to BioTechniques free email alert service to receive content updates.
DNA and General PCR Methods: Restriction Digests
Kristie Nybo, Ph.D.
BioTechniques, Vol. 47, No. 5, November 2009, pp. 915–917
Full Text (PDF)

This month's question from the Molecular Biology Forums (online at molecularbiology. comes from the “DNA and General PCR Methods” section. Entries have been edited for concision and clarity. Mentions of specific products and manufacturers have been retained from the original postings, but do not represent endorsements by, or the opinions of, BioTechniques.

Molecular Biology Techniques Q&A

What could cause unexpected band sizes following restriction digestion? (Thread 22839)

Q I would like to do restriction fragment length polymorphism (RFLP) analysis on a segment of the fish cytochrome b gene from mtDNA. I used MboII to cut the PCR products amplified from this segment. The enzyme cut the sequence, but it looks like it cut at the wrong site. According to New England BioLabs (NEB) (NEBcutter;, the digest should have resulted in 155-, 88-, and 91-bp fragments. Sequencing verified that I got the correct 334-bp PCR product, but when I ran the gel after digesting with MboII and compared with a 100-bp marker, the bands were not the expected sizes.

Can anyone advise me about what went wrong and how I can correct it?

A Common problems with restriction digests include methylation sensitivity of the enzyme, use of an unmatched buffer that leads to star activity, or too much enzyme in too little reaction volume, which increases the amount of glycerol and causes star activity.

A MboII is blocked by overlapping Dam methylation. If you try to cut DNA that was isolated from bacteria encoding the dam gene, the DNA will be methylated and the MboII enzyme probably won't be able to recognize its cut site. This will depend on the proximity of the MboII recognition site to the Dam methylase recognition site. If the methylation occurs within the MboII recognition site, the enzyme won't be able to cut the DNA. Not all MboII sites will necessarily be methylated, so that might explain the aberrant restriction pattern.

A After PCR, the products involving the original parental strands will likely be hemimethylated, but the amount of that product in your overall reaction mixture will be minimal. You could use DpnI on your PCR product to get rid of double-helixes involving the original template.

A Another possibility is that the enzyme remains associated with your product and causes a migration shift in the band. An article on the Fermentas web site ( says: “MboII may remain associated with the cleaved DNA. This may cause DNA band shifting during electrophoresis. To avoid an atypical DNA band pattern, use the 6X Loading Dye & SDS Solution (#R1151) for sample preparation or heat the digested DNA in the presence of SDS prior to electrophoresis.”

A Is it possible that multiple copies of the gene are present within the fish genome? If so, a mutation may have eliminated the restriction site in one copy but not the other.

A Is there any chance that you have additional DNA flanking the gene, making the PCR product larger? If you had incomplete digestion or started with a larger product than expected, that might explain the results.

A Did you remember to include the primer sites in the fragment you used to search NEBcutter?

A You should try to purify each band on your gel and sequence with forward and reverse primers from the ends.

A There are a couple of things to remember if you try sequencing. The first is that sequencing will only read from the location of sequencing primer binding. The second is that MboII doesn't necessarily add weight to the PCR product, but might change its shape or cause charge shielding that could affect migration.

Q Is there a way to disassociate the endonuclease from the DNA prior to the gel run? If I heat the products to 65°C at the end of the digestion period to inactivate the enzyme, will that solve the problem? Can the inactivated MboII still associate with the product?

A Incubating at 65°C for 10 minutes will heat denature most (but not all) proteins and inactivate most enzymes. The enzymes will not disappear, but will lose their function. An inactive enzyme might still associate with the cut DNA.

A The NEB catalog says that heating MboII at 65°C for 20 minutes will heat inactivate the enzyme. Heating, especially with SDS, could remove the enzyme from the DNA. NEB also recommends not digesting with MboII for longer than 1 hour. This suggests that MboII tends toward nonspecific activity.

I noticed that MboII has a non-palindromic recognition site. I once had problems with FokI (which also recognizes a non-palindromic site) degrading my DNA. When I called, the technical support representative told me that the enzyme did not have much turnover, so I cut with much less enzyme for a shorter amount of time and that solved my degradation problem.

A The orientation of the DNA strand is essential to a restriction enzyme's recognition specificity. All computer programs into which you put a DNA sequence assume that you are giving a 5′-to-3′ strand orientation. Since your enzyme recognizes a non-palindromic recognition site, it won't cut the same sequence in the complementary strand. For example, the recognition sequence of MboII is 5′-GAAGA-3′. It will not cut 3′-GAAGA-5′, but it will cut 5′-TCTTC-3′. You cannot take the complement of a sequence without also reversing it to reflect the two complementary DNA strands in opposite orientation. What this means is that you need to be certain of the strand sequence you use for NEBcutter analysis.

A If you used mitochondrial genomic DNA to create your cDNA, then both strands will be present as templates.

A Considering all of the information mentioned in this discussion thread, it looks like the fragment lengths for your sequence following digestion with MboII will be 160 bp, 86 bp, and 88 bp. If you disassociate the MboII from the fragment, you will see two bands on an agarose gel, one at 160 bp and the other at about 88 bp, which will include both smaller products.