Specific DNA duplex formation at an artificial lipid bilayer: fluorescence microscopy after Sybr Green I staining

Summary The article describes the immobilization of different probe oligonucleotides (4, 7, 10) carrying each a racemic mixture of 2,3-bis(hexadecyloxy)propan-1-ol (1a) at the 5’-terminus on a stable artificial lipid bilayer composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC). The bilayer separates two compartments (cis/trans channel) of an optical transparent microfluidic sample carrier with perfusion capabilities. Injection of unlabeled target DNA sequences (6, 8, or 9), differing in sequence and length, leads in the case of complementarity to the formation of stable DNA duplexes at the bilayer surface. This could be verified by Sybr Green I double strand staining, followed by incubation periods and thorough perfusions, and was visualized by single molecule fluorescence spectroscopy and microscopy. The different bilayer-immobilized complexes consisting of various DNA duplexes and the fluorescent dye were studied with respect to the kinetics of their formation as well as to their stability against perfusion.


Introduction
The post-biosynthetic lipophilization of various biomolecules such as of proteins and carbohydrates is of decisive importance for the correct function of the cell [1]. With the recent discovery of geranylated tRNAs in bacteria [2] the interest in so-called lipo-oligonucleotides (LONs) has grown tremendously [3,4]. The ability to form complex nano-architectures [5,6] as well as self-assembling aggregates such as micelles, vesicles [3] and bilayer formation of nucleolipids [7] offers numerous possibili-ties, e.g., for drug delivery [3]. Simultaneously, the interaction of such nanostructures with lipid membranes becomes an evergreater focus [5,6,[8][9][10][11].
The study of the interactions of single-and double-stranded nucleic acids with lipid bilayers is, therefore, of significant importance, particularly for the following reasons: 1. for the optimization of the in vivo delivery of lipophilic siRNAs [12][13][14], 2. for the development of analytical techniques for the detection of nucleic acids [15][16][17], 3. for structure elucidation of complex aggregates formed by such natural nanostructures [5,6] and 4. for cell-surface engineering [18].
In a preceding manuscript [19] we reported the lipid bilayer immobilization of lipo-oligonucleotides carrying a racemic bis(hexadecyloxy)propan-1-yl tag (1a) at the 5'-termini at an artificial lipid bilayer-water phase boundary. These were prepared using the cyanoethyl phosphoramidite 1b. A specific duplex formation with complementary cyanine-5 (Cy5)-labelled DNA strands (2b) -prepared by using compound 2a -was proven by fluorescence microscopy. Now, we simplify this technique by hybridizing an unlabelled DNA target strand to the bilayer-immobilized lipo-oligonucleotide and by using Sybr Green I (3, SG) [20][21][22] as a fluorescent double strand indicator [20][21][22]. This bears the advantage that the target DNA -the presence or absence of which is going to be analysed -should not be labelled separately with a fluorochrome tag such as cyanine-5 (Cy5) or TAMRA, neither by chemical synthesis nor by a polymerase chain reaction (PCR).
The chemical formulae (1)(2)(3) as well as the lipo-oligonucleotide sequences the paper is dealing with are shown in Figure 1. In the following the results of these experiments regarding the kinetics of incorporation into the bilayer (or its formation) as well as the release of the ternary complex (or its disaggregation) upon perfusion of the cis channel is described. Figure 3 illustrates a scheme of the experimental setup (for details see Experimental).

Figure 2:
Schematic illustrations (experiments A-F) of the specific DNA duplex formation at artificial lipid bilayers forming structures of various membrane-bound nucleic acid-dye complexes. A) Covalent bound Cy5-fluorophore and intercalated SG with DNA duplex were irradiated simultaneously at 635 nm and 470 nm. B) 100% match of the duplex formation with 12 bp-target oligomer. C) 50% match (6 bp) of the duplex formation with 12 bp-target oligonucleotide. D) 24-mer target oligonucleotide with AT-rich tails, each tail with 6 AT-bases. E) 18-mer probe oligonucleotide with a free AT-rich tail of 6 bases as a spacer covalently linked to the double-tailed lipid, forming a 12 bps duplex with the target sequence. F) The same lipooligonucleotide as described in (E) forms a 12 bps duplex in combination with a target oligonucleotide, resulting in free overhanging AT-rich tails at the 5'-as well as 3'-ends, each tail containing 6 nucleobases.
As a negative control it had been shown that a membrane-bound duplex formation between the lipo-oligonucleotide 7 (5'-d(1a-ATC CAG TTA TGA)-3') and the oligomer 5 failed. Now, we repeated the first experiment (4 + 5) successfully and run a z-scan of the bilayer after 45 min of incubation with a laser irradiation of the Cy5 dye at 635 nm ( Figure 2). Then, a Sybr Green I (3) solution in dimethyl sulfoxide (≈1 µg/mL) was added to the cis compartment of the slide. In the first experiment (A) laser irradiation of the intercalated dye was performed at 470 nm. In orienting experiments Sybr Green I was irradiated at 470 nm, in order to see if this dye stains the complementary DNA strands 4 + 5, which are known to form a duplex at the bilayer membrane [19]. Firstly, we evaluated the optimal measuring conditions such as the volume, the Sybr Green I concentration as well as of the lipo-oligonucleotides. At this point we first wanted to find out if a non-optimal wavelength of 470 nm would be sufficient for an efficient irradiation of the dye. It turned out, however, that an irradiation of Sybr Green I needs an irradiation wavelength which meets the optimal absorption wavelength to a higher extent. In subsequently performed experiments (B to F) the intercalated Sybr Green I was irradiated at 488 nm because this wavelength is close to its absorption maximum at 494 nm [20][21][22]. The experiments confirmed a duplex formation in an antiparallel mode on the bilayer surface being stable towards perfusion ( Figure 4). Figure 5 demonstrates the difference in the bilayer brightness upon irradiation of either cyanine-5 (635 nm) [19] or of the ternary complex with SG (4 + 5 + SG) (470 nm). For the determination of the bilayer brightness a region-of-interest (ROI) was defined around the bilayer, and subsequently the density of the intensity counts were summarized (for details, see Experimental).
Besides the bilayer brightness, also the diffusion time, t D , of the lipo-oligonucleotide duplex-Sybr Green I complex (i) within the bilayer (location 1) and (ii) above the bilayer (location 2) was determined ( Figure 6). Table 1 and Table 2 summarize the diffusion times [t D (ms)] of the complexes 4·5·SG and 4·6·SG before and after a perfusion at the two locations 1 and 2 as illustrated in Figure 6. It can be clearly seen that the diffusion time of the aggregates is significantly longer within the bilayer compared to the region slightly above (location 2), proving the strong immobilization of the lipophilized DNA duplex within the bilayer via the lipophilic head group. The fastest diffusion of the aggregates occurs in free solution without bilayer (Table 2) [19]. Experiment B: 4 + 6 + SG and control experiment: 7 + 6 + SG. In a second series of experiments we used an unlabelled target DNA (6) for duplex formation at the lipid bilayer and added Table 1: Diffusion times [t D (ms)] of the 4·5·SG complex in the presence of a lipid bilayer, either by irradiation of the cyanine-5 dye or by irradiation of intercalated Sybr Green I within the 4·5·SG complex, before and after perfusion of the cis compartment. Location 1: bilayer; location 2: solution in close proximity to the bilayer ( Figure 6).  only a DMSO solution of Sybr Green I to the cis compartment of the bilayer slide. As can be seen from Figure 7 also the hybridization of an unlabelled target oligomer 6 with the lipo-  complementary unlabelled oligomer 6 (negative control experiment), no duplex formation occurs (data not shown).
From Figure 7 it can be seen that a stepwise mixing of the lipooligonucleotide 4 and the complementary strand 6 with an inter- mediary incubation of 43 min for an optimal insertion of 4 into the bilayer, followed by a further waiting period (41 min) for duplex formation leads to a bilayer brightness of 70% upon addition of Sybr Green I. Two subsequent incubation periods of totally 40 min enhances the normalized brightness of 100% (≈15.2 × 10 4 kHz). In the following 7 perfusion steps (see Experimental), interrupted by several incubation periods over a period of about 2 h, led to an almost total decrease of the brightness dropping to about 5%. This decrease might be due to either a release of the dye from the intact duplex or to a full disintegration of the ternary complex (4 + 6 + SG).
Experiment C: 4 + 8 + SG. In this experiment the lipo-oligonucleotide 4 is inserted into the lipid bilayer, followed by a dodecamer (8) which matches the sequence of 4 only in its innermost part to 50%. An addition of Sybr Green I does not lead to a fluorescent bilayer indicating an unstable ternary complex at room temperature. This is in contrast to the experiment B where a 100% match of both dodecamers exists.
Experiment D: 4 + 9 + SG. In a further series of experiments we studied the duplex formation of the lipo-oligonucleotide 4 and the oligomer 9 at the lipid bilayer-water phase boundary layer. The oligomer 9 matches 4 to 100% within its innermost part but contains hexameric overhangs at both termini. Duplex formation was again indicated using Sybr Green I.
In this case an interesting phenomenon was observed: after immobilization of oligomer 4 within the bilayer and after addi-tion of the complementary strand 9 30 min later as well as of Sybr Green I, further 30 min later, a full development of maximal fluorescence (normalized maximal brightness of 100% ≈ 5.6 × 10 4 kHz) of the bilayer could only be observed after ≈1 h ( Figure 8). This time of formation of the ternary complex is significantly slower than in the experiments with blunt-ended duplex formation reactions described before, which show an almost spontaneous complex formation with the intercalating dye. Moreover, the complex between 4 and 9 as well as of Sybr Green I at the bilayer seems to be highly labile because already a single perfusion step of the cis compartment (1 min, 1.1 mL of buffer at t = 143 min) leads to an almost 90% disappearance of the fluorescence. Interestingly, however, is the finding that a renewed addition of a Sybr Green I solution to the cis compartment (at t = 168 min) leads again to the appearance of fluorescence 40 min later (brightness, 70%), indicating a partial reconstruction of the complex consisting of 4, 9, and the intercalating dye.
The results described above offer several possibilities of interpretation: The target strand 9 carries 3'-terminally to the recognition site of the lipo-oligonucleotide 4 an overhang of 6 nucleotides in length. Therefore, duplex formation between 4 and 9 over a full length of 12 base pairs leads inevitably to a clash of the overhang with the bilayer surface. As the head groups of the neighbouring bilayer molecules (POPC, POPE) are both positively charged and the internucleotide residues of the overhanging oligonucleotide are negatively charged, a strong Coulomb attraction should exist which can only be shielded by solvation of the lipid head groups and of the nucleic acid phosphodiester groups as well as by the metal cations of the surrounding buffer. The supramolecular assembly of all reaction partners might look like as shown in Figure 9. From this figure it can be deduced that two different tilt angles (θ 1 and θ 2 ) probably exit between the bilayer surface and the double helical part (θ 1 ) as well as between the single stranded sequence and the surface (θ 2 ). A stretched-out, parallel association of the full length nucleic acid with the bilayer surface can be most probably ruled out. In such a case a conformational constraint would arise between the lipid head group and the appending oligonucleotide. Moreover, a Cy-5-labelled oligonucleotide such as compound 5 should be bound tightly to the bilayer surface leading to a strong and stable fluorescence of the bilayer which could not be observed.
It is, however, obvious that a severe disturbance of either the complex formation between the probe lipo-oligonucleotide 4 and the target nucleic acid 9 and/or the intercalation of the Sybr Green dye by the overhang take place. This might be the reason for the instability of the assembled complex at the lipid bilayer. Experiment E: 10 + 6 + SG as well as in a reversed order of addition (SG + 10 + 6). Next, the oligomer 10 was prepared which contains the lipophilic 5' head group 1a and the recognition sequence separated by an oligonucleotide spacer having the same length as the target sequence overhang (6 bp).
From Figure 10 it can be seen that in this case the ternary complex formation occurs spontaneously upon addition of a first portion of the dye solution. This is, however, not a stable situation because an incubation of only 6 min (without perfusion!) leads to a reduction of the bilayer brightness of just 20%. Several further waiting periods, interrupted by 6 perfusions, reduced the bilayer brightness from the original 100% (100% normalized, ≈6.6 × 10 5 kHz) to less than 5% within 2.5 h. A further addition of Sybr Green I at t = 212 min, however, enhances the bilayer brightness again to about 20%, indicating the presence of a substantial amount of the intact DNA duplex at the lipid bilayer surface on the cis side. Subsequent incubation periods, followed by single perfusion steps reduced the bilayer brightness each time to about 5%. A repeated Sybr Green I addition at t = 234 min leads again to an enhancement of the brightness value of 18%. A 4 th addition of Sybr Green I at 251 min results in an only very slight brightness enhancement of ≈5% ( Figure 11).
In Figure 12 the z-scans of the Experiment E -before and directly after Sybr Green I additions -are displayed. As already shown in Figure 10 and Figure 11 the bilayer brightness increases with each Sybr Green I addition. After 6 perfusions and incubation steps the brightness is reduced to almost zero (t = 211 min), however, with the next additions of the dye the brightness increases again. Only after seven perfusion and incubation periods the brightness does no longer increase significantly, which might be traced back to the fact that the immobilized DNA duplex is meanwhile dissociated, and that the nonlipophilized strand is has been washed out.
In a further experiment using the lipo-oligonucleotide 10 we changed the order of the addition of the three components of the ternary complex to the cis compartment of the bilayer slide ( Figure 13). Within the first 11 min we studied the interaction between Sybr Green I and the empty bilayer. As can be seen no ponderable brightness of the bilayer could be detected after addition of the dye. However, after addition of the 24-mer 10 (at t = 13 min) the brightness increased to about 60% (normalized brightness, 100% ≈6.5 × 10 4 kHz) and stays almost constant (35-50%) until t = 28 min indicating a strong interaction between the dye and the single stranded 24-mer 10 [20]. Such an interaction could not be observed between the dye and the duplex of the 12-mers 4 and 7. A subsequent perfusion reduces the brightness significantly to 5%. Interestingly, a subsequent waiting period of 10 min leads again to an increase of the bilayer brightness back to about 25%. This is probably due to an additional delivery of the complex 10·SG from the cis compart-ment or from the cis side of the hydrophobic Teflon-made annulus into the lipid bilayer.
At t = 63 min a second portion of the dye solution and at t = 73 min the oligomer 6, being partly complementary to the lipo-oligonucleotide 10, were added. Within an incubation time of 22 min after addition of 6 the full bilayer brightness (100 % at t = 95 min) was reached. Subsequent perfusion steps (1-4) interrupted by several incubation periods reduced the bilayer brightness from 100 to 10% (Figure 14). These results clearly indicate a significantly higher stability of the ternary complex at the bilayer compared to the complex of experiment D.
Experiment F: (10 + 9 + SG). In the following the lipo-oligonucleotide 10 (24-mer) was immobilized on the lipid bilayer. After an incubation time of 30 min, an equimolar amount of the oligomer 9 -an oligomer of the same length and a central recognition sequence of 12 bases -was added, followed by addition of Sybr Green I after a further waiting period (30 min). Figure 15 displays the full protocol of mean brightness values (both, in % and kHz) as a function of time or the various events (incubation periods or perfusion). After the addition of the dye as last component of the ternary complex at t = 66 min, a maximal brightness is slowly developed within about 2 h. In this case, however, the maximal brightness (normalized 100% ≈6 × 10 4 kHz) does not reach the value which had been observed in the experiments A, B, and E. Only in case of experiment D the maximal brightness is even lower than in case F (2.5 × 10 4 kHz).
18 Subsequent perfusions, interrupted by waiting periods of 10 min each, gave after 382 min a bilayer brightness of ≈5% (= 1 × 10 4 kHz). This indicates that the formation of the ternary complex 10·9·SG occurs slowly, but when it is once formed, it remains stable for more than 6 h. Figure 16 shows a comparison of the kinetics of complex formation at the lipid bilayer surface for 5 experiments within the first 60 min (brightness [kHz] vs time [min]).

Conclusion
As can be seen, the fastest ternary complex formation occurs for 10 + 6 + SG (Experiment E) where the complex formation occurs at a distance of a hexamer spacer between the bilayer surface and the DNA duplex region with intercalated Sybr Green I. It can also be observed that, in this case, a maximal brightness at ≈6.5 × 10 5 kHz is reached.
If the spacer is missing, the complex (4 + 6 + SG, Experiment B) formation occurs significantly slower; the maximal brightness is reached at 1.5 × 10 5 kHz. The same results are found in case of experiment A (4 + 5 + SG) with two different fluorescent dyes -intercalating Sybr Green I and pending cyanine-5 ( Figure 5).
For the scenario of experiment F ( Figure 2) the complex formation kinetics drops even further, but the brightness reaches a plateau at ≈2 × 10 5 kHz. The by far slowest complex formation is observed in case of experiment D (Figure 1) in which a steric clash between the ternary complex and the bilayer surface occurs. Here, no plateau value of the bilayer brightness is reached, and the complex proved to be highly labile.
Further experiments concerning a multiple compartment chamber with a common aqueous sub-phase, as well as a thermostated device for a suppression of nonspecific base pairing are underway [15]. Moreover, ab initio molecular-dynamics-(AIMD)-and ab initio Monte-Carlo-(AIMC)-calculations of  lipid-nucleic acid complexes with MC and MD simulations of semi-quantitative, "course-grained" models (CGM) are going to be performed (Prof. Dr. Philipp Maass, Department of Statistical Physics, University of Osnabrück).
Studies described in this manuscript as well as of those of a forthcoming paper which deals with the interaction of a defined oligonucleotide single strand, carrying nucleolipid head groups of different lipophilicity with lipid bilayer membranes, might be of importance for the optimization of the in vivo delivery of lipophilic siRNA [13] (e.g., by DNA trafficking as shown in Figure 17), as well as of lipid derivatives of nucleoside antimetabolites [23,24]. Moreover, studies of a transdermal application of lipo-oligonucleotides through the human Stratum corneum by iontophoresis techniques are underway.
Bilayer fabrication and incorporation of lipo-oligonucleotides therein. Similar as described in [19], in the following the automated bilayer fabrication is described. Horizontal bilayers were fabricated automatically using a lipid mixture of 1-palmitoyl-2-oleyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleyl-sn-glycero-3-phosphocholine (POPC) (8:2, w/w, 10 mg/mL of n-decane) within the "Bilayer Slides" (Figure 18C,D) and an add-on for the inverted confocal microscope (Bilayer Slides and Ionovation Explorer, Ionovation GmbH, Osnabrück, Germany). After pre-filling with buffer (250 mM KCl, 10 mM MOPS/Tris, pH 7), the slide was inserted into the stage unit mounted on an inverted confocal microscope ( Figure 18A). Ag/AgCl electrodes were mounted and after the addition of 0.2 µL of POPE/POPC lipid to the cis compartment using a 1 µL bended Hamilton syringe (CH-Bonaduz) ( Figure 18E), the automated bilayer production was started; a modified painting technique, in which the air-water interface paints the lipid across the aperture was applied. The bilayer formation was monitored optically and electrically.
As soon as a stable bilayer has been established (C > 50 pF) the lipophilized oligonucleotide solution (8 µL, 500 nM) was injected into the cis compartment of the "Bilayer Slide". During the incubation time of 30-40 min the bilayer integrity was monitored by continuous capacitance measurements. After the incubation of the first sample the second sample of a complementary or non-complementary oligonucleotide was injected into the cis compartment and incubated for the next 30 minutes. Subsequently, the Sybr Green I solution (8 µL, undiluted) was injected and incubated until the highest brightness amount was reached. In all graphs the brightness was given either in kHz or it was normalized (0-100%). In the meantime every 10 minutes a z-scan was performed (see below). After the incubation steps the cis compartment was perfused repeatedly for 30 s (1.1 mL/min, each), and the bilayer was inspected by a confocal fluorescence microscope. Each perfusion step using the Ionovation perfusion unit ( Figure 18B) was followed by an incubation step of 10 minutes. This procedure was repeated until the brightness of the bilayer finished decreasing continuously and turned out to be stable. In summary, each measuring protocol was carried out as follows: (I) a reference scan of the stable pure bilayer was performed; (II) then the lipophilized oligonucleotide sample was added and incubated for 30 min, (III) this step was followed by the addition of the complementary and non-complementary sample, respectively, and further incubation of 30 min; (IV) finally the cyanine-dye Sybr Green I solution (DMSO) was added to allow the quantification of double stranded DNA, and the solution was incubated until the brightness in the bilayer reached the highest amount; (V) afterwards additional scan series were performed after each perfusion of the cis compart- On the left, electrodes are visible, connected with the amplifier; on the right hand side tubes connected with the perfusion unit can be seen. F) Vertical cut of a buffer-filled bilayer slide, demonstrating its general design. The bilayer slide encloses two microfluidic channels (cis and trans) which are separated by a thin medical-grade PTFE (= polytetrafluoroethylene, Teflon) foil. This foil hosts a central 100 µm aperture which is located 120 µm above the cover slip and thus within the working distance of high NA (= numerical aperture) objectives. It is the only connection between the trans and the cis channels. When a lipid solution is painted across the aperture, a bilayer is formed spontaneously. The electrodes in the cis and trans channels allow an online monitoring of the bilayer integrity, as well as electrophysiological recordings.
ment (for 30 s; 1.1 mL/min) and (VI) followed by an incubation of 10 min.
In order to analyze the 2D images they were previously edited with Image JA 1.44 and the obtained data was evaluated with OriginPro 8 (OriginLab Corporation). For an evaluation of the z-scan data, the brightness within the bilayer was measured; the mean area of the bilayer cross section amounts to 353 ± 19 counts per pixel (N = 201). Next, the brightness was determined by summing up the number of pixels. 1 Pixel equals 1 kHz or 1 µm.
All devices ( Figure 18) as well as the general techniques used in this paper have been described in detail in a preceding manuscript [19].