br Discussion The biochemical separation of exonuclease acti

Discussion The biochemical separation of exonuclease activity from DNA-PK dependent endonuclease activity reported in this manuscript is consistent with genetic separation of Artemis enzymatic activity. Mutations in the protein which result in disruption of endonuclease activity have no effect on exonuclease activity, consistent with our interpretation that Artemis does not contain intrinsic exonuclease activity [13], [14], [18], [22], [19], [30]. Many of these mutational studies were conducted to determine how Artemis and DNA-PKcs interact, and what role this interaction and DNA-PK mediated phosphorylation of Artemis play in endonucleolytic cleavage activity [13], [14], [22], [30]. Differing results from these studies have left the mechanism of Artemis endonuclease activation an open question. Analysis of Artemis phosphorylation site mutants and large scale Tranexamic Acid receptor of sites of phosphorylation have led to the conclusion that Artemis must be phosphorylated by DNA-PKcs to gain endonuclease activity [13], [22]. Analysis of DNA-PKcs phosphorylation mutants led others to conclude that autophosphorylation of DNA-PKcs is required to facilitate Artemis-catalyzed endonuclease activity [14], [15]. Importantly, the analysis of an extensive collection of N-terminal Artemis mutations located in the enzymatically important metallo-β-lactamase and β-CASP domains resulted in identification of a sub-set of mutants which functionally abrogated endonuclease activity via disrupting metal coordination [18]. Analysis of these mutants resulted in no loss of exonuclease activity. A recent paper generated two additional mutations, also in the N-terminus domain, which have reduced and inactive endonuclease activity, respectively and these endonuclease deficient mutants also retained exonuclease activity [19]. An additional phosphorylation mutant, associated with partial immunodeficiency in a mouse model, exhibits reduced endonuclease activity but nearly complete retention of exonuclease activity [29]. While the potential exists that the exonuclease activity could be located in another active site other than those identified by generating mutants, this seems unlikely, as it is thought that metallo-β-lactamase fold enzymes have one active site that is responsible for all enzymatic processing [20]. This is further supported by data published regarding SNM1, a 5′–3′ mammalian exonuclease classified in the metallo-β-lactamase superfamily. SNM1 is characterized by having only exonuclease activity on single-strand DNA, with no accompanying endonuclease activity, and a mutation of a conserved aspartate (D736) in the β-lactamase domain functionally disrupts the exonuclease activity [30], [31]. Paradoxically, mutagenesis of the conserved aspartate in Artemis (D37) eliminated endonuclease activity, but the exonuclease activity remained [18]. This indicates that the exonuclease activity is not located within the same active site as the endonuclease activity, and the extensive mutational analysis performed to date has yet to locate an exonuclease active site within Artemis. The combination of these genetic studies coupled with our biochemical analysis indicates that not only is the exonuclease activity separate from the endonucleolytic active site, but is not part of the Artemis polypeptide at all. Separation of the nuclease activities, as presented in this paper, was achieved with multiple protein purification preparations. However, it is important to note that the separation of exonuclease activity from Artemis did vary between protein preparations. As we continued to improve our purification procedures, specific changes, albeit small, in the protocol resulted in subtle differences in separation of exonuclease activity from endonuclease activity. In separating [His]6-Artemis over the HAP column, we found that greater separation of activity was achieved on a 5mL HAP column compared to a 2mL HAP column, despite more than enough protein-binding capacity on the 2mL column. However, the residual exonuclease activity that flowed through the 2mL column could be separated from Artemis by re-running the flow-through on a second HAP column (data not shown). This suggests that saturation of the hydroxyapatite column with exonuclease activity, at least to a certain level, can occur, leading to sub-optimal separation. These variations are largely a result of the specifics of our protocol, and can be impacted by numerous factors, including specificity of the matrix used, MacroPrep Ceramic Hydroxyapatite (Biorad), or relatively high pH (7.85) of the buffer used in the purification. Interestingly, we did observe a nominal amount of Artemis in many of the fractions collected from the HAP column, including the wash and elution. The diminutive levels of [His]6-Artemis were often only observed by Western blot analysis. This phenomenon was also observed during fractionation over the nickel–agarose column, indicating a certain degree of spreading of the fusion protein during all fractionation steps.