6.6 Comparison of Chemical and Enzymatic DNA Sequencing Methods

In comparing and evaluating the chemical and enzymatic DNA sequencing methods, one must find answers to the following questions:

(1) how easy is it to obtain the DNA in the form suitable for sequencing?

(2) what are the number of DNAs necessary for sequencing and the specific activity of the labeled sample?

(3) how time-consuming is the method?

(4) how accurate is the method?

The chemical method is simpler when the DNA under analysis is not too large (200-500 b.p.). When it comes to sequencing a high-molecular weight DNA, preference should be given to the polymerase copying method in order to avoid the restriction procedure yielding individual fragments. In enzymatic sequencing of large single-stranded DNAs (e.g., bacteriophages), one can use a set of priming oligonucleotides whose synthesis today does not call for any significant time and labor expenditures. The ideal method for double-stranded high-molecular weight DNAs is blind enzymatic sequencing using a universal primer (currently offered by many suppliers) and a computer for data processing. The chemical method can also be used, but then one must cut out the DNA fragments of interest from the vector, which makes the entire procedure too complicated. The next criterion is the amount and specific activity of the DNA to be analyzed. One of the reasons for the great popularity of the chemical degradation method is that it allows DNA to be sequenced from any end after the incorporation of a labeled (³²P) terminal phosphate. The terminator is kept standard by using restriction endonucleases while preparing the sample for analysis. A drawback of this approach is the impossibility of introducing a highly active label (only one molecule of labeled phosphate per DNA molecule can be inserted). One of the attractions of the enzymatic method is the insertion of a multiple label into the strands being synthesized, which means that a much smaller amount of DNA is necessary for analysis, as compared to chemical sequencing. Simple calculations have shown that for sequencing a 200-membered polynucleotide and obtaining autoradiogram of the same clarity at the same exposure time, 2 picomoles of DNA are needed if the sequencing is done chemically and only 0.1 picomole in the enzymatic case. The chemical method also calls for gamma-labeled ATP with an activity of 3000 Ci/mole, whereas in the enzymatic method the precursor of the label may have a specific activity of 300 Ci/mole, which is more economical.

As regards the labor requirements, both methods seem to be comparable. From the purely experimental standpoint, chemical sequencing is more time-consuming (because of the repetitive alcohol precipitation, lyophilization and other procedures). In the case of sequencing by the polymerase copying method, on the other hand, the most time-consuming operations are gene engineering of the sample and preparation of the primer.

In terms of accuracy and effectiveness, the close similarity of both methods has been proved by numerous experiments. In both cases, the two characteristics are largely dependent on the structure of extended single-stranded DNA fragments that may form hairpins and have varying mobility during PAGE. Short fragments tend to "run away" from the sequencing gel, which is why one must always have the possibility to check the results of separation in polyacrylamide gel by a cross experiment - that is, sequencing the same portions in other fragments of the starting DNA or sequencing the complementary strand. The latter is usually a must in both methods. Good results in gel electrophoresis of high-molecular weight fragments are obtained at elevated temperatures because they render formation of double-stranded structures difficult. Significantly, good possibilities are being opened up at present for increasing the sequencing rate by either method. This is borne out by new developments in solid-phase chemical sequencing and advent of automatic sequencers based on the enzymatic method. Moreover, the role of computers in processing experimental data is steadily increasing. Today it is hard to make any predictions as regards sequencing rates. The mere fact that work aimed at complete sequencing of the human genome which contains nearly three billion nucleotides is already under way testifies to the spectacular strides being made in this area.