[36] Detection and quantification of mitochondrial DNA deletions

[36] Detection and quantification of mitochondrial DNA deletions

[36] DETECTION AND QUANTIFICATION OF m t D N A DELETIONS 421 the neighboring peaks (Fig. 3). Therefore, to obtain sequence data of 500 bp, the dis...

591KB Sizes 22 Downloads 27 Views

[36]

DETECTION AND QUANTIFICATION OF m t D N A

DELETIONS

421

the neighboring peaks (Fig. 3). Therefore, to obtain sequence data of 500 bp, the distance between primers FLz and H2 should be about 600 bp. In the sequencing results, large peaks are sometimes observed at 40-50 nucleotides from the primer peak. These are due to primer dimers formed between primer FL2 and primer H2 (or Ha). To reduce the precipitation of such primer dimers, it is recommended to precipitate the second PCR products with ammonium acetate and 2-propanol. Resolution of the sequencing gel depends largely on the quality of polyacrylamide gels. The acrylamide/bisacrylamide mixture (19:1, w/w) from Bio-Rad (Hercules, CA) is recommended by Applied Biosystems. A premixed acrylamide solution for automated sequencing (Pageset SQC6A, from Biomate, Tokyo, Japan) is also useful. A 24-well square-tooth comb (originally designed for GeneScan, Applied Biosystems), instead of a conventional shark-tooth comb for the sequencer, is convenient both for gel preparation and for sample application.

[36] D e t e c t i o n a n d Q u a n t i f i c a t i o n of M i t o c h o n d r i a l DNA D e l e t i o n s

By NAY-WEI SOONGand NORMANARNHEIM Introduction In diseases such as chronic progressive external ophthalmopelgia, Kearn-Sayre's syndrome, and Pearson's syndrome, a significant proportion (20-80%) of the mitochondrial DNA (mtDNA) carry large deletions of up to 10 kb (reviewed in Ref. 1). The proportion of deleted mtDNA populations in diseased tissues can usually be measured using standard Southern blotting techniques. However, detection and measurement of low levels of deletion in tissues from nondiseased individuals present challenges not only because of the high sensitivity required but also because of the need to discriminate between the few molecules of deleted mtDNA and the overwhelming excess of wild-type mtDNA molecules. We describe in the following sections PCR (polymerase chain reaction)-based techniques used in our laboratory2-4 for the detection and semiquantitative and quantitative 1 D. C. Wallace, flnnu. Rev. Biochem. 61, 1175 (1992). 2 G. A. Cortopassi and N. A r n h e i m , Nucleic Acids Res. 18, 6927 (1990). 3 N. W. Soong and N. A r n h e i m , in " P C R in Neuroscience" (G. Sarkar, ed.) p. 105 (1995). A c a d e m i c Press, New York. 4 N. W. Soong, D. R. Hinton, G. A. Cortopassi, and N. A r n h e i m , Nat. Genet. 2, 318 (1992).

METHODS IN ENZYMOLOGY,VOL. 264

Copyright © 1996by AcademicPress, Inc. All rights of reproduction in any form reserved.

422

MITOCHONDRIAL DISEASES AND AGING

[36]

measurement of the levels of the m t D N A 4977deletion, the so-called common deletion that removes 4977 bp of m t D N A sequence between two 13-bp repeats.

P o l y m e r a s e C h a i n Reaction A s s a y for D e t e c t i o n of Mitochondrial DNM 977 Deletion The m t D N A 4977deletion removes a section of m t D N A between nucleotide positions 8470 and 13,447 in the mitochondrial genomic sequence and occurs at a presumed deletion " h o t s p o t " involving two 13-bp direct repeats beginning at these positions. I Two primers are designed to lie just outside the 13-bp repeats (Fig. I). The P C R assay relies on the deletion to bring the two primers sufficiently close together to enable efficient amplification. The use of short P C R cycle times preferentially amplifies the deleted molecules over the nondeleted wild-type molecules presumably because, for the wild-type m t D N A , there is insufficient time for the extension of each primer through 5 kb of sequence. Thus, exponential amplification cannot take place. However, when the deletion removes this 5 kb of intervening sequence, the primers are brought close enough together such that each primer can be extended through the binding site for the other primer, a prerequisite for PCR.

OH

MT1A

i!i' FIG. 1. A 4977-bp deletion between two 13-bp direct repeats brings primers MTIA and MT2 sufficientlyclose together to allow preferential amplification under short cycle times. Primers MTIC and MT2 amplifya section of undeleted, wild-typemtDNA and are used for normalization of total mtDNA. Ou and OH denote the mitochondrial origins of replication.

[36]

DETECTION AND QUANTIFICATION OF

mtDNA

DELETIONS

423

PCR Analysis The PCR is carried out in 50-/xl volumes in 1x PCR buffer (50 mM KC1, 10 mM Tris-HC1, pH 8.3, 2.5 mM MgCI2, 0.1 mg/ml gelatin). The primer and nucleotide concentrations are 0.5 and 187.5 tzM per deoxynucleoside triphosphate (dNTP), respectively. For detection of mtDNA molecules with the 4977-bp deletion, primers M T I A and MT2 are used. Their sequences and the regions they correspond to on the mtDNA sequence map 5 are listed below. MT1A: MT2:

G A A T T C C C C T A A A A A T C T T T G A A A T , nucleotides 8224-8247 A A C C T G T G A G G A A A G G T A T T C C T G C , nucleotides 13,501-13,477

An initial denaturation step of 3 min at 92° is used. Cycling is carried out with a 20-sec denaturation segment at 92° followed by a combined annealing and extension segment of 20 sec at 60°. A final extension step of 72 ° for 3 min is performed at the end of the appropriate number of cycles. Deletion PCR is typically carried out for 30 cycles with 2 units of Taq (Thermus aquaticus) polymerase using 1 /xl of total genomic D N A (100 ng-1 /zg) isolated from tissues. Some tissues may contain extremely low levels of deletion, and it may be necessary to run more cycles of amplification using two rounds of amplification (see below). In all cases, however, proper contamination controls are mandatory. The reaction products can be resolved using 2% (w/v) agarose gels and visualized by ethidium bromide staining. With primers MT1A and MT2, a 303-bp product is amplified from mtDNA 4977 templates. We estimate that 103-104 deleted mtDNA molecules will result in detectable levels of the 303-bp product with ethidium bromide staining at 30 cycles of PCR. Using the short cycling parameters described above, no detectable amplification of undeleted mtDNA molecules occurs even though these sequences are in vast excess.

Considerations In practice, it should be possible to detect any large deletion using this strategy by proper design of primers located just outside the deletion break points and optimization of PCR parameters. However, the ability to amplify a small fragment should not be taken as proof of the presence of a specific 5 S. Anderson, A. T. Bankier, B. G. Barrell, M. H. L. D e Brujin, A. R. Coulson, J. Drouin, I. C. Eperon, D. P. Nierlich, B. A. Roe, F. Sanger, P. H. Schreier, A. J. H. Smith, R. Staden, and I. G. Young, Nature (London) 290, 457 (1981).

424

M I T O C H O N D R I A L DISEASES A N D A G I N G

[36]

deletion. The possibility of PCR artifacts should be explored carefully and eliminated. 2-4 Occasionally, we encounter samples that give no detectable deletion product even with increased cycle numbers. In these cases, the levels of deletions may be extremely low, or inhibitors that interfere with PCR may be present. To eliminate the latter possibility, a tested positive control D N A sample is mixed with the questionable sample and PCR performed. If there is a significant decrease in the signal of the deletion product from that of the unmixed positive sample, then the presence of inhibitors should be suspected. Further purification of the questionable D N A sample should then be performed. Detection of extremely low levels of deletion may still be achieved by concentrating the D N A sample, using radioactively labeled primers or nested PCR. For the mtDNA 4977deletion, we perform a primary PCR step of 25 cycles using primers MT1A and MT3 (sequence: G C G A T G A G A G TAATAGATAGGGCTCAGGCG, nucleotides 13,580-13,551), 50 bp upstream of MT2. One microliter of this primary reaction is then used as template for the secondary reaction using primers MT1A and MT2 for another 25 cycles. The PCR conditions for both primary and secondary amplifications are as described above. It is vital that stringent contamination controls be implemented in this procedure, as it is extremely sensitive and in principle can detect single molecules.

Semiquantitative Comparisons of Mitochondrial D N A

4977 Levels

Any differences in the intensity of the deletion product from PCR of different samples may be due to varying amounts of total mtDNA added to the reaction. Therefore, for a valid comparison of mtDNA 4977levels to be made between different samples, the samples first have to be normalized for total mtDNA content. 2-4 This is done by using another set of primers (MT1C and MT2, Fig. 1) to amplify a region of wild-type mtDNA of similar size (324 bp) to the deleted product (303 bp). Primer MT1C has the sequence A G G C G C T A T C A C C A C T C T T G T T C G and corresponds to nucleotides 13,176-13,198 on the Anderson mtDNA sequence. 5 The PCR conditions are identical to those for the deletion-specific reaction except that, typically, deletion-specific PCR are run for 30 cycles while control PCR for wildtype mtDNA are run for 15 cycles; the difference in cycle numbers reflect the rare occurrence of mtDNA 4977molecules. Also, only 1 U of Taq polymerase is required for each control reaction. The signals of the different samples from this control reaction are visually compared on ethidium bromide-stained 2% agarose gels. Samples that give more intense signals are diluted and the control reaction performed again. Normalization is achieved

[36]

DETECTION AND QUANTIFICATION OF m t D N A

DELETIONS

425

through this iterative process of adjusting the DNA concentration of different samples such that they produce roughly equal intensities of wild-type PCR signals. The PCR is then performed on the normalized samples using the deletion-specific primers (MTIA and MT2). A semiquantitative comparison of the relative mtDNA 4977 levels between different samples can then be made by a visual comparison of the deletion product signals. It is important not to run the reactions into saturation by starting with too much template or running too many cycles. A quick way to check is to run two successive dilutions of the same sample. The product signals obtained should correspond to the dilutions. If different dilutions of the same sample give similar intensities, then the reaction is saturated and higher dilutions of the sample ought to be made. In general, 2-fold dilutions should give detectable differences in signal intensities of the PCR products.

Estimating Proportion of Deleted Molecules by Limiting Dilutions An estimate of the proportion of mtDNA 4977 relative to wild-type mtDNA can be obtained by using the method of limiting dilutions. Control and deletion PCR are performed separately on serial dilutions of a DNA sample using the same number of cycles (we use 30 cycles). Comparison of the different dilutions at which the control and deletion signals become undetectable gives an idea of the proportion of mtDNA 4977 present. For example, if the deletion signal disappears at 10 -2 dilution, and the control signal disappears at 10 -6 dilution, then the proportion of deleted to wildtype mtDNA molecules would be roughly 1 in 10,000. Using this procedure on adult heart DNA, 2 it was found that the limiting dilution for the detection of wild-type mtDNA was 1000 times less concentrated than the limiting dilution for the detection of mtDNA 4977. Thus, there appears to be 1 mtDNA 4977 molecule for every 1000 wild-type mtDNA molecules in this tissue. It is assumed that the efficiencies of the deletion and control PCRs were similar as the sizes of the amplified regions were similar (324 bp for control product versus 303 bp for the deletion product).

Quantitative Analyses of Deleted Mitochondrial DNA Levels

Principle The proportion of deleted mtDNA to wild-type mtDNA can be more precisely determined by comparing, quantitatively, the signals of deleted and control PCR products amplified from a sample to external standard curves. 4 Thus, final amount of template present in the sample can be correlated to the initial amount of starting template present. Two standard curves

426

MITOCHONDRIAL DISEASES AND AGING

[36]

are used, one for deletion PCR and the other for control PCR. For each sample, deletion and control PCR are performed separately and the products quantitated relative to the respective standard curve. The use of radioactively end-labeled primers allows for sensitive quantitation of PCR products. The standard curves are prepared by amplification of a known dilution series of the corresponding template under the same conditions as the sample. It is imperative that quantitation be performed in the exponential phase of both amplifications, as this is the phase when both deletion and control products will be accumulating with similar kinetics. The accumulation of products in the exponential phase can be modeled by Eq. (1): N=n(1

+ E) k

(1)

where N is the number of molecules at the end of the reaction, n is the number of template molecules, k is the number of cycles, and E is the average efficiency. The value of E has a theoretical maximum of 1 where the amount of product doubles with each cycle. This value declines during late cycles of amplification as the reaction proceeds past the exponential phase. There is a linear relationship between the logarithm of the amount of starting template n, and the logarithm of the final amount of the specific, amplified product N [Eq. (2)]. log N = log n + k log(1 + E)

(2)

A plot of log N against log n for a fixed number of cycles should therefore generate a line with a slope of 1. For this to be true, E has to remain fairly constant over k numbers of cycles, although E does not necessarily have to have a value of 1. If the value of E is decreasing substantially with each cycle as the reaction progresses past the exponential phase, a plot will be generated with a slope significantly less than 1. Linearity is not a sufficient condition for exponential amplification. A more stringent indicator is the slope of the plot, which should be close to 1. The standard curves are used to define the exponential range over which quantitation will be valid by ensuring that the above criterion is met. The various steps in the procedure are now described in detail for the quantitation of the mtDNm 4977 deletion. The procedure should be adaptable for the quantitation of other mtDNA deletions.

End Labeling of Primers, PCR, and Quantitation of Products Primers are end-labeled with [T-32p]ATP using T4 polynucleotide kinase. Unincorporated nucleotides are removed by spinning through P4 columns. End-labeled primer lots are prepared to give an approximately 10X concentration for PCR (5 tzM) and are diluted directly into the PCR

[36]

DETECTION AND QUANTIFICATION OF m t D N A

DELETIONS

427

mix. The PCR are performed with both primers end-labeled using conditions described earlier. Primers MTIC and MT2 are used for the control reaction, and primers MT1A and MT2 are used for the deletion reaction. The PCR can be performed in 96-well microtiter plates to accommodate many samples at the same time. After PCR, 10% (5/zl) of each reaction is electrophoresed through 8% polyacrylamide gels. The gel is dried, and the counts from each specific product band are quantified with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) after 15-24 hr of exposure. In our hands, using end-labeled primers, better sensitivity and lower noise were achieved than with direct incorporation of [a-32p]dATPduring PCR.

Preparation of External Standards For the construction of standard curves for deletion and control PCR, the respective PCR products are purified as a source of templates for the amplification reactions. We use genomic DNA from aged heart tissues as the initial template for these preparative PCRs. Each of the deletion and control reactions are performed and the products electrophoresed on 8% w/v polyacrylamide gels separately. The product bands are excised, electroeluted, and concentrated by centrifuging through Centricon-10 (Amicon, Denvers, MA). Stock solutions of both purified deleted and control products are prepared in this way and stored at -20 °. As a precaution against degradation, contamination, and loss of standards due to adherence to vessel walls during prolonged storage, the samples are diluted in solutions of Escherichia coli tRNA as carrier. These dilutions are then distributed into smaller volumes and stored at - 2 0 °. Individual aliquots are used when necessary, minimizing handling of the rest of the standard stock.

Definition of Exponential Range The use of external standards requires that quantification be performed during the exponential phase of the reaction. The use of radiolabeled primers allows sensitive quantification of products before the exponential range has been exceeded. From our experience, when the products become clearly visible by ethidium bromide staining, the PCR is at or on the verge of surpassing the exponential range. The range of dilutions for both deletion and control external standards over which amplification was exponential has to be determined. Operationally, this is determined as the range over which a plot of log counts versus log dilutions give a good linear fit with a slope close to 1. The incorporated counts provide a measure of the amount of accumulated product, whereas the dilution of the standard stock is proportional to the amount of starting

428

MITOCHONDRIAL DISEASES AND AGING

[361

template. Thus, this plot preserves the relationship in Eq. (2). The ranges are optimized for 15 cycles of control amplification and 30 cycles of deletion amplification. These cycle numbers are found from earlier experiments to give good, quantifiable signals for most genomic DNA samples for mtDNA

4977.

First, 10-fold serial dilutions of the external standard stocks are used as templates over a 10 log range. This allows the different phases of the amplification to be clearly distinguishable (Fig. 2 shows the plot for the deletion standards). The next step consists of amplifying 2-fold serial dilutions encompassing the suggested exponential phase. The range of dilutions that gives a slope closest to 1 (Fig. 3 shows the exponential range of deletion standards) can thus be determined with better resolution. We choose to work with known dilutions of our standards rather than quantifying the absolute amount of D N A as this may be subject to systematic errors if the initial quantification of the stock is inaccurate. For our system the exponential dilution "window" for deletion amplification using 30 PCR cycles was determined as ranging from 10 -6 down to 1/256 × 10 -6, all dilutions being expressed relative to the initial standard deletion stock. The window for control amplification using 15 PCR cycles ranged from 10 -3 down to 1/128 x 10 -3 relative to the standard control stock. The 2-fold difference in the ranges (256-fold for deletion PCR versus 128-fold for control PCR) reflects small disparities in background noise and amplification efficiencies. By spectrophotometric measurements, we estimate that this corresponds to a maximum of 6.2 × 10 7 input molecules

[]

m

m

Im

i



i

m

8 E 0 0

7 []

I

m 6



-12

i

-10

-

i



i



-8 -6 -4 -2 log DNA dilution

FIG. 2. Amplification of 10-fold serial dilutions of the deletion standard stock over a 10 log range. T h e reaction is exponential between dilutions of 10 -6 and 10 -l° and enters a plateau phase above a dilution of 10 -5. Twofold dilutions are m a d e within this suggested exponential range and amplified as in Fig. 3.

[36]

DETECTION AND QUANTIFICATION OF m t D N A

9~

DELETIONS

429

y=13.572 + 0.96755x R^2 = 0.993

8 0 o

7 6 5 -9

-8

-7

-6

-5

Log DNA dilution

FIG. 3. Amplificationof 2-fold serial dilutions of the deletion standard stock starting from 10-6. A linear regression analysis is performed and shows that the amplificationis exponential in this range with a slope close to 1 (0.97).

for control amplification and a maximum of 104 input molecules for deletion amplification. These ranges are reproducible. Thus, theoretically, we can quantitate differences in mtDNA4977/wild-type m t D N A ratios over a 3.3 × 104 (128 × 256) range.

Quantitation of Samples Preliminary deletion and control PCR with unlabeled primers are performed on dilutions of D N A samples whose levels of m t D N g 4977 are to be quantitated. The product signals on ethidium bromide-stained gels are then visually compared with those generated by amplification of the most concentrated standard dilution in the exponential window (10 -3 for control, 10 -6 for deletion). The samples are then diluted if necessary such that the amplification signals do not exceed that of the most concentrated standard dilution. Two different dilutions of each sample D N A that conform to this criterion are then used for quantitation to ensure that amplification remains in the exponential phase. Performing these preliminary steps eliminates handling of radioactivity until the actual quantification is to be done. Aliquots of these sample D N A dilutions are then amplified with 32p end-labeled primers in both control and deletion PCR. Both series of standards are amplified in parallel with the samples. The products are run in an 8% polyacrylamide gel, and the specific deletion or control product bands are quantified using a phosphor imager. The signal generated by each sample is then extrapolated from the appropriate standard curve to obtain the equivalent dilution of the standard stock that would have given the same signal. The ratio of the equivalent dilution for the deletion standards to that for the control standards can then be used to derive the

430

MITOCHONDRIAL DISEASES AND AGING

[36]

ratio of mtDNA4977/wild-type mtDNA. The calculation is illustrated in the example below. All standard curves are inspected for conformity to the conditions that the slope should be close to 1 and for good linear regression fit. This check should warn of any significant interreaction variation in amplification efficiencies. The two different dilutions of each sample should also give similar ratios. Samples for which they are significantly different are reanalyzed. Each experiment typically consists of 30 assay reactions for tissue samples (15 samples, 2 dilutions each) and 10 serial dilutions for the external standards. This is done for the control and deletion amplifications.

Relative Calibration of Standard Curves To convert the ratio of equivalent dilutions to the ratio of mtDNA4977/ wild-type mtDNA, the relative concentration of the control standard to the deletion standard must be known exactly. This is determined by the following experiment. Eight 2-fold dilutions of both control and deletion standards starting from a 10 -3 dilution of each stock are amplified separately using their respective primers for 15 cycles. To avoid differences in specific activities, only the common primer, MT2, is end-labeled. The same lot of labeled MT2 is used for both PCRs. Log plots of both PCR series are made to ensure that the amplifications remain in exponential phase using the criteria described above. The ratio of counts of the control PCR product to that of the deletion PCR product are calculated for each dilution in the range over which both PCR are exponential. For our system, the average of the eight ratios worked out to 5.4. This means that for an equivalent dilution Standard Control PCR Curve

Standard Deletion PCR (~urv~ 9

y= 13.6o3+0.9497~

Log coun~

9

y = 10.384+0.98632xRA2= 0.993

8 7

Del ~PCR

7

6on PCR

I

-9

-8

6

6

5

5

................

-7 -6 log DNA dil

-5

Brain DNA

-6

•...~ .....,.......~ -5

-4 -3 log DNA dil

-2

FIG. 4. Quantification of the deletion level in a brain D N A sample using the constructed deletion and control external standard curves. See text for further explanation.

[36]

DETECTION AND QUANTIFICATION OF m t D N A

DELETIONS

431

of control and deletion standards, the control standard would be 5.4 times more concentrated. This value is factored into the ratio of equivalent dilutions for each sample to obtain the ratio of mtDNAa977/wild-type mtDNA. This calibration exercise need only be performed once and thereafter periodically checked to ensure that degradation of the standards has not occurred. We performed the identical experiment nearly 1 year after our first calibration and obtained a similar value of 5.2. Example: Calculation of Ratio of mtDNA 4977 to Total Mitochondrial D NA Two standard calibration curves that relate the dilution of the standard template to the amount of incorporated radioactivity are shown in Fig. 4. The standard curve specific for mtDNA 4977 uses known dilutions of a 303bp template (deletion), and similarly the calibration curve for undeleted mtDNA uses known dilutions of a 324-bp template (control). For each standard curve, the linear regression and correlation coefficient (r 2) are included. In the example shown in Fig. 4, the deletion (Del) and control (Con) PCR signals for a brain DNA sample are extrapolated from the respective standard curves to obtain the dilution of the standard templates which would have given an equivalent signal. From the deletion curve, a value of -6.75 for the Del PCR is obtained from the x axis. The antilogarithm of this value is taken to give an equivalent dilution of 1.76 × 10-7. Similarly, for the control curve, a value of -3.93 for the Con PCR is obtained, which gives an equivalent dilution of 1.18 × 10-4. Dividing the Del dilution by the Con dilution gives 0.149%. At identical dilutions, the control standard is 5.4 times more concentrated in amplifiable templates than the deletion standard (see section on relative calibration of standard curves). This value is therefore factored in to give a final ratio of 0.027% (0.149/5.4). Conclusion We describe PCR techniques to measure the proportion of mtDNA 4977 to wild-type mtDNA. The quantitative method has been used to measure the levels of the deletion in different regions of the brains of aged individuals. 4 The method could be used for the quantitation of other rare mtDNA deletions in brain or other tissues of humans or other species. The parameters for PCR need to be individually optimized and exponential ranges of amplification determined for each PCR target. Once these conditions are determined, the technique is quite reproducible.