucleic acid. In a single step the filter-bound first probe is exposed to a sample to be searched for the target nucleic acid sequence and to a second, labelled nucleic acid probe complementary to a second portion of the target nucleic acid which portion is separate from (i.e., non-overlapping with) the portion of the target to which the first probe is complementary. Ranki, et al., U.S. Pat. No. 4,486,539. This one-step technique eliminates the delay caused by immobilization of a sample on a filter; eliminates differences between the types of treatment required for binding ribonucleic acid (RNA) and deoxyribonucleic acid (DNA) to certain types of support inasmuch as the first probe may be selected to suit the support; and is even less sensitive to contaminating materials in the sample, e.g., mucus, than is a direct hybridization assay where the target is bound to the support. Ranki, et al., Curr.Top. Microbiol.Immunol., 104: 307-318 (1983). Nevertheless, leakage of the first probe from the support during hybridization occurs frequently and significantly diminishes the sensitivity of the assay. Because hybridization occurs more readily where both partners to the hybridization are in solution rather than where either is bound to a support, the target preferentially binds to leaked first probe. Inasmuch as the amount of leaked first probe is likely to be in excess of the amount of target present, a significant amount of target nucleic acid may be consumed by the leaked probe rather than be bound to the support.

In another approach to hybridization, called blot hybridization, nucleic acids within a sample are separated according to size by electrophoresis through a gel and are transferred to a nitrocellulose filter on which they are immobilized in their relative positions on the gel. Because any target in the sample is confined to a distinct band on the filter, even weak signals resulting from small amounts of target may be distinguished from non-specific background after exposure to a radiolabelled probe. Bornkamm, et al., Curr.Top.Microbiol. Immunol., 104: 288-298 (1983). Nevertheless, the added difficulty and expense of performing an electrophoretic separation on a sample limits the practicality of applying a blot hybridization technique in a clinical setting.

Where a sample is in the form of a touch smear of a fluid, a section through cells, or chromosomal squashes from cells on slides, hybridization may be performed in situ. Generally, a radioactively labelled probe is applied to the sample which is bound to the slide in a histological preparation. After coating the slide with a photographic emulsion, autoradiographic procedures reveal the location of target-probe hybrids by means of clusters of silver grains formed in the emulsion over the hybridization site. However, where only a few grains are observed, it is difficult if not impossible to prove that the hybridization is specific. Bornkamm, et al., supra.

An attempt has been made to improve upon the in situ hybridization technique by using as a label albumin-coated gold spheres which are cross-linked to a nucleic acid having a poly(dT) tail. Chromosomes are hybridized in situ with a probe having a poly(dA) tail which is in turn hybridized to the poly(dT) tail attached to the gold sphere in order to mark the site of hybridization. Wu, et al., Proc.Natl.Acad.Sci. (USA), 78: 7059-7063 (1981). Despite having advantages over other in situ hybridization techniques, this procedure shares a disadvantage with the other in situ hybridization methods in that it requires microscopic examination of large numbers of slides on which is spread only a small amount of sample. As a result, this approach is difficult and time consuming to apply in a clinical setting.

Thus there exists a continuing need in the art for improved nucleic acid hybridization methods for minimizing the effect of probe leakage on the accurate detection of target molecules in a sample.

A method according to the present invention involves searching a sample solution for a selected polynucleotide target. In this method, the sample is exposed to a mediator polynucleotide which has a first portion with a nucleotide sequence complementary to an immobilized polynucleotide and has a second portion, separate and distinct from the first portion, with a nucleotide sequence complementary to a first portion of the target polynucleotide. A probe polynucleotide is introduced to the sample. The probe polynucleotide has an attached reporter group and has a portion with a nucleotide sequence complementary to a second portion of the target polynucleotide which is separate and distinct from the first portion of the target polynucleotide. The target is hybridized with the mediator polynucleotide and with the probe polynucleotide. The mediator polynucleotide is immobilized by hybridization with the immobilized polynucleotide.

A kit according to the present invention is useful for performing a hybridization assay on a sample to search for a selected target polynucleotide. The kit includes a polynucleotide immobilized on a support and a mediator polynucleotide associated with the immobilized polynucleotide. The mediator polynucleotide has a first portion with a nucleotide sequence complementary to the immobilized polynucleotide and has a second portion, separate and distinct from the first portion, with a nucleotide sequence complementary to a first portion of the target. A polynucleotide probe is associated with the immobilized polynucleotide and has a portion with a nucleotide sequence complementary to a second portion of the target polynucleotide. The second portion of the target polynucleotide is separate and distinct from the first portion of the target polynucleotide. A reporter is attached to the probe polynucleotide.

Other aspects and advantages of the present invention will become obvious to one skilled in the art upon consideration of the present invention.

In an examination of the hybridization assay of Ranki, et al., supra, polythymidine (poly T) was synthesized on amine-functionalized controlled pore glass (CPG) (available from Pierce Chemical Company, Rockford, Ill.) and on amine-functionalized Fractosyl-500 glass beads (available from Polysciences, Inc., Warrington, Pa.).

An amine function was attached to F-500 by treating 500 mg of F-500 with a solution of 700 .mu.l of H.sub.2 N(CH.sub.2).sub.3 Si(OCH.sub.2 CH.sub.3).sub.3 in 10 ml of 95% ethanol at room temperature for 3 hours. The treated F-500 was washed once with methanol and then once with ethyl ether. The F-500 was dried at room temperature and then baked at 110.degree. C. for 15 hours. It was then washed with water, methanol and water, and then dried. The product was (F-500)O(CH.sub.2).sub.3 NH.sub.2, an amine-functionalized F-500.

Next, the amine-functionalized CPG and the amine-functionalized F-500 were prepared for poly T synthesis.

500 mg of the amine-functionalized CPG was reacted for 30 minutes at room temperature with 250 mg (1 millimole) of phthallic anhydride in the presence of 2 ml of anhydrous pyridine and 61 mg of 4-dimethyl amino pyridine to give a product, ##STR1##

The product was rinsed with methylene dichloride, ethyl alcohol and ether, and then dried. 450 mg of the product was reacted with 330 mg of dicyclohexylcarbodiimide (DCC) for 30 minutes at room temperature. The solution was decanted and replaced with a solution of 117 mg of 6-amino-1-hexanol in 2 ml of methylene dichloride and then left at room temperature for approximately 8 hours.

Amine-functionalized F-500 was prepared for poly T synthesis by treatment with 400 mg of succinic anhydride and 244 mg of 4-dimethyl aminopyride in 3 ml of anhydrous pyridine for 18 hours at room temperature. The treated F-500 was then washed with N,N-dimethylformamide (DMF), methanol and ethyl ether. A ninhydrin test showed that 98% of the free amino groups had reacted to produce approximately 400 mg of (F-500)O(CH.sub.2).sub.3 NHCO(CH.sub.2).sub.2 COOH. This product was suspended in 2 ml of DMF containing 3 millimoles (330 mg) of DCC and 3 millimoles (420 mg) of p-nitrophenol at room temperature overnight. The product, ##STR2## was washed on a sintered glass funnel with DMF, CH.sub.3 CN, CH.sub.2 Cl.sub.2 and ethyl ether. A solution of 2 millimoles (234 mg) of H.sub.2 N(CH.sub.2).sub.6 OH in 2 ml of DMF was reacted with product (2) overnight. The product of this reaction was a support,

which was washed with DMF, CH.sub.3 CN, methanol and ethyl ether.

The functionalized ester resulting from the preparation of CPG, as outlined above, and product (3) were used as supports for the synthesis of a poly T chain. Each thymidine (T) residue was added as a phosphoramidite according to the procedure of Caruthers, et al., U.S. Pat. No. 4,415,732, in a cycle of phosphoramidite synthesis as described therein.

After 51 cycles of phosphoramidite synthesis for each support, (CPG) p(Tp).sub.50 T-OH 5' (hereinafter referred to as CPG-T.sub.51) and (F-500) p(Tp).sub.50 T-OH 5' (hereinafter referred to as F-500-T.sub.51) were respectively obtained A 25 mg unit of each of these two poly T supports was phosphorylated in 100 .mu.l of 1 X ligation buffer [0.05 M Tris, pH 7.4; 0.01 M MgCl.sub.2 ; 0.01M dithiothreitol (DTT); 1 mM spermidine; 1 mg/ml bovine serum albumin (BSA)]containing 50 nanomoles of .sup.32 P-.gamma.-ATP (specific activity, 300 cpm/picomole) and 6 units of T-4-polynucleotide kinase for 2 hours at 37.degree. C. The supports were then washed with water, 5 X SSPE (0.9M NaCl; 0.06M NaH.sub.2 PO.sub.4, pH 7.5; and 5 mM EDTA) at 90.degree. C. The supports were then washed with 90.degree. C. water and dried. The loading of the CPG beads was determined to be 760 picomoles of poly T per milligram and the loading of the F-500 support was determined to be 260 picomoles of poly T per milligram.

For use in hybridization procedures, a single-stranded phage containing the minus (-) or anticoding strand of the Herpes Simplex Virus Type I (HSV-I) glycoprotein D (gD) gene was employed as the target nucleic acid. The entire nucleotide sequence of the double-stranded HSV-I gD gene sequence appears in Table I, the top strand being the plus (+) coding strand and the bottom strand being the anticoding strand. Portions of the plus strand were employed as probe and as a portion of the immobilized nucleotide according to the present invention. These single-stranded sequences have been designated on Table I by the lines drawn above the top, coding strand of the gene.

Four synthetic polynucleotides were synthesized using the techniques of Caruethers, et al., supra, and were hybridized and ligated in 1 X ligation buffer containing 7 nM ATP and 4.5 units of T-4 ligase to form ##STR34##

The poly A 5' end of the (-) strand of double-stranded polynucleotide (4) (approximately 250 picomoles) was incubated overnight in ligation buffer containing 7 mM ATP and 4.5 units of T-4 DNA ligase. The supports were then washed three times with 100 .mu.l of ligation buffer, three times with 100 .mu.l of water at room temperature, and three times with 100 .mu.l of water at 90.degree. C. Thus, the (+) strand of polynucleotide (4) was ligated to each of the supports CPG-T.sub.51 and F-500-T.sub.51, while the (-) strand of polynucleotide (4) was removed by denaturation in the course of the washings at 90.degree. C. to leave a complex (denoted as CPG-T.sub.51-51 and F-500-T.sub.51-51) having the (+) strand of polynucleotide (4) as an immobilized probe.

A hybridization solution was prepared containing 12.5 .mu.l of 20 X SSPE; 0.5 picomoles of HCl-cut pHgD 2.9 (a plasmid-containing 2.9 kilobase insert of HSV-I plus 2.7 kilobases of plasmid sequence) in 13.6 .mu.l; 1 picomole of .sup.32 P-labelled probe A (labelled by phosphorylation with .sup.32 P-.gamma.-ATP according to the procedure described above for T.sub.51) in 10 .mu.l; 1.0 .mu.l of 10% sodium dodecyl sulfate (SDS); and 9.9 .mu.l of H.sub.2 O. The hybridization solution was boiled for 5 minutes and then incubated at 60.degree. C. Samples were taken at 15 minutes, 30 minutes, 60 minutes, and 7 hours after boiling, washed three times at room temperature in 5 X SSPE, and counted on a scintillation counter. The results appear in Table II.

As indicated in Table II, the control (unhybridized CPG-T.sub.51-51*) produced a higher level of counts than did the hybrid (CPG-T.sub.51*-51 hybridized with pHGd 2.9 and probe A*), suggesting that the immobilized poly T separated from the bead.

In order to localize the point of separation from the bead, the leakage probe was 5' .sup.32 P-labelled, as described above, and attached to a labelled CPG-T.sub.51 support to produce CPG-T.sub.51*-51*. The 5' .sup.32 P-label was removed from a portion of the CPG-T.sub.51*-51* by phosphatase treatment to produce CPG-T.sub.51*-51. Samples of these two immobilized, labelled polynucleotides were counted and then boiled in 500 .mu.l of 5 X SSPE for periods of 15 minutes, 30 minutes, 45 minutes and 75 minutes to obtain the results illustrated in Table III.

Comparison of the two columns in Table III indicates that dramatic leakage occurs whether only T.sub.51 is labelled or whether both T.sub.51 and the leakage probe are labelled, so that separation from the support apparently occurs at the point of attachment to the CPG.

Similar experiments were run on an aliphatic amide linkage of T.sub.51 to F-500. These experiments indicated that support by an aliphatic amine linkage resulted in two to three times slower leakage than the aromatic amine linkage examined in Tables II and III. Specifically, CPG attached to T.sub.51 by an aromatic amine linkage showed 20% leakage at 65% C. after three hours, while F-500 linked to T.sub.51 by an aliphatic amine linkage displayed only 7.8% leakage at 65.degree. C. after three hours.

A different sort of linkage system, wherein poly T was directly synthesized on F-500, was examined for leakage. In this system, 500 mg of glass beads (available from Polysciences, Inc., Warrington, Pa., as Catalog No. 5483) were exposed to 2 ml of 0.5 molar tetrazole and 0.5 molar T-phosphoramidites for 2 hours in anhydrous CH.sub.3 CN, and then washed with CH.sub.3 CN and treated with an iodine solution for 5 minutes. Thereafter, phosphoramidite synthesis according to Caruthers, et al., supra, was used to obtain F-500-T.sub.25, which was .sup.32 P-labelled The system was examined for leakage as above, with the result that after 3 hours at 65.degree. C. in 500 .mu.l of 5 X SSPE, only 0.5% of the poly T leaked from the support.

These results have significant implications for the practicality of one-step sandwich assays according to Ranki, et al., supra. In such assays, an immobilized single strand of DNA is attached to a solid support. A single-stranded DNA target is immobilized by hybridization to the strand attached to the support. A labelled probe is then hybridized with the target to permit detection of the immobilized strand-target-probe sandwich.

In order for such a single-step model to be practical, it is necessary: that the immobilized strand exhibit high loading to the support (1 to 5 picomoles of DNA per assay mixture, assuming that detection of as low as 10.sup.6 target molecules is desired); and that no leakage of DNA from the support occur.

Assuming 0.05% leakage of 3.times.10.sup.12 molecules (5 picomoles) of an immobilized strand, 1.5.times.10.sup.9 molecules of immobilized strand are leaked into the hybridization medium. If the number of target molecules to be detected is 3.times.10.sup.6 and, as may be shown, the efficiency of hybridization between the immobilized strand and the target is an order of magnitude lower than the efficiency of hybridization between the leaked strand and the target, 75% of the target molecules hybridize with leaked strands rather than immobilized strands, causing a loss of 3/4 of the potential signal.

Inasmuch as the lowest amount of leakage actually measured was 10 times this assumed figure (i.e., 0.5%), it is likely that more than 96% of the potential signal is lost. Therefore, there is a need for some means for preventing the leaked strands from drastically reducing the detection efficiency of the target.

In the method according to the present invention, a single-stranded polynucleotide is immobilized on a solid support. The immobilized polynucleotide is hybridized with a first portion of a single-stranded mediator polynucleotide. A second portion of the mediator polynucleotide is hybridized with a first portion of a single-stranded target polynucleotide, thereby immobilizing it. Of course, the first and second portions of the nucleotide sequence of the mediator are preferably separate and distinct (i.e., non-overlapping) in order to prevent mutual interference. The target polynucleotide is made detectable by hybridization at a second portion to a probe polynucleotide attached to a reporter group. The first and second portions of the nucleotide sequence are preferably separate and distinct (i.e., non-overlapping) in order to prevent mutual interference.

Conventional methods may be used to detect or quantify the target by measuring the amount of label on the immobilized "sandwich" hybrid and may also involve separating the immobilized hybrid from the solution. The detectable label may be radioactive, such as .sup.125 I, .sup.35 P, and the like, or non-radioactive (e.g., fluorescent or immunological).

The method of the present invention may be employed where the target polynucleotide is a deoxyribonucleic or ribonucleic acid. In either case, depending on preference for a DNA-DNA, RNA-RNA, or DNA-RNA hybridization between immobilized polynucleotide, labelled probe polynucleotide, mediator and target, the mediator and probe polynucleotides may be deoxyribonucleic or ribonucleic acids.

Any solid support to which a sequence may be bound is useful in this method, including both porous and non-porous supports, e.g., such as silica gel, controlled pore glass, and nitrocellulose paper. Most commercially available supports contain or may be provided with amine or carboxylic acid functional groups to which DNA may be linked. Alternatively, a bead coating having multiple points of attachment for immobilizing a polynucleotide to the support may be employed.

One immobilized polynucleotide for use according to the present invention is a poly-thymidine strand which may be synthesized on a support, as described above, without the need for deprotection at the end of the synthesis. The complementary mediator polynucleotide may be poly-adenosine. Where poly T/poly A hybridization is employed, the hybridization assay for the target polynucleotide is preferably conducted at a temperature lower than the melting point of AT pairs to prevent the immobilized polynucleotide from disengaging from the mediator.

The use of a huge excess of mediator polynucleotide sequence minimizes the inaccuracies present in sandwich assays of the prior art. Leakage of immobilized polynucleotide from the support, a major defect in presently available sandwich assays, does not affect the accuracy of the present method because the immobilized polynucleotide is not complementary to the target polynucleotide. Use of excess amounts (i.e., in great excess of the anticipated amount of leaked immobilized probe) of the mediator polynucleotide will "buffer" any "immobilized" polynucleotide which leaks into solution by binding to the leaked polynucleotide. This prevents withdrawal of target polynucleotide upon washing of the support after hybridization. Thus the method of the present invention provides greatly increased sensitivity (approximately 90% efficiency) in detecting small amounts of target polynucleotide.

The following examples illustrate the practice of the method according to the present invention. Specifically, hybridization assays are demonstrated which employ the mediator system according to the present invention to detect and to quantitate the amount of desired target sequence in a solution.

Example 1 describes a hybridization assay according to the method of the present invention.

A chain of fifty-three thymidine bases (poly-T chain) was synthesized on a solid support, as described above, for use as the immobilized polynucleotide according to the present invention. Both F-500 and CPG were employed as the supports. The loading of the poly-T chains on the supports was determined by the demethoxytrityl assay as follows: ##EQU1## wherein A.sub.498 is the absorbence of the solution containing the removed demethoxytrityl components at 498 nm. CPG-T.sub.53 was found to be loaded at 3.6 nanomoles/mg, while Fractosyl-T.sub.53 was loaded at 2.8 nanomoles/mg.

In the assay, 10 picomoles of probe polynucleotide were phosphorylated using .sup.32 P-labelled adenosine triphosphate, resulting in a specific activity of approximately 6.times.10.sup.6 cpm/picomole. Table IV sets forth the experimental protocol and the results of the hybridization experiment. To each of nine assay tubes was added 1.5 mg of either CPG-T.sub.53 or F-500-T.sub.53. The reaction buffer employed was 6X SSPE and the final volume in each tube was 25 .mu.l. No foreign DNA was added. In the first step of the assay, the target (.PHI.2, a single-stranded bacteriophage containing 1.3 kilobases of HSV-I including the sequence set forth in Table I) was hybridized at 45.degree. C. with either the labelled probe polynucleotide (having the sequence set forth in Table I) and the mediator polynucleotide (having the sequence as indicated in Table I, but having a 27-nucleotide poly A tail at the 5' end), the probe polynucleotide alone, or mediator polynucleotide alone for 90 minutes. This hybridization mixture was thereafter reacted with the support system at 45.degree. C. for 30 or 60 minutes. After both hybridization steps were completed, each tube was washed three times with buffer at room temperature and the immobilized and labelled target sandwich on the bead counted using a scintillation counter.

The results in Table IV indicate that hybridized probe polynucleotide/mediator polynucleotide is capable of immobilizing the target polynucleotide, and that the longer the reaction time and the smaller the ratio of target to mediator, the greater the immobilizing efficiency.

The following example illustrates modification of the assay procedures to decrease non-specific DNA binding and provide a lower load capacity on the support.

Poly T was synthesized on untreated glass beads (GB) as above in 66 cycles and detritylated to obtain GB-T.sub.66 having a loading of 12 picomoles/mg.

In order to determine the loading of poly-T accessible for hybridization to the mediator polynucleotide sequence, 50 picomoles of the mediator polynucleotide of Example 1 were phosphorylated with .sup.32 P to a specific activity of 30,000 cpm/mole following the procedure of Example 1. 50 picomoles of the probe polynucleotide of Example 1 were similarly labelled to the same specific activity.

Table V indicates the protocol of the experiment conducted to ascertain appropriate loading conditions for the assay. The beads with attached immobilized probe were incubated in the presence of labelled mediator polynucleotide or target probe at 45.degree. C. for 2 hours in 6X SSPE buffer, washed 3 times at room temperature with buffer, and counted.

The results, as shown in Table V, indicate that the loading of the accessible poly-T chain is approximately 1.84 picomoles/mg and that the non-specific binding is less than 1% under the conditions of the assay.

The following example illustrates the sensitivity of a hybridization assay according to the method of the present invention.

In the first step of the assay, various amounts of the target of Example 1 were hybridized at 48.degree. C. with 0.9 picomoles of the .sup.32 P-labelled polynucleotide probe of Example 1 (specific activity 4.times.10.sup.6 cpm/picomole), 1 picomole of the mediator polynucleotide of Example 1, and 5 .mu.g of human placental DNA (HP-DNA) (available from Sigma Chemical Company, St. Louis, Mo.) for 100 minutes. The final volume of each tube was 25 .mu.l, employing 6X SSPE as buffer. Each resulting hybrid was thereafter reacted with 5 mg of the GB-T.sub.66 of Example 2 for 100 minutes at 39.degree. C. The beads were then separated from solution, washed 3 times with buffer at room temperature, and the amount of immobilized, labelled sandwich was counted. Table VI below sets forth the results.

These results demonstrate the sensitivity of the probe/mediator "sandwich" assay of the present invention in detecting target concentrations of 1 femtomole or less. It is expected that better sensitivity could be achieved through use of different detection systems, e.g., antigenic systems, or other non-isotopic systems.

It is expected that numerous modifications and variations will occur to those skilled in the art upon consideration of the present invention. For example, it is obvious to one skilled in the art that the component elements necessary to test a sample for the presence of a particular target polynucleotide may be assembled in advance in the form of a kit. Specifically, a mediator polynucleotide, a probe polynucleotide bound to a reporter group and complementary to a separate and distinct, different portion of the target polynucleotide than the first probe, and an immobilized polynucleotide bound to a support may be associated in such a kit as separately packaged components. Such a kit may be used to detect the presence of and to quantify the target for which it was designed by combining the probes and support with a sample to be tested for a target polynucleotide prepared, for example, according to the procedure of Ranki, et al., Curr.Top. Microbiol.Immunol., 104: 317-318 (1983). Consequently, it is intended that the present invention be given the full scope of the appended claims.