I've found this article, which is a very old one (from the time when nucleoside analogs where researched as a possible way to prevent replication of virus genetic material, before the HIV epidemics). In the last page, the authors suggest that cellular repair enzymes could remove azidothymidine (AZT, zidovudine) from cellular DNA if it's incorporated, but viruses' polymerases lack this ability. As it's an old article, I would like to know if this hypothesis was confirmed. As both azidothymidine and thymidine can form normal base pairings, how could human enzymes detect an azidothymidine was incorporated?

If someone is not familiar with AZT mechanism: basically, it's an analog of thymidine which has an azide group instead of 3'-OH


It can be phosphorylated by cell kinases and incorporated to the growing DNA chain; however, this will stop the polymerization because it depends on free 3'-OH. The natural question which follows is: why wouldn't AZT also block the replication of human cells' DNA?

  • 1
    $\begingroup$ it does and that is why the side effects $\endgroup$
    Commented Mar 1, 2015 at 19:37
  • 2
    $\begingroup$ tandfonline.com/doi/abs/10.1080/15257771003741018 It has been shown to poison mitochondrial thymidine kinase, amongst other side effects. $\endgroup$
    – March Ho
    Commented Mar 1, 2015 at 20:18
  • 1
    $\begingroup$ @WYSIWYG it has side effects, but they're not compared to nucleotide analogs that target cancer cells, for example cytosine arabinoside. $\endgroup$
    – El Cid
    Commented Mar 1, 2015 at 21:09
  • 1
    $\begingroup$ Now that I'm getting a closer look at that structure, the AZT looks wrong, shouldn't that 3' amine be an azide? $\endgroup$
    – user137
    Commented Mar 2, 2015 at 16:40
  • 1
    $\begingroup$ @User137 indeed, it's $N_{3}$ $\endgroup$
    – El Cid
    Commented Mar 2, 2015 at 20:48

1 Answer 1


Retroviruses depend on being able to convert their RNA genome into a DNA copy, and have a reverse transcriptase enzyme to do that. This unique activity is not found in human cells, allowing for potential antiviral therapy if a drug can be used to inhibit the reverse transcriptase while not affecting the human enzymes.

AZT is one such drug, by mimicking the nucleotide thymidine, it can it can act as a substrate for nucleotide kinases and become 5' phosphorylated. AZT is phosphorylated by host cell enzymes, specifically thymidine kinase, this means that the levels of AZT phosphates are similar between HIV infected and non-infected cells. Once the 5' triphosphate is formed, the drug is a substrate for the viral reverse transcriptase. If an AZT is incorporated into the DNA backbone, it terminates the DNA because the 3' OH is not available for further polymerization.

However, at high enough doses AZT can be toxic to healthy cells. One mechanism for toxicity is the depletion of thymidine, because AZT competes with thymidine for the kinase enzymes, but is phosphorylated at a slower rate. At high doses, >1mM, AZT can act as a substrate for human DNA polymerases. Thankfully, AZT is about 100 fold more selective for HIV reverse transcriptase than human DNA polymerase.

The mitochondrial DNA polymerase seems to be a little more sensitive to AZT than other human DNA polymerases. This can lead to toxicity in the heart, skeletal muscles, hepatocytes, or other cells that use a lot of mitochondria. This toxicity is generally reversible if doses are reduced.

For more information, read about other nucleotide analogs, such as acyclovir, ganciclovir, and others which work through similar mechanisms.

  • 1
    $\begingroup$ All you wrote is relevant, and thanks about the suggestion of the other nucleotide analogs. However, I still don't know the molecular mechanism that would explain that "AZT is about 100 fold more selective for HIV reverse transcriptase than human DNA polymerase". There are other nucleotide analogs (such as cytosine arabinoside) that actually impair human DNA replication, so they're used to target cancer cells. $\endgroup$
    – El Cid
    Commented Mar 1, 2015 at 22:30
  • $\begingroup$ Maybe (just a hypothesis) the higher rate in which the virus proliferate could make it more susceptible (by a similar mechanism for which the traditional chemoterapeutic drugs target cancer cells). However, just that doesn't appear to be sufficient, since AZT or acyclovir have fewer side effects than chemoterapeutic drugs... $\endgroup$
    – El Cid
    Commented Mar 1, 2015 at 22:34
  • 4
    $\begingroup$ @ElCid Even though the nucleotide analogs have the same general mechanism, they have different structures. This could explain why some bind to human enzymes more efficiently than others and account for the side effects. I'm not aware of any binding studies or crystalography to support this, but it probably exists. $\endgroup$
    – user137
    Commented Mar 1, 2015 at 23:09

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .