(A) TK6-DRGFP cells depleted for RAD51AP1 (shRNA #2 or shRNA #3) or XRCC3 show ∼2-to 2.5-fold lower levels of GFP+ cells (i.e. homologous recombinants) than control cells transfected with either pRNAi or mutated shRNA #2. In four experiments, both negative controls were used and both gave very similar numbers (for example, see Fig. S3C). Therefore, their values were averaged and this average was set to 1.0. In the remaining experiments only mut. shRNA was used as a negative control and this number was set at 1.0. Data for the depleted cells are the mean of the relative fraction of GFP+ cells from 5-7 independent experiments ± 1 SE (see also Fig. S3C).
(B and C) Western blot analyses to show the extent of EGFP-RAD51AP1 (here: AP1)-depletion (B) or XRCC3 (here: X3)-depletion (C) in TK6-DRGFP cells. Two hairpins were tested for each gene (lanes 2 and 3 corresponding to shRNA #2 and shRNA #3, respectively for RAD51AP1) and compared to control cells (lanes 1: transfected with mutated shRNA #2). QM: loading control.
(D) RAD51 foci form normally in RAD51AP1-depleted HeLa cells after 8 Gy X-rays. Control transfected HeLa cells (mutated shRNA #2; panels A and B) and RAD51AP1-depleted cells (panels C-F) display bright RAD51 foci in ∼ 80 % of the cells analyzed. In contrast, HeLa cells depleted for either RAD51C (panels G and H) or XRCC3 (panels I and J) show impaired RAD51 foci formation (i.e. only some cells form overall smaller and less intense RAD51 foci). The single cell with large foci in panel J may not have been transfected with the XRCC3 shRNA. Cells were fixed and stained at 8 h after 8 Gy X-rays. Two panels for each sample from different areas of the chamber slide are shown.
(E) Western blot analysis demonstrating the depletion of RAD51AP1, XRCC3 or RAD51C for the experiment shown in panel D. RAD51 acts as both a non-depleted negative control and as a loading standard. XRCC3 and Rad51C exist as a complex (Wiese et al., 2002a), and depletion of XRCC3 reduces the level of RAD51C and vice versa, as has previously been reported (Lio et al., 2004).

(A) Schematic of the D-loop assay.
(B) D-loop reactions mediated by combinations of RAD51 (0.8 µM) and RAD51AP1 (0.05 to 1 µM). ATP was omitted from the reaction in lane 10. The results were plotted.
(C) D-loop reactions mediated by combinations of RAD51 K133R (0.8 µM) and RAD51AP1 (0.05 to 1 µM). ATP was omitted from the reaction in lane 10. The results were plotted.
(D) D-loop reactions mediated by combinations of RAD51 or RAD51 K133R (0.8 µM each) and the MBP tagged form of wild type or mutant RAD51AP1 (0.2 µM each). The results were plotted.

(A) Survival curves of RAD51AP1-depleted HeLa cells treated with MMC. Non-depleting negative controls: pRNAi, GFP shRNA and mutated shRNA #2. All points with error bars in panels A-E represent the average of at least three different experiments ± 1 SD. * Note: shRNA #1 was tested in triplicate but with 2 µM MMC only, and resulted in the same sensitization as shRNA #2 and #3.
(B) Complementation of RAD51AP1-depletion in HeLa cells. Cells expressing both shRNA #2 and EGFP-RAD51AP1res are more resistant to MMC than cells expressing shRNA #2 only.
(C) Epistasis between Rad51AP1 and XRCC3. HeLa cells depleted for both XRCC3 and RAD51AP1 show the same sensitivity to MMC as cells depleted for RAD51AP1 only.
(D and E) Survival curves of RAD51AP1-depleted asynchronous HeLa cells treated with camptothecin or X-rays. Non-depleting negative controls are GFP shRNA and mutated shRNA #2 for D, or just mutated shRNA #2 for E. Note: shRNAs #1 and #3, which do not sensitize asynchronous HeLa cells to X-rays, do sensitize cells synchronized in S-phase (Fig. S2A).
(F) Representative western blot analysis to show the extent of RAD51AP1-depletion observed in HeLa cells with three different shRNAs; probed with anti-GFP antibody. The EGFP-RAD51AP1 fusion protein is a surrogate marker for RAD51AP1 expression. pRNAi is a negative control and QM (a transcription factor) is a loading control. Additional western blots, also demonstrating depletion of endogenous RAD51AP1, are presented in Figs. 2, S1 and S2.
(G and H) MMC (50nM)-induced and spontaneous chromatid breaks in RAD51AP1-depleted cells. Mutated shRNA #2 is a non-depleting negative control. In panel H, spontaneous ctbs were scored 72 h and 96 h after RAD51AP1-depletion. Error bars: ± 1 SE for two (G) or three (H) independent experiments.
(I) Giemsa-stained metaphase spread of MMC-treated RAD51AP1-depleted HeLa cells. Black arrows: chromatid breaks; red arrow: chromatid-type exchange aberration.

(A) Purified GST-RAD51AP1, RAD51AP1, MBP-RAD51AP1, MBP-RAD51AP1 L319Q, MBP-RAD51AP1 H329A, and MBP-RAD51AP1 CΔ25 were analyzed by SDS-PAGE and Coomassie Blue staining.
(B) GST-tagged RAD51AP1 or GST was incubated with RAD51 or yRad51 and glutathione Sepharose beads were used to capture any protein complex that had formed. The beads were washed and treated with SDS to elute the bound proteins. The supernatant (S), wash (W) and SDS eluate (E) were analyzed by SDS-PAGE with Coomassie Blue staining.
(C) MBP-tagged wild type, L319Q, H329A, or CΔ25 RAD51AP1 protein was incubated with RAD51 and amylose agarose beads were used to capture any protein complex that had formed. The analysis was as in (B).
(D) RAD51AP1 (0.03 to 1.5 µM) was incubated with ssDNA and dsDNA (panel I) or with dsDNA and the D-loop substrate (panel III). The mobility shift of the DNA substrates was analyzed in a 10% (panel I) or 5% (panel III) polyacrylamide gel. The asterisk denotes the position of the 5′ ³²P label. The results from panels I and III were plotted in panels II and IV, respectively.