If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

Main content

Restriction enzymes & DNA ligase

Restriction digestion. Sticky ends and blunt ends. Ligation reactions.

Key points:

  • Restriction enzymes are DNA-cutting enzymes. Each enzyme recognizes one or a few target sequences and cuts DNA at or near those sequences.
  • Many restriction enzymes make staggered cuts, producing ends with single-stranded DNA overhangs. However, some produce blunt ends.
  • DNA ligase is a DNA-joining enzyme. If two pieces of DNA have matching ends, ligase can link them to form a single, unbroken molecule of DNA.
  • In DNA cloning, restriction enzymes and DNA ligase are used to insert genes and other pieces of DNA into plasmids.

How do you cut and paste DNA?

In DNA cloning, researchers make many copies of a piece of DNA, such as a gene. In many cases, cloning involves inserting the gene into a piece of circular DNA called a plasmid, which can be copied in bacteria.
How can pieces of DNA from different sources (such as a human gene and a bacterial plasmid) be joined together to make a single DNA molecule? One common method is based on restriction enzymes and DNA ligase.
  • A restriction enzyme is a DNA-cutting enzyme that recognizes specific sites in DNA. Many restriction enzymes make staggered cuts at or near their recognition sites, producing ends with a single-stranded overhang.
  • If two DNA molecules have matching ends, they can be joined by the enzyme DNA ligase. DNA ligase seals the gap between the molecules, forming a single piece of DNA.
Restriction enzymes and DNA ligase are often used to insert genes and other pieces of DNA into plasmids during DNA cloning.

Restriction enzymes

Restriction enzymes are found in bacteria (and other prokaryotes). They recognize and bind to specific sequences of DNA, called restriction sites. Each restriction enzyme recognizes just one or a few restriction sites. When it finds its target sequence, a restriction enzyme will make a double-stranded cut in the DNA molecule. Typically, the cut is at or near the restriction site and occurs in a tidy, predictable pattern.
As an example of how a restriction enzyme recognizes and cuts at a DNA sequence, let's consider EcoRI, a common restriction enzyme used in labs. EcoRI cuts at the following site:
When EcoRI recognizes and cuts this site, it always does so in a very specific pattern that produces ends with single-stranded DNA “overhangs”:
If another piece of DNA has matching overhangs (for instance, because it has also been cut by EcoRI), the overhangs can stick together by complementary base pairing. For this reason, enzymes that leave single-stranded overhangs are said to produce sticky ends. Sticky ends are helpful in cloning because they hold two pieces of DNA together so they can be linked by DNA ligase.
Not all restriction enzymes produce sticky ends. Some are “blunt cutters,” which cut straight down the middle of a target sequence and leave no overhang. The restriction enzyme SmaI is an example of a blunt cutter:
Blunt-ended fragments can be joined to each other by DNA ligase. However, blunt-ended fragments are harder to ligate together (the ligation reaction is less efficient and more likely to fail) because there are no single-stranded overhangs to hold the DNA molecules in position.

DNA ligase

If you’ve learned about DNA replication, you may already have met DNA ligase. In DNA replication, ligase’s job is to join together fragments of newly synthesized DNA to form a seamless strand. The ligases used in DNA cloning do basically the same thing. If two pieces of DNA have matching ends, DNA ligase can join them together to make an unbroken molecule.
How does DNA ligase do this? Using ATP as an energy source, ligase catalyzes a reaction in which the phosphate group sticking off the 5’ end of one DNA strand is linked to the hydroxyl group sticking off the 3’ end of the other. This reaction produces an intact sugar-phosphate backbone.

Example: Building a recombinant plasmid

Let's see how restriction digestion and ligation can be used to insert a gene into a plasmid. Suppose we have a target gene, flanked with EcoRI recognition sites, and a plasmid, containing a single EcoRI site:
Our goal is to use the enzyme EcoRI to insert the gene into the plasmid. First, we separately digest (cut) the gene fragment and the plasmid with EcoRI. This step produces fragments with sticky ends:
Next, we take the gene fragment and the linearized (opened-up) plasmid and combine them along with DNA ligase. The sticky ends of the two fragments stick together by complementary base pairing:
Once they are joined by ligase, the fragments become a single piece of unbroken DNA. The target gene has now been inserted into the plasmid, making a recombinant plasmid.

Restriction digests and ligations involve many molecules of DNA

In the example above, we saw one outcome of a ligation between a gene and plasmid cut with EcoRI. However, other outcomes could happen in this exact same ligation. For instance, the cut plasmid could recircularize (close back up) without taking in the gene. Similarly, the gene could go into the plasmid, but flipped backwards (since its two EcoRI sticky ends are identical).
Restriction digests and ligations like this one are performed using many copies of plasmid and gene DNA. In fact, billions of molecules of DNA are used in a single ligation! These molecules are all bumping into one another, and into DNA ligase, at random in different ways. So, if multiple products can be made, all of them will be made at some frequency – including ones we don't want.
How can we avoid the "bad" plasmids? When we transform bacteria with DNA from a ligation, each one takes up a different piece of DNA. We can check the bacteria after transformation and use only the ones with the correct plasmid. In many cases, plasmid from transformed bacteria is analyzed using another restriction digest to see if it contains the right insert in the right orientation.

Explore outside of Khan Academy

Do you want to learn more about restriction enzymes? Check out this scrollable interactive and this simulation from LabXchange.
Do you want to learn more about DNA ligase? Check out this scrollable interactive and this simulation from LabXchange.
LabXchange is a free online science education platform created at Harvard’s Faculty of Arts and Sciences and supported by the Amgen Foundation.

Want to join the conversation?

  • starky ultimate style avatar for user alina
    Why do restrictive enzymes that do blunt cuts even exist if they are so inefficient?
    (28 votes)
    Default Khan Academy avatar avatar for user
    • blobby green style avatar for user Dr Kalpesh
      Restriction enzymes are found in bacteria and they have some biological role (explained below), but we are exploiting it in our way to use in experiment.

      Biological role of restriction enzymes in bacteria: when restriction enzyme is present in a given bacterium, such bacterium can prevent (restrict) the growth of certain bacterial viruses (bacteriophages) and this is the reason also to call it as RESTRICTION enzymes. in this way it is the defensive enzyme that protects the host bacterial DNA from the DNA genome of foreign organism (bacteriophage) by specifically inactivating the invading bacteriophage DNA by digestion

      Explanation: if particular bacterium has restriction enzyme, it must have companion site specific DNA methylase which methylates DNA of host bacterium in site specific manner and methylated DNA is not the substrate for restriction enzyme. so host bacterium DNA is not cut by restriction but when new DNA is inserted by bacteriophage, it is not methylated and so it chopped by restriction enzyme and bacteria can survive (i.e. bacteria's innate immunity !)
      Hope everything is clear...
      (19 votes)
  • blobby green style avatar for user astephenson1
    How long does the process of cutting DNA take?
    (10 votes)
    Default Khan Academy avatar avatar for user
    • blobby green style avatar for user 😊
      It depends on the enzyme and the lab that produces them, but the rule of thumb for digestions is 1 hour at the appropriate temperature: For example, SmaI works at 25C, while EcoRI works at 37C.

      If you want to know more then know that enzymes are sold in certain 'sizes', which are the units present on the vial. A unit, according to NEB is:

      "One unit is defined as the amount of enzyme required to digest 1 µg of λ DNA (HindIII digest) in 1 hour at 25°C in a total reaction volume of 50 µl."

      Hope this helps
      (12 votes)
  • starky tree style avatar for user loganbro45
    what would happen if the gap never closes?
    (8 votes)
    Default Khan Academy avatar avatar for user
    • leafers tree style avatar for user emilyabrash
      Although the other answer is funnier, what would actually happen if the gap never closed during a ligation is that the DNA fragments would come apart again. The sticky ends will only hold them together briefly, and if ligase doesn't connect them during that time, they will go back to floating around and bumping into other pieces of DNA and enzymes in the reaction mix.
      (15 votes)
  • aqualine tree style avatar for user suncoats1
    I did not understand how to differentiate between plasmids in which the gene of interest has been correctly inserted and those in which it isn't.
    (8 votes)
    Default Khan Academy avatar avatar for user
  • blobby green style avatar for user Tania Pogue
    What happens to the restriction enzyme once the recombinant plasmid has been formed. Is it destroyed?
    (6 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      You must remove or destroy the restriction enzymes (REs) before you ligate. Otherwise the REs will just recut your newly ligated DNA.

      This is often done by purifying the cut DNA — usually by running the digest (cut DNA) on an agarose gel and then cutting out the band of interest.
      (9 votes)
  • leaf green style avatar for user SV
    How do scientists make sure that the bases of the plasmid are complementary to the bases of the inserted DNA?
    (7 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      The easy way is to use the same restriction enzyme(s).

      Sometimes this won't be possible§ — in these cases you could try to find enzymes that leave the same overhang (i.e. have compatible cohesive ends).

      There are other techniques that can be done if this isn't possible such as partially filling in the ends to create compatible ends or "blunting" where you fill in and/or chew of the overhangs and then do a blunt end ligation.

      Another very common alternative is to use primers with restriction sites at their 5' ends and then PCR amplify the insert you want — this creates a copy of the insert DNA with whatever restriction sites you want at the ends.

      §For example there might not be restriction sites for the same enzymes in the correct places in both the vector and insert.
      (4 votes)
  • leafers seedling style avatar for user Hafsa Abdinur
    I am quite confused as to the strand of the target gene, when we are cutting the gene are we pasting the whole gene into the plasmid or are we just pasting the part that the restriction enzyme has cut from the whole targeted gene?
    (3 votes)
    Default Khan Academy avatar avatar for user
  • leafers seedling style avatar for user Catcher Salazar
    What if there are not restriction enzymes on either side of the target DNA? Would it still be possible to use restriction enzymes?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      First, most vectors will have a region known as the "Multiple Cloning Site" (MCS) that can be cut with many different restriction enzymes† — this gives you more choices of enzyme and makes it more likely that you can find one that cuts near the ends of the region you wish to clone.

      Second, we often don't care if we clone a small amount of extra DNA , this means that we can search over a larger area than you might expect to find appropriate restriction enzymes.

      If the regions flanking the sequence you want to clone don't contain any useful restriction sites you can use primers with restriction sites added to their 5' ends and then amplify the sequence using PCR§.

      This amplifies the insert you want and creates a copy of the insert DNA with whatever restriction sites you want added at the ends.

      There are many more tricks that have been developed, but adding sites at the ends of primers almost always works, so that is a very good one to know!

      Does that help?

      †Note: There are hundreds of commercially available restriction enzymes recognizing many different sequences (many of which are palindromes, but not all).

      Among these the most commonly used are six-cutters (with 6 bp recognition sites — if you make a bunch of simplifying assumptions you can calculate that these enzymes on average will cut once every 4096 bp.

      §Note: Polymerase chain reaction — you can learn more about this technique here:
      (3 votes)
  • aqualine seed style avatar for user Methmi Peiris
    If you put a same restriction enzymes to two samples of the same person's DNA. The resulting DNA strands after the restriction enzymes cutting should be the same size, right?

    So, Can a forensic scientist use gel electrophoresis after this to determine if the DNA of the suspect matches the DNA found or to determine if the found DNA belongs to the criminal or the victim?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      That is true, but for a typical restriction digest of human DNA you will get around a million different bands with a range of different sizes§ — on a gel this just looks like a smear of DNA and is of no use in identifying individuals.

      In addition, two DNA molecules could be exactly the same size, but have different sequences — even if you isolated them (or detected them with a specific probe) you would probably not be able to distinguish them using any form of gel electrophoresis.

      Individual humans are about 99.9% identical at the nucleotide level, so telling us apart by DNA requires relatively sophisticated techniques!

      There are many different techniques to get information from this sea of DNA, if you want to know more about how this is currently done:

      § ~6.4 billion base pairs in a diploid human genome and a typically six cutter enzyme will on (a very rough) average cut every 4096 bp.
      (2 votes)
  • blobby green style avatar for user Ben Hitchcock
    Dystrophin is one of the longest genes, with 2.4 million base pairs. Does size have any impact on the size of the plasmid that needs to be used (does it have to be big enough to be able to cut a 2.4 million base pair section out of it?), does the plasmid simply expand to accomodate the gene? Isn't it quite likely that the gene itself would be cut up by the restriction enzymes? How could you assure that the gene would remain in tact and recircularize in the plasmid successfully with such a large gene?
    (2 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      A typical plasmid can accommodate inserts of any size up to total size of around 50 kb, but plasmids that are more than 20 kb are very difficult to work with and may require special transformation techniques. The efficiency of ligation and transformation tends to decrease with extremely large inserts.

      For large inserts there are different kinds of vectors (not plasmids) that can be used.

      For an enormous insert like you are asking about you would need to use a type of vector known as an artificial chromosome. These are specific to the type of organism in which you wish to grow the vector with insert. For example you could use a YAC (yeast artificial chromosome) for the dystrophin gene.

      To clone the entire dystrophin gene you would probably have to screen through a library of large fragments made by cloning randomly made fragments of genomic DNA (for example this can be done mechanically by passing the DNA solution through a needle).

      In practice you would probably get clones from an already made cDNA library (made by reverse transcribing mRNA) — this means you wouldn't need such a giant vector. There are actually many different transcripts made from the dystrophin gene that produce different versions of the protein (known as isoforms) — the transcripts range from about 5-14 kb, much more manageable!

      If you want to get an idea of the complexity of transcription from the dystrophin gene:
      (4 votes)