WO1999018236A1

WO1999018236A1 - Obtaining nucleic acid sequences

Info

Publication number: WO1999018236A1
Application number: PCT/GB1998/002958
Authority: WO
Inventors: Ian Noel Hampson; Lynne Hampson
Original assignee: CHRISTIE HOSPITAL NHS TRUST; Victoria University of Manchester
Current assignee: CHRISTIE HOSPITAL NHS TRUST; University of Manchester
Priority date: 1997-10-03
Filing date: 1998-10-05
Publication date: 1999-04-15
Anticipated expiration: 2000-04-03
Also published as: GB9720992D0; EP1017852A1; AU9449098A

Abstract

A method of obtaining target nucleic sequences derived from genes differentially expressed in a test cell sample as compared to a reference cell sample by the use of subtractive hybridisation. The duplex molecules formed during the subtractive hybridisation step are subjected to selective cross-linking and the unsubtracted, single stranded cDNA is used as a template to generate nucleic acid strands complementary thereto incorporating a binding ligand. These complementary strands are then immobilised on a solid phase support system having a specific binding partner for the ligand. The solid phase support system is then treated to remove non-bound nucleic acid and leave a pure sample of the bound complementary strand. This bound strand may be used as, or to generate, the target nucleic acid molecules.

Description

Obtaining Nucleic Acid Sequences

The present invention relates to a method of obtaining nucleic acid sequences and more particularly such a method employing the technique of subtractive hybridisation to obtain nucleic acid sequences derived from genes differentially expressed in one cell sample as compared to another.

The technique of subtractive hybridisation is a known method in the field of microbiology for investigating differential gene expression between two related cell samples, sec for example Nucleic Acids Research, vol 20, No 1 1 , 2899 (Hampson et al) and CA-A-2 093 567. Such differential expression may occur, for example, in cells from a patient having a particular medical condition, as compared to related cells from a person not having that condition.

The subtractive hybridisation technique allows the generation of nucleic acid molecules derh ed from genes expressed in one of the cell samples but not the other. Such molecules may be used, for example, as highly specific probes in a diagnostic technique for identifying a particular medical condition or for investigating the sequences of the differentially expressed genes.

For

estigating differential gene expression as between test and reference samples of cells to investigate genes which are expressed in the former but not the letter, subtractive hybridisation involves the steps of

(a) obtaining cDNA derived from transcript RNA of a test cell sample,

(b) obtaining "hybridising driver" nucleic acid derived from transcript RNA of a reference cell sample, and

(c) treating the cDNA of (a) (in one or more steps) with an excess of the "hybridising driver" nucleic acid of (b) under hybridising conditions.

As a result of these steps, the cDNA (of (a)) and the "hybridising driver" nucleic acid (of (b)) which are derived from genes expressed in both the test and reference cell samples hybridise together to form duplexes. Any remaining non- hybridised cDNA of (a) is referred to as "unsubtracted cDNA" and is derived from genes expressed in the test sample but not the reference sample.

The cDNA of (a) is typically derived from mRNA. The "hybridising driver" nucleic acid is also typically mRNA but in certain circumstances may be cDNA.

(It will be appreciated that the procedure only generates unsubtracted cDNA from genes expressed in the test but not the reference sample. Consequently the procedure may be repeated by treating the sample originally designated as the test sample as the reference sample for the purpose of the repeat and vice versa. In this way a cDNA library may be established representative of genes expressed in one cell type but not the other.)

The unsubtracted cDNA generated in step (c) may be amplified to produce probes for investigating genes expressed in one cell sample but not the other. A potential problem with said amplification reaction is the presence, in the product of step (c), of the duplex nucleic acid (i.e. cDNA/"hybridising driver") (and possibly also excess "hybridising driver^" nucleic acid) since unless precautions are taken this can also lead to the production of unwanted probes.

The prior art has proposed that the duplexes initially be separated from the product of step (c) prior to amplification. More particularly, an eco-physical separation was used (hydroxy apatite, adivin-biotin) to remove the duplexes. Subsequently probes could then be generated under conditions that only the unsubtracted cDNA (and not excess "hybridising driver", e.g. mRNA) was amplified. There was however a disadvantage in that the eco-physical separation step led to incomplete separation and/or potential loss of low abundance unsubtracted cDNA sequences.

To overcome this problem, CA-A-2 093 567 proposes that the duplexes generated in step (c) be rendered inert by treatment of the reaction mixture with a cross-linking agent which is effective selectively to cross-link the nucleotide strands in the duplex molecules. The non-subtracted single stranded unique cDNA that remains can then be used to produce probes by a random priming reaction. This may for example be effected in the presence of a radiolabelled (e.g. ³²P) using a DNA polymerase which lacks exonuclease activity and which is inactive with respect to said duplex molecules and RNA. Upon termination of the labelling reaction, the double stranded product can then be denatured, for example by boiling, to release the single strand length or lengths of radio-labelled DNA from the template strand, ready for use as a probe or probes. The presence of the radioactive label allows the total reaction mixture to be used subsequently, as only the probe sequences in the reaction mixture are radioactively labelled.

In one embodiment of the procedure disclosed in CA-A-2093567, the cDNA of step (a) above incorporates biotin. As a result, the unsubtracted cDNA of step (c) and the cross-linked cDNA/mRNΛ duplexes of that step are both biotin labelled and may be immobilised on a solid phase support having streptavidin bound thereto. This allows mRNA to be washed from the support to leave immobilised unsubtracted cDNA and immobilised cross-linked duplexes. The cDNA may then be subjected to a single stage amplification reaction (under condition in which the immobilised duplex does not partake in the amplification) to generate probes which are initially hybridised to the cDNA but which may then be denatured, separated from the support and subjected to further rounds of solution phase application. This technique results in double stranded DNA which may be used for cloning.

There is however a problem associated with the techniques disclosed in CA- A-2 093 567 in that we have ascertained that, under certain circumstances the crosslinks do not maintain their integrity so that the nucleic acid strands (of the original duplex) may partake in the amplification reactions thereby resulting in the generation of unwanted probes nucleic acid sequences. Such circumstances including, for example, healing the cross-linked duplex molecules at temperatures greater than 90°C in the denaturation phase of an amplification reaction. As a result, subsequent stages of the PCR amplification may involve copying of previously cross-linked sequences (i.e. sequences derived from genes expressed in both the test and reference samples).

It is therefore an object of this present invention to obviate or mitigate the abovementioned disadvantages.

According to the present invention there is provided a method of obtaining target nucleic acid sequences derived from genes differentially expressed in a test cell sample as compared to a reference cell sample, the method comprising the steps of

(i) effecting subtractive hybridisation using cDNA derived from transcript RNA of the test cell sample and an excess of a hybridisation driver nucleic acid derived from the reference cell sample so as to generate unsubtracted cDNA derived from genes expressed in the test but not the reference sample and duplexes comprised of cDNA and "hybridising driver" nucleic acid derived from genes expressed in both the test and reference cell samples.

(ii ) effecting selective cross-linking of the strands in the duplex molecules,

(iii) using from the unsubtracted cDNA as a template to generate nucleic acid strands complementary thereto incorporating a binding ligand,

(iv) immobilising said complementary strands on a solid phase support system having a specific binding partner for said ligand as to generate immobilised complementary strands,

(v) treating the solid phase support system so as to remove non-bound nucleic acid and leave a pure sample of the bound complementary strand, and (vi) using said pure sample of the bound complementary strand as, or to generate, said target nucleic acid molecules.

Thus in the method of the invention the unsubtracted cDNA is used for the generation of complementary copy strands incorporating a ligand (e.g. biotin) which is specific for a binding partner therefor immobilised on a solid phase support system. These strands are generated under conditions such that there is no copying of single stranded "hybridising driver" nucleic acid. Consequently, it is only the ligand labelled complementary strands which become immobilised on the solid phase support, allowing the cDNAZhybridising driver" duplexes and single stranded "hybridising driver" nucleic acid to be removed (e.g. by elution) from the system, e.g. by a high stringency wash, leaving a pure sample of the complementary cDNA immobilised DNA immobilised on the support.

As a result, problems associated with "break-down" of the cross-links in the cDNA/mRNA duplexes encountered in the prior art technique are avoided.

The complementary cDNA immobilised in the support will generally be hybridised to the original, unsubtracted cDNA. The support may then be subjected to a high stringency wash so as to leave the complementary cDNA immobilised on the support with original unsubtracted cDNA hybridised thereto. This original unsubtracted cDNA may then be denatured from the immobilised complementary cDNA and, in one embodiment of the invention, may provide (possibly after amplification) the target nucleic acid. Alternatively, after denaturation of the original unsubtracted cDNA and removal thereof from the solid support system, the immobilised complementary cDNA (for simplicity referred to in the subsequent description as "the immobilised cDNA") may itself provide the nucleic acid molecule of interest. For example, the immobilised cDNA may be eluted from the support system and used as a probe. It is however more preferred that the immobilised cDNA is used as an "intermediate" for generating nucleic acid sequences of interest in greater quantities by means of an amplification reaction. In this respect it is particularly preferred that the immobilised cDNA is subjected to a random priming amplification reaction (preferably a Degenerate Oligonucleotide Primed (DOP) PCR reaction) in situ on the support to generate amplification products that may be eluted from the support and collected, e.g. for use as probes. The initial cycles (e.g. 5 cycles) of the DOP PCR reaction may be effected in the presence of the solid phase (with immobilised cDNA). Subsequently it may be desirable to separate the solid phase system from the amplification reaction mixture and effect further cycles of amplification wholly in the solution phase.

A further possibility to add the total cDNA from the test sample (possibly after solution phase, non-specific amplification reaction, e.g. by DOP PCR) to the solid support system with immobilised complementary cDNA under hybridising conditions. As a result, complementary immobilised and solution phase sequences become hybridised to each other. The support may then be washed to remove non- hybridised sequences. The (original solution phase) sequences hybridised to the immobilised cDNA are those derived from genes expressed in the test sample but not the reference sample. These sequences may be denatured from the support, denatured therefrom, collected and used in a solution phase amplification reaction to generate further copies, e.g. by random priming.

The cDN A required for step (i) of the present invention may be obtained from the poly-A+ selected total mRNA of the test sample by well known techniques. Similarly the "hybridising driver^" nucleic acid required for step (i) may also be obtained by conventional techniques. For preference the "hybridising driver" nucleic acid is total mRNA from the reference cell sample.

The subtractive hybridisation may be effected using known conditions. Thus for example the hybridisation may be effected for 30-40 hours at 68°C in the presence of a hybridising buffer using an excess of the "hybridising driver" nucleic acid. The hybridisation reaction may, if desired, be carried out in more than one stage, adding "hybridising driver" nucleic acid at each stage to ensure that the only remaining cDNA from the test sample is derived from genes expressed in that sample but not he reference sample.

In the next step of the procedure, the reaction mixture is treated with a cross- linking agent which is effective selected to cross-link the hybridised strands of cDNA/"hybridising drive" duplexes. The preferred cross-linking agent is an aziridinylbenzoquinone and most preferably such a compound of the formula.

in which R is H. alkyl, thiol. alcohol or halide

It is particularly preferred that the cross-linking agent is 2.5-diaziridinyl-l ,4- benzoquione.

The use of aziridinylbenzoquinones as agents for cross-linking duplexes in subtractive hybridisation is disclosed in CA-A-0 093 567 and the conditions disclosed therein are equally applicable to the present invention.

In the next step of the process, the unsubtracted cDNA present in the reaction mixture is used to prepare ligand labelled complementary strands in a copying reaction in which there is no copying of single-stranded "hybridisation driver" nucleic acid. This can conveniently be achieved by using mRNA as the "hybridisation driver" so that the ligand labelled complementary strands may be produced by random priming and using a DNA polymerase having no exo-nuclease activity or RNA copying ability for effecting the copying reaction.

The ligand may be incorporated in the primer employed in the reaction. Alternatively or additionally, a ligand labelled deoxy-nucleotide may be used, for example ligand labelled deoxy-uridine tri phosphate (dUTP). If ligand labelled dUTP is used to provide the ligand, preferably the dUTP is provided at a lower concentration that the other deoxy-nucleoside triphosphates. Keeping the concentration of ligand labelled dUTP lower than the concentration of unlabelled dTTP acts to prevent too much ligand being incorporated into the DNA strands, which may stop the random oligonuleotide priming reaction (as dUTP and dTTP are interchangeable) and also to prevent ligand being introduced into non-desirable cDNA products. Preferably, ligand labelled dUTP is provided at between 5% to 15% of the final concentration of dTTP in the random priming reaction mixture.

The preferred ligand is biotin.

Complete Random Primed DNA labelling kits suitable for producing DNA strands complementary to the unsubtracted DNA are supplied by United States Biochemical of Cleveland, Ohio, U.S.A. (Product No. 70150).

DNA polymerases with the required characteristics specified above for the random oligonucleotide priming reaction are available commercially, such as for example that supplied under the Trade Mark "Sequenase" by United States Biochemical of Cleveland, Ohio, U.S.A. which not only lacks exonuclease activity but is also unable to utilise RNA as a template for synthesis strands.

The procedure as described results in a reaction mixture comprised of the ligand labelled complementary cDNA strands, the cDNA/"hybridising driver" duplexes, and single stranded "hybridising driver". A solid phase support system having, bound to the support a specific binding partner for the ligand is then used to separate the complementary cDNA.

More particularly, the complementary strands become immobilised on the support by virtue of the interaction of the ligand incorporated in the complementary strands and its binding partner.

The solid support system may then be separated from the reaction mixture and/or subjected to a wash procedure, so as to leave a pure sample of the complementary strands bound to the support system.

In the case where the ligand in the complementary strand is biotin the specific binding partner therefor can the support may be avidin or streptavidin.

The solid support system may comprise beads, e.g. magnetic beads. Beads which have been pre-coated with (strept)avidin (for use when the ligand is biotin) are commercially available under the Trade Mark Dynabeads M-280 Streptavidin.

There is thus obtained on the solid support system a pure sample of the complementary cDNA which may be employed as described above. Alternatively, this cDNA may be used for any of the purposes described in CA-A-2 093 567.

The method according to the present invention is particularly useful for obtaining molecules to be used in studying differential gene expression between diseased cells and normal cells to determine the effect of the disease on gene expression. Furthermore, the method according to the present invention also finds utility in studying the effect of a drug or other agent on gene expression, by comparing the genes expressed in treated samples with those expressed in normal samples. The method could also be used to determine the effect of environmental factors on gene expression, for example pollution and radioactivity. Any cell type is suitable for performing the method of the present invention. Preferred cell types upon which to carry out the method include epithelial cells, especially those from the normal metaplastic and native squamous epithelium.

A preferred system that may be studied using the method of the invention is to look at the effect of cancer on gene expression. Cancers that could be investigated include cervical cancer. It is also possible to utilise the method of the present invention to investigate any link between cancer and smoking on the differential expression of genes.

A methodology which should lead to the isolation of genes differentially expressed in patients with cervical cancer who are either non-smokers or smokers would be to cam' out the method of the present invention on the following four sets:

1. Target non-smoking native squamous epithelial cell cDNA to an excess of driver non-smoking metaplastic cell RNA.

2. Target non-smoking metaplastic cell cDNA to an excess of driver nonsmoking native squamous epithelial cell RNA.

3. Target smoking metaplastic cell cDNA to an excess of driver smoking native squamous epithelial cell RNA.

4. Target non-smoking metaplastic cell cDNA to an excess of driver nonsmoking metaplastic cell RNA.

Note: Both RNA and its derivative cDNA are referred to as target material when the objective is to clone sequences that are expressed in this target material and not the driver. The term driver is used to describe RNA that is used to subtract all the "house-keeping^" genes that are commonly expressed between the target and driver. Thus set (1) above will allow DNA sequences specific to native squamous epithelial cells and not metaplastic cells to be obtained. Set (2) above will allow DNA sequences specific to metaplastic cells and not native squamous epithelial cells obtained. Sets (3) and (4) above will allow DNA sequences specific to smokers metaplastic cells but not non-smokers metaplastic cells and vice versa to be obtained. The invention will be further described by way of example only with reference to the following non-limiting Example and the accompanying drawings, in which:

Fig. 1 shows an agarose gel illustrating the result of the Example; and

Fig. 2 shows a Northern blot illustrating a further result of the Example.

EXAMPLE

This Example illustrates the method of the invention in investigating differential gene expression in oestradiol treated myeloid cells as compared to non- treated cells.

The myeloid stem cell line FDCP-mix (Nature, vol. 310, page 228 (1985)) was cultured in Iscove's Medium containing 20%) horse serum and 100 units/ml of interleukin 3 both in the presence (driver) and absence (target) 0.2 μm oestradiol for two days.

1 RNA Isolation

RNA was isolated from the two cell samples by the acid phenol method (Chomczynski and Sacchi, 1987, Anal. Biochem. 162, 156-9).

2. DNA'ase Treatment

50μg of both target and driver total RNA were dissolved in 30μl of 40mMTris/HCl pH7.5, 20mM MgCl₂, 50mM NaCl, 5mM dithiothreitol (DTT). 1 unit of placental ribonuclease inhibitor (Boehringer Mannheim) and 1 unit of RNA'ase free DNA'ase (Promega, UK) were then added and the mixture incubated for 10 minutes at 37°C. This was then extracted with an equal volume of water saturated phenol/chloroform and the RNA precipitated by addition of one tenth volume of 3M sodium acetate pH 5.5 and 3 volumes of ethanol. The RNA was then pelleted by centrifugation at 21,000 x g. 3. Poly A+RNA Selection

50μg of DNA'ase treated target RNA was mixed with lOOμl of oligo dT Dyna beads (Dynal, UK) and poly A+ isolation carried out as described by the manufacturer. This material was precipitated and pelleted as described above.

4. I^s' Strand cDNA Synthesis

The total poly A+ RNA (approx. 2μg) isolated from 50μg of total RNA was dissolved in 5μl of sterile distilled water (SDW). O.5μl of O.I M methyl mercuric hydroxide was added to this and the mixture left at room temperature for 5 minutes. 1 .25μl of O. IM β-mercaptoethanol was then added and the tube immediately placed on ice. To this was added; l μl of 1 μg/μl of a suitable oligonucleotide primer (oligo dT is _mer or random hexamer). l Oμl of 5 x reverse transcription buffer (250mM Tris/HCl pH8.3, 375mM KCl. 15mM MgCl₂, l μl of O.IM DTT, 0.5μl of placental ribonuclease inhibitor, 0.5μl of a stock deoxynucleotide triphosphate solution at 25mM each (ACGT), 30μl of SDW, 2.5μl of RNA'ase H minus reverse transcriptase (Superscript I or II, GIBCO BRL Life Technologies) and the mixture incubated at 37°C for 1 hour. This was then extracted with phenol/chloroform, ethanol precipitated and pelleted as previously described.

3. Alkaline Hydrolysis of the Target Parent RNA

The first strand product was dissolved in 30μl of O.IM NaOH. ImM ethylene diamine tetra-acetic acid (EDTA) and incubated at 37°C for 15 minutes. 4μl of 3M sodium acetate pH5.5 and 1.5μl of 10%o sodium dodecyl sulphate (SDS) were then added and the hydrolysed product spun through a 1ml spin column packed with Sephadex G50. The resulting single stranded cDNA was then co-precipitated and pelleted with 50μg of the DNA'ase treated driver RNA using sodium acetate and ethanol as described.

4. Target cDNA Driver RNA Hybridisation

The co-precipitated target cDNA and driver RNA were carefully dissolved in 5μl of hybridisation buffer (50mM Na/HEPES, pH7.0, 80% formamide, 5mM ι:

EDTA). This was incubated for 48 hours in a 0.2ml tube using a hot lid PCR machine (Perkin Elmer GeneAmp 2400) then precipitated by the addition of lOμl of SDW plus 50μl of ethanol and pelleted by centrifugation (21 ,000 x g).

5. Chemical Cross-Linking

This was carried out as described in the RNA subtractor kit (Amersham/Pharmacia, Cat No. RPN 2240) (Hampson et al. 1992). Briefly, the hybridised cDNA/RNA was dissolved in 39μl of SDW plus 5μl of 250mM Tris/HCl pH 7.0, 5μl of XXX ascorbic acid and lμl of a lOmM solution of 2.5-diaziridinyl, 1 ,4-benzoquinone (DZQ) and incubated for 20 minutes at 42°C. This was then extracted with an equal volume of chloroform and the aqueous phase cross-linked material precipitated and pelleted as described.

6. Random Primed Biotin Labelling

The cross-linked material was dissolved in 12μl of SDW to which was added 4μl of 200mM TrisHCl pH7.5. l OOmM MgCl₂, 250mM NaCl plus lμl of a l μg/μl solution of random hexanucleotide primer (Promega, UK), Deoxynucleotide triphosphates (ACGT) were then added to a final concentration of 40μM each and biotin-16-2^*-deoxyuridine-5^,-triphosphate (Boehringer Mannhein) added to give a final concentration of 4μM. l μl of Sequenase II was added to 7μl of enzyme dilution buffer (l OmM TrisHCl pH7.5, 5mM DTT, 0.5mg/ml bovine serum albumin) and the 2μl of this was added to the random primed reaction mixture. Incubation at 37°C was for 20 minutes. The product was spun through a microcon 100 spin centrifugal concentrator (Amicon) at 500 x g for 5 minutes then lOOμl of phosphate buffered saline (PBS), 50μg/ml of ultrapure BSA (GIBCO BRL Life Technologies) added followed by a respin. This operation was repeated twice more. 20μl of PBS/BSA was added to the microcon membrane, inverted and spun (5000 x g) for 2 minutes to recover the retentate. 7. Binding to Streptavidin Dynabeads

A 50μl aliquot of steptavidin dynabeads (Dynal, UK) was magnetically separated and washed twice with lOOμl of PBS/BSA buffer. The biotinylated retentate was added to the dynabeads in a volume of 50μl and left at room temperature for 40 minutes with occasional shaking. The beads were separated and washed with l OOμl of PBS/BSA three times at room temp. These were then washed twice with l OOμl of 50mM TrisHCl pH 7.5, ImM EDTA, 0.1% SDS at 85°C for 2 minutes and finally washed twice more with lOOμl 1 x Taq DNA polymerase buffer (lOmM Tris HCI pH8.3, 50mM KCl, 3mM MgCl₂).

8. Degenerate Oligonucleotide Primed (POP) PCR Amplification

The washed dynabeads from step 9 were added to a lOOμl DOP reaction and PCR amplification carried out as described by the manufacturer (Protocol 1. Boehringer Mannheim) except that, after the initial five low-temperature annealing cycles, the Dynabeads were removed from the reaction tube before allowing the amplification to proceed for the remaining cycles. The products were then spun through a centrifugal concentrator (microcon 100) as above, washed with lOOμl of 50mM TrisHCl pH7.5, ImM EDTA (TE), 20μl of TE added to the concentrator and the retentate recovered by inverting and spinning as described in step 8. The cDNA fragments range in size from 100>1000 base pairs and should represent only those cDNA's that are present in the target cell RNA population and absent or reduced in the driver cell RNA population.

An example of a typical unsubtracted DOP product separated by agarose gel electrophoresis is shown in the agarose gel of Figure 1.

This DOP amplified unsubtracted material can then be labelled by any convenient method and used to probe a Northern blot of target and driver RNAs. It can be seen in Figure 2 that the target RNA has strong multiple signals when compared to the driver RNA population indicating that the unsubtracted DOP product contains cDNA's that are differentially expressed in the target compared to the driver RNA. This material can be cloned into a suitable PCR product cloning vector (T/A, Invitrogen) to produce a unsubtracted library of differentially expressed genes which can then be individually picked and sequenced or subjected to additional round of screening.

Claims

1. According to the present invention there is provided a method of obtaining target nucleic acid sequences derived from genes differentially expressed in a test cell sample as compared to a reference cell sample, the method comprising the steps of

(i) effecting subtractive hybridisation using cDNA derived from transcript RNA of the test cell sample and an excess of a hybridisation driver nucleic acid derived from the reference cell sample so as to generate unsubtracted cDNA derived from genes expressed in the test but not the reference sample and duplexes comprised of cDNA and "hybridising driver" nucleic acid derived from genes expressed in both the test and reference cell samples,

(ii) effecting selective cross-linking of the strands in the duplex molecules,

(iii) using from the unsubtracted cDNA as a template to generate nucleic acid strands complementary thereto incorporating a binding ligand.

(v) treating the solid phase support system so as to remove non-bound nucleic acid and leave a pure sample of the bound complementary strand, and

(vi) using said pure sample of the bound complementary strand as, or to generate, said target nucleic acid molecules.

2. A method as claimed in claim 1 wherein the immobilised cDNA provides the nucleic acid sequence of interest.

3. A method as claimed in claim 1 wherein the immobilised cDNA is used to generate target nucleic acid sequences.

4. A method as claimed in claim 3 wherein, to generate the target nucleic acid sequences, the immobilised cDNA is subjected to a random priming amplification reaction in situ on the support to generate amplification products which provide the target nucleic acid sequences of interest.

5. A method as claimed in claim 3 wherein, to generate target nucleic acid sequences of interest, total cDNA from the test sample is added to the solid support system with immobilised complementary cDNA under hybridising conditions, the support is washed to remove non-hybridised sequences, the hybridised sequences are denatured, an the denatured sequences are amplified to generate the target nucleic acid molecules of interest.

6. A method as claimed in any one of claim 1 to 5 wherein the "hybridising driver^" nucleic acid required for step (i) is mRNA.

7. A method as claimed in anyone of claims 1 to 6 wherein the cross-linking agent is an aziridinylbenzoquinone.

8. A method as claimed in claim 7 wherein the aziridinylbenzoquinone is a compound of the formula:

in which R is H, alkyl, thiol, alcohol or halide.

9. A method as claimed in claim 8 wherein the cross-linking agent is 2,5- diaziridenyl-1 ,4-benzoquinone.

10. A method as claimed in anyone of claims 1 to 9 wherein the ligand is biotin.

11. A method as claimed in claim 10 wherein the binding partner is avidin or streptavidin.