Tracking DNA Synthesis with Single-Molecule Strand Displacement

Oct 4, 2018 - *E-mail: [email protected]. ... By monitoring strand displacement, we track the motion of a single ϕ29 DNA polymerase without ...
2 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCB

Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX

Tracking DNA Synthesis with Single-Molecule Strand Displacement Charles E. Wickersham†,¶ and Everett A. Lipman*,†,‡ Department of Physics and ‡Biomolecular Science and Engineering Program, University of California, Santa Barbara, California 93106, United States

J. Phys. Chem. B Downloaded from pubs.acs.org by UNIV OF SUNDERLAND on 10/05/18. For personal use only.

†

ABSTRACT: We have previously shown that double-stranded DNA labeled with a periodic series of fluorescent dyes can be used to track a single helicase. Here we demonstrate how this technique can be adapted to follow processive DNA synthesis. By monitoring strand displacement, we track the motion of a single ϕ29 DNA polymerase without labeling or altering the enzyme or the template strand, and without applying any force. We observe a wide range of speeds, with the highest exceeding by several times those observed in earlier in vitro single-molecule experiments. Because this method enables repeated observations of the same polymerase traversing identical segments of DNA, it should prove useful for determining the effects of sequence on DNA replication and transcription. In addition, future measurements of this type may allow us to examine in detail the interactions of individual DNA polymerases with other components of the replisome.

■

INTRODUCTION Over the past two decades, several techniques have been developed that enable high-resolution tracking of single-motor proteins. Mechanical assays, which include optical and magnetic tweezers, can achieve extreme spatial resolution provided that sufficient tension, often on the order of several tens of piconewtons, is applied to the motor or its substrate.1,2 These techniques also require that the motor dwell between steps for approximately 500 ms in order for signal averaging to make subnanometer motion detectable. Typical fluorescencebased techniques do not require tension to damp out Brownian fluctuations, but often suffer from limited temporal resolution,3,4 and most require that fluorescent dyes or large quantum dot labels be attached directly to the motor. Recently, single-molecule studies of DNA polymerases have revealed new information about specific mechanisms,5,6 and there has been significant growth in the application of these methods in living cells.7−10 One of the simplest model systems for DNA replication is that of the Bacillus subtilis phage ϕ29. The phage encodes its own DNA polymerase, which is able to carry out DNA replication in vitro without any accessory proteins. Moreover, this polymerase is one of the most processive motor proteins ever discovered, capable of incorporating more than 70 000 nucleotides before dissociating from the template strand.11 An ensemble measurement typically reports little more than the average speed of a population of such enzymes, with recovery of the underlying distribution being difficult or impossible. © XXXX American Chemical Society

Tracking this and other DNA polymerases at the singlemolecule level will help shed light on the mechanisms by which they move and function. Here we demonstrate a fluorescence-based assay that can detect translocation of a single unhindered wild-type ϕ29 polymerase as it conducts strand-displacement-coupled replication of unstretched DNA. This new technique complements existing methods by providing high-resolution tracking information at natural speeds, without alteration of the polymerase structure or perturbation of the substrate. Potential applications include the study of how DNA sequence and modification might affect replication and how various replisome components and their combinations interact with a DNA polymerase.

■

METHODS Single-Molecule Displacement Encoder. In our previous work,12 we created a double-stranded DNA construct with long-wavelength visible dyes (“acceptors”) attached periodically to one strand. A helicase was labeled with a short-wavelength visible dye (the “donor”), and as it processively separated the two DNA strands, Förster resonance Special Issue: William A. Eaton Festschrift Received: August 1, 2018 Revised: August 31, 2018

A

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 1. Synthesis of the displacement encoder. Oligonucleotides designed to polymerize create a double-stranded DNA construct with fluorescent dyes placed periodically on one strand.13 The segments on the labeled strand lack 5′ phosphate groups and are therefore not ligated. Terminator oligos provide a biotin linkage for attaching the encoder to a treated coverglass, and a gap (g) where the polymerase can load.

energy transfer (FRET) produced a periodic fluorescence signal, with a pulse of red light generated each time the externally excited moving donor dye passed a fixed acceptor. Because of the analogy to macroscopic optical encoders, which translate motion into high-precision periodic signals in printers, scanners, disk drives, and numerous other types of machinery, we refer to our periodically labeled DNA constructs as single-molecule “encoders.” In the present experiment, we give up the exceptional spatial resolution of FRET in order to track the motion of a DNA polymerase without having to label it, and without having to alter or manipulate the template strand that it replicates. To do this, we take advantage of strand displacement, wherein the polymerase removes any residual DNA bound to the template strand before adding nucleotides to the strand it is creating. As shown in Figure 1 and described below, we use polymerizing self-assembly13 to create an encoder with an unligated dye-labeled strand. As ϕ29 DNA polymerase replicates the ligated strand, it displaces the dye-labeled segments (Figure 2). One downward step is produced in the resulting fluorescence signal each time a dye is displaced and diffuses out of the focal volume in approximately 1 ms. DNA Oligonucleotides. The oligonucleotide sequences used to synthesize displacement encoders are listed in Table 1. A-B63 and a-b63 anneal to form a 47 bp double-stranded DNA molecule with 16-nucleotide complementary 5′ overhangs. AB63 has a Cy5 fluorophore between bases 32 and 33. This oligo was used in three different forms: phosphorylated, unphosphorylated, and acyclo-terminated. The phosphorylated form facilitates ligation between the dye-labeled oligos by T4 ligase, producing a fully duplex DNA molecule with two contiguous backbones. When A-B63 lacks the 5′ phosphate group, nicks remain in the backbone between the segments of the dye-labeled strand, allowing them to be displaced one at a time. The last form, A-B63-acyclo, contains an additional acyclo-thymine nucleotide at the 3′ terminus (denoted as t), which acts a chain terminator, preventing ϕ29 DNA polymerase from initiating replication from anywhere but the 3′ end of P2. P2 serves as the primer for replication and carries a 5′ terminal biotin for immobilization. Buffers and Reagents. ϕ29 DNA polymerase experiments were conducted in a buffer (“P100”) that consists of 50 mM Tris-HCl (pH 7.9), 10 mM (NH4)2SO4, 10 mM MgCl2, 4 mM dithiothreitol, 100 μg/mL bovine serum albumin, and 100 μM dNTPs, with an oxygen scavenging system containing 35 μg/ mL catalase (Sigma-Aldrich C1345), 0.2 mg/mL glucose oxidase (Sigma-Aldrich G7141), 4.5 mg/mL β-D-glucose

Figure 2. Schematic of the ϕ29 DNA polymerase strand-displacement measurements. A single encoder, immobilized via a biotin− NeutrAvidin interaction on a PEG-coated coverslip, is centered in the confocal detection volume and directly excited with a 633 nm laser reduced to 1 μW. A 1.3 numerical aperture oil-immersion objective12 is used for both excitation and detection, resulting in a peak irradiance of approximately 4 × 107 W/m2 at the center of focus. The detection system is as described in ref 12, but using only a single channel. ϕ29 DNA polymerase is introduced, and the ensuing displacement and subsequent diffusive departure of each dye-labeled strand segment from the confocal volume is detected as a stepwise loss of fluorescence.

(Sigma-Aldrich G8270), and 1.2 mM Trolox antioxidant (Sigma-Aldrich 238813).14 DNA and NeutrAvidin (Pierce 31000) were stored in 10 mM Tris-HCl (pH 8.0) containing 50 mM NaCl (“T50”). A buffer consisting of 10 mM Tris-HCl (pH 8.0) with 1 mM EDTA (“TE”) was used during encoder synthesis. Encoder Synthesis. Encoders were synthesized using a polymerization-based approach.13 1 μL each of 100 μM a-b63 and A-B63 oligo (monomers), 1 μL of 10 μM a-g-p2 (terminator) and 2 μL of 10 μM P2 (primer) were combined with 4 μL of TE and 1 μL of TE containing 1 M NaCl. The final solution contained 10 μM of each of the monomer oligos, 1 μM of the terminator oligo, 2 μM of the primer, and 100 mM NaCl. The component mixture was heated to 75 °C and cooled slowly to room temperature over the course of several hours. 1 μL of T4 ligase (400 units) and 1.22 μL of 10× T4 ligase buffer (B0202S, New England Biolabs) were added, and the reaction was incubated at 16 °C for 1 h. The DNA was separated on a 1.2% agarose gel, and encoders ranging in size from approximately 8−20 periods were excised and purified using the Wizard Gel Clean-Up kit (Promega, Madison, WI) B

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B Table 1. Single-Stranded Oligonucleotide Sequences Used for Synthesis of Displacement Encoders name

nucleotide sequence (5′ → 3′)

a-b63 phos-A-B63 A-B63 A-B63-acyclo a-g-p2 P2

TTCTTGCACCAGTGTCCTCTCCGCGTCGTCCTACAGTGACTAGTCTGAACGTATAGTATAAGC GACACTGGTGCAAGAAGCTTATACTATACGTTCAGACTAGTCACTGTAGGACGACGCGGAGAG GACACTGGTGCAAGAAGCTTATACTATACGTTCAGACTAGTCACTGTAGGACGACGCGGAGAG GACACTGGTGCAAGAAGCTTATACTATACGTTCAGACTAGTCACTGTAGGACGACGCGGAGAGt TTCTTGCACCAGTGTCGTACCGGAATGACGGTTGGTCGATTAAGTAGAGCCGATACTGCT biotin−AGCAGTATCGGCTCTACTTAATCGACCAACCGTC

Figure 3. Strand displacement results in a stepwise decrease of the fluorescence signal, indicating progress of ϕ29 DNA polymerase. As it synthesizes a new strand, the polymerase displaces unligated DNA segments with fluorescent labels from encoders having nicks at 63 base pair intervals. We attribute each downward step to the translocation of the polymerase by 63 base pairs, or, rarely, to photobleaching.

μL of T50 buffer, after which the channel was filled with a 10 pM encoder DNA solution. The biotinylated encoders were allowed to bind to adsorbed NeutrAvidin molecules for 20 min. Excess DNA was washed out with 200 μL of P100 buffer. To verify that a sufficient quantity of spatially resolvable molecules had been deposited on the surface, another 20 × 20 μm scan was conducted. Surfaces prepared in this manner had approximately 100 molecules in any 20 × 20 μm area. Strand-Displacement Measurements. To verify that ϕ29 DNA polymerase can bind to and displace the dye-labeled strand from immobilized encoders, we scanned a 20 × 20 μm area of a flow cell surface before the protein was introduced, and then again after ϕ29 DNA polymerase, and any displaced DNA strands, had been washed out of the flow cell. A reduction in the number of fluorescent spots on the surface indicated that strand displacement had occurred. Fluorophore photobleaching is insignificant in this assay, because each dye is exposed to the laser for less than 100 ms, whereas on average, individual dyes required nearly 5 min to bleach at the power used here (Figure 5). After scanning a 20 × 20 μm area, we introduced 10 units of ϕ29 DNA polymerase in 20 μL of P100 buffer into the flow

following the manufacturer’s instructions, except that the gel slice was dissolved at room temperature. The resulting DNA contains a 10 nucleotide gap between the 3′ and 5′ termini of oligo P2 and the first A-B63 oligo, respectively. ϕ29 DNA polymerase loads at this gap,15 and translocates in the 3′ to 5′ direction on the unmodified DNA strand, using it as a template for replication. The continuity of the labeled strand backbone depends on the choice of A-B63 oligo used. It is fully ligated in the case of phos-A-B63 and nicked with A-B63 or nicked and terminated (no 3′ OH groups) if A-B63-acyclo was used. Flow Cell Preparation. Prior to measurements, surfaces were checked for nonspecific binding of DNA by incubating a 1 nM solution of encoder DNA in a flow cell12 for 5 min. After washing the channel with 200 μL of P100, we scanned a 20 × 20 μm area with our confocal microscope and counted nonspecifically immobilized molecules. If there were fewer than seven molecules, the flow cell was considered clean enough for subsequent single-molecule measurements. The channel was functionalized by drawing a 200 μg/mL solution of NeutrAvidin in T50 buffer into the channel and waiting 1 min. Excess NeutrAvidin was washed out with 200 C

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 4. Fluorescence loss occurs abruptly when encoders are fully ligated, because only after the entire labeled strand has been displaced can it diffuse out of the excitation volume.

Figure 5. (A) Distribution of photobleaching times for 148 Cy5 molecules excited with a 633 nm laser reduced to 1 μW. The data are fit well by a single exponential decay proportional to e−t/τ with τ = 295 ± 14 s (red line). (B) Representative trace showing the gradual photobleaching of a single encoder with approximately 10 dyes. Note the much longer time scale relative to the events in Figure 3.

cell and incubated for 3 min. The channel was washed with 200 μL of P100, after which the same area was scanned again. Immobilized molecules on the surface were counted in both scan images. Typically, 30% of labeled strands were displaced from the surface. When this procedure was repeated without ϕ29 DNA polymerase, or without nucleotides, no significant reduction in the number of immobilized encoders was observed. Real-time measurements of ϕ29 DNA polymerase strand displacement were accomplished as follows. Encoders were immobilized as described above. A solution containing 10 units of ϕ29 DNA polymerase in 10 μL of P100 buffer was loaded into the channel inlet, but not drawn into the channel. Then, a single encoder was located with a raster scan and centered using an automated centration algorithm,16 after which the laser shutter was immediately closed. The shutter was reopened, and the plunger on the channel outlet syringe was pulled gently, drawing polymerase into the cell, while the encoder fluorescence was monitored. This procedure allowed ϕ29 DNA polymerase events to be observed in real time. Introducing ϕ29 DNA polymerase before an encoder was located made observing an event too difficult, because all competent encoders are displaced too quickly. Lowering the concentration of ϕ29 DNA polymerase was also impractical, because the frequency of events became too low. Using the above procedure, ϕ29 DNA polymerase activity was seen in approximately 30% of trials, consistent with the observation

that only about 3 in 10 encoders appear to vanish when these measurements are made.

■

RESULTS Figure 3 shows four distinct events. ϕ29 DNA polymerase was drawn into the flow cell just prior to 0 s. The data exhibit stepwise loss of fluorescence, which we attribute to either displacement of dye-labeled strand segments by the polymerase or to photobleaching (see discussion below). To verify that the fluorescence steps are each correlated with the loss of a single dye-labeled DNA strand segment, and not simply the result of an unexpected change in solution conditions upon flow, another version of the encoder was synthesized with a fully ligated labeled strand. In this configuration, the dye-labeled strand would be expected to remain attached to the template strand until it has been displaced in its entirety. Only then will it diffuse away. In Figure 4, the sudden loss of all fluorescence from two separate ligated encoders, each having a multitude of dyes, confirms that ϕ29 DNA polymerase is responsible for the observed activity. Furthermore, repeating the experiment without ϕ29 DNA polymerase or without nucleotides results in long traces such as the one shown in Figure 5B. There the stepwise loss in fluorescence occurs over a much longer period of time, which is consistent with gradual photobleaching. In events such as the ones in Figure 3, time intervals between steps were calculated using the change-point D

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 6. (A) Single-molecule speeds of ϕ29 DNA polymerase measured by Schwartz and Quake.18 In that experiment, the camera frame rate placed an upper bound on observed speeds. (B) Distribution of translocation speeds of 15 individual polymerase molecules measured in this work (gray). The average speed seen by Ibarra et al.19 at 3 pN strand tension is shown in red.

Figure 7. Photobleaching complicates the interpretation of stepped fluorescence data. Most events will fall into one of the four categories shown. Panel A represents the ideal case, in which none of the dyes bleach. Red circles indicate emitting dyes, and the red lines are the noiseless fluorescence signals they would ideally produce. Panels B−D show three possible cases in which a single dye bleaches (gray circles) during (B,C) or before (D) the event begins. For each class of photobleaching event, the expected effect on the distribution of measured velocities is shown in a histogram to the right. In the absence of photobleaching, the width of the velocity distribution reflects intrinsic variability in the speed of the motor. Photobleaching during an event broadens the distribution.

detection algorithm described by Kalafut and Visscher17 modified for Poisson-distributed noise, then converted to speeds by multiplying their reciprocals by 63 base pairs. A

histogram of the resulting speeds is shown in Figure 6B. The mean speed of 85.9 ± 8.3 bp/s (mean ± s.e.m.) is consistent with optical tweezers measurements by Ibarra et al.,19 but E

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B differs from FRET measurements by Schwartz and Quake,18 who report a rate of just 28 ± 5 bp/s. However, in their analysis the latter authors discarded events with replication rates higher than 50 bp/s because of their limited time resolution. When we discard rates higher than 50 bp/s from our data, we obtain a mean rate of 24.0 ± 2.5 bp/s, indicating that the measurement technique of Schwartz et al. probably biased their result.

imply constant translocation speed, and the width of the resulting velocity distribution will reflect intrinsic variability, averaged over our 63-base pair resolution, in replication speed. However, if one of the dyes bleaches during an event, one of the steps will be misplaced as shown in panel B. Bleaching of the second dye in the encoder before it is displaced causes the second step to arrive early, while the duration of the third plateau becomes longer. The two measured time intervals will thus have a greater spread, producing a broader velocity distribution but with the mean unaffected. Less likely cases where the last dye bleaches during the event (panel C) will only contribute a high-velocity tail to the speed distribution, since there will be no longer time interval measured. If one of the interior dyes bleaches before the event begins (during the initial alignment of the encoder before the polymerase is introduced into the flow cell), then one of the steps will be missed, yielding a low-velocity tail. In addition to complications arising from photobleaching, shot noise hinders our ability to locate the discrete downward steps in fluorescence resulting from strand displacement. To detect steps in noisy data, Kalafut and Visscher17 have developed a fully objective algorithm that requires no user input other than the data, and was shown to outperform several other algorithms in common use.20 The algorithm makes the sole assumption that the data have a Gaussian distribution with fixed width, but with a mean that changes in sudden steps. In its original form, it is thus not optimal for Poisson-distributed photon counts. We therefore adapted the algorithm by using a Poisson likelihood function in the Schwartz Information Criterion hypothesis testing for each iteration. To remove small spurious steps caused by overfitting the data, upward steps, steps that were less than half of the expected height obtained from independent photobleaching measurements, and steps that were less than 200 ms in duration were rejected. In this context, missed and false steps, which concatenate or bifurcate time intervals and yield inaccurate velocities, have more severe consequences than slightly misplaced steps or steps with incorrect heights. To minimize problems of this nature, speeds were only calculated from pairs of adjacent steps whose sizes were at least 1.5 times greater than the standard deviation of the data in the plateaus flanking each step. Some events, such as the one in Figure 3D, appear to exhibit very slow bleaching of one or a few remaining dyes following rapid strand displacement. Dissociation of the polymerase before reaching the last few dyes on the DNA is unlikely owing to its extreme processivity.11 Another explanation is that the polymerase initiates replication not from the end of the primer but from a downstream nick between two of the dye-labeled strand segments. In fact, Truniger et al.21 demonstrated that an exonuclease-deficient mutant of ϕ29 DNA polymerase is capable of initiating strand displacement from as small as a two-nucleotide gap, which may transiently form on our encoders if any two nucleotides adjacent to a nick fray apart thermally, allowing the polymerase to bind. In such a case, the upstream dyes would never be displaced and would emit until bleached. Alternatively, gaps in the dye-labeled strand, produced by the incorporation of truncated versions of the A-B63-Cy5 oligo left over from its synthesis, could provide unintended downstream loading sites for the polymerase. Considering the polymerization-based encoder synthesis employed here, it is difficult to imagine a scenario in which

■

DISCUSSION On an individual basis, it is impossible to determine whether a given downward step shown in Figure 3 is caused by stranddisplacement activity of ϕ29 DNA polymerase or by photobleaching of one of the dyes. However, since photobleaching occurs randomly at a fixed rate that depends on excitation power (Figure 5A), we can compute the likelihood that a particular number of dyes will remain unbleached during a given time interval. The probability that a single dye will not yet have bleached after time t is e−t/τ, where τ = 295 ± 14 s is the measured photobleaching lifetime under our experimental conditions. Each encoder has multiple dyes that may all bleach independently of one another. As a given event progresses, dyes are removed from the encoder, and the number that can affect the measurement by bleaching decreases. In order to calculate the probability that an observed event contains no bleaching, we will assume that the dyes are removed by the polymerase at a constant rate. This is clearly not the case in many actual observations (see Figures 3 and 6), but the assumption will provide us with a good first-order estimate that can be further refined if necessary. In an event lasting for time t, during which N dyes are displaced from an encoder, one dye will be lost in every interval of t/N. To find the desired probability, we note that the first dye will be illuminated for t/N, the second for 2t/N, and so on. The probability of no observed bleaching during the event is then 70(t ; N ) = e−t /(Nτ)e−2t /(Nτ) ... e−Nt /(Nτ) = e−[t /(Nτ)](1 + 2 + ... +N ) = e−[t /(Nτ)][N(N + 1)/2] = e−[(N + 1)/2](t / τ)

(1)

With typical encoders having 8 ± 2 dyes initially, event durations averaging 25.1 s, and using the measured value for τ, we obtain 70(t , N ) ≈ 0.68. Since only ≈32% of all events have one or more dyes that bleach while being recorded, we expect that approximately 112 of the 117 observed steps were caused by strand-displacement activity, and the effect of photobleaching on the average measured speeds is therefore minimal (in this approximation, we have neglected the small probability of two or more bleachings during a given event). Nevertheless, it is important to understand the potential effects of photobleaching. It is tempting to assume that photobleaching will bias the measurement toward higher velocities, but this is not necessarily the case. Figure 7 shows four distinct fluorescence signals (red lines) that could be produced by hypothetical events in which ϕ29 DNA polymerase displaces three dye-labeled strand segments at constant velocity. Panel A shows the ideal case, in which none of the three fluorophores bleach. Equally spaced steps F

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Under our typical experimental conditions, P = 1 μW, tb = 40 ms, and s = 100 counts. This gives a = 2.5 × 109 counts per joule. From eq 5, we then have

an entire A-B63-Cy5 oligo is missing, as its absence would terminate the encoder. To test the possibility that ϕ29 DNA polymerase can initiate replication from a site other than the primer, we capped oligo A-B63 with an acyclonucleotide acyTTP, which acts as a chain terminator, on its 3′ end using terminal transferase. With terminated encoders, we still observed long dwells at nonzero fluorescence levels following an initial rapid strand-displacement phase, indicating that replication initiation from a site other than the primer is unlikely. An alternate explanation for the lingering fluorescence signals is that the displaced strand does not immediately leave the confocal volume. Magnesium-mediated interactions between the sugar−phosphate backbones of the replicated duplex and displaced strand could keep it localized within the laser focus.22 The dye-labeled oligo, having several homopurine tracts capable of forming stable base triads under our experimental conditions, could possibly become part of a Hoogsteen triple-helical structure.23−26 The wide range of speeds seen in Figure 6B is striking, given the nominally identical conditions under which the measured polymerase molecules were operating. In addition to possible artifacts discussed above, we note that a DNA polymerase working in vitro without any other replisome components or additional factors present in a living cell may be operating in the manner of an unregulated engine. Indeed, the method presented here may in the future shed light on how such regulation is realized. Even individual polymerase molecules seem to display a wide range of dwell times while displacing our 63-base pair fragments (Figure 3). While the average observed speeds are consistent with other measurements and a careful analysis of enzyme kinetics,27 an explanation of the broad distribution will require further study. It is worth considering how signal-to-noise constraints limit the displacement encoder technique. So long as the laser irradiance is kept well-below the level at which saturated excitation becomes a problem, the average measured fluorescence signal from a single dye molecule s = aPtb

σB =

SNR B =

The dominant source of noise in this experiment is shot noise, in comparison with which the background fluctuations are negligible. The one standard deviation noise associated with : is therefore Ns

(4)

SNR d =

As one dye is displaced from the encoder, we expect a drop of s in the measured signal : . It follows that for one step, the signal-to-noise ratio SNR =

s = σ:

s = Ns

s = N

aPtb N

(7)

aPtb aPtb Ptb = =a ≈ 32 σB b bPtb

(8)

with the signal as in eq 2 and the parameters given above. Clearly, this will not change if Ptb is held constant. The degree to which the power can be reduced may then be limited by intrinsic detector noise, which produces a background of approximately 25tb counts per bin with our equipment. The associated detector noise σd ≈ 5 tb . A calculation similar to eq 8 shows the corresponding

(3)

: =

bPtb

where the constant b ≈ 2.5 × 108 counts per joule. We then have signal to background noise ratio

where a is a constant that depends on the details of the apparatus, P is the laser power, and tb is the time during which we accumulate signal in one bin. s is measured in photocounts. With N dyes present on the encoder, we expect in one bin a total average fluorescence signal

σ: =

(6)

This result suggests that we can make useful measurements on encoders with up to about 10 dyes (SNR ≈ 3) and that we could improve that limit somewhat by increasing the laser power or by decreasing time resolution (increasing tb). It turns out, however, that at least for our present encoder design, we bypass the shot noise limit discussed here to some degree because of the finite size of the focal volume. The dyes are spaced by 63 base pairs, or about 20 nm. A 10 dye encoder therefore has a contour length around 200 nm. The laser irradiance falls off to 1/e2 = 0.14 of its peak value about 130 nm from the center of focus, so an encoder much longer than 10 dyes is unlikely to be illuminated in such a way that the shot noise calculated above is realized. We conclude that experiments with very long displacement encoders should be possible; however, the DNA substrate may need to be immobilized, and the focal volume will need to actively track the polymerase. It may also be useful to use this method to track slower DNA-associated molecular motors. From eq 1, it is apparent that if the photobleaching lifetime τ is increased by the same factor as the event duration t, the probability of bleaching during an event will not change. At the low excitation power P used in these experiments, τ is proportional to 1/P. From eq 5, we see that the signal-to-noise ratio from shot noise will remain constant so long as the bin width tb is increased to keep the product Ptb constant if P is reduced to increase τ. Detector-independent background is presently approximately 10 counts in a 40 ms bin, with associated noise σB = 10 ≈ 3 counts. In general, if the experiment is wellisolated from ambient light,

(2)

: = Ns

Ptb 10 = N N

SNR = 5 × 104

P

a Ptb 5

(9)

which varies as P when Ptb is held constant. At present SNRd ≈ 100, so detector noise will become significant only if the power is reduced by a factor of several hundred. If other factors such as focal drift do not come into play, it may thus be possible to use the present technique for reliable tracking at speeds less than one base pair per second.

(5) G

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

■

(8) Uphoff, S.; Reyes-Lamothe, R.; Garza de Leon, F.; Sherratt, D. J.; Kapanidis, A. N. Single-molecule DNA repair in live bacteria. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 8063−8068. (9) Liao, Y.; Li, Y.; Schroeder, J. W.; Simmons, L. A.; Biteen, J. S. Single-molecule DNA polymerase dynamics at a bacterial replisome in live cells. Biophys. J. 2016, 111, 2562−2569. (10) Ghodke, H.; Ho, H.; van Oijen, A. M. Single-molecule live-cell imaging of bacterial DNA repair and damage tolerance. Biochem. Soc. Trans. 2018, 46, 23−35. (11) Blanco, L.; Bernad, A.; Lazaro, J. M.; Martin, G.; Garmendia, C.; Salas, M. Highly efficient DNA-synthesis by the phage phi-29 DNA-polymerase − symmetrical mode of DNA replication. J. Biol. Chem. 1989, 264, 8935−8940. (12) Wickersham, C. E.; Cash, K. J.; Pfeil, S. H.; Bruck, I.; Kaplan, D. L.; Plaxco, K. W.; Lipman, E. A. Tracking a molecular motor with a nanoscale optical encoder. Nano Lett. 2010, 10, 1022−1027. (13) Wickersham, C. E.; Kerr, D. H. S.; Lipman, E. A. Synthesis of extended nanoscale optical encoders. Bioconjugate Chem. 2010, 21, 2234−2238. (14) Rasnik, I.; McKinney, S. A.; Ha, T. Nonblinking and longlasting single-molecule fluorescence imaging. Nat. Methods 2006, 3, 891−893. (15) Rodriguez, I.; Lazaro, J. M.; Blanco, L.; Kamtekar, S.; Berman, A. J.; Wang, J. M.; Steitz, T. A.; Salas, M.; de Vega, M. A specific subdomain in Phi29 DNA polymerase confers both processivity and strand-displacement capacity. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 6407−6412. (16) Ha, T.; Chemla, D.; Enderle, T.; Weiss, S. Single molecule spectroscopy with automated positioning. Appl. Phys. Lett. 1997, 70, 782−784. (17) Kalafut, B.; Visscher, K. An objective, model-independent method for detection of non-uniform steps in noisy signals. Comput. Phys. Commun. 2008, 179, 716−723. (18) Schwartz, J. J.; Quake, S. R. Single molecule measurement of the “speed limit” of DNA polymerase. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 20294−20299. (19) Ibarra, B.; Chemla, Y. R.; Plyasunov, S.; Smith, S. B.; Lazaro, J. M.; Salas, M.; Bustamante, C. Proofreading dynamics of a processive DNA polymerase. EMBO J. 2009, 28, 2794−2802. (20) Carter, B. C.; Vershinin, M.; Gross, S. P. A comparison of stepdetection methods: How well can you do? Biophys. J. 2008, 94, 306− 319. (21) Truniger, V.; Lazaro, J. M.; Salas, M.; Blanco, L. ϕ29 DNA polymerase requires the N-terminal domain to bind terminal protein and DNA primer substrates. J. Mol. Biol. 1998, 278, 741−755. (22) Duckett, D. R.; Murchie, A. I. H.; Diekmann, S.; Vonkitzing, E.; Kemper, B.; Lilley, D. M. J. The structure of the Holliday junction, and its resolution. Cell 1988, 55, 79−89. (23) Frankkamenetskii, M. D.; Mirkin, S. M. Triplex DNA structures. Annu. Rev. Biochem. 1995, 64, 65−95. (24) Morgan, A. R.; Wells, R. D. Specificity of 3-Stranded complex formation between double-stranded DNA and single-stranded RNA containing repeating nucleotide sequences. J. Mol. Biol. 1968, 37, 63− 80. (25) Lee, J. S.; Johnson, D. A.; Morgan, A. R. Complexes formed by pyrimidine purine DNAs on lowering the pH are 3-stranded. Nucleic Acids Res. 1979, 6, 3073−3091. (26) Soengas, M. S.; Gutierrez, C.; Salas, M. Helix-destabilizing activity of phi-29 single-stranded-DNA binding-protein - effect on the elongation rate during strand displacement DNA-replication. J. Mol. Biol. 1995, 253, 517−529. (27) Morin, J. A.; Cao, F. J.; Lázaro, J. M.; Arias-Gonzalez, J. R.; Valpuesta, J. M.; Carrascosa, J. L.; Salas, M.; Ibarra, B. Mechanochemical kinetics of DNA replication: identification of the translocation step of a replicative DNA polymerase. Nucleic Acids Res. 2015, 43, 3643−3652.

CONCLUSIONS The measurements presented above demonstrate that singlemolecule fluorescence can be used to track a single unlabeled DNA polymerase replicating an unperturbed template strand. While the technique in its present nascent form cannot for a given step distinguish between photobleaching and forward motion, and cannot detect the departure or binding of polymerase molecules during replication, we expect that in the near future several enhancements will improve both spatial resolution and reliability. So long as the encoders remain thermally stable, dye-labeled oligos can be made shorter to increase resolution. More than one dye on each labeled segment will increase our signal-tonoise ratio. By using different lengths for successive labeled oligos and by introducing a second dye color, we will be able to measure recognizable phases in each of two signal channels, eliminating photobleaching artifacts. Simultaneous shifts in the measured phases will indicate whether and when a polymerase stalls or is replaced. Finally, a second dye color may enable us to take advantage of the extreme spatial resolution afforded by FRET between encoder dyes as they are displaced.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Everett A. Lipman: 0000-0002-3157-9233 Present Address ¶

Charles E. Wickersham, First Solar Inc., Perrysburg, Ohio 43551, United States

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Prof. Omar Saleh for helpful suggestions and access to laboratory equipment, and the late Dr. Armand Vartanian for advice regarding DNA electrophoresis. Financial support for this work was provided by the Hellman Family Foundation and the California Nanosystems Institute.

■

REFERENCES

(1) Greenleaf, W. J.; Woodside, M. T.; Block, S. M. High-resolution, single-molecule measurements of biomolecular motion. Annu. Rev. Biophys. Biomol. Struct. 2007, 36, 171−190. (2) Neuman, K. C.; Nagy, A. Single-molecule force spectroscopy: optical tweezers, magnetic tweezers and atomic force microscopy. Nat. Methods 2008, 5, 491−505. (3) Yildiz, A.; Forkey, J. N.; McKinney, S. A.; Ha, T.; Goldman, Y. E.; Selvin, P. R. Myosin V Walks Hand-over-Hand: Single fluorophore imaging with 1.5-nm localization. Science 2003, 300, 2061−2065. (4) Joo, C.; Balci, H.; Ishitsuka, Y.; Buranachai, C.; Ha, T. Advances in single-molecule fluorescence methods for molecular biology. Annu. Rev. Biochem. 2008, 77, 51−76. (5) Berezhna, S. Y.; Gill, J. P.; Lamichhane, R.; Millar, D. P. Singlemolecule Forster resonance energy transfer reveals an innate fidelity checkpoint in DNA polymerase I. J. Am. Chem. Soc. 2012, 134, 11261−11268. (6) Markiewicz, R. P.; Vrtis, K. B.; Rueda, D.; Romano, L. J. Singlemolecule microscopy reveals new insights into nucleotide selection by DNA polymerase I. Nucleic Acids Res. 2012, 40, 7975−7984. (7) Li, G.-W.; Xie, X. S. Central dogma at the single-molecule level in living cells. Nature 2011, 475, 308−315. H

DOI: 10.1021/acs.jpcb.8b07440 J. Phys. Chem. B XXXX, XXX, XXX−XXX