Technical Comments

Response to Comment on "Uracil DNA Glycosylase Activity Is Dispensable for Immunoglobulin Class Switch"

Science  17 Dec 2004:
Vol. 306, Issue 5704, pp. 2042
DOI: 10.1126/science.1105225

The major argument presented by Stivers in (1) is that the residual uracil (U) removal activities of single mutants, but not double mutants, of uracil DNA glycosylase (UNG) are sufficient to introduce double-strand breaks (DSBs) and induce class switch recombination (CSR). First, there are no published data for the relative catalytic activities of the single versus double mutants of mouse UNG (mUNG). In our cleavage assay, both have negligible activities (2). Stivers' argument is based on speculation from studies on E. coli UNG, which differs from full-length mammalian UNG not only in size, but also in catalytic activity (3). It is not even clear whether UNG plays the same role in E. coli and mammals; accumulation of spontaneous mutations in UNG–/– mice is not strikingly enhanced (4), which suggests that UNG–/– cells remove U efficiently by alternative enzymes. Second, if the residual activity of a single mutant such as D145N— the human counterpart of which has roughly 0.05% of the wild-type activity (5)—is responsible for DSB and CSR, one would expect the wild-type UNG activity to be supersaturating in the experiments shown in figure 4 in (2). However, we found that titrations of UNG wild-type and D145N mutant proteins expressed in UNG–/– B cells showed almost parallel curves when plotted against CSR activity. A similar curve was obtained for the N204V mutant, which cannot bind U (5). These results strongly indicate that the CSR rescue activities of mUNG wild-type and single mutants are equivalent per given amount of protein. Third, although Msh2 (a mismatch-repair protein) and UNG are proposed to complement each other to remove U, Msh2–/– mice show reduced CSR efficiency (about one-quarter of wild type), which indicates that UNG is not saturating in B cells.

To explain the result that Ugi blocked CSR but not the introduction of DSBs, Stivers proposes that Ugi cannot completely block the enzymatic activity of mUNG for promoting DSBs, but can block complex formation with other proteins required for DNA repair. He also cites a study showing that Ugi has a relatively weaker affinity to human UNG than to E. coli UNG (6). However, Ugi has previously been shown to completely block human UNG activity (6, 7). We also confirmed that Ugi-expressing CH12 cell extracts can inhibit recombinant UNG in vitro, which indicates that Ugi is present in excess. Furthermore, if Stivers' view is correct, it would demand reevaluation of the original paper proposing the DNA deamination model, in which Neuberger et al. used Ugi to inhibit UNG in a DT40 cell line (8).

Stivers also dismisses the possibility that double mutants have lower affinity for another protein than single mutants have, because single E. coli UNG mutants do not show structural alterations. However, gross structural change is not required for loss of interaction with another protein. For example, thymine-DNA glycosylase (TDG) potentiates the transcription of estrogen-regulated genes through direct interaction with estrogen receptor α (ERα) (9). Single or double mutants within the glycosylase domain abolish TDG interaction with ERα. Moreover, the TDG-N140A mutant, which lacks glycosylase activity, stimulates ERα activity as well as wild-type TDG. Given this and other examples that mutations in a few residues can change one protein's affinity for another, it is possible that the UNG mutations effectively reduce the affinity of UNG to DNA or other repair enzymes that are critical for CSR (2, 10).

References and Notes

Related Content

Navigate This Article