WO1993002214A1

WO1993002214A1 - Normalised pcr method

Info

Publication number: WO1993002214A1
Application number: PCT/GB1992/001303
Authority: WO
Inventors: David Stephen Charnock Jones; Sydney Brenner
Original assignee: Medical Research Council
Priority date: 1991-07-17
Filing date: 1992-07-16
Publication date: 1993-02-04
Also published as: GB9115407D0

Abstract

A method for selectively increasing the relative abundance of rare cDNA species in a population is disclosed. The method involves denaturing the cDNA population and allowing the more abundant species to reanneal before amplifying the remaining, single stranded species using non-selective primers.

Description

NORMALISED PCR METHOD

This invention relates to the selective amplification of cDNA sequences using the Poly erase Chain Reaction (PCR) , and more particularly to the preferential amplification of rare cDNAs in inhomogenous cDNA populations in order to equalise the abundance of each cDNA within that population. This process is an example of a normalisation process.

The analysis of genes encoded in organisms with complex genomes, such as man, is greatly facilitated by analysing cDNA populations in preference to geno ic DNA. Analysis of the genomic DNA will often entail redundant analysis of repetitive DNA or non-coding DNA, and is therefore potentially inefficient. It is often desirable to sequence individual cDNA species in a population. This is straight forward as far as the more abundant cDNAs are concerned, but presents great difficulties with the rarer species. As the relative abundance of different cDNA species in a given population can vary by factors of as much as 10⁵ as a result of the variation in levels of gene expression, it will be appreciated that the attempt to sequence the rarer species presents a formidable problem.

This problem has been previously addressed in a number of ways.

Firstly, a standard PCR protocol can be used to amplify selectively cDNAs which are present at extremely low levels, if there is information about the sequence of those cDNAs. If not, a primer specific to the desired cDNA cannot be constructed and the desired cDNA cannot be selectively amplified. The standard PCR method is therefore insufficient if it is desired to characterise a number of unknown genes.

A second approach involves hybridisation of cDNA to genomic DNA. At saturation, the cDNAs recovered from genomic/cDNA hybrids will be present in the same abundance as the genes encoding them. This will provide a much more homogenous population than the original cDNA library, but does not entirely solve the problem. In order to reach saturation in respect of the very rare sequences, it will be necessary to use huge quantities of cDNA, which need to be allowed to anneal to large amounts of genomic DNA over a considerable period of time. Furthermore, cDNAs which are homologous to genes which are present in multiple copies in the genome will be over-represented.

A third approach exploits the second order reassociation kinetics of cDNA annealing to itself. After a long period of annealing, the cDNAs which remain single stranded will have nearly the same abundance, and can be recovered by standard PCR. This method has been employed in two recent publications (Patanjali, S.R. et al . , PNAS USA 88. 1991, pp. 1943-1947; Ko, M.S.H., NAR 19, no.18, 1991, pp. 5705-5711) . The methods disclosed in these two publications, however, suffer from notable disadvantages. They are entirely dependent on the stringent physical separation of single stranded and double stranded DNA, require an elevated number of manual manipulations in each reaction, and necessitate protracted hybridisation times (up to 288 hours in the method of Patanjali et al . )

It is therefore an aim of the invention to provide a method by which cDNA populations can be readily normalised in order to allow analysis of the individual cDNAs present in these populations. The method of the invention employs PCR in a novel and inventive way, thereby normalising and amplifying a cDNA population in a single step.

According to a first aspect of the invention, there is provided a method for normalising a DNA population comprising the steps of:

(a) preparing a mixture comprising a heterogenous DNA population and oligonucleotide primers suitable for use in a PCR process, in which the DNA is denatured; (b) altering the conditions to allow the denatured strands of the more common DNA species to reanneal, .while preventing the annealing of the primers to the DNA strands; (c) further altering the conditions of the mixture in order to allow the primers to anneal to the remaining single-stranded DNA comprising the rarer DNA species; and (d) carrying out an extension synthesis in the mixture produced in step (c) . Advantageously, the method will consist in the cyclic application of the above four steps.

Preferably, the DNA population is a cDNA population. The principle embodied in the method of the invention is to interpose a reannealing step between the denaturation step and the priming step of a standard PCR protocol. The reannealing step is carried out under conditions sufficiently stringent to prevent annealing of the primers to the DNA, which will only occur under less stringent conditions as the primers are considerably shorter than the DNA strands.

Preferably, the conditions may be altered by the alteration of the temperature of the reaction mixture. However, it will be appreciated that any conditions which affect the hybridisation of complementary DNA strands to one another may be varied to achieve the required result.

Because the reannealing efficiency of any given DNA species will depend on the product of its concentration and time, CoT, the more abundant the sequence the greater the extent to which it will reanneal in any given time period. Once a DNA species has reached a certain threshold concentration it will no longer be amplified exponentially, as a significant amount will have annealed to the double stranded form before the priming step. Thus, as each individual DNA species is amplified by the PCR process to its threshold concentration, the rate of amplification of that species will start to tail off. Eventually, therefore, all DNA species will be present at the same concentration. The length of the reannealing step will determine how much DNA is present at the threshold concentration. Preferably, therefore, the duration of the reannealing step will be determined empirically for each DNA population. Advantageously, the duration of the reannealing step may be between 5 and 30 minutes, preferably between 10 and 20 minutes, and most preferably 16 minutes.

The PCR primers used should preferably be effective to prime non-selectively all DNA species in the population. This will lead to a normalised DNA population containing equal amounts of all the DNA species. Alternatively, in the case of a cDNA population, they may be specific for the DNA vector 'arms' used in the cDNA cloning process by which the cDNA population was obtained, and which will be common to all cDNA species in the population.

It has been found that the method described above is effective in producing normalised DNA populations when starting with a population of a few hundred sequences. However, when it is desired to normalise a cDNA library consisting of tens of thousands of distinct cDNA species, it has been found that the use of primers complementary to the arms of the cDNA clones or which otherwise prime all the species present is impracticable as excessive amounts of cDNA are generated.

The total quantity of DNA generated by the method of the invention is dependent on the threshold concentration which is selected. The lower this threshold concentration, the lower the total amount of DNA which will synthesised, as each individual species will cease to be amplified exponentially when a lower concentration is reached. However, in a population of 10,000 or more individual cDNA species, the threshold concentration required to limit the total cDNA produced to within a feasible limit would require an excessively long cycling time (CoT being a constant) .

According to a second aspect of the invention, therefore, there is provided a method according to the first aspect of the invention, in which the DNA primers are adapted to prime selectively a sample of the total DNA population.

By using primers which will only prime a sample of the population, only that sample will be amplified and normalised. The total quantity of DNA generated will thereby be reduced, which means that the cycling times can be kept low. This ensures that the method of the invention is applicable to complex DNA populations such as cDNA populations. Advantageously, it is envisaged that the method of the second aspect of the invention could be carried out using a first primer adapted to selectively prime a sample of the total cDNA population and a second primer which is a general primer. Advantageously, the general primer is oligo dT. This will ensure that each primed cDNA will be replicated in its entirety, as the oligo dT primer will anneal to the poly-A tail at the end of the cDNA.

The first primer can have any sequence which is desired. Advantageously, however, it will be GC-rich. The longer the primer, the more specific it will be: for example, if the complexity of a cDNA library derived from an organ such as the human placenta is assumed to be in the region of 100 million base pairs, a 10-nucleotide sequence would be expected to occur one hundred times in the entire library. Only 100 sequences, therefore, would be expected to be amplified and normalised if a 10-nucleotide first primer were used.

The priming efficiency of the first primer can be increased by the inclusion of any number of degeneracies in the DNA sequence, each additional degenerate nucleotide doubling the expected number of sequences primed. In a preferred aspect of the invention, an 11-base pair primer with four degeneracies may be used.

According to a third aspect of the invention, there is provided a normalised DNA population in which the individual DNA species have approximately the same abundance.

The invention is now described, by way of example only, with reference to the accompanying drawings, which are graphical representations of the PCR cycle. Figure 1A represents a normal PCR cycle, in which the denatured template/primer mixture is rapidly.cooled to the annealing temperature. Figure IB represents the modified PCR cycle of the invention, which includes a hybridisation step to allow the templates to reanneal before the annealing step.

Example 1 Two recombinant plasmids with unrelated inserts were mixed in different ratios between 1:10^_1 and 1:10~⁵. In theory, normalisation over a range of 10⁵ in difference of abundance will require an additional 16.6 PCR cycles (2^16*6 - 10⁵) for the least abundant sequence to reach the same threshold concentration as the most abundant. It was empirically determined that an annealing time of 16 minutes resulted in a threshold concentration giving a good total yield of DNA and added only 8 hours to a 30 cycle reaction.

The following mixture was prepared, containing buffer (lOmM Tris-HCl pH8.3, 25°C, 40mM KCl, 1.5mM MgC_₂, 0.01% gelatin) , 200μm dNTPs, 10μCiα³²P dCTP, lμM each of M13 universal forward and reverse primers, Taq Polymerase and DNA from the first plasmid which was present at a constant concentration. Variable concentrations of DNA from the second plasmid were then added to this mixture.

The procedure was initiated with three cycles of standard PCR, and followed by 27 cycles of modified PCR according to the invention (hereafter referred to as CoT PCR) . A final cycle of standard PCR was carried out to remove any incomplete strands trapped in annealed structures. The condition profile for the reactions was as follows:

Standard PCR: 95°C 90s, [95°C 30s, 30°C 30s, 72°C 30s) x 3; CoT PCR: [95°C _{30S/ 72}°_c 16 min, 30°C 30s, 72°C 30s] x 27; Standard PCR: 95°C 30s, 30°C 30s, 72°C 3 min. The heating rate from 30°C to 72°C was set 10°C min^"1, all other heating and cooling rates being set at the maximum achievable with the PCR heating block. The products of the reaction were analysed by gel electrophoresis on a 1% agarose gel, which was subsequently dried and autoradiographed. Quantitation was achieved by densitometric analysis of the autoradiograph and the abundance ratios of the of the plasmids at the end of the reaction calculated, taking into account their molecular weights. The results of this experiment are shown in Table 1.

Table 1

Ratio before CoT PCR 10^"22 11C0^"3 10^"4 10^"5 Ratio after CoT PCR 1.6 1.1 1.02 1.07

These results demonstrate the effectiveness of CoT PCR in equalising the concentrations of two plasmids which were initially present at very different concentrations.

Example 2

Human placental cDNA was obtained by synthesising cDNA from 0.5μg of placental mRNA under standard conditions using AMV reverse transcriptase and the primer LNotdt (5' TACGTCGACAAGCTTGAATTCGCGGCCGC(T)₂₆3') at lμM. The reverse transcriptase was heat-inactivated and the cDNA diluted to lOOμl. lμl of this single-stranded cDNA population was then used for PCR.

Two reaction mixtures were set up exactly as described in Example 1, except that the plasmid DNA was replaced by the synthesised cDNA and the M13 primers were replaced by the general primer LNotdT and the specific primer 11AD1 (5'GCC(TA) (GC)CGCCGA3') .

The specific primer 11AD1 was selected to give approximately 100 amplified sequences from the DNA population.

One mixture was subjected to a standard PCR procedure (95°C 90s, [95°C 30s, 45°C 30s, 72°C 30s] x 35; 72°C 3 min) while the other was subjected to CoT PCR (95°C 90s, [95°C 30s, 45°C 30s, 72°C 30s] X 3; [95°C 30s, 72°C 16 min, 45°C 30s, 72°C 30s] X 27; 90°C 30s, 45°C 30s , 72°C 3 min]. The heating and cooling rates were set as in Example 1.

The reaction products were end-repaired with T4 DNA poly erase (5 units were added to the PCR mix and incubated for 10 minutes at 37°C) and then extracted with phenol, ethanol precipitated and- resuspended in Notl digestion buffer. An excess of Notl (20 units) was added, and after digestion the DNA was again extracted with phenol and ethanol precipitated. About 10% of the DNA was ligated into EcoRV/Notl digested pBluscript (Stratagene) and transformed into E.Coli DH5α. White (recombinant) colonies were picked to microtitre plates, grown and the sequence of the inserts determined either by double stranded dideoxy sequencing (Jones, D.S.C. and Schofield, J.P., NAR 18., no.24, 1991, pp. 7463-7464) or by preparation of a single stranded template by PCR and binding to a solid support (Jones, D.S.C. et al . , 1991, DNA Sequence, in press) . Analysis of the products of both normal and CoT PCR procedures on agarose gels showed that whereas the normal PCR reaction gave distinct bands of products, the CoT PCR reaction produced a smear, indicating that a greater number of different sequences had been amplified. This was confirmed by the sequencing results, which are set out in Table 2. A total of 62 clones from the standard protocol and 73 clones from the CoT protocol were sequenced. The sequences were sorted into families of similar sequences using the nonhierarchical clustering program ICA tools (Parsons, J.D. et al . , 1991, Genomics, submitted) It was found that, of the 73 CoT PCR sequences, 44 were different (60%) as opposed to 24 different sequences from the 62 normal PCR sequences (39%) .

In order to extend these results, the abundance of 23 of the sequences was assayed by hybridisation to a large cDNA placental library. The results (Table 3) show that the clones from the normal PCR procedure are biased to the abundant cDNAs, while the clones from the CoT PCR procedure are evenly distributed across the range. In addition, most of the new sequences isolated by CoT PCR are rare ones, confirming that the CoT PCR procedure is effective. This is further confirmed by Poisson distribution analysis of the results shown which, after the anomalies have been eliminated, show that the actual results conform to the expected distribution extremely faithfully.

Table 2

The figures in brackets indicate the number of sequences common to both standard and CoT PCR.

+ These sequences have an internal Not l site and have therefore cloned with high efficiency.

Table 3 Abundance of sequences in a placental cDNA library.

Probe Identity Standard CoT frequency lo-

1 1.9

10 1 1.6

0 1.5

9 i.5

1 1

0 0.8

15 3 0.7

1 0.7

1 0.5

1 0.4

2 0.4

1

0.3-0.8

It will be appreciated that the present invention is described above by way of example only, and that modifications and variations may be effected by a person skilled in the art without departing from the scope of the invention as defined in the appended claims.

Claims

^{•
l. A method for normalising a heterogenous DNA population comprising the steps of: a) preparing a mixture comprising a heterogenous DNA population and oligonucleotide primers suitable for use in a PCR process, in which the DNA is denatured; b) altering the conditions to allow the denatured strands of the more common DNA species to reanneal while preventing the annealing of the primers to the DNA strands; c) further altering the conditions in order to allow the oligonucleotide primers to anneal to the single- stranded DNA comprising the rarer DNA species; and d) carrying out an extension synthesis in the mixture produced in step (c) .}

2. A method according to claim 1 in which the heterogenous DNA population is a heterogenous cDNA population.

3. A method according to claim 1 or claim 2, involving the cyclical application of steps (a) - (d) .

4. A method according to any one of claims 1 to 3, further comprising at least one cycle of a normal PCR process.

5. A method according to claim 4, which is initiated by three cycles of a normal PCR process and terminated with a single cycle of a normal PCR process.

6. A method according to any one of claims 1 to 5, wherein the duration of step (b) is empirically determined.

7. A method according to any one of claims 1 - 5, where the duration of step (b) is between 5 and 30 minutes, preferably between 10 and 20 minutes, and most preferably 16 minutes .

8. A method according to any one of the preceding claims in which the DNA primers comprise a primer adapted to specifically prime a sample of the DNA population.

9. A method according to claim 8, in which the primer contains at least one base degeneracy.

10. A method according to claims 8 or claim 9, in which the primer is GC rich.

11. A method according to any one of claims 8 to 10, in. which the primer comprises eleven nucleotides.

12. A method according to any one of claims 8 to 11, in which the primer which has the sequence:

5• (GCC(TA) (GC)CGCCGA) 3^» .

13. A method as claimed in any one of claims 8 to 12, further comprising the use of a general primer.

14. A method as claimed in claim 13, where the general primer is oligo dT.

15. A normalised DNA population.

16. A normalised DNA population prepared by the method of any one of claims 1 to 14.

17. A normalised DNA population according to claim 15 or claim 16 which is a cDNA population.

18. A primer which has the sequence: 5' (GCC(TA) (GC)CGCCGA) 3'.