Change of DNA concentration due to restriction digest?

Question

Assume that you perform a restriction digest in a molecular biology lab: you combine genomic DNA, a restriction endonuclease (e.g., EcoRI), and the optimal buffer for that endonuclease and are about incubate the reaction at the optimal temperature for the endonuclease (e.g., 37 C).

Before you incubate the reaction, you measure the reaction's DNA concentration via a spectrophotometer. Then, you incubate the reaction for one hour and, afterwards, measure the DNA concentration again via a spectrophotometer again.

Q: Has the DNA concentration of the reaction changed through the restriction digest?

Most people would argue that the overall DNA concentration does not change due to the restriction digest, whereas the fragment distribution does, with the average fragment length decreasing through the digest. However, every time I have run this experiment with my students, we have observed a slight increase in the DNA concentration after the restriction digest. What could be the reason for this observation?

Answer 1

Summary

The poster wishes to know why spectrophotometric readings (presumably absorbance at 260 nm) on a sample of genomic DNA from an unspecified organism appear “slightly” greater after incubation with a restriction endonuclease, EcoRI. Assuming trivial explanations can be excluded — the effect is not statistically significant or results from components of the added enzyme preparation — it is evident that, under these conditions, absorbance measurement is not a measure of the concentration of DNA. The change would reflect an effect of the change of structure of the DNA on the absorbance, either as an intrinsic difference between intact and linearized species, or in their ability of the two strands to separate (melt) at the temperature and ionic conditions employed. Suggestions are made for cheap and simple experiments to clarify the situation.

More Detailed Consideration

Let me consider some different possibilities in turn.

1. The readings do not actually show a statistically significant difference. The poster cites neither the magnitude of the effect nor the results of statistical tests done to show that the increase is significant. Without the latter we cannot be sure that there is any question to answer, and without the former we cannot judge whether a particular explanation could account for the observed change.

2. An increase in absorbance results from something in the preparation of restriction enzyme added. The poster argues against this on the basis of purification following the reaction, but more direct experiments are both possible and preferable. Comparison could be made between the reaction mixture at the end of the incubation and at the start, or after incubation with inactivated enzyme, and other controls are possible. A time course of the values of absorbance throughout the experiment would be useful to see whether the increase in absorbance correlated with the cleavage, as judged by the electrophoretic behaviour of the products.

3. DNA is being created during the experiment. As the law of conservation of matter has been shown to hold at the molecular level, I reject this possibility. The obvious conclusion is then that ultraviolet absorbance is not a valid measure the amount of DNA, especially when comparing DNA of different structural forms. This point is developed further in 4, below. The question can be resolved by using an alternative method to measure the concentration of DNA — preferably a chemical method that is not affected by the physical form of the DNA†.

“When you have eliminated the impossible, whatever remains, however improbable, must be the truth.”
[Sir Arthur Conan Doyle — Sherlock Holmes]

4. The change reflects a change in the ultraviolet absorbance properties of different structural forms of DNA. It has long been known that on heating double-stranded DNA so that the strands separate (melt) the absorbance at 260nm increases: the hyperchromic effect. This is thought to reflect the greater energy required for the π→π* transitions in the aromatic bases of double-stranded DNA, where the base stacking involves π–π interactions. It seems to me most likely that the changes observed reflect a change in the spectral behaviour of the bases in the DNA after the structure has been changed by cleavage into small fragments.
   This could conceivably be an intrinsic difference between the linear cleavage products and the original form of the DNA (supercoils etc.), unfortunately not specified in the question. I am not aware of reports of such differences, but this does not mean that they do not exist. Alternatively ‘melting’ of DNA could be occurring at the temperature and ionic strength used in the reaction, and the supercoiled or higher-order structure of the undigested DNA might be hindering strand separation and suppressing the hyperchromic effect.
   In relation to the latter possibility, experiments are required with linear DNA (in the absence of enzyme) to see if there is indeed a hyperchromic effect on heating to 37C at the ionic conditions used. I would suggest three other main lines of approach to clarifying the situation. The first would be to employ different types of “genomic” DNA — animal, bacterial, viral, plasmid — in the digestion. The second would be to do a clean structural comparison without restriction enzymes. Plasmid DNA can be separated by gel electrophoresis and the supercoiled and linear form extracted. The spectrophotometric and chemical ‘concentration of DNA’ could be compared in the two cases, as could be their maximum percentage increase on melting at increasing temperatures. The third would be to prepare linear DNA of different sizes, using either different restriction endonucleases or non-enzymatic physical shearing of the DNA. Indeed, experiments with sheared DNA would exclude the possibility of effects from the endonuclease preparations and allow experiments at temperatures and ionic strengths not possible with the latter.

Of course this is not Sherlock Holmes’ “all that remains”, as I am mindful of Shakespeare’s:

“There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy.”
[Hamlet]

Concluding suggestion

I would respectfully suggest that questions of this sort are framed in an explicit manner that clearly distinguishes between what was done, what was observed, and what conclusions were drawn.

––––––––––––––––

† This is something of a problem as the only chemical method I am aware of is the diphenylamine method, with the advantage that the reaction is with the deoxyribose ring. However this is several orders of magnitude less sensitive than spectrophotometric methods. The fluorometric method suggested in another answer depends on the interaction of fluorophore with the bases, and its access could well be different for single-stranded and double-stranded DNA.

Answer 2

As you did a clean up, then you lost some DNA in this process. There are always losses from clean up steps.

Note also that spectrophotometry is notoriously unreliable for DNA quantitation; it's more of a "there's roughly this amount of DNA there" sort of reading, especially as they are particularly prone to organic compound contamination (e.g. guanidine, phenol from your extraction) throwing your reading off. Interestingly, the original use for 280 nm measurement for DNA was for determining DNA contamination in protein samples. Have a look in Sambrook's Molecular Cloning: a laboratory manual for further detail ¹.

If you want a more reliable reading, then you should use something like a fluorophore that is known to bind at a certain ratio to the DNA, as these are less prone to contamination problems and are more sensitive².

Refs:

1. Sambrook, Joseph & Russell, David W. (David William), 1957- & Cold Spring Harbor Laboratory. (2001). Molecular cloning : a laboratory manual / Joseph Sambrook, David W. Russell. Cold Spring Harbor, N.Y. : Cold Spring Harbor Laboratory

2. Green MR, Sambrook J. Estimating the Concentration of DNA by Fluorometry Using Hoechst 33258. Cold Spring Harb Protoc. 2017 May 1;2017(5). doi: 10.1101/pdb.prot093567. PMID: 28461657.