WO1999001575A1

WO1999001575A1 - Method for semiquantization of alleles

Info

Publication number: WO1999001575A1
Application number: PCT/FR1998/001413
Authority: WO
Inventors: Marie-Christine Chartier-Harlin; Jean-Charles Lambert; Philippe Amouyel
Original assignee: Institut Pasteur De Lille; Institut National De La Sante Et De La Recherche Medicale (Inserm)
Priority date: 1997-07-01
Filing date: 1998-07-01
Publication date: 1999-01-14
Also published as: AU8445698A; FR2765590B1; FR2765590A1

Abstract

The invention concerns a method for the relative quantization of a first allele, called allele 1, with respect to a second allele, called allele 2, comprising in particular the following steps: a) amplifying the DNA by PCR; b) digesting the amplified DNA by at least one restriction enzyme, for differentiating the alleles; c) separating the DNA fragments; d) evaluating the amount of resulting fragments by means of a marker emitting a signal capable of detection; e) determining the initial percentage in the different alleles, using formula (I).

Description

Method for semi-quantification of alleles

The subject of the present invention is a method of semi-quantifying alleles.

Several methods for quantifying mRNA exist.

Hybridization by Northern Blot, by "dot and slot" hybridization, or the protection test by S1 are techniques which make it possible to estimate the total quantity of a particular species of mRNA in a preparation.

However, techniques require large amounts of mRNA to be valid, which is a major drawback. In particular, this drawback becomes critical when the mRNA preparation comes from tissue removed by autopsy, the integrity of the sample being compromised. In addition, these methods do not provide any information as to the level of expression of the various alleles of a locus.

Other methods take advantage of the sensitivity of RT-PCR to increase the concentration of specific mRNA and are widely used to measure the amount of mRNA in a preparation.

Thus, techniques have been developed from RT-PCR to study the difference in expression level of the alleles of a specific gene, using the combination of allele-specific PCR amplification and quantification by PCR (Horiskoshi et al, 1994, Leuk. Res., 18, 693-702) or even using electrophoresis on an automatic DNA sequencer with fluorescent labeled primers (Jensen et al., 1996, Hum. Mut. , 8, 126-133).

However, these methods can be difficult to implement and require significant equipment.

The authors of the present invention have developed a method for the relative quantification of a first allele, called alele 1, with respect to a second allele, called allele 2, comprising the steps consisting in: a) subjecting the DNA to a amplification by PCR in the presence of primers allowing the specific amplification of at least one polymorphic sequence of the alleles; b) subjecting the amplified DNA to the action of at least one restriction enzyme, allowing the differentiation of the alleles; c) separating the DNA fragments; d) evaluating the quantity of fragments obtained using a marker emitting a detectable signal; e) determine the initial percentage in the various alleles, by the following formula:

No. allele 1 Aα 'allele 1

N ₀ allele 1 ⁺ N ₀ allele 2 Aα 'allele 1 ⁺ Ct'allele 2

in which N ₀ is the initial number of DNA molecules, A is the proportionality coefficient making it possible to correct the difference in size between the different restriction fragments and is equal to the ratio of the lengths of restriction fragments characteristic of each allele, and α 'is determined

either by the following formula

. S „α = 'max

K 'in which S _max is the maximum signal that can be measured, and K' is a constant,

S _max and K 'being determined by the following function f:

S = f (V) = ₇ - ^ ^{!! ≥} _τ . V (K '+ V)

in which V is the volume of the sample obtained by PCR subjected to step c) and S is the measured signal emitted by the marker;

• either by the following function g:

1 1 1

^ 7J = ^ V ⁺ S _m „ in which S is the measured signal emitted by the marker, V is the volume of the sample obtained by PCR subjected to step c) and S _max is the maximum signal which can be measured; provided that the amplification coefficient Ei of the DNA containing allele 1 is identical to the amplification coefficient E ₂ of DNA containing allele 2.

This new method of semi-quantification of alleles is valid regardless of the efficiency of the PCR, the current signal variations from one experiment to another, and takes into account the possible phenomenon of signal saturation.

It is also particularly simple. It suffices, in fact, either to transfer the value S of the signal emitted as a function of the volume V of the sample, the curve representative of the function f such that S = f (V) can be plotted using the method of least squares , and to determine S _ma χ and K 'from this curve; either to carry the value 1 / S as a function of the inverse of the sample volume, namely 1 / V, the curve representative of the function g such that - = gl - I can be plotted by linear regression, the slope of the straight line obtained being equal to the inverse of α '.

Said DNA amplified according to step a) can be a genomic DNA or a cDNA comprising the allelic sequences of interest, this cDNA can be obtained from mRNA by the usual RT-PCR technique. The alleles are differentiated according to steps b) and c) described above, which demonstrate the restriction polymorphisms (RFLP).

The separation of the DNA fragments according to step c) can be carried out in particular by gel electrophoresis, preferably by electrophoresis on polyacrylamide gel.

The method according to the invention as previously written is only applicable if the condition E ^ E _∑ is satisfied.

Indeed, if, theoretically, PCR is an exponential process in which, at the end of each cycle, the quantity of amplified DNA is directly linked to the quantity of DNA molecules initially present, it is however not excluded that, in practice, the ratio between two similar amplified targets is not preserved at the end of the amplification process. The quantity of PCR products increases exponentially according to the following equation:

N = N ₀ (1 + E) ⁿ

where N is the number of PCR products,

No is the initial number of DNA molecules, n is the number of amplification cycles; and

E is the amplification coefficient, which is representative of the efficiency of the PCR for each amplified population.

No, the initial number of DNA molecules, is therefore estimated from N, the number of PCR products, this relationship depending on the coefficient E.

To compare the N ₀ values of two different matrices, namely DNA, whether genomic DNA or complementary DNA, corresponding to an allele 1 and to an allele 2, from the N values, it is therefore necessary to first evaluate the variation of the coefficient E for one matrix (allele 1) compared to the other (allele 2).

The percentage of allele 1 PCR products compared to all PCR products can be expressed by the following formula:

^N 0allèlel (1 + Eallèlel) ⁿ ^ allèle-! OallόleU ' ⁺ ^ allèlell ⁺ ^ 0allèle ₂ V' + ^ allèle∑ N _a iièle1 ⁺ allèle ₂

either, if Ei≈E _∑

oallèlel ^ allèlel oallèlβl ⁺ θallèle2 allèlel ⁺ ^ allόle2

The final percentage in allele 1 is then equal to the initial percentage in allele 2.

The coefficients Ei and E ₂ can be calculated, in practice, by kinetic experiments, using in particular dilutions of genomic DNA from homozygous individuals making it possible to obtain initial percentages of known alleles. The PCR reaction is then stopped after a variable number of cycles, preferably from 25 and up to 35 cycles. The PCR products are then subjected to digestion with restriction enzymes allowing the differentiation of the alleles and the fragments obtained are separated, preferably by gel electrophoresis. A marker emitting a quantifiable signal makes it possible to detect the fragments.

As a first approximation, the signal S emitted by the marker is directly proportional to the number N of amplified molecules, according to the relation S = αN, where α is the coefficient of proportionality. Knowing that N and n are linked by the relation: log N = niog (1 + E) + N ₀ then it suffices to report, for each allele, the logarithm of the value of the signal S emitted by the marker as a function of the number of PCR cycles, the corresponding curve can be obtained by linear regression.

This allows the determination and comparison of the amplification coefficients E of the two alleles.

The validity of this determination is confirmed by representing the values S _a ιièiei as a function of S _a ιièiβ2-

The slope of the line obtained by linear regression is called p and is such that Saiièiβ 2 = pSaiiôiβ 1- Using the formula

θallèlβ-1 allèlel θallèle-1 ⁺

N, allelel ⁺

with S = αN

we obtain

^ allèle-l θallèle-1 α ₁ θallèle1 ⁺ N _0a || èle2 ^ allèlel £> _a lléle2 α ₁ α ₂

The signal also depends on the length of the labeled DNA fragments, which is expressed according to the relationship: α _a n _è ι _{θ 2} = Aα _a n _{è | θ} ι, A being the proportionality coefficient to correct the difference in size between the restriction fragments and is equal to the ratio between the lengths of the fragments.

The initial percentage of allele 1 can then be expressed by the following formula:

is

N Oallèlei

N _0allele1 + N _{0allè | β2} A + p

N The ratio ^0a " ^èlθ1 being known, the determination of p

Nθallèl ^β 1 ⁺ ' ^S * 0allèle ₂ allows the verification of the above equality, which validates the kinetic conditions.

The equality between the amplification coefficients Ei and E ₂ being verified, the method of the invention as described above can be applied.

This method was developed to account for the observed variation in the coefficient α, due in particular to a phenomenon of signal saturation. It therefore applies to any processing of a signal emitted by a marker, whether or not the latter is subject to a saturation phenomenon. In accordance with the invention, the saturation phenomenon is described by analogy with the ligand-receptor or enzyme-substrate interaction and can be expressed as equivalent to the Michaelis-Menten equation, according to the following formulas:

S ^ ^ N _{or S =} S ^ V

K + N ° K '+ V in which S _max is the maximum signal emitted by the marker which can be measured, V is the volume of the sample obtained by PCR subjected to the separation of the fragments,

K and K 'are constants, with K' = -, where C is the

final concentration in each PCR product corresponding to the alleles.

The value of the coefficient α can be expressed as a function of the volume V, as being the derivative of the function f such that

^{s = f} ( ^v ) = | ^

is

S _mav K 'α' = f '(V) =' max '

(K '+ V) ²

with α '= Cα.

When the values of V tend towards zero

_α ' ₌ | j _m f (V) =% ≈- v → OK

According to the invention, the relationship between the dilution of the PCR sample and the measured signal is studied. The measured signal values make it possible to obtain a curve representative of the function f as defined above. This method makes it possible to estimate the constant K 'and the value Smax and thus to calculate the coefficients α' _a n _è | _{θ 1} and α ' _a ιi _è ie ₂ . According to another preferred embodiment of the invention, the values of the measured signal make it possible to obtain a curve representative of the function g, which follows from the function f as follows: SV

S ≈ ^ 7- = f ()

5 hence, as α '= -rϋr ^~ K

This method gives directly the value of the coefficients α ' _a ii _è i _{e 1} and α' _a ιi _è i _e2 and makes it possible to evaluate by linear regression the validity of the equation describing the phenomenon of saturation.

The allele semi-quantification method according to the invention has several advantages:

It makes it possible in particular to take account of the saturation phenomenon. The introduction of the constants K and K ′ furthermore makes it possible to normalize the signal, whatever the inevitable variations from one gel to another due to the technique used.

Finally, as the method aims at evaluating a DNA ratio for each allele, the effectiveness of the PCR has no impact on the final results.

This method can be applied to RT-PCR products. A specific band of an allele can be used as internal control compared to another allele, this application eliminates the major technical difficulty of RT-PCR to have comparative data. In addition, this method requires small amounts of biological material, which is an important advantage when analyzing mRNAs. The method according to the invention is particularly suitable for normalizing the phenomenon of saturation. The marker used can in particular be a DNA coloring agent such as a DNA intercalating agent, or a DNA complexing agent, a fluorescent agent, a radioactive agent or a luminescent agent, in particular a agent chemiluminescent, which can be detected on autoradiographic films. The preferred marker is silver nitrate or ethidium bromide.

The following examples and figures illustrate the invention without limiting it.

LEGEND OF FIGURES

Figure 1 shows an example of kinetic experiment. FIG. 1A represents a digital gel of the digested PCR products corresponding to the alleles ε3 and ε4 of the APOE gene after staining with silver. This example corresponds to the initial ratio in ε4 allele of 27.3%.

FIG. 1B represents the quantification by densitometry of the quantity of restriction fragments. A linearity zone (25-30 amplification cycles) was established between the logarithm of the staining intensity of two restriction fragments (72 bp, 91 bp) and the number of PCR cycles. The values corresponding to the 72 bp fragment (base pairs) were obtained by multiplying the intensity corresponding to the 72 bp fragment measured on the gel multiplied by the coefficient A equal to 91/72. Figure 1C shows a linear domain between the optical density corresponding to the 72 bp and 91 bp fragments, which was established by including PCR amplification. The ratio calculated in ε4 allele was 27.6%. FIG. 2 represents an example of a semi-quantification experiment and of the different mathematical treatments developed to interpret the experimental measurements. This example corresponds to an initial ratio in ε4 allele of 20%.

FIG. 2A represents a digitized gel of digested PCR products corresponding to the alleles ε3 andε4 after staining with silver.

FIG. 2B represents the relationship between the intensity of the colorings of two restriction fragments (72 bp O, 91 bp O) and the volume of the PCR products applied to the gel. The curve was performed using the method of least squares for the optical density values corresponding to the 72 bp (O) and 91 bp (O) fragments. For the 72 bp fragment (allele ε4), S _max = 2.881 and K "= 6.448 μl. For the fragment of 91 bp (allele ε3), the calculated values are S _max ⁼ 4.327 and K '= 1.781 μl. The final percentage of ε4 allele calculated from these data is 18.87%, Figure 2C represents the relationship between the inverse of the coloring intensity of the two restriction fragments (72 bp O, 91 bp O) and the inverse of the volume of PCR products. The linear regressions were performed for an optical density value corresponding to the fragments of 72 bp and 91 bp. The equation obtained for the ε4 allele (optical density value corresponding to the fragment of 72 bp) was y = 1.702x + 0.615 (^ = 0.995) and y = 0.348x + 0.281 (^ = 0.999) for the ε3 allele (value corresponding to the 71 bp fragment). The final percentage in ε4 allele calculated from this data was 20.53%.

EXAMPLES

Semi-quantification of the PCR products corresponding to the APOE alleles by silver staining.

The apolipoprotein E gene is involved in the development of Alzheimer's disease. The variability of the locus changes the risk of developing the disease. But, although there is evidence that the specific effects of isoforms can influence the development of the disease, the fact that these effects are not seen on all carriers of a specific isoform prompted the authors of the present invention to seek additional variabilities for this locus. In the context of this study, it was suggested that differences in the level of expression of the different alleles could explain the differences observed between individuals sharing the same genotype on this locus. It was therefore necessary to quantify the level of expression of the alleles by RT-PCR followed by an RFLP analysis, this technique being the only way to prove the specific differences between the alleles. EXAMPLE 1: Verification of the conditions of the semi-quantitative method: kinetics

a) PCR conditions for the validation of the semi-quantitative method Dilutions of genomic DNA from individuals homozygous ε3 and ε4 were used to obtain a known initial percentage of ε4 allele (ranging from 10 to 70%). The PCR sample contained 20 ng of genomic DNA. The genomic DNA was amplified by PCR in a DNA thermocycler (Equibio Thermojet Eurogentec) with the primer F4 5'- ACAGAATTCGCCCCGGCCTGGTA-3 'and the primer F6

5TAAGCTTGGCACGGCTGTCCAAGGA-3 '. (Hixon et al., 1991, J. Lipid Res., 31, 545-548). Briefly, the PCR was carried out according to RT-PCR conditions in a total volume of 25 μl, containing 2.5 units of Taq DNA polymerase, 2 μM of each primer, 0.2 mM of each dNTP, 4 mM of DTT, 0.1 mM MgCI ₂ , 0.04% Triton X-100, 15% (V / V) glycerol, 1X RT-PCR buffer (Gibco / BRL), 0.5 μl of storage buffer M-MLV reverse transcriptase (Gibco / BRL). The reaction was stopped in ice after 25, 26, 27, 28, 29, 30, 31, 32 and 35 cycles (1 minute at 94 ° C, 1 minute at 58 ° C and 1 minute at 72 ° C for each cycle).

b) Electrophoresis

After digestion by adding 12 units of Cfo I endonuclease (Promega), the DNA fragments were resolved in an 8% polyacrylamide gel, for four hours at 12 V / cm. 1 to 2 μl of digested samples were loaded onto the gel.

c) Coloring by arαent The gel was fixed for 90 minutes in 10% (VA /) of ethanol, 0.5% (VA) of acetic acid. After two washes with deionized water, the gel was placed in a silver nitrate solution (1 mg / ml) for 25 minutes. Then, the gel was washed twice with deionized water. Polymorphic DNA fragments were stained for 30 minutes in 0.037% formaldehyde (VA), and sodium hydroxide (15 mg / ml). Finally, the reaction was stopped with a sodium carbonate solution (15 mg / ml). The developer and the silver nitrate solutions were prepared immediately. The gel was digitized on a high resolution Sharp JX-325 color scanner and the intensity of the fragments was measured using Image Master software (Pharmacia) by subtracting background noise.

d) Mathematical and statistical calculation method

The curve adjustments using the least squares method were carried out by matrix calculation with Matlab software. Linear regression was performed with SAS 6.04 software (SAS Institute Inc., Cary, NC, USA).

e) Validation of the conditions by kinetic experiments The coefficients E were calculated by kinetics (FIG. 1A), the quantity of allele ε4 ranging from 10 to 50%, using the formula:

log N = n log (1 + E) + N ₀

Up to thirty cycles, the exponential amplification has been verified (Figure 1B, Table 1). The coefficients E are heterogeneous among the PCR samples but homogeneous between the ε3 and ε4 alleles in a PCR sample (Table 1). However, if we estimate the initial percentage of ε4 from the coefficients E (Table 1) we make an error of 14% to 74% after thirty amplification cycles. To limit point-to-point variations due to the experimental approximation, a linear regression was carried out using the formula DO _a ιièi _{β 2} = pDO _a n _è ι _θ ι, DO being the measured optical density. N / A

The equation ^0a " ^high = - - allows to calculate the

Nûεi _ll è _l e _l ⁺ N _{0a || è} | _θ2 A + P initial percentage in ε4 allele in populations ε2 / ε4 and ε3 / ε4.

In the case of the APOE gene, the ε3, ε2 and ε4 alleles can be characterized by restriction fragments of 91 bp, 83 bp and 72 bp respectively.

The proportionality coefficient A can therefore be calculated as A = 91/72 for the alleles ε3 / ε4, A being 83/72 for the individuals ε2 / ε4 and

A being 91/83 for individuals ε2 / ε3. For this last genotype, since the two alleles ε2 and ε3 give a restriction fragment length of 91 bp then we have the relationship A DOaiièie 1 + DO _a nèi _θ 2 = DO 91 _bp .

The initial percentage of ε2 allele in the ε2 / ε3 population can be calculated as

No ^β llèlel ⁺ 0aHèle2 P

By this method of calculating the percentage of the ε4 allele, an error of 0.01% to 11.5% in the linearity zone (Table 1) was found. By this approach, it has thus been demonstrated that the coefficients E were similar whatever the allele studied in the same PCR sample: the efficiency is not a limit for retaining the initial percentage in ε4. The initial report being preserved, all the conditions necessary for the experiment of semi-quantification are checked.

EXAMPLE 2 Semi-quantitative method taking into account a saturation phenomenon

a) PCR conditions The genomic DNA (20 ng to 100 ng) was amplified by PCR according to the conditions described in example 1.

The PCR reaction was performed up to 30 cycles.

b) Electrophoresis

The digested PCR samples were deposited on the gel at a rate of 4 μl to 0.1 μl.

c) Silver staining The silver staining was carried out according to the protocol described in Example 1.

d) Validation of the semi-quantification method

The relationship between the dilution of a PCR sample and the optical density measured on the gel for different initial ratios in the ε4 allele was studied. The optical density resulting from the treatment of the gels has been reported according to two methods: i) method 1 corresponds to the adjustment of curves by the use of the method of least squares (FIG. 2B), this method allows the estimation of the constant K 'and the maximum optical density DO _max and thus makes it possible to calculate the coefficients α'. ii) method 2 consists in processing the measured data according to the

1 (λ 1 1 1 curve g representative of the function - - ^• = d - = - - x— + - - (Figure 2C).

DO VJ α 'v S

This method gives directly the values of the coefficients α 'and makes it possible to evaluate by linear regression the validity of the equation describing the phenomenon of saturation.

Method 1 mainly emphasizes the importance of the points corresponding to a high quantity of PCR products, while method 2 emphasizes more particularly the low dilutions. The major interest of

use of these two methods is to verify that the formula S = ^nw is

K '+ V able to describe the saturation phenomenon whatever the points used.

For these two methods, the curves obtained by regression can follow the distribution of the points (Figures 2B and C). This was confirmed by the linear regression obtained from method 2 and for which the linear regression coefficients were always found to be very close to 1 (Figure 2C).

Regardless of the mathematical processing, the final ratio calculated corresponds to the initial ratio in allele ε4 whatever the percentage studied (table 2). However, method 2 gave better approximations than method 1. In fact, the narrowness of the bands allows more efficient quantification for low dilutions than for large quantities of PCR products. Consequently, it seems that method 2 is more suited to the treatment of these results. In addition, different concentrations of genomic DNA (20 ng, 50 ng and 100 ng) were used to verify that the conditions of application of the semi-quantitative method were verified for different concentrations of genomic DNA. The quality of the quantification method was not changed by the initial concentration of genomic DNA (Table 2).

Table 1

Comparison of PCR parameters as a function of the initial percentage of ε4 allele

(a) the final estimated percentage was calculated from the values of the E coefficients in the amplification linearity zone

(b) the final calculated percentage was calculated from formula 8 in the linearity area of the amplification Table 2

Comparison between the final ratio calculated and estimated in ε4 allele as a function of the quantity of genomic DNA in the PCR sample and the known initial ε4 ratio

The experiments were carried out in triplicate for 20 ng and in duplicate for

50 and 100 ng.

(a) method 1 corresponds to the adjustment of the curve using the method of least squares and makes it possible to calculate the values K 'and DO _max

(b) method 2 consists in the direct calculation of the coefficients α '.

Claims

1. Method for the relative quantification of a first allele, called allele 1, compared to a second allele, called allele 2, comprising the steps consisting in: a) subjecting the DNA to amplification by PCR in the presence of primers allowing specific amplification of at least one polymorphic sequence of the alleles; b) subjecting the amplified DNA to the action of at least one restriction enzyme, allowing the differentiation of the alleles; c) separating the DNA fragments; d) evaluating the quantity of fragments obtained using a marker emitting a detectable signal; e) determine the initial percentage in the various alleles, by the following formula:

No '_; allele 1 Aα 'allele 1

NOaiièle 1 + NOgiièie 2 Aα ' _a || èle 1 + CC'allèle 2

where No is the initial number of DNA molecules,

A is the proportionality coefficient making it possible to correct the difference in size between the different restriction fragments and is equal to the ratio of the lengths of restriction fragments characteristic of each allele, and α ′ is determined

• either by the following formula:

. S „α- ^max

K '

where S _max is the maximum signal that can be measured, and K 'is a constant,

S _max and K 'being determined by the following function f:

S = f () = 7 ^ ≈ .V '(K' + V) in which V is the volume of the sample obtained by PCR subjected to step c) and S is the measured signal emitted by the marker;

"Or by the following function g

in which S is the measured signal emitted by the marker, V is the volume of the sample obtained by PCR subjected to step c) and S _max is the maximum signal which can be measured;

provided that the amplification coefficient E1 of the DNA containing allele 1 is identical to the amplification coefficient E2 of DNA containing allele 2.

2. Method according to claim 1 in which S _max and K "are determined using the curve representative of the function f such that S = f (V) plotted using the method of least squares.

3. Method according to claim 1, in which α ′ is determined using the line representative of the function g such that:

plotted by linear regression.

4. Method according to any one of the preceding claims, in which said DNA amplified according to step a) is a genomic DNA or a cDNA comprising the allelic sequences of interest.

5. Method according to any one of the preceding claims, in which the signal is subjected to a saturation phenomenon.

6. Method according to any one of the preceding claims, in which the label used is a coloring agent, a fluorescent agent, a luminescent agent or a radioactive agent.

7. Method according to any one of the preceding claims, in which the label used is silver nitrate.

8. Method according to any one of the preceding claims, in which the DNA fragments are separated according to step c) by gel electrophoresis.