oriD structure controls RepD initiation during rolling-circle replication

Bacterial antibiotic resistance is often carried by circular DNA plasmids that are copied separately from the genomic DNA and can be passed to other bacteria, spreading the resistance. The chloramphenicol-resistance plasmid pC221 from Staphylococcus aureus is duplicated by a process called asymmetric rolling circle replication. It is not fully understood how the replication process is regulated but its initiation requires a plasmid-encoded protein called RepD that nicks one strand of the parent plasmid at the double-stranded origin of replication (oriD). Using magnetic tweezers to control the DNA linking number we found RepD nicking occurred only when DNA was negatively supercoiled and that binding of a non-nicking mutant (RepDY188F) stabilized secondary structure formation at oriD. Quenched-flow experiments showed the inverted complementary repeat sequence, ICRII, within oriD was most important for rapid nicking of intact plasmids. Our results show that cruciform formation at oriD is an important control for initiation of plasmid replication.

= 1 2 (2 n) 2 ; where  = (C/lo), and at low longitudinal force (i.e. before the magnetic field is applied), C~160 pN.nm 2 .rad -1 (19) and lo= 3400 nm (10kb template) and 1360nm (4 kb template). This gives an expected r.m.s. angular deviation; < n > = 1.47 and 0.93 turns for 10-kb and 4-kb templates resp. Our measured r.m.s. deviation in angular offset for the 10-kb template was 1.48 turns and for 4-kb, 1.18 turns. So, our measured variation is slightly larger than expected and this may be explained by additional rotational noise due to the random starting orientation of the paramagnetic bead easy axis relative to the magnetic tweezers which would contribute an additional (~0.25) turn variation. The reproducibility of the measurement was determined by paired replicates (Fig. S1B) which indicate the error for a given estimate was less than 0.5 turns (i.e. deviation from the unitary gradient indicated by the dotted line in Fig. S1B). The magnet rotation vs. z-displacement plots were also used to calculate the length change per turn of writhe (which was 69 ± 1.3 nm/turn over the range 8 to 32 turns, shown shaded on the plot) (Fig. S1C).
Supplementary Figures S2-S3 are referred to in the main text.

Supplementary figures: FIGURE S1. DNA tether length-dependence on supercoiling
In order to determine the tether length-dependence upon superhelical density, , the magnetic field was rotated by a known number of revolutions while recording bead height above the coverslip surface. Because the current study requires measurement of the absolute superhelical density we measured the initial offset due to thermal motion for each bead before the start of every experiment.

(A)
The example shown (10-kb DNA tether, held at 0.4 pN, 23 o C) illustrates the fact that DNA-bead constructs exhibit variable starting offset in super-helical density due to thermal rotation of the DNAbead template before the magnetic field is applied to the sample. Six examples are shown, which were collected from a single field of view and data for each bead is plotted in a different color and offset with respect to the y-axis for clarity. The shape of the z-displacement vs rotation plots are used to determine the starting offset value; the data for each bead was fitted empirically to a Gaussian function (solid lines) and center values (indicated by arrows) were determined for two replicates for each bead. The fitted center values in the top record are labelled (i) and (ii) see (B) below. Raw data for the two experimental runs are overlaid in the figure. (B) Reproducibility of the protocol is demonstrated by plotting the estimated center values (i.e. resting offsets due to rotational thermal motion before application of the magnetic field) for each bead for two replicates. The dotted line of unity gradient allows correlation between repeat measurements to be judged. The data (brown record) labelled (i) and (ii) in (A) represent the "worst case" and are shown for clarity as an example and show the largest deviation between the two replicates.
(C) A plot of z-displacement vs magnet rotation enables the relationship between length change (zdisplacement) and turns of writhe to be determined. The gradient of the graph (solid lines) over the region (±) 8 to 32 turns (i.e. the linear, shaded, part of the plot) gives a length change of 69 ± 1.3 nm per turn of writhe (n=24).   where the amplitude, A1, is proportional to the fraction of time spent at the short tether length (x0), A2 is proportional to the fraction of time spent at the extended tether length (x0+x) where the length change x = 71.5 nm (see main text Fig. 7) and the root mean squared deviation in DNA tether length (<w>) due to thermal motion is approximated by: Where DNA contour length, L0 = 1360nm (4kbp) and persistence length, Lp = 50nm, the relative extension, Lrel, at a force of 0.4 pN, is 0.76: gives <w> = 42 nm and at 0.67pN Lrel = 0.82: gives <w> = 28 nm. The ratio of time spent in the two states (A1/A2) gives the equilibrium constants (tabulated in the main text Fig. 6).