Single Molecule Analysis of Laser Localized Interstrand Crosslinks

Abstract

DNA interstrand crosslinks (ICLs) block unwinding of the double helix, and have always been regarded as major challenges to replication and transcription. Compounds that form these lesions are very toxic and are frequently used in cancer chemotherapy. We have developed two strategies, both based on immunofluorescence (IF), for studying cellular responses to ICLs. The basis of each is psoralen, a photoactive (by long wave ultraviolet light, UVA) DNA crosslinking agent, to which we have linked an antigen tag. In the one approach, we have taken advantage of DNA fiber and immuno-quantum dot technologies for visualizing the encounter of replication forks with ICLs induced by exposure to UVA lamps. In the other, psoralen ICLs are introduced into nuclei in live cells in regions of interest defined by a UVA laser. The antigen tag can be displayed by conventional IF, as can the recruitment and accumulation of DNA damage response proteins to the laser localized ICLs. However, substantial difference between the technologies creates considerable uncertainty as to whether conclusions from one approach are applicable to those of the other. In this report, we have employed the fiber/quantum dot methodology to determine lesion density and spacing on individual DNA molecules carrying laser localized ICLs. We have performed the same measurements on DNA fibers with ICLs induced by exposure of psoralen to UVA lamps. Remarkably, we find little difference in the adduct distribution on fibers prepared from cells exposed to the different treatment protocols. Furthermore, there is considerable similarity in patterns of replication in the vicinity of the ICLs introduced by the two techniques.

Keywords: DNA damage; DNA fiber; interstrand crosslinks; laser; replication; single molecule.

FIGURE 1

Distribution of Dig-TMP on DNA fibers from cells after photoactivation by UVA lamp. (A) The structure of trimethylpsoralen (TMP) linked to Digoxigenin (Dig). (B) Cells were incubated with Dig-TMP and exposed to UVA in the Rayonet lamp chamber. They were then immunostained against the Dig tag. (C) Cells were prelabeled with CldU, then incubated with Dig-TMP and exposed to the UVA lamp. DNA fibers were prepared and immunostained for CldU (green) and Dig-TMP (red dots). (D) Cells were treated with the indicated concentration of Dig-TMP and DNA fibers prepared and stained, and fields photographed as in (C). The fraction of fibers in a field containing a Dig-TMP signal was determined. From cells treated with 5 μM Dig-TMP the number of fibers with a Dig signal and the total fibers scored in individual experiments were: 33 and 297; 25 and 279; 32 and 224. From cells treated with 20 μM Dig-TMP: 39 and 285; 21 and 208; 35 and 206 (total fibers 742, 604). The Chi squared test showed no significant difference (p = 0.114). (E) Fibers were prepared and stained from cells treated as in (C). The lengths of all fibers in a field, and the total number of Dig-TMP signals on fibers, were determined and used for the calculation of the number of Dig-TMP signals per 10³ kb of fiber. The number of Dig-TMP signals scored and the total number of fibers examined were: at 5 μM- 44 and 78; 70 and 98; 54 and 65; at 20 μM- 97 and 69; 81 and 62; 80 and 61 (N = 245, 188). The t-test showed significant difference (p < 0.001) between the two means (2.69, 4.48, SD 0.095, 0.285).

FIGURE 2

Laser localized Dig-TMP induces the DDR. (A) Cells were incubated with 20 μM Dig-TMP and exposed to the UVA lamp. They were fixed and immunostained for γ-H2AX. (B) Cells were incubated with Dig-TMP and a defined region of interest was exposed to the 365 nm laser light. Cells were fixed and immunostained for the Dig tag. (C) Cells were incubated with Dig-TMP, exposed to the laser in an ROI and after 15 min fixed and immunostained for γ-H2AX (green) and FANCD2 (red).

FIGURE 3

Distribution of UVA lamp induced and laser localized ICLs. The experimental design is presented in the schematic. The ICLs, indicated in red, are localized to stripes after laser photoactivation of Dig-TMP at ROIs (A), and distributed throughout the nucleus after exposure of cells to Dig-TMP and a UVA lamp (B).

FIGURE 4

The distribution of Dig-TMP ICLs on DNA molecules after photoactivation by laser or UVA lamp. (A) Cells were prelabeled with CldU, seeded in Mattek plates marked with a cross on the growth surface of the glass. The next day cells were incubated in 20 μM Dig-TMP and the ROI in nuclei in cells close to the cross were photoactivated with the laser. Cells in the vicinity of the cross were harvested and placed on a slide and fibers spread, fixed, and immunostained for CldU (green) and Dig-TMP (red). (B) Cells were prelabeled with CldU, incubated with 20 μM Dig-TMP and exposed to the UVA lamp. (C) The distribution of Dig-TMP signals on fibers from cells with laser or lamp photoactivation. The plot represents results from 3 independent experiments (number of laser localized Dig-TMP signals used for distance measurements in the individual experiments-111, 85, 90; number of Dig-TMP signals induced by UVA lamp used for the measurements-102, 101, 102).

FIGURE 5

Replication in the vicinity of laser localized Dig-TMP ICLs. Cells were seeded on plates marked with a cross, incubated with 20 μM Dig-TMP, and ICLs introduced by laser photoactivation in ROI in nuclei in cells close to the cross. The cells were then incubated at 37°C for 1 h in the presence of 10 μM IdU. Cells near the cross were recovered, placed on a slide and fibers spread, and stained as before for CldU (purple), IdU (green), and Dig-TMP (red). Four kinds of fibers were scored: (A) those containing a Dig-TMP without a replication tract; (B) those with a replication tract and, at some distance, a Dig-TMP; (C) those in which replication had occurred immediately adjacent to one side of the Dig-TMP; and (D) those on which replication tracts were on both sides of the Dig-TMP. (E) Quantitation of the relative frequencies of fibers with replication on one or both sides of the Dig-TMP. Results from three independent experiments are represented (numbers of Dig-TMP signals with: replication on one side- replication on both sides; 9-55; 7-42; 3-21). The z-test showed significant difference (p < 0.001) between the two proportions (averaged proportions 0.138, 0.862, N = 137).