In attempts to further purify His Artemis numerous matrices
In attempts to further purify [His]6-Artemis, numerous matrices were assessed including anion and cation exchange, and Methoxyresorufin interaction chromatography. The results from these matrices were universally poor (data not shown). Fractionation via adsorption chromatography on a hydroxyapatite (HAP) matrix at a relatively high pH of 7.8 did however result in substantial purification of Artemis. The majority of total protein applied (greater than 90%) was retained on the HAP column while the majority of [His]6-Artemis was not retained on a HAP column under these conditions but was identified in the flow-through fractions. Western blot analysis of the flow-through fractions (Fig. 3A) and elution fractions (Fig. 3B) demonstrates the vast majority of Artemis was recovered in the flow-through fraction. Coomassie Blue stained SDS-PAGE of the HAP load and flow-through fractions confirmed the decrease in complexity of the HAP FT pool (Fig. 3C), though definitive determination of the Artemis polypeptide remains. The Artemis containing flow-through fractions were pooled and Artemis levels in each were determined by the more quantitative in vitro phosphorylation analysis which confirmed that the vast majority of the Artemis protein was present in the flow-through pool and was capable of being phosphorylated by DNA-PK (Fig. 3D, lane 1) compared to control reactions without the HAP FT (lane 2). To determine the retention of exonuclease activity upon HAP fractionation, we assayed the column fractions for 5′–3′ exonuclease activity using a 34-base single-strand oligonucleotide with a 5′ [32P]-label. The release of the 5′ radiolabeled nucleoside monophosphate is a direct measure of 5′–3′ exonuclease activity. The majority of exonuclease activity was resolved in six fractions that bound to the column and were subsequently eluted during the phosphate gradient. Minimal exonuclease activity was observed in the flow-through fractions, which contain the majority of Artemis protein (Fig. 4A and B). The exonuclease containing fractions in the HAP phosphate elution were pooled. These data demonstrate a small portion of the overall exonuclease activity loaded onto the HAP column was identified in the flow-through, while the majority was located in the eluate (Fig. 4B). Importantly, measurement of the bound exonuclease activity is potentially an underestimation, as all of the substrate was completely degraded in the peak elution fractions (fractions 48–50, Fig. 4A and B). These results suggest that at specific conditions of this fractionation, [His]6-Artemis does not bind to a HAP column while the majority of exonuclease activity remains bound under the same conditions. Further confirmation of the fractionation of Artemis into the HAP FT separated from the majority of exonuclease is revealed by the presence of DNA products consistent with a sequence specific, single-stranded endonuclease activity recently attributed to Artemis in the HAP FT fractions . In an effort to determine if the minimal exonuclease activity found in the HAP FT and the major exonuclease in the elution was specific to the [His]6-Artemis purification, we designed an experiment to determine whether the exonuclease activity is a contaminating nuclease that has a high affinity for a nickel–agarose column. Using recombinant baculovirus, we overexpressed another His-tagged DNA repair protein, XPA, with no intrinsic nuclease activities and minimal protein interaction domains. A whole cell extract was prepared in the same fashion as extracts of overexpressed [His]6-Artemis, and the identical purification protocol was followed. Following fractionation over the nickel–agarose column and HAP column, fractions were assayed for exonuclease activity. Robust exonuclease activity was seen in the whole cell extract, as well as in protein pooled from the nickel column (data not shown). Fractions assayed from the HAP column again revealed a minimal amount of exonuclease activity flowing through the column (Fig. 4C and D). Furthermore, the peak of exonuclease activity eluted from the HAP column with the phosphate gradient coincides exactly with the peak of exonuclease activity eluted from the HAP column in the [His]6-Artemis prep. Interestingly, the low level of exonuclease activity flowing through the HAP column coincides with the low level of exonuclease activity seen flowing through the HAP column from the [His]6-Artemis prep. Finally, pools of protein from the nickel–agarose elution, HAP flow-through and HAP elution were examined for DNA-PK dependent endonuclease activity, and all pools were completely devoid of this activity, as expected (Supplemental Fig. 1). Analysis of the HAP elution pool for DNA-PK phosphorylation of Artemis reveals an extremely low level of Artemis in this pool of protein (Fig. 3D, lane 3), consistent with the spreading out of the Artemis proteins over the entire gradient as assessed by western blot analysis of the fractions (Fig. 3B).