Locked nucleic acid oligomers as handles for single molecule manipulation

Single-molecule manipulation (SMM) techniques use applied force, and measured elastic response, to reveal microscopic physical parameters of individual biomolecules and details of biomolecular interactions. A major hurdle in the application of these techniques is the labeling method needed to immobilize biomolecules on solid supports. A simple, minimally-perturbative labeling strategy would significantly broaden the possible applications of SMM experiments, perhaps even allowing the study of native biomolecular structures. To accomplish this, we investigate the use of functionalized locked nucleic acid (LNA) oligomers as biomolecular handles that permit sequence-specific binding and immobilization of DNA. We find these probes form bonds with DNA with high specificity but with varied stability in response to the direction of applied mechanical force: when loaded in a shear orientation, the bound LNA oligomers were measured to be two orders of magnitude more stable than when loaded in a peeling, or unzipping, orientation. Our results show that LNA provides a simple, stable means to functionalize dsDNA for manipulation. We provide design rules that will facilitate their use in future experiments.

Supplemental Figure S3. PNA handles have similar mechanical stability to traditional methods. The lifetimes of enzymatically labeled DNA constructs (blue, 34 initial tethers) and PNA-hybridization labeled DNA constructs (red, 24 initial tethers) are statistically equivalent. Both constructs use the DNA-2 substrate; the PNA constructs were created with PNA-2 incubated for 1 hour with 110 mM Na + .

Psoralen LNA handle
An 11 bp LNA oligomer with a 3' biotin moiety and an additional 5' modification of psoralen was designed to target an 11 bp sequence within the lambda genome. The oligomer (Exiqon) had the following sequence: 5' -/PsoC6/+TT+T/iMe-dC/+CT+TT+C/iMe-dC/+C/3bioTEG/ -3'. This design was similar to that used before to target DNA where the psoralen moiety targeted a TpA position, allowing the formation of a covalent bond upon UV irradiation (12,(26)(27)(28). We then measured the specificity using the same techniques as described in the full text. The results (Supplemental Figure  S1) show that the binding was highly non-specific. Because of this non-specific binding, which we attributed to the stable non-specific binding to the binding of psoralen to TpA sites throughout the DNA, we did not find psoralen-modified handles to be useful for site-specific DNA functionalization.

DNA for PNA targeting
For the PNA experiments, lengths of the lambda genome (NEB) were amplified using the Long Template PCR kit from Roche. One primer from each pair carried a digoxigenin label, such that the completed PCR products were digoxigenin-labeled on one end and unlabeled on the opposite end (primers from IDT). DNA-1 was 2,189 nm in contour length and extended from 37,585 to 44,059 bp in the lambda genome; DNA-2 was 5,449 nm and extended from 37,585 to 47,728 bp (thus containing the DNA-1 sequence). DNA-2fl was identical in sequence to DNA-2, but with digoxigenin at the opposite end. DNA-3 was 7,097 nm and extended from 40,632 to 47,709 bp.

PNA-DNA complex formation
PNA oligomers and digoxigenin labeled DNA were incubated in volumes of 10-50 µL at 37°C for various times (2 min to 1 week) in reaction buffer (1 mM EDTA, 10 mM sodium phosphate buffer (pH 7), 10 to 500 mM NaCl, and 10% acetonitrile). The concentration of DNA was between 1 and 5 nM and the PNA ranged from 1 to 200 times in excess of DNA. Once the reaction time was complete, the reaction was diluted 100x to 500 mM NaCl to stop the reaction and stored at 4°C.

Psol-LNA target DNA
The DNA used for psol-LNA:DNA hybridization was prepared in the same manner as the lambda DNA in the normal section of the paper. Simply, a cos sequence with a digoxigenin label was ligated onto one end of the lambda genome DNA.

PNA results
PNA/DNA binding was tested with four DNA substrates: three in which the PNA targeted an internal sequence (DNA-1, DNA-2, DNA-2fl), and one in which an extremity sequence was targeted (DNA-3). In all cases, substrates incubated with PNA oligomers form a significant number of tetherable constructs. Supplemental Figure S2 shows histograms of compiled contour lengths from several experiments. These plots show extensive non-specific binding, particularly at the DNA extremities. This non-specific end-binding moves with the end of the DNA as the DNA substrate is changed.

Psol-LNA results
Psol-LNA oligomers were hybridized with lambda DNA. Tests of similar LNA designs without psoralen modifications suggest that LNA oligomers of 11 bp are very unstable in response to force and the application of 3-5 pN is highly destabilizing. Nonetheless, the psoralen modified LNA oligomers formed stable handles on the dsDNA and we used the magnetic tweezers to collect force-extension data, extracted the contour lengths from WLC fits and compiled histograms of tether binding positions. These histograms of binding sites show that the psol-LNA oligomers bound to the DNA non-specifically (Supp. Figure S1). There were no positions within the DNA at which the LNA oligomers seemed to bind preferentially, indicating non-specific binding to the dsDNA substrate.
Analysis of the mismatch sites within the substrate DNA could not explain the observed non-specific binding. We attribute the stable non-specific binding to the binding of psoralen to TpA sites throughout the DNA, forming covalently bound complexes at these sites.

PNA Tether lifetime
Using techniques described in the full text, we studied the lifetime of tethers composed with PNA oligomers. We prepared two sets of substrates, one with a biotin label carried by PNA and the other with a biotin label covalently attached by standard enzymatic methods. Both substrates were tethered to magnetic beads, and subjected to a constant force. Distributions of tether lifetimes under 13 pN for both substrates display a roughly exponential decay with a 40-50 minute characteristic time (Supp. Figure S3). Indeed, with 99% certainty, the two populations are statistically identical, as determined by a Kolmogorov-Smirnov test; this indicates both sets of tethers share the same weakest link (likely the dig/anti-dig bond), and thus that the PNA/DNA bond is not limiting.