Efficient preparation of internally modified single-molecule constructs using nicking enzymes

Abstract

Investigations of enzymes involved in DNA metabolism have strongly benefited from the establishment of single molecule techniques. These experiments frequently require elaborate DNA substrates, which carry chemical labels or nucleic acid tertiary structures. Preparing such constructs often represents a technical challenge: long modified DNA molecules are usually produced via multi-step processes, involving low efficiency intermolecular ligations of several fragments. Here, we show how long stretches of DNA (>50 bp) can be modified using nicking enzymes to produce complex DNA constructs. Multiple different chemical and structural modifications can be placed internally along DNA, in a specific and precise manner. Furthermore, the nicks created can be resealed efficiently yielding intact molecules, whose mechanical properties are preserved. Additionally, the same strategy is applied to obtain long single-strand overhangs subsequently used for efficient ligation of ss- to dsDNA molecules. This technique offers promise for a wide range of applications, in particular single-molecule experiments, where frequently multiple internal DNA modifications are required.

Figure 1.

Schematic representation of the internal labeling method. (a) A DNA sequence which incorporates five equally spaced BbvCI recognition sites (black triangles) is nicked only at one of the two strands using either the nicking enzyme Nt.BbvCI or Nb.BbvCI. This results in the formation of short 15–16 bases long fragments. Denaturation and subsequent hybridization, in the presence of a DNA strand (shown in red) that is complementary to the resulting 63 bp gap and that carries the desired internal modifications (e.g. two or six biotins as depicted), lead to an efficient replacement of the original fragments with the labeled fragment. The spacing between the internal modifications of 10–11 bp ensures that they extrude in the same direction from the DNA. (b) Model of a streptavidin tetramer bound to an internally biotinylated DNA molecule (Streptavidin PDB id: 1MK5; the bound monomer is illustrated as a yellow ribbon while for the other subunits the surface representation was used. DNA PDB id: 2BNA). The attachment of the streptavidin tetramer to only one of the biotins was arbitrarily chosen.

Figure 2.

Internal modification and religation efficiencies. (a) Polyacrylamide gel electrophoresis of DNA samples in the course of the internal labeling procedure. (Lane 1) pNLrep after digestion with MluI and AatII yielding a 0.8-kb fragment (red line) that carries the region to be replaced as well as a 0.4-kb (light blue line) and a 5-kbp fragment (dark blue line). (Lane 2) pNLrep after simultaneous digestion with MluI, AatII and Nt.BbvCI. The nicking of the 0.8-kb fragment (represented by the fragmented red line) can be seen as a slight mobility decrease. (Lane 3) Sample from lane 2 after column purification, which leads to gap formation within the 0.8-kb fragment causing a large mobility alteration (gapped red line). (Lane 4) Sample from lane 2 after the replacement reaction with oligo biotinx2, during which the 0.8-kb fragment becomes internally biotinylated, and subsequent column purification. The inserted oligo is stably bound and therefore displays the same mobility as the nicked fragment in lane 2. (Lane 5) Sample from lane 4 with >10-fold molar excess of streptavidin added. (Lane 6) Pulldown assay with sample from lane 4 (see ‘Materials and Methods' section). (Lanes 0, 7) 100 bp step DNA ladder, starting at 400 bp with an additional 517 bp band. (b) (Lane 1) pNLrep. (Lane 2) pNLrep after nicking and internal biotinylation with oligo biotinx2. (Lane 3) Sample from lane 2 after ligation. (Lane 4) pNLrep after internal biotinylation with 5′-phosphorylated biotinx2 oligo and religation. (Lane 5) pNLrep after nicking with Nt.BbvCI and religation. Positions of supercoiled, nicked and linearized plasmid species are indicated by corresponding symbols at the right side. (Lane 0) 1 kb step DNA ladder with the shortest fragment starting at 1 kb.

Figure 3.

Site-specific attachment of Q-dots to internally biotinylated DNA. (a) Band-shift assay of Q-dot binding to DNA. (Lane 0) 1 kb step DNA ladder with the shortest fragment starting at 1 kb. (Lane 1) pNLrep after digestion with BamHI, PspOMI and Nt.BbvCI and internal biotinylation with oligomer biotinx2 (Figure 2a). (Lane 2) Sample from Lane 1 with 5-fold molar excess of streptavidin coated Q-dots added. (Lane 3) Sample containing Q-dots only. Symbols on the right side indicate Q-dots (yellow spheres), the short biotinylated fragment (red line) and the long non-biotinylated fragment (blue line). (b and c) AFM images of Q-dots bound to DNA. The colour scale corresponds to a height-range of 1 nm, and the scale bar corresponds to 100 nm. (d) Histogram of the Q-dot position measured from the nearest DNA end. The expected Q-dot position at 920 bp (310 nm) (blue dashed line) is within the double confidence interval (light grey band) of the experimentally determined mean (290 ± 20 nm, red dashed line).

Figure 4.

Magnetic tweezers experiment with internally attached Q-dot. DNA molecules with an attached Q-dot bound through internal biotinylation were stretched in the vertical direction within a magnetic tweezers setup (see Inset in Figure 4b). Q-dots were detected by fluorescence. (a) Fluorescence images taken at different focal (z-) positions above the surface of the flow cell. At 0, 1.5 and 3 μ m a surface-bound Q-dot, the internally attached Q-dot and the weakly auto-fluorescent magnetic bead were within the focal position of the objective. This was in agreement with the expected positions, since the Q-dot was internally attached at 4.4 kb (≈1.5 μm) while the full length of the molecule was 9 kb (≈3.0 μm). (b) z-stack of a linear section across the surface- and the DNA-bound Q-dot at different z-positions. (Inset) schematic representation of the experimental configuration. DNA was drawn as a white line, Q-dots as small yellow spheres and the magnetic bead as a large orange sphere.

Figure 5.

Inserting structural modifications. Schematic representation: (a) a hairpin (shown in red) containing a BstXI site (black triangle) was introduced via the single-step replace reaction. (b) After digestion with BstXI, a 500-bp DNA fragment (shown in green) was ligated to the junction arm. (c) AFM images of the resulting Y-shaped molecules; the lengths of the three branches (355 ± 5, 250 ± 15 and 140 ± 14 nm) agree with the expected values (390, 240 and 130 nm) for the exception a discrepancy of the longest branch. The colour scale corresponds to a height range of 1 nm, the scale bar corresponds to 100 nm.

Figure 6.

ssDNA to dsDNA ligation at nicking-enzyme-generated overhangs. (a) Schematic representation of overhang generation. A BbvCI recognition site (blue letters) was incorporated near the DNA end in such a way that nicking with Nt.BbvCI generates a 10-bp fragment at the 5′-end. (b) Agarose gel of DNA fragments, ligation products and streptavidin-induced band shifts. The biotinylated 40-bp hairpin, the 430-bp dsDNA handle and streptavidin are represented by a blue, red and green symbol, respectively. `Lig.' indicates where a ligation for 1 h at room temperature was carried out. Positions of the reaction products are marked at the right side. (Lanes 1–5) reaction products for the 4 nt overhang generated by BstXI. (Lanes 6–10) reaction products for the 10 nt overhang generated by Nt.BbvCI. The lane in the middle is a 100-bp size marker ladder with the shortest fragment starting at 100-bp and 100-bp size difference between all subsequent fragments (and an additional band at 517 bp). The success of the ssDNA to dsDNA ligation was confirmed by the streptavidin-induced band-shift, in which the desired product specifically shifted only in the case where a 10 nt 3′-overhang had been used. (c) Magnetic tweezers experiment with the generated hairpin construct. The molecule was held at the critical force where the closed and the opened states of the hairpin (as illustrated by the sketches) were nearly equally populated. The change in height between the two states was ≈38 nm as expected for a 40-nt hairpin.