Technical Comments

Is CdtB a Nuclease or a Phosphatase?

Science  26 Jan 2001:
Vol. 291, Issue 5504, pp. 547a-547
DOI: 10.1126/science.291.5504.547a

Lara-Tejero and Galán (1) reported deoxyribonuclease (DNase) I–like activity of cytolethal distending toxin B (CdtB), which causes arrest in the G2/M transition phase of the cell cycle (2). Even though the authors noted that CdtB is a poor nuclease in vitro, they proposed that CdtB-inflicted DNA damage could trigger the checkpoint system that prevents cell entry into mitosis.

A relationship between CdtB and DNase I similar to that observed by Lara-Tejero and Galán has recently been documented (3,4). In those studies, however, both proteins were also identified as members of a broad metalloenzyme superfamily that includes, in addition to nucleases, various phosphatases of lipid second messengers such as sphingomyelin and inositol polyphosphate 5-phospate. In general, then, any protein of uncertain function belonging to this superfamily might be either a nuclease or a phosphatase. The exclusively nuclear localization of CdtB observed by Lara-Tejero and Galán argues against its phosphatase activity on lipid second messengers. A more likely group of CdtB targets are proteins that control tyrosine phosphorylation of the cell cycle regulator Cdc2 kinase, most notably the kinase Wee1 and the phosphatase Cdc25. These two proteins are regulated by distinct phosphatase/kinase networks, but they have one thing in common: dephosphorylation of either leads to increased Cdc2 phosphorylation (5), which is a hallmark of cells treated by CdtB (2). Clearly, an unexplored possibility remains—that CdtB could be a phosphatase for either Wee1 kinase or Cdc25 phosphatase, or for other proteins in the upstream part of their regulatory networks (2).

Finally, many unrelated proteins from pathogenic bacteria, most of which are of unknown function, have been identified as members of this nuclease/phosphatase superfamily (3). Elucidating molecular details of the CdtB-caused cell arrest could help generalize the mechanism or mechanisms of eukaryotic cell invasion by other pathogens as well.


  1. F1.
  2. F2.
  3. F3.
  4. F4.
  5. F5.

Response: Dlakic correctly points out that CdtB, like other nucleases, shares a common fold with other proteins, including some phosphatases of lipid second messangers (indeed, it is not uncommon that proteins of different function share a similar fold). He argues that CdtB could then be a phosphatase that may act on Wee1, Cdc25, or other cell cycle regulatory proteins, because its in vitro nuclease activity is poor. This possibility is extremely unlikely, however, and is not supported by the available evidence.

First, CdtB does show nuclease activity in vitro; the low level of this activity is consistent with the requirement of either a cellular cofactor or a particular topology of the DNA for its activity (1). The requirement of a cellular cofactor is a common theme in bacterial toxins (2, 3) and may be necessary to protect the bacteria from the potentially harmful effects of the nuclease. In addition, a Escherichia coli homolog of CdtB was also shown recently to have nuclease activity in vitro (4).

Moreover, microinjection of 0.5 mg/ml solution of purified CdtB into the cytoplasm of cultured cells resulted in catastrophic changes in the chromatin only ∼2 hours after microinjection (1). These changes are consistent with a nuclease activity and are difficult to explain in the context of a potential phosphatase activity of CdtB. Indeed, induction of cell cycle arrest and cytoplasm distention, instead of catastrophic chromatin changes, required microinjection of extremely low concentrations (0.001 mg/ml solution) of CdtB (1). This agrees with the observations that one double-stranded break is capable of triggering DNA damage checkpoints (5) and that CdtB-intoxicated cells exhibit identical features as those treated with a DNA-damaging agent (6).

Finally, it is unclear in the context of the current knowledge of cell cycle checkpoints how a lipid phosphatase could induce G2/M arrest, particularly because CdtB localizes to the nucleus. In light of all of these considerations, we believe that CdtB controls cell cycle progression as a nuclease.

Navigate This Article