PICH: a DNA translocase specially adapted for processing anaphase bridge DNA

Abstract

The Plk1-interacting checkpoint helicase (PICH) protein localizes to ultrafine anaphase bridges (UFBs) in mitosis alongside a complex of DNA repair proteins, including the Bloom's syndrome protein (BLM). However, very little is known about the function of PICH or how it is recruited to UFBs. Using a combination of microfluidics, fluorescence microscopy, and optical tweezers, we have defined the properties of PICH in an in vitro model of an anaphase bridge. We show that PICH binds with a remarkably high affinity to duplex DNA, resulting in ATP-dependent protein translocation and extension of the DNA. Most strikingly, the affinity of PICH for binding DNA increases with tension-induced DNA stretching, which mimics the effect of the mitotic spindle on a UFB. PICH binding also appears to diminish force-induced DNA melting. We propose a model in which PICH recognizes and stabilizes DNA under tension during anaphase, thereby facilitating the resolution of entangled sister chromatids.

Figure 1. PICH is an ATP-dependent translocase protein

(A) Representative time course (as indicated above the lanes) of DNA triplex disruption by native PICH at either 2 nM (top panel) or 8 nM (middle), and of PICH-K128A at 8 nM (bottom). Lane ‘H’ denotes heat-denatured substrate. The positions of the triplex substrate and the free ssDNA oligonucleotide are shown on the right. The red asterisk denotes the radioactive label. The experiment was repeated 3 times with comparable results. (B) Quantification of the data from panel A. See also Figures S1, S2, S3, S4 and S5.

Figure 2. Single molecule analysis

(A) Outline of single molecule experimental set-up. Experiments were carried out in a four-channel flow cell (top left: cut-out image of cell). In section I (expanded on the right), the laminar flow (indicated by right arrows) prevents the contents in the combined channels from mixing. By moving the flow cell with the aid of a stage, the channels can be rapidly switched to pick up beads (1), capture a single DNA molecule between the beads (2), and check the integrity of the DNA (3). The trapped DNA is subsequently moved into the fourth channel (4) (section II; expanded below), which contains fluorescent PICH in the same buffer. (B) Unprocessed fluorescence image depicting five single PICH proteins bound to λ DNA (invisible) stretched between the beads. See also Movie S1. (C) The fluorescence signal (red boxed area) of a fully PICH-eGFP-coated DNA divided by the average brightness of a single molecule yields the number of bound molecules (~2400) from which the size of the footprint can be derived. (D) The majority of observed PICH-DNA interactions fall into a narrow brightness range that is fitted well by a Gaussian distribution. The small fraction (<5%) of events at significantly higher brightness can be attributed to dimer interactions. (E) The course of the fluorescence signal over time for single binding events under bleaching conditions can be used to distinguish between monomer (bleaching occurs in single steps, traces I-III) and dimer binding (two steps, IV).

Figure 3. Translocation by PICH

(A) Kymographs of eGFP-labeled PICH (~50 pM) on DNA molecules at medium tension (25 pN) show that under low salt conditions (25mM NaCl, top), interactions last much longer than in high salt (100 mM, bottom). See also Movies S1 and S2. (B/C) Three independent traces of PICH-eGFP interactions on DNA using high (B) or low (C) salt reaction conditions to demonstrate typical translocation characteristics of PICH. In all cases, the displacement scales linearly with time, as expected for an ATP-dependent translocase. Note the frequent reversal of translocation direction, which it is especially prominent under low salt conditions. Occasional pausing is also apparent. (D) The histograms of observed translocation speeds display a Gaussian distribution in both low and high salt conditions. (E) The cumulative distribution of switch times (i.e., the time that elapses between each reversal of translocation direction) determined under low salt conditions is well fitted by an exponential decay with a characteristic time of ~39 ± 3 s. (F) Because in low salt the switch time is much shorter than the average interaction time, PICH motion resembles random diffusion. See also Figures S6 and S7.

Figure 4. PICH responds to DNA stretching

(A/B) Kymographs of PICH-EGFP interactions at different DNA tension. The red dotted line denotes where a change of the tension was applied, the respective value of which is given in the upper section in yellow. Note that only the duration, but not the frequency, of interactions changes significantly with DNA tension. See also Movie S3. (C) Cumulative distributions of the interaction time on DNA for PICH-eGFP at different forces follow a mono-exponential decay. (D) The interaction time of PICH on DNA increases by ~10-fold as the force is increased from 1 to 30 pN, and displays a mono-exponential tension dependence (red triangles). At higher tensions, >30 pN, the interaction decreases slightly again. For comparison, the interactions at equivalent PICH-eGFP concentrations after addition of a 15-fold excess of non-labeled PICH (see Figure 6 below) follow the same trend, albeit with an about 3-fold longer interaction time (black circles).

Figure 5. hRAD54 has different DNA-binding properties from PICH

(A) hRAD54-eGFP binding to DNA. Snapshots of hRad54-eGFP binding to DNA in low salt buffer either without ATP (left) or with ATP (right). (B) Comparison of hRAD54 and PICH shows that, in low salt, hRAD54 (Ia, left) displays rapid diffusion without ATP, while PICH (IIa, right) remains static. In the presence of ATP and Mg²⁺, hRad54 shows a combination of diffusion and directed motion (Ib), whereas PICH translocates solely with a directed motion (IIb). (C) Comparison of PICH (I, left) and hRAD54 (II, right) interactions before (a, upper) and after (b, lower) DNA overstretching demonstrates that only PICH displays a force-dependent binding; interactions of PICH are significantly depleted by overstretching, while hRad54 binding is unaffected.

Figure 6. Effects of PICH on the physical properties of DNA

(A/B) Kymographs of PICH-eGFP (80 pM) interactions with DNA, either before (A) or after (B) addition of a 15-fold molar excess of non-labeled PICH. Note the clear increase of the interaction time at higher PICH concentrations. (C) PICH-eGFP together with non-labeled PICH display a similar force dependence of the interaction time to PICH-eGFP alone, such that, up to 30 pN, higher DNA tension results in a longer interaction time. (D) The force-extension curve of a PICH-coated DNA (red) shows that, compared to normal DNA (black), there is, an earlier force increase combined with a shift to longer extension beyond 30 pN. This effect is much more pronounced in the presence of ATP (blue). Furthermore, the overstretching plateau of normal DNA (black) at >60 pN force has given way to a strong force-increasing regime, showing that PICH binding stabilizes DNA against force-induced melting.

Figure 7

Proposed model for the action of PICH on UFB formation and resolution. (I) At the start of mitosis, the chromatin is still in the 30 nm fibre configuration. Since the nucleosomes at least partially shield the phosphate backbone of the DNA, PICH is not able to bind stably. (II) Once sister chromatid disjunction is triggered, there is a build-up of tension leading to DNA extension (denoted by the upward facing arrows), which causes fibre elongation, without exposing bare DNA. (III) Beyond a certain force threshold, the area close to the entanglement is forced to extend further into a beads-on-a-string configuration; PICH will accordingly bind to the exposed linker DNA (~50 bp) and is then able to translocate in either direction. (IV) Aided by the tension on the corresponding DNA section, PICH may be able to assist in overcoming the energy barrier for the unwrapping of nucleosomes with a ‘weakened grip’ on the DNA adjacent to the exposed section. This would expose more free DNA, thus recruiting more PICH mono- and oligomers to the exposed region of DNA. (V) Nucleosomes are expelled from the DNA, which becomes coated with PICH. This has the dual effect of stabilizing the DNA against stretching-induced strand unwinding, and allowing the DNA tension to be maintained at a constant level, as denoted by the horizontal arrows. Finally, PICH recruits the BTRR complex and possibly other DNA metabolising enzyme to permit the disentanglement of the intertwined DNA that had prevented normal disjunction in the first place.