Strand displacement and duplex invasion into double-stranded DNA by pyrrolidinyl peptide nucleic acids †

The so-called acpcPNA system bears a peptide backbone consisting of 4’-substituted proline units with (2’R,4’R) configuration in an alternating combination with (2S)-amino-cyclopentane-(1S)-carboxylic acids. acpcPNA forms exceptionally stable hybrids with complementary DNA. We demonstrate herein (i) strand displacements by single-stranded DNA from acpcPNA–DNA hybrids, and by acpcPNA strands from DNA duplexes, and (ii) strand invasions by acpcPNA into double-stranded DNA. These processes were studied in vitro using synthetic oligonucleotides and by means of our concept of wavelength-shifting fluorescent nucleic acid probes, including fluorescence lifetime measurements that allow quantifying energy transfer efficiencies. The strand displacements of preannealed 14mer acpcPNA–7mer DNA hybrids consecutively by 10mer and 14mer DNA strands occur with rather slow kinetics but yield high fluorescence color ratios (blue : yellow or blue : red), fluorescence intensity enhancements, and energy transfer efficiencies. Furthermore, 14mer acpcPNA strands are able to invade into 30mer double-stranded DNA, remarkably with quantitative efficiency in all studied cases. These processes can also be quantified by means of fluorescence. This remarkable behavior corroborates the extraordinary versatile properties of acpcPNA. In contrast to conventional PNA systems which require 3 or more equivalents PNA, only 1.5 equivalents acpcPNA are sufficient to get efficient double duplex invasion. Invasions also take place even in the presence of 250 mM NaCl which represents an ionic strength nearly twice as high as the physiological ion concentration. These remarkable results corroborate the extraordinary properties of acpcPNA, and thus acpcPNA represents an eligible tool for biological analytics and antigene applications.


Introduction
Peptide nucleic acid (PNA) represents a very important nucleic acid analog and chemical-biological tool based on its important features like affinity, sequence selectivity and stability against degradation by nucleases and proteases. 1Especially the uncharged pseudopeptide backbone makes PNA an attractive scaffold for diagnostic applications.2][3][4] The attempt to increase affinity and, more importantly, sequence selectivity of PNA by incorporation of cyclic derivatives in the peptide backbone failed in many cases since the nucleic acid hybridization properties are deteriorated. 5,6mong several of those PNA architectures reported to date, and especially among those with cyclic structures as part of the backbone, [7][8][9] the so-called acpcPNA system (Fig. 1) exhibits significant features.The peptide backbone consists of 4′-substituted proline units with (2′R,4′R) configuration in an alternating combination with (2S)-amino-cyclopentane-(1S)carboxylic acids.1][12][13] The recognition and binding to DNA shows high sequence selectivity and does not tolerate single base mismatches.The sequence selectivity can be visualized by fluorescent dyes such as pyrene, 14,15 thiazole orange 16 and nile red, 17 that were synthetically incorporated into acpcPNA.Most importantly, selfpaired hybrids are not formed by acpcPNA. 11This feature stands in contrast to many other PNA scaffolds that bind to themselves equally or even better than to naturally occurring nucleic acids. 4ase and/or PNA backbone modifications are traditionally required to destabilize PNA-PNA duplexes in order to allow efficient double duplex invasion of DNA. 18,191][22][23] To prevent extensive or biologically harmful modifications, a chiral derivative of PNA called γPNA that exhibits even higher DNA affinity than normal PNA with extended strand length was employed.However, 20 equivalents of γPNA and incubating times of 8 hours are necessary to accomplish quantitative invasion and complex formation with double-stranded DNA. 24Moreover, limitations for applications in vivo still remain due to the high affinity of γPNA to both DNA and RNA.In order to demonstrate the advantages of acpcPNA in this regard we report herein (i) strand displacements by single-stranded DNA from acpcPNA-DNA hybrids, and by acpcPNA strands from DNA duplexes, and (ii) strand invasions by acpcPNA strands into doublestranded DNA with high efficiency.These processes have high biological significance; however, we studied them in vitro with model PNA strands by means of fluorescence color changes as well as fluorescent intensity enhancements using our concept of wavelength-shifting nucleic acid probes ("DNA traffic lights"). 25

Strand displacements
7][28][29] Furthermore, these processes play a crucial role in DNA nanotechnology. 30onsecutive annealing and strand displacement steps represent the methodological basis to dynamically control the formation of highly ordered and self-assembled DNA nanoarchitectures, such as DNA switches, DNA tweezers, [31][32][33][34] DNA-based logic gates, [34][35][36][37][38] DNA walkers [39][40][41][42] and others.In order to readout the strand displacement dynamics with acpcPNA not only by conventional fluorescence intensity changes at a single emission wavelength, we followed our concept of wavelength-shifting nucleic acid probes. 25Five years ago, we introduced this concept that is based on energy transfer between thiazole orange and thiazole red as fluorescent DNA base substitutions and applied it for molecular beacons. 25,43Moreover, we showed that this concept works also for pairs of fluorescent dyes attached to the 2′-positions in two different DNA strands and yields even better emission color contrasts as readout. 44n order to apply such fluorescence color changes to directly report the dynamics of strand displacements in vitro we synthesized the 14mers PNA1 and PNA2 that carry the blue emitting dye 1 in the middle of the sequence (Fig. 2).These acpcPNA modifications were introduced by post-synthetic copper(I)-catalyzed cycloaddition between the precursor of dye 1 bearing an azide group and the synthesized precursors of PNA1 and PNA2 that carry an alkyne functionality.The only difference between the two PNA strands is the position of dye 1 as modification of the aminopyrrolidinecarboxylic acid spacers either on the carboxy or on the amino side of the central thymine unit in the sequence.All synthesized PNA strands bear five lysine residues at the C-terminus to enhance solubility.Five DNA counterparts of different lengths were synthesized.The shortest one, 7mer DNA1, was not labelled.The 10mers DNA2 and DNA4 were labelled by the yellow emitting dye 2 attached to the 2′-position of uridine next to the central adenosine unit (either on the 5′or 3′ side).The longest counterparts, 14mers DNA3 and DNA5, were modified at corresponding positions by the red emitting dye 3 (for the syntheses of dyes 1-3 and PNA/DNA labelling procedures see ESI †). 44,45he dye 1-labeled PNA1 and PNA2 were separately annealed with the unmodified 7mer oligonucleotide counterstrand DNA1 to obtain the starting duplexes for strand displacement experiments.Both samples were consecutively titrated with 1.0 equivalent of DNA2 or DNA4, followed by DNA3 or DNA5, respectively (Fig. 3).The first titrations with DNA2 or DNA4 place the energy acceptor dye 2 into proximity to donor dye 1.
If the excitation wavelength is kept constant at 389 nm, the efficient (vide infra) energy transfer between these two dyes causes significant fluorescence color changes from blue to yellow which were observed in all four cases (PNA1-DNA1 and PNA2-DNA1, each with DNA2 or DNA4).The second titrations with DNA3 or DNA5, respectively, remove dye 2 from the hybrids and place the energy acceptor dye 3 into close proximity to dye 1.As a result the fluorescence color changes from yellow to red; again at the same excitation wavelength (389 nm).
From these titrations the following two observations have been made and further investigated.(i) The first strand displacements of 7mer DNA1 by 10mer DNA2 or DNA4 take place relatively fast (within 5 minutes).The second strand displacements of 10mer DNA2 or DNA4 as part of the intermediate acpcPNA hybrids by 14mer DNA3 or DNA5 occur significantly slower than the first displacement processes (approximately 1 h).This can be explained by the increasing affinity of acpcPNA to longer DNA pieces during the strand displacements and the exceptionally high stability of acpcPNA-DNA hybrids in general.Melting temperatures of the starting, intermediate and final acpcPNA-DNA hybrids were determined (Table 1).The representative titration of hybrid PNA2-DNA1 with DNA2 and DNA3 is shown in Fig. 3   in the presence of an attached dye.In case of the majority of acpcPNA-10mer DNA hybrids, e.g.PNA1-DNA2, the T m values show that the fluorescent dye labels are destabilizing and require carefully chosen placements in complementary hybrids. 17Especially the dye attachment in the hybrid PNA1-DNA4 causes less destabilizing effects and gives a T m of >90 °C.(ii) Large fluorescence color ratios blue : yellow/blue : red and fluorescence intensity enhancements were accomplished by an efficient energy transfer (vide infra) from donor dye 1 to both acceptor dyes 2 and 3, respectively (Table 1).The highest blue : red contrast (I 478 nm /I 600 nm ) of 1 : 223 is obtained during displacement of 10mer DNA4 by 14mer DNA5 yielding hybrid PNA2-DNA5.The enhancements of the yellow and red fluorescence signals were quantified by the fluorescence intensity increase factors (Table 1) that were determined for both acceptor dyes 2 or 3, respectively, at their corresponding emission maximum.The factors represent the ratios of the fluorescence intensities of the acceptor dyes after excitation at the absorption maximum of donor dye 1 to the fluorescence intensities of the acceptor dyes upon excitation at the same wavelength but in the absence of the donor dye 1.The latter fluorescence intensities were obtained with single-stranded DNA1 to DNA5.The largest fluorescence enhancement factor of 21.1 was obtained during the strand displacement of PNA1-DNA2 by DNA3.
In order to calculate energy transfer efficiencies, fluorescence lifetimes of dye 1 (λ exc = 389 nm, λ em = 478 nm) were determined for all modified PNA-DNA hybrids (see ESI †).In the absence of energy acceptor, dye 1 in hybrids of PNA1 and PNA2 with full length DNA counterstrands (see ESI †) exhibits lifetimes of approximately 1.6 ns and 1.5 ns, respectively.Due to the presence of acceptor dyes 2, or 3, respectively, these lifetimes are significantly shortened to the sub-ns range in all doubly labeled acpcPNA-DNA hybrids yielding excellent energy transfer efficiencies between 75% and 99%.
The following strand displacement experiment representatively shows the ability of acpcPNA to displace one DNA strand out of the 14mer duplex DNA6-DNA3.The recorded fluorescence of dye 3 shows a significant intensity increase due to the fact that the unmodified DNA6 is replaced by PNA1 bearing the energy donor which is dye 1 (Fig. 4).This clearly demonstrates that strand displacement of DNA6 by PNA1 occurs efficiently within several minutes due to the large increase in stability from 32.7 °C (DNA6-DNA3) to >90 °C (PNA1-DNA3).This kind of strand displacement is an important prerequisite for the in vitro strand invasion experiments described in the second part below.

Strand invasions
Antigene targeting of double stranded (ds) DNA by PNA is an important goal for biological applications inside living cells.It was shown 10,11 that the conformationally restricted acpcPNA is not able to form triplex structures like conventional PNA that binds to dsDNA in several different binding modes, including triplex formation, triplex invasion, duplex invasion and double duplex invasion. 4,46In order to confirm that acpcPNA is able to undergo double duplex invasion and to follow this process by fluorescence spectroscopy, dsDNA7 to dsDNA10 were synthesized (Fig. 5).They were labeled by the blue emitting dye 1 or the red emitting dye 3, respectively, at the central A (dsDNA7, dsDNA9) or at the U on the 3′-side (dsDNA8, dsDNA10).PNA3 to PNA5 contain 14mer sequences that are complementary to the central part of the sequences of dsDNA7 to dsDNA10, and again 5 lysine residues to get enough solubility.Based on studies in other PNA systems 47,19 it can be assumed that the lysine residues facilitate the strand invasion process by stabilizing the PNA-DNA duplex and destabilizing the PNA-PNA self duplex, although whether their presence is essential for the invasion is difficult to confirm due to the poor solubility of dye-labeled PNA without lysine.The stability of these double strands is only slightly diminished, if at all, by the attached dyes: the T m difference is only between −1.7 °C and +0.1 °C, if compared to a 30mer, completely unmodified DNA double strand (DNA16-DNA15, T m = 70.4°C, see Table S5 in ESI †).PNA3 was modified with the yellow emitting dye 2 that serves as energy acceptor for dye 1 in dsDNA7 and dsDNA8, whereas PNA4 is modified with the green emitting dye 4 that represents the energy donor for dye 3 in dsDNA9 and dsDNA10 (for synthesis of dye 4, see ESI †).With respect to the biological context, PNA3 and PNA4 are complementary to the TIAM1 sequence that is regulated by miRNA-21 and miRNA-31 in colon carcinoma cells. 46The unmodified PNA5 serves as the fourth strand that was originally thought to be essential for the double duplex invasions.PNA6 bears a nonsense sequence for control experiments.Invasions were detected by blue : yellow or green : red fluorescence color contrasts (e.g.I 475 nm /I 585 nm ) and by fluorescence intensity increase factors (Table 2) that were determined similarly as described above, for both acceptor dyes 2 or 3, respectively, at their corresponding emission maximum.The factors represent Energy transfer efficiencies (Table 2) of all invasion complexes were calculated from the corresponding fluorescence lifetimes of donor dyes 1 (λ exc = 370 nm, λ em = 475 nm) and 4 (λ exc = 435 nm, λ em = 528 nm).All strand invasion experiments were performed in presence and absence of 250 mM NaCl since we assumed that duplex invasions by acpcPNA might be salt-dependent processes.At first, we studied kinetics of strand invasion into dsDNA7 and dsDNA8 by 1.0 equiv. of PNA3 together with 1.0 equiv. of PNA5.Constant fluorescence intensities are reached within 60 minutes with NaCl and 90 minutes without NaCl (see ESI †).Although NaCl-free conditions show comparable kinetics, they yield stronger increase of emission intensities and considerably higher color contrasts blue : yellow.Obviously the dye interactions with DNA in the hybrids with DNA are salt dependent and alter the fluorescence intensities.Further titrations revealed that the fluorescence color change is maximized upon addition of 3.0 equiv. of acpcPNA (see ESI †).Hence, full double-duplex invasions into dsDNA7 and dsDNA8, respectively, were carried out by addition of 3.0 equiv. of PNA3 together with PNA5.In absence of acceptor dyes, the blue emitting dye 1 exhibit lifetimes of approximately 2.4-2.6 ns in dsDNA7 and dsDNA8.The lifetimes are shortened down to 0.6-0.9ns in the invasion complexes with PNA3 and PNA5, which yields energy transfer efficiencies in the range from 66% to 76% (see ESI †).Moreover, strand invasion efficiencies could be determined based on these energy transfer efficiencies.As fluorescence reference for complete invasion, reference hybrids of PNA3 with fulllength complementary DNA strands bearing dye 1 at the central A were annealed.The energy transfer efficiencies of these hybrids were measured and represent 100% reference values.Remarkably, invasion efficiencies of approximately 99% could be accomplished in all cases (see ESI †).Best results with this dye combination were accomplished e.g. with dsDNA7 in absence of NaCl (Fig. 6).The invasion by PNA3 together with PNA5 exhibits an intensity increase of the yellow emission by a factor of 19.6, a color contrast blue : yellow (I 477 nm /I 580 nm ) of 1 : 7.5 and an energy transfer efficiency of 74%.
Similarly, double-duplex invasion kinetics of dsDNA9 and dsDNA10 after adding 1.0 equivalent PNA4 together with PNA5 were recorded.It is important to note that titrations revealed that 1.5 equiv.PNA4 and PNA5 were sufficient for this type of strand invasions to obtain maximal fluorescence change (see ESI †).Hence, quantification of energy transfer efficiencies,  fluorescence intensity increase factors and color contrasts green : red were carried out under these conditions, in absence and presence of NaCl, respectively (Table 2).Remarkably, energy transfer from donor dye 4 to acceptor dye 3 occurs with 97-99% efficiency which yields very large increase factors and color contrasts.Here again, invasion efficiencies of 99% were achieved in all cases (see ESI †).Especially the modification at the central adenosine by dye 3 (dsDNA9) and invasion in the absence of NaCl turned out to be the optimal condition for double duplex invasion (Fig. 6).After addition of 1.5 equiv. of PNA3 together with PNA5 an increase factor of 23.3 and a color contrast green : red (I 528 nm / I 600 nm ) of 1 : 87 could be accomplished.The observed variation of intensity enhancements and color contrasts could be assigned to the different anchor point of the post-synthetic DNA labeling by dye 3. Reference measurements with corresponding DNA duplexes bearing the dye 3 modification at an adenosine (dsDNA9) vs. uridine (dsDNA10) clearly show approximately 20% stronger fluorescence intensities (see ESI †).Importantly, the possibility of strand invasion of dsDNA9 by using only one PNA strand (PNA4) yields a D-loop and could be evidenced by energy transfer efficiency of 94% and invasion efficiency of 96% (see ESI †) which is only slightly lower compared to 99% invasion efficiencies obtained for double duplex invasions.Hence, the latter result demonstrates that single stranded acpcPNA may be sufficient to open double stranded DNA without requiring the other acpcPNA strand.
The duplex of the 14mer PNA with DNA containing a single mismatch base was still quite stable under the experimental conditions, therefore the invasion also proceeded with dsDNA target containing a mismatch at the invasion site (data not shown).In order to rule out that unspecific PNA aggregation to dsDNA occurs instead of the expected invasion, a random sequence and non-sense PNA6 was synthesized and modified with the blue donor dye 1 (Fig. 5).A rather small fluorescence enhancement after addition of 1.5 equiv.non-sense PNA6 to dsDNA9 was observed (Fig. 6, Table 2).As proven in strand displacement experiments above, an efficient energy transfer is only possible between dye 1 and dye 3 if they get in close proximity to each other.Thus, the very small intensity increase factor of 1.8 in case of PNA6 clearly excludes non-specific PNA-DNA aggregation.
In order to support the duplex invasions by PNA3 and PNA4, melting temperatures of dsDNAs, acpcPNA-DNAhybrids, their DNA-DNA-analogs and of complexes invaded by 1.0 equiv.acpcPNA were determined, in the absence or presence of NaCl, respectively (see ESI †).Fig. 7 representatively displays the normalized absorption changes during T m determination of the complex dsDNA8 invaded by only 1.0 equiv.PNA3 together with PNA5 without NaCl.It should be noted here that strand invasion with 1.0 equiv. of PNA3 is not complete.Accordingly, discrete melting temperature regions for each part of the invading complex are explicitly visible: (i) Comparison 8mer dsDNA13 (overhang on both ends of dsDNA8) that invasion takes place without dissociation of the remaining duplex regions on both sides of invasion, and with a locally displaced complementary strand.(ii) The melting transition of dsDNA8 which corresponds to the non-invaded

Experimental
All experimental details, syntheses and spectroscopic data are described in the ESI.†

Conclusions
In conclusion to the first part of experiments, it became clear that displacements of DNA strands from acpcPNA-DNA hybrids can be followed by fluorescence color changes using our concept of wavelength probes ("DNA traffic lights") and are possible with longer oligonucleotides than the preannealed ones.The kinetics of these strand displacements are rather slow due to the high stability of the acpcPNA-DNA hybrids.Moreover, acpcPNA is able to displace a DNA strand from double stranded DNA of the same length.There is important significance of these in vitro experiments for cell biological applications.The intracellular transport could potentially be facilitated by annealing acpcPNA to a short DNA counterstrand.In combination with attached positively charged lysine residues on the termini of acpcPNA strands, such acpcPNA/ DNA hybrids may potentially be easier delivered into the cells.
In the second part, we demonstrated that acpcPNA is able not only to displace one of the strands in double-stranded DNA of the same length, but also to invade into long doublestranded A-T rich DNA, remarkably with quantitative efficiency.Again, these processes can be followed by means of fluorescence color changes.Compared to conventional PNA systems which require 3 or more equiv.PNA, only 1.5 equiv.acpcPNA are sufficient to get efficient invasion.Although initial DNA duplex invasion experiments were performed with two non-self-pairing acpcPNA strands, single stranded acpcPNA was also able to invade into DNA duplex without requiring the other PNA strand, at least in the A-T rich sequence context being studied.Beyond that, our findings clearly confirm that the invasion takes place even at high ionic strength, a condition that is known to decrease the PNA duplex invasion efficiency.Therefore, acpcPNA exhibits sophisticated properties which are highly desirable in the field of biological analytics and antigene applications.

Fig. 5
Fig. 5 Schematic drawing of double duplex invasion, sequences of PNA3-PNA6, structure of dye 4 and sequences of dsDNA7-dsDNA10 (from top to bottom).

Table 1
Color contrasts, fluorescence intensity increase factors, energy transfer efficiencies and melting temperatures T m of PNA/DNA hybrids a See definition in text.b Calculated by fluorescence lifetimes (see ESI). c λ = 260 nm, 5-95 °C, interval: 0.5 °C min -1 , 2.5 μM duplex in 10 mM Na-P i , pH 7. d I 478 nm /I 560 nm .e I 478 nm /I 606 nm .f I 478 nm /I 579 nm .g I 478 nm / I 600 nm .

Table 2
Color contrasts, fluorescence intensity in-crease factors and energy transfer efficiencies of acpcPNA strand invasion into dsDNA I sample /I dsDNA (definition see text).cCalculated by fluorescence lifetimes (see ESI). d I 475 nm /I 585nm .eI 477 nm /I 580 nm .fI 472 nm /I 583 nm .g I 475 nm /I 574 nm .h I 528 nm /I 604 nm .i I 528 nm /I 600 nm .j I 477 nm /I 613 nm .k I 529 nm /I 611 nm .l I 528 nm /I 608 nm .
a 250 mM.b