Impact of non-nucleotidic bulges on recognition of mixed-sequence dsDNA by pyrene-functionalized Invader probes

Abstract

Invader probes, i.e., DNA duplexes modified with +1 interstrand zippers of intercalator-functionalized nucleotides like 2’-O-(pyren-1-yl)methyl-RNA monomers, are energetically activated for sequence-unrestricted recognition of double-stranded DNA (dsDNA) as they are engineered to violate the neighbor exclusion principle, while displaying high affinity towards complementary DNA sequences. The impact on Invader-mediated dsDNA-recognition upon additional modification with different non-nucleotidic bulges is studied herein, based on the hypothesis that bulge-containing Invader probes will display additionally disrupted base-stacking, more extensive denaturation, and improved dsDNA-recognition efficiency. Indeed, Invader probes featuring a single central large bulge – e.g., a nonyl (C₉) monomer – display improved recognition of model DNA hairpin targets vis-à-vis conventional Invader probes (C₅₀ values ~1.5 μM vs. ~3.9 μM). In contrast, probes with two opposing central bulges display less favorable binding characteristics. Remarkably, C₉-modified Invader probes display perfect discrimination between fully complementary dsDNA and dsDNA differing in only one of eighteen base-pairs, underscoring the high binding specificity of double-stranded probes. Cy3-labeled bulge-containing Invader probes are demonstrated to signal the presence of gender-specific DNA sequences in fluorescent in situ hybridization assays (FISH) performed under non-denaturing conditions, highlighting one potential application of dsDNA-targeting Invader probes.

Graphical Abstract

Invader probes featuring non-nucleotidic bulges are energetically activated for highly specific recognition of complementary double-stranded DNA targets.

INTRODUCTION

Development of non-protein-based probes for recognition of specific double-stranded DNA (dsDNA) sequences is a long-standing goal that has been fuelled by the prospect of gaining the ability to detect, regulate, and manipulate chromosomal DNA. Two main types of probes have been developed towards this end, i.e., those recognizing nucleotide-specific features from one of the grooves in the dsDNA target region, and those capable of disrupting the existing Watson-Crick base-pairs in the dsDNA target region through formation of more stable base-pairs between probe and complementary DNA (cDNA) strands. Triplex forming oligonucleotides (TFOs)^1,2 and peptide nucleic acids (PNAs),³ and pyrrole-imidazole polyamides⁴ are examples of the first probe type. Triplex-forming probes require extended polypurine stretches for stable binding, which limits the number of suitable target sites, whereas polyamides typically are used to recognize shorter dsDNA regions (<8 bp), which renders recognition of unique genomic regions challenging. These limitations have spurred development of duplex-invading approaches that are based on high-affinity single-stranded probes such as Locked Nucleic Acids (LNAs)^5,6 or γ-functionalized PNAs,⁷ or energetically activated double-stranded probes such as pseudo-complementary (pc) DNA,⁸ pcPNA,^9–11 heteroduplexes between intercalator-modified oligonucleotides and RNA, LNA or PNA strands,^12–15 or related approaches.^16,17 While these probes have been shown to enable sequence-unrestricted recognition of DNA via invasion of Watson-Crick base-pairs, they are not without limitations. For example, single-stranded LNA and γ-PNA probes display a proclivity for self-hybridization in certain sequence contexts, which impedes their ability to recognize dsDNA.¹⁸ Moreover, duplex invasion by single-stranded probes only results in recognition of one target strand, leading to the formation of an unbound D-loop, which represents unrealized binding potential. Energetically activated double-stranded probes offer the promise of more favorable gains in free energy upon DNA recognition since both strands are targeted, and of improved binding specificity due to stringency clamping effects;¹⁹ recognition of non-target dsDNA regions would necessitate dissociation of two complementary duplexes to form two mismatched duplexes, an inherently unfavorable process.²⁰ However, recognition via double-duplex invasion necessitates both the probe and dsDNA target region to be partially denatured or – at a minimum – display localized regions of base-pair breathing,^19,21 which may slow down recognition kinetics.

As indicated above, two main strategies have been explored in the design of energetically activated double-stranded probes for dsDNA-recognition. Pseudo-complementary DNA/PNA rely on the presence of weak base-pairs between 2-thiouracil and 2,6-diaminopurine for energetic activation.^8–11 Base-pairing is weakened due to steric clashes between the modified nucleobases, whereas dsDNA-recognition is facilitated by stable base-pairing between 2-thiouracil:adenine and 2,6-diaminopurine:thymine, which ensues when individual pcDNA/PNA strands hybridize with cDNA regions. Conversely, the use of heteroduplexes between intercalator-modified oligonucleotides and RNA, LNA or PNA strands for recognition of dsDNA targets, relies on the fact that intercalators are poorly accommodated in A-type duplexes,^22,23 whereas the individual probe strands display prominent affinity towards cDNA regions, resulting in favorable gains in free energy upon dsDNA-binding.

Our laboratory has explored so-called Invader probes as an alternative strategy towards generating energetically activated double-stranded probes for sequence-unrestricted recognition of dsDNA. Invader probes are short DNA duplexes that are modified with 2’-intercalator-functionalized nucleotides arranged in +1 interstrand zipper motifs (Figure 1).^18,24–26 This particular arrangement forces the covalently linked intercalators from opposing strands into the same region, prompting localized unwinding and destabilization of the probe duplex,^27,28 as the neighbor exclusion principle²⁹ – which asserts that intercalation is anti-cooperative at adjacent sites³⁰ – is violated. Recognition of complementary dsDNA is thermodynamically favored since duplex formation between individual probe strands and cDNA reverses the effects of violating the neighbor exclusion principle as intercalators no longer are vying for the same region and instead stabilize the resulting duplexes through π-π stacking interactions with neighboring base pairs (Figure 1).^{15,18,24–28}

Figure 1.

Illustration of bulged Invader probes and modifications used herein.

In addition to demonstrating recognition of various dsDNA model targets,^18,24–28 Invader probes have also been used to detect mixed-sequence DNA fragments from specific food pathogens,³¹ stain telomeric DNA of individual chromosomes in metaphasic spreads,¹⁸ and signal the presence of Y-chromosome-specific targets in isolated nuclei from male bovine kidney cells under non-denaturing conditions.²⁶ Recent results suggest that optimized Invader probes recognize partially self-complementary dsDNA targets more efficiently than LNA-modified ONs or single-stranded PNA probes that are functionalized at the γ-position with (R)-diethylene glycol (i.e., ^MPγPNAs), which speaks to the promise of Invader probes for biomedical applications.¹⁸

We have explored different monomer chemistries and probe architectures to improve the binding characteristics of Invader probes.^{15,26–28,32,33} As part of these efforts, we have explored Invader probes that are additionally modified with non-nucleotidic bulges,^34,35 The underlying hypothesis motivating this work is that bulge-containing Invader probes will display additionally disrupted base stacking,³⁶ and consequently denature and nucleate more readily with the dsDNA target. Indeed, Invader probes with central +1 interstrand zipper arrangements of 2’-O-(pyren-1-yl)methyl-RNA monomers and nonyl bulges (C₉) near the edges of the probe duplex, display more efficient (>5-fold) and faster (>4-fold) recognition of model dsDNA targets than conventional probes.³⁴

To study the effect of the bulge chemistry on Invader-mediated dsDNA-recognition in more detail, we, herein, evaluated a library of different Invader probe designs entailing three bulge monomers (Figure 1).

RESULTS AND DISCUSSION

Design of modified ONs

Fourteen different oligodeoxyribonucleotides – each of them eighteen nucleotides long and featuring i) two incorporations of 2’-O-(pyren-1-yl)methyl-RNA monomers (U or C), one within 3–4 nucleotides of each terminus, and ii) no, one, or three consecutive incorporations of commercially available two-, four- or nine-atom linker monomers 2, 4, or 9 in the central region – were synthesized on a DNA synthesizer using established protocols (Table S1^†, Figs. S1–S4^†). Access to these ONs enabled construction of nineteen different double-stranded probes, i.e., twelve probes with a central non-nucleotidic bulge in one of the strands, six probes with two identical central bulges positioned opposite of each other, and conventional Invader probe ON1:ON2, serving as a reference (Table 1). The initial probe set was not designed with a biological target or maximal affinity in mind. The latter requires more densely modified probes, which was not pursued to conserve material. The emphasis was, rather, placed on delineating the impact of different bulge designs on the dsDNA-targeting properties of model Invader probes that are of a relevant length for biological applications. Bulges were positioned centrally to maximize their disruptive effect, whereas the intercalator-based hotspots were positioned at some distance from the ends to augment duplex fraying and probe denaturation (Figure 1).

Table 1.

Thermal denaturation temperatures (T_ms) of Invader probes and duplexes between individual probe strands and cDNA, and $\Delta G_{\text{rec}}^{298}$ and $\Delta G_{\mathit{\text{rec}}}^{310}$ values.^a

Sequence 5’-GGUGGTCAA X1 CTATCUGGA 3’-CCACCAGTT X2 GATAGACCT
				Tm [ΔTm]/(˚C)
Probe	X1	X2	Invader duplex	5’-ON:cDNA	3’-ON:cDNA	$\Delta G_{\text{rec}}^{298}$ (kJ/mol)	$\Delta G_{\mathit{\text{rec}}}^{310}$ (kJ/mol)
1:2	-	-	60.5 [+1.5]	65.5 [+6.5]	68.5 [+9.5]	−39	−35
3:2	2	-	58.0 [−1.0]	63.0 [+4.0]	68.5 [+9.5]	−27	−26
1:4	-	2	54.0 [−5.0]	65.5 [+6.5]	64.5 [+5.5]	−35	−31
3:4	2	2	52.0 [−7.0]	63.0 [+4.0]	64.5 [+5.5]	−17	−18
5:2	222	-	52.0 [−7.0]	58.0 [−1.0]	68.5 [+9.5]	−43	−34
1:6	-	222	47.5 [−11.5]	65.5 [+6.5]	59.0 [0]	−44	−39
5:6	222	222	41.0 [−18.0]	58.0 [−1.0]	59.0 [0]	−39	−32
7:2	4	-	51.5 [−7.5]	58.0 [−1.0]	68.5 [+9.5]	−36	−33
1:8	-	4	47.5 [−11.5]	65.5 [+6.5]	60.0 [+1.0]	−28	−28
7:8	4	4	40.5 [−18.5]	58.0 [−1.0]	60.0 [+1.0]	−23	−22
9:2	444	-	44.0 [−15.0]	53.0 [−6.0]	68.5 [+9.5]	−38	−34
1:10	-	444	39.5 [−19.5]	65.5 [+6.5]	53.5 [−5.5]	−41	−35
9:10	444	444	37.0 [−22.0]	53.0 [−6.0]	53.5 [−5.5]	−17	−15
11:2	9	-	44.0 [−15.0]	53.5 [−5.5]	68.5 [+9.5]	−36	−32
1:12	-	9	42.0 [−17.0]	65.5 [+6.5]	55.5 [−3.5]	−40	−34
11:12	9	9	44.0 [−15.0]	53.5 [−5.5]	55.5 [−3.5]	−10	−6
13:2	999	-	45.0 [−14.0]	53.0 [−6.0]	68.5 [+9.5]	−36	−32
1:14	-	999	44.5 [−14.5]	65.5 [+6.5]	56.0 [−3.0]	−47	−40
13:14	999	999	45.0 [−14.0]	53.0 [−6.0]	56.0 [−3.0]	−15	−12

^aThermal denaturation curves (Fig. S5†) were recorded in medium salt phosphate buffer ([Na⁺] = 110 mM, [Cl⁻] = 100 mM, pH 7.0 (NaH₂PO₄/Na₂HPO₄), [EDTA] = 0.2 mM) using each ON at 0.5 μM concentration. ΔT_m is calculated relative to the corresponding unmodified DNA duplex (T_m = 59.0 °C). ΔG_rec = ΔG_5’-ON:cDNA + ΔG_3’-ON:cDNA - ΔG_{Invader probe} - ΔG_dsDNA. The corresponding ΔG²⁹⁸ and ΔG³¹⁰ values used to calculate $\Delta G_{\text{rec}}^{298}$ and $\Delta G_{\text{rec}}^{310}$ are listed in Tables S4^† and S5^†, respectively. A = adenin-9-yl, C = cytosin-1-yl, G = guanin-9-yl, T = thymin-1-yl. U/C = 2’-O-(pyren-1-yl)methyl-RNA U or C monomers. X₁ and X₂: 2= 1-amino-3-hydroxyprop-2-yl monomer, 4= 4-hydroxybutyl monomer, 9= 9-hydroxynonyl monomer (see Fig. 1 for structures of modified monomers).

Thermal denaturation properties

Thermal denaturation temperatures (T_ms) were determined for double-stranded probes and for duplexes between individual probe strands and cDNA (Table 1). Conventional Invader probe ON1:ON2 is only minimally stabilized relative to the corresponding unmodified DNA duplex (ΔT_m = +1.5 °C), strongly suggesting that the two +1 interstrand zipper arrangements of 2’-O-(pyren-1-yl)methyl-RNA monomers result in the expected²⁸ violation of the neighbor exclusion principle.^29,30

Probe duplexes featuring a single bulge are progressively destabilized as the size of the non-nucleotidic monomer increases, in some cases by as much as 18.5 °C relative to ON1:ON2 (e.g., compare ΔT_ms for ON1:ON2 < ON3:ON2 < ON7:ON2 < ON11:ON2). Incorporation of a bulge in the 3’-CCACCAGTT-X₂-GATAGACCT strand (= 3’-ON) results in greater relative destabilization than incorporation in the 5’-GGUGGTCAA-X₁-CTATCUGGA strand (= 5’-ON), indicating that the destabilizing effect is dependent on the sequence context (e.g., compare ΔT_ms of ON3:ON2 and ON1:ON4). Expansion of the bulges through incorporation of three consecutive small monomers (i.e., X = 2 or 4) results in further destabilization of the double-stranded probes (e.g., compare ΔT_ms of ON3:ON2 and ON5:ON2). Presumably, the additional destabilization is due to increased local helical flexibility, decreased base-pair cooperativity, and/or increased electrostatic repulsion between the strands since the number of negatively charged phosphodiester linkages is increased. Interestingly, incorporation of three consecutive large monomers (i.e., X = 9) does not result in further destabilization, suggesting that there is a saturation point for bulge-mediated destabilization (e.g., compare ΔT_ms of ON11:ON2 and ON13:ON2).

Probe duplexes with two small bulges opposite of each other display additive destabilization relative to the corresponding single-bulge probes (e.g., compare ΔT_ms of ON3:ON2 and ON1:ON4 with ON3:ON4), whereas no further destabilization is observed upon introduction of a second large bulge (e.g., compare ΔT_ms of ON11:ON2 and ON1:ON12 with ON11:ON12). Expansion of the bulge regions through incorporation of consecutive non-nucleotidic monomers, has an additionally destabilizing effect on dual-bulge probes when monomer 2 is used but not when larger monomers are used (e.g., compare ΔT_ms for ON3:ON4 and ON5:ON6 with ON11:ON12 and ON13:ON14). This lends support to the hypothesis that there is a saturation point for duplex destabilization that is dependent on the size and flexibility of the bulge.

Thus, the results suggest that introduction of a single bulge comprised of a 2, 4, 9, 222, or 444 segment or introduction of two opposing bulges comprised of 2, 4, 222 or – to a lesser degree – 444 segments, are efficient strategies for labilizing an Invader probe as no additional destabilization is observed with larger and/or more flexible bulges.

Introduction of non-nucleotidic bulges also decreases the T_ms of duplexes between individual probe strands and cDNA, with similar monomer-specific impacts as observed for probe duplexes. Thus, probe strands with a large bulge form less stable duplexes with cDNA than probe strands with a small bulge (e.g., compare ΔT_ms for duplexes between ON3, ON7, or ON11 and cDNA). Probe strands with expanded bulges, constructed using consecutive incorporations of small non-nucleotidic monomers, display progressively lower cDNA affinity (e.g., compare ΔT_ms for ON3:cDNA and ON5:cDNA), whereas consecutive incorporations of large non-nucleotidic monomers do not further reduce cDNA affinity (e.g., compare ΔT_ms for ON11:cDNA and ON13:cDNA).

Thermodynamic parameters associated with formation of bulge-containing double-stranded Invader probes and duplexes between individual probes strands and cDNA were determined through baseline fitting of thermal denaturation curves (Fig. S5^† and Tables S4–S8^†).³⁷ The observed trends in ΔG values largely mirror the T_m-based trends discussed above; for full details, see the ESI^†.

Thermodynamic driving force for dsDNA-recognition

Since introduction of non-nucleotidic bulges decreases the T_m and increases the ΔG of probe duplexes (likely desirable for dsDNA-recognition), as well as of duplexes between individual probe strands and cDNA (likely undesirable for dsDNA-recognition), the overall impact on the thermodynamic driving force for Invader-mediated recognition of dsDNA regions is not immediately clear. We express the driving force for recognition of complementary dsDNA targets (ΔG_rec) at a specific temperature as the difference in Gibbs free energy associated with the formation of the “product” duplexes (i.e., 5’-ON:cDNA and 3’-ON:cDNA) and the “reactant” duplexes (i.e., Invader probe and dsDNA). In other words, ΔG_rec = ΔG_5’-ON:cDNA + ΔG_3’-ON:cDNA - ΔG_{Invader probe} - ΔG_dsDNA, with strongly negative ΔG_rec values indicating a prominent driving force for dsDNA-recognition (Table 1). For additional background discussion of the ΔG_rec term, see the ESI^†. Thermal advantage, TA, is the equivalent T_m-based term, which we define as TA = T_m (5’-ON:cDNA) + T_m (3’-ON:cDNA) - T_m (Invader probe) - T_m (dsDNA) (Table S3^†).

Invader probes featuring a small bulge comprised of monomer 2 or 4 display slightly less favorable thermodynamic driving forces for dsDNA-recognition at 298 K than conventional probe ON1:ON2 (e.g., compare $\Delta G_{\text{rec}}^{298}$ values for ON3:ON2, ON7:ON2, and ON1:ON2, Table 1), whereas Invader probes with a large bulge on one of the probe strands - constructed using a 222, 444, 9, or 999 segment - display similar or slightly more favorable driving forces (e.g., see $\Delta G_{\text{rec}}^{298}$ for ON1:ON6, ON1:ON10, ON1:ON12, and ON1:ON14). Probes with two opposing bulges generally display driving forces that are substantially less favorable than the corresponding single bulge probes (e.g., compare $\Delta G_{\text{rec}}^{298}$ values for ON13:ON2, ON1:ON14, and ON13:ON14). As previously discussed, this trend ensues because incorporation of a second bulge only results in limited additional probe destabilization, whereas each bulge-containing Invader strand displays low cDNA affinity. ON5:ON6, featuring two bulges comprised of 222 segments, constitutes an exception hereto, which in large part is due to a surprisingly enthalpically destabilized probe duplex (Table S6^†).

Similar trends are observed at 310 K except that the driving forces, on average, are reduced by ~4 kJ/mol (compare $\Delta G_{\text{rec}}^{298}$ and ΔG$\Delta G_{\text{rec}}^{310}$ values, Table 1). This is a result of bulge-containing probe duplexes being ~8 kJ/mol less destabilized at 310 K than at 293 K relative to the unmodified reference duplex, and probe-target duplexes being ~2 kJ/mol more stable (i.e., ΔΔG³¹⁰ < ΔΔG²⁹⁸, Tables S4^† and S5^†).

Recognition of mixed-sequence dsDNA model targets – design and initial screen

The dsDNA-targeting properties of the bulge-containing Invader probes were evaluated using an electrophoretic mobility shift assay,²⁶ in which a 200-fold molar probe excess was incubated with a digoxigenin (DIG)-labelled DNA hairpin (DH) model target at room temperature or 37 °C (Fig. 2). DH1 is comprised of an 18-mer double-stranded stem that is isosequential relative to the Invader probes and connected by a decameric thymidine (T₁₀) loop. Successful recognition of this high-melting target (T_m = 73 °C, Table S10^†) is expected to result in the formation of a ternary complex with lower electrophoretic mobility on non-denaturing polyacrylamide gels than DH1 (Fig. S6^†).

Figure 2.

Recognition of a DNA hairpin model target. Illustration of electrophoretic mobility shift assay used to evaluate dsDNA-recognition of bulge-containing Invader probes (upper left). Histograms depict averaged results from at least three recognition experiments in which a 200-fold molar excess of Invader probes containing one (upper and lower right) or two bulges (lower left) were incubated with DIG-labeled DNA hairpin DH1 (34.4 nM, 5’-GGTGGTCAACTATCTGGA-T₁₀-TCCAGATAGTTGACCACC-DIG-3’) at room temperature (red bars) or 37 °C (blue bars) for 17 h in HEPES buffer (50 mM HEPES, 100 mM NaCl, 5 mM MgCl₂, pH 7.2, 10% sucrose, 1.44 mM spermine tetrahydrochloride). For representative electrophoretograms and tabulated data, see Fig. S6^† and Table S9^†, respectively. Error bars represent standard deviation.

The screen suggests that Invader probes featuring a single large bulge comprised of a 9, 222, 444, or 999 segment display improved room temperature recognition of DH1 vis-à-vis the conventional probe ON1:ON2 (35–65% vs. ~20% recognition, Fig. 2 and Table S9^†). It is interesting to note that these probes display $\Delta G_{\text{rec}}^{298}$ values ≤ −36 kJ/mol and T_ms ≤ 52 °C (Table 1). Probes with a single small bulge comprised of a 2 or 4 monomer and probes with two adjacent bulges comprised of 2, 4, 222 or 444 segments, recognize DH1 with similar efficiency as conventional probe ON1:ON2 (i.e., 15–35% recognition, Fig. 2 and Table S9^†). These probes display less favorable thermodynamic gradients (−36 kJ/mol ≤ $\Delta G_{\text{rec}}^{298}$ ≤ −17 kJ/mol, Table 1) and/or T_ms ≥ 51.5 °C.³⁸ Probes with two adjacent bulges comprised of 9 or 999 segments display poor room temperature recognition of DH1 and are characterized by less favorable thermodynamic driving forces $\Delta G_{\text{rec}}^{298}$ values ≥ −15 kJ/mol, Table 1).

Recognition of DH1 is considerably more efficient at 37 °C, presumably as the higher experimental temperature facilitates denaturation of probe duplexes and renders the stem region of DH1 more dynamic. Thus, conventional probe ON1:ON2 results in ~60% recognition (Fig. 2 and Table S9^†). Invader probes witha single large bulge comprised of 444 – and, especially – 9 or 999 segments, recognize DH1 very efficiently (75–90%; e.g., see ON11:ON2, Fig. 2 and Table S9^†). These probes display very favorable thermodynamic gradients $\Delta G_{\text{rec}}^{310}$ ≤ −32 kJ/mol) and are highly labile (T_ms ≤ 45 °C). Invader probes with a single 222 bulge, two opposing bulges comprised of 2 or 4 monomers, and ON1:ON4 and ON1:ON8, which feature a single small bulge in the 5’-ON strand, recognize DH1 with similar efficiency as ON1:ON2. These probes display less favorable thermodynamic driving forces (−22 kJ/mol ≤ $\Delta G_{\text{rec}}^{310}$ ≤ −18 kJ/mol) and/or are relatively stable (T_ms ≥ 47.5 °C). All other probes, especially those with two large opposing bulges comprised of 222, 444, 9, and 999 segments, display far less efficient recognition of DH1 at 37 °C. These probes are characterized by even less favorable thermodynamics $\Delta G_{\text{rec}}^{310}$ ≥ −15 kJ/mol) and/or high stability (T_ms ≥ 51.5 °C).³⁸

Based on the results from the preliminary screen, we decided to carry out dose-response experiments at 37 °C for Invader probes ON1:ON4, ON1:ON8, and ON1:ON12, which feature a single 2, 4, or 9 bulge, using the electrophoretic mobility shift assay described above to determine C₅₀ values, i.e., the probe concentrations that result in 50% recognition of DH1 (Fig. 3). These probes were selected because they i) display moderate-to-high dsDNA-recognition efficiency, ii) facilitate a comparison between different bulge chemistries, and iii) consume a minimal amount of bulge monomers.

Figure 3.

Dose-response curves for DH1-recognition using conventional and select single-bulge Invader probes at 37 °C. For experimental conditions, see Figure 2. For representative gel electrophoretograms, see Fig. S7^†.

The single-bulge Invader probes display C₅₀ values of 1.5, 3.3, and 10.2 μM for ON1:ON12, ON1:ON8 and ON1:ON4, respectively, whereas the conventional Invader probe ON1:ON2 displays a C₅₀ value of 3.9 μM. Interestingly, while displaying similar C₅₀ values, ON1:ON8 comprising a 4 bulge, results in more efficient DH1-recognition at low probe excess than conventional Invader probe ON1:ON2 (i.e., ~24% versus ~6% for ON1:ON8 and ON1:ON2, respectively, at 10-fold molar probe excess). Importantly, the results suggest that introduction of a single central 9 bulge is advantageous for recognition of mixed-sequence dsDNA targets by Invader probes.

Given the increased recognition of these bulge-containing Invader probes, the binding specificity was evaluated. Probes were incubated with DIG-labelled DNA hairpins DH2-DH4, which have fully base-paired stem regions that only differ in sequence at either the 4-, 6- or 9-position vis-à-vis the probes (Fig. 4). For recognition to occur, two singly mismatched probe-target duplexes must form, with mismatched base-pairs located either next to the bulge (DH2), in the unmodified region between the bulge and Invader monomers (DH3), or near one of the Invader monomers (DH4). Gratifyingly, no recognition of these DNA hairpins is observed even when probes are used at 400-fold molar excess (Fig. 4). This level of binding specificity is remarkable considering that seventeen out of eighteen base pairs are complementary to the probes. It stands to reason that double-duplex invasion is a far more specific process than simple duplex formation, as both the dsDNA region and the double-stranded probes would need to be disrupted to form two mismatched duplexes.^19,20 This is underscored by the observation that the binding specificity of individual bulge-containing probe strands is not improved vis-à-vis the corresponding unmodified strands (Table S2^†).

Figure 4.

Binding specificity of conventional and single-bulge Invader probes at 37 °C. Probes were incubated with DH1-DH4 at 400-fold molar excess. Experimental conditions are otherwise as described in Fig. 2. For sequences of DH1-DH4 and their T_ms, see Table S10^†.

Detection of Y chromosomal-specific repeat sequence under non-denaturing conditions

Encouraged by the promising binding characteristics, we sought to utilize bulge-containing Invader probes in non-denaturing in situ fluorescence hybridization (nd-FISH) assays. Towards this end a set of 14-mer Cy3-labeled Invader probes, featuring three separated energetic hotspots comprised of 2’-O-(pyren-1-yl)methyl-RNA monomers and a single 4 or 9 bulge, were designed to target a complementary, highly repeated sequence in the DYZ-1 satellite gene (~6 × 10⁴ tandem repeats of a ~1175 bp region) on the bovine (Bos taurus) Y chromosome (NCBI code: M26067, positions: 562–575)³⁹ (Tables 2 and S11^†, Figs. S8–S10^†). Conventional PNA FISH probes have been used to recognize this model target under denaturing conditions.^40–42 Locked Nucleic Acid (LNA), PNA, and polyamides FISH probes have proven refractory in affording target-specific signals under non-denaturing conditions.^26,43 The bulge-modified Invader probes were compared to the corresponding conventional Invader probe ON15:ON16, which we have previously demonstrated results in efficient and specific recognition of this highly GC-rich heterochromatic target (71% GC-content),^26,43 as well as other control probes. Bulges were placed either near the center (i.e., ON22:ON16, ON15:ON17, and ON15:ON18) or near a terminal of the probe (i.e., ON23:ON16, ON15:ON24, ON15:ON25, and ON15:ON26).

Table 2.

Thermal denaturation temperatures of DYZ-1 targeting Invader probe duplexes and duplexes between individual probe strands and cDNA.^a Also shown are TA values, signal intensity, and proportion of nuclei presenting signal for the different Invader probes.

		Tm [ΔTm]/(˚C)
Probe	Sequence	Probe duplex	5’-ON:cDNA	3’-ON:cDNA	TA (˚C)	Signal intensity	% Coverage
15:16b	5’-Cy3-AGCCCUGTGCCCTG 3’-TCGGGACACGGGAC-Cy3	66.0 [+5.5]	69.5 [+9.0]	74.0 [+13.5]	+17.0	+++	77
22:16	5’-Cy3-AGCCCUGTG 4 CCCTG 3’-TCGGGACAC GGGAC-Cy3	50.5 [−10.0]	61.5 [+1.0]	74.0 [+13.5]	+24.5	+++	69
15:17	5’-Cy3-AGCCCUGTG CCCTG 3’-TCGGGACAC 4 GGGAC-Cy3	41.5 [−19.0]	69.5 [+9.0]	62.5 [+2.0]	+30.0	+	49
15:18	5’-Cy3-AGCCCUGTG CCCTG 3’-TCGGGACAC 9 GGGAC-Cy3	32.5 [−28.0]	69.5 [+9.0]	57.5 [−3.0]	+34.0	+	25
23:16	5’-Cy3-AGCCCUGTGCC 4 CTG 3’-TCGGGACACGG GAC-Cy3	ntc	71.0 [+10.5]	74.0 [+13.5]	nd	++	76
15:24	5’-Cy3-AGCCCUGTGCC CTG 3’-TCGGGACACGG 4 GAC-Cy3	56.0 [−4.5]	69.5 [+9.0]	62.0 [+1.5]	+15.0	++	47
23:24	5’-Cy3-AGCCCUGTGCC 4 CTG 3’-TCGGGACACGG 4 GAC-Cy3	54.0 [−6.5]	71.0 [+10.5]	62.0 [+1.5]	+18.5	+	26
15:25	5’-Cy3-AGCCCUGTGCC CTG 3’-TCGGGACACGG 9 GAC-Cy3	52.0 [−8.5]	69.5 [+9.0]	65.0 [+4.5]	+22.0	++	70
15:26	5’-Cy3-AGC CCUGTGCCCTG 3’-TCG 9 GGACACGGGAC-Cy3	~30d [−30.5]	69.5 [+9.0]	62.0 [+1.5]	~ +41	++	76
19:20e	5’-Cy3-AGCGCUGAG GCCTG 3’-TCGCGACTC 4 CGGAC-Cy3	43.5 [−20.0]	78.5 [+15.0]	64.5 [+1.0]	+36.0	-f	4
19:21e	5’-Cy3-AGCGCUGAG GCCTG 3’-TCGCGACTC 9 CGGAC-Cy3	42.0 [−21.5]	78.5 [+15.0]	60.5 [−3.0]	+33.5	-f	0

^aΔT_m = change in T_m relative to unmodified dsDNA (T_m = 60.5 °C, except for ON19:ON20and ON19:ON21for which the corresponding unmodified dsDNA has a T_m = 63.5 °C). TA = T_m (5’-ON:cDNA) + T_m (3’-ON:cDNA) - T_m (Invader probe) - T_m (dsDNA). For experimental conditions, see Table 1. Cy3 = Cy3 monomer, 4= 4-hydroxybutyl monomer, 9= 9-hydroxynonyl monomer, and A/U/C= 2’-O-(pyren-1-yl)methyl-RNA monomers. For structures, see Fig. 1. nt = no transition. nd = not determined.

^bData previously published in Reference 26.

^cBased on a single measurement due to low probe quantities.

^dBroad transition.

^eON20and ON21do not form stable duplexes with the cDNA of ON16, whereas the duplex between ON19and the cDNA of ON15displays a T_m = 44.0 °C. Thus, ON19:ON20and ON19:ON21display TA << 0 °C for recognition of the DYZ-1 target region.

^fSignal was not observed.

In line with our previous observations (Table 1), incorporation of a 4 or 9 monomer results in dramatic destabilization of the Invader probes (note the lower T_ms in the “Probe duplex” column relative to ON15:ON16, Table 2). Decreased stability, albeit to a lesser degree, is also observed for the corresponding duplexes between individual bulge-modified probe strands and cDNA (observe the lower T_ms in the “5’-ON:cDNA/3’-ON:cDNA” column relative to ON15:cDNA and ON16:cDNA, Table 2). As a result, the bulge-containing Invader probes generally display more favorable TA values than conventional Invader probe ON15:ON16, indicating that the thermodynamic driving force for dsDNA-recognition is more pronounced (Table 2).

Fixed isolated nuclei from a male bovine kidney cell line were incubated with different Invader probes under non-denaturing conditions (3 h, 37 °C, Tris-Cl/EDTA buffer, pH 8.0). All of the sequence-matched, bulge-containing Invader probes produced a single punctate Cy3 signal against a DAPI-counterstained chromosomal DNA background, indicating that recognition of the mixed-sequence Y-chromosome-specific dsDNA target is feasible at mild incubation conditions (Figs. 5 and S8–S10^†). However, the signal intensity, percentage of nuclei presenting signal, and degree of non-specific background signal varied among the probes (Table 2). For example, ON22:ON16, which features a single centrally positioned 4 bulge in the C-rich strand, displays a strong signal that is present in ~70% of the nuclei and thus has comparable staining characteristics to the conventional Invader probe ON15:ON16 (Fig. 5 and Table 2). Surprisingly, incorporation of a 4 or 9 bulge in the equivalent position of the G-rich strand yields probes with lower signal intensity (see data for ON15:ON17 and ON15:ON18, Table 2 and Figs. S11^† and S12^†), even though they display more favorable TA values (Table 2). This indicates that recognition is not thermodynamically limited. Along similar lines, incorporation of a single 4 bulge closer to the 5’-terminus of the G-rich strand, yields a probe that produces a moderately intense signal in ~47% of nuclei (see data for ON15:ON24, Table 2 and Fig. S13†). Probes featuring a single 4 bulge that is positioned closer to the 3’-terminus of the C-rich strand, or a single 9 bulge positioned closer to one of the termini of the G-rich strand, also produce moderately intense signals against a stronger background but in a higher proportion of nuclei (70–75%, see data for ON23:ON16, ON15:ON25, and ON15:ON26, Table 2 and Figs. S13^† and S14^†). As expected, ON23:ON24, in which two 4 bulges are opposite of each other, and ON19:ON20 and ON19:ON21, which are mismatched controls of ON15:ON17 and ON15:ON18 differing in sequence at three positions relative to the DYZ-1 target, display very weak or no signal (Table 2 and Figs. S11–S13^†), corroborating that Invader-mediated recognition of chromosomal DNA is sequence-specific. This is further underlined by the lack of signals when the various bulge-containing Y-chromosome-targeting Invader probes are incubated with isolated nuclei from a female bovine fibroblast cell line (nuclei presenting signal is <5%, Figs. S11–S14^†). The lack of improvement in signal intensity and coverage of the bulge-modified Invader probes in the nd-FISH assays vis-à-vis conventional Invader probe ON15:ON16 is unexpected given the initial observations, i.e., greater thermodynamic driving force and improved Invader-mediated recognition of dsDNA model targets (e.g., Table 1 and Fig. 2). Several factors may underlie these observations. Perhaps most importantly, ON15:ON16 is already an efficient probe. In fact, the detection efficiency displayed by ON15:ON16 is surprising given that the probe is designed to target a region with a high GC-content (~71%). Partial probe and target denaturation do not appear to be a limiting factor with this target. Indeed, results from ongoing studies indicate that the target region is unusually accessible to Invader probes.⁴⁴ Hence, the decreased cDNA-affinity of individual bulge-modified Invader strands may be disadvantageous for dsDNA-recognition and result in reduced signal intensity in this sequence context. Nonetheless, incorporation of non-nucleotidic bulges into Invader probes to improve recognition of chromosomal DNA targets is an approach that deserves further study.

Figure 5.

Representative images of bulged Invader probe ON22:ON16 and conventional probe ON15:ON16 incubated (1.2 ng of probe, 3 h, 37.5 °C, Tris-Cl/EDTA buffer, pH 8.0) with fixed nuclei (MeOH:AcOH, 3:1, v/v) from a male bovine kidney cell line (MDBK) or a female bovine endothelial cell line (CPAE) as indicated. Images are overlays of Cy3 (red) and DAPI channels (blue). Scale bar denotes 10 μm.

Conclusion

The results reported herein underscore the potential for using bulge-containing Invader probes^34,35 for mixed-sequence recognition of dsDNA targets, and thus represent an interesting addition to a growing toolbox of probes that encompasses conventional Invader probes,^{18,24–28,31,32,43,46} pcDNA,⁸ pcPNA,^9–11 pc-Invader probes,³³ zorro LNA,⁶ γ-PNA,^7,18,21 heteroduplexes between intercalator-modified oligonucleotides and RNA, LNA or PNA strands,^12–15 among other approaches.^16,17 Introduction of internal alkyl bulges destabilizes double-stranded Invader probes but also reduces the cDNA affinity of individual probe strands. Invader probes with a large bulge on one of the probe strands are shown to be more energetically activated for dsDNA-recognition than conventional Invader probes, whereas probes with two opposing bulges are less suitable for dsDNA-recognition. Energetically activated probes with low T_ms result in particularly efficient recognition of model mixed-sequence dsDNA targets. Binding by bulge-modified Invader probes is highly sensitive to the presence of mismatched base-pairs as demonstrated by their ability to perfectly discriminate complementary dsDNA targets from dsDNA that share seventeen out of eighteen base pairs with the probe. As an illustrative example of possible applications, bulge-modified Invader probes are shown to detect specific chromosomal DNA regions when used in FISH assays under non-denaturing conditions. Other biological applications for bulge-modified Invader probes will be explored.

Experimental section

Synthesis and purification of modified ONs

ONs were synthesized using an automated DNA synthesizer (0.2 μmol scale) using standard protocols. Thus, long chain alkyl amine-controlled pore glass (LCAA-CPG, 500 Å pore size), pre-loaded with the 3’-nucleotide, was used as a solid support. The corresponding phosphoramidites of 2’-O-(pyren-1-yl)methyl-RNA monomers U⁴⁵ or C⁴⁶ were incorporated into ONs via hand-coupling using 4,5-dicyanoimidazole (DCI, 0.01 M in acetonitrile, 15 min) as an activator with extended oxidation (45 s). The non-nucleosidic linkers were incorporated in a similar manner using the corresponding commercially available protected phosphoramidites (ChemGenes: CLP-1661, CLP-9775, and CLP-9009), with the exception of monomer 2 where 5-ethyltio-1H-tetrazole was used as the activator (15 min coupling performed at 55 °C). ONs were cleaved from the solid support and base-labile groups were removed by treatment with 32% ammonia (55 °C, 17 h). DMTr-protected ONs were purified using ion-pair reverse-phase HPLC (0.05 M triethylammonium acetate and acetonitrile gradient), followed by detritylation (80% AcOH, 20 min), and precipitation (NaOAc, NaClO₄, acetone, −18 °C, 16 h). The purity was determined via analytical HPLC (at least 85%), while the identity of the modified ONs was confirmed by MALDI-MS using 2,4,6-trihydroxyacetophenone as a matrix.

Cy3-labeled ONs were labeled using a commercially available Cy3 phosphoramidite (Glen Research), which was coupled as described above, with the following exceptions: coupling time was 5 min, detritylation was performed on the synthesizer, followed by subsequent deprotection in 32% ammonia (55 °C, 4 h) and purification as described.

Thermal denaturation experiments

ON concentrations were estimated using the following extinction coefficients (OD₂₆₀/μmol): G (12.01), A (15.20), T (8.40), C (7.05), pyrene (22.4),⁴⁷ and Cy3 (4.93).⁴⁸ ONs (each strand used at a final concentration of 0.5 μM) were mixed in quartz optical cells with a pathlength of 1.0 cm and annealed by heating (85 °C, 2 min) in medium salt buffer ([Na⁺] = 110 mM, [Cl⁻] = 100 mM, pH 7.0 (NaH₂PO₄/Na₂HPO₄), [EDTA] = 0.2 mM), followed by cooling to the starting temperature of the thermal denaturation experiment. Thermal denaturation curves of duplexes were recorded on a Cary 100 UV/VIS spectrophotometer equipped with a 12-cell Peltier temperature controller. Thermal denaturation curves were recorded from at least 15 °C below to 15 °C above the T_m. T_ms were determined as the first derivative maximum of denaturation curves (A₂₆₀ vs. T, averaged from at least two experiments within 1.0 °C and rounded to the nearest 0.5 °C). A temperature ramp of 1.0 °C/min was used, which resulted in minimal hysteresis for most duplexes (T_ms determined from heating and cooling cycles differed by < 2.5 °C). More pronounced hysteresis, consistent with broader transitions (Figure S5†), was observed for Invader probes with two opposing bulges (< 5 °C), potentially resulting in an overestimation of probe T_ms.

Thermodynamic parameters

Thermodynamic parameters were obtained through baseline fitting of denaturation curves using software provided with the UV-Vis spectrophotometer (van’t Hoff method). Bimolecular reactions, two-state melting behavior, and constant heat capacity were assumed.³⁷ Two denaturation curves per duplex were analyzed at least three times to minimize errors arising from baseline choice.

Electrophoretic mobility shift assay

DNA hairpins were obtained from a commercial vendor and used without further purification. Hairpins were digoxigenin (DIG)-labeled using the 2^nd generation DIG Gel Shift Kit (Roche Applied Bioscience) as recommended by the manufacturer. Briefly, following annealing of the DNA hairpins (95 °C for 2 min, followed by cooling), digoxigenin-ddUTP was added along with recombinant DNA 3’-terminal transferase (15 min, 37 °C). At this point, the reaction was quenched with EDTA (0.05 M) and the hairpin used without further purification. Invader probes were pre-annealed (95 °C for 2 min, followed by cooling) and subsequently incubated with the DIG-labeled DNA hairpin (34.4 nM) in HEPES buffer (50 mM HEPES, 100 mM NaCl, 5 mM MgCl₂, pH 7.2, 10% sucrose, 1.44 mM spermine tetrahydrochloride) at either 22 or 37° C (±2 °C). Following addition of loading dye, the reaction mixtures were loaded onto 12% non-denaturing polyacrylamide gels (45 mM tris-borate, 1 mM EDTA; acrylamide:bisacrylamide (19:1)). Following electrophoresis (~4 °C, 70 V, 2 h), the bands were electroblotted onto positively charged nylon membranes (100 V, 30 min, ~4 °C) and cross-linked (254 nm, 5 × 15 W bulbs, 5 min). Membranes were incubated with anti-DIG alkaline phosphatase F_ab fragments, according to the manufacturer’s recommendations. After incubation with a chemiluminescent substrate (CSPD) at 37 °C for 10 min, images of the chemiluminescent bands were captured on X-ray film and quantified using densitometry software. The percentage of DNA recognition was calculated as the ratio of intensity of the recognition complex relative to the intensity of the total lane, reported as the average of three independent experiments. Non-linear regression was used to fit data points from dose–response experiments. A script written for the “Solver” module in Microsoft Office Excel,⁴⁹ was used to fit data points from dose–response experiments to the following equation: y = C + A (1 - e^−kt) where C, A, and k are fitting constants. The resulting equation was used to calculate C₅₀ values by setting y = 50 and solving for t.

Cell culture and nuclei preparation

Male bovine kidney cells (MDBK (NLB-1), ATCC: CCL-22, Bethesda, MD) were maintained in DMEM with GlutaMax (Gibco, 10569–010) and 10% fetal bovine serum (Invitrogen). Female bovine endothelial cells (CPAE, ATCC: CCL-209) were maintained in Eagle’s Minimum Essential Medium (ATTC, 30–2003) and 20% fetal bovine serum (Invitrogen). Cells were cultured in separate 25 mL or 75 mL flasks at 37.5 °C in a 5% CO₂ atmosphere for 72–96 h to achieve 70–80% confluency. At this point, 65 μL of colcemid for every 5 mL of cell growth medium was added and the cells were incubated at 37.5 °C and 5% CO₂ for additional 20 minutes. The media was then replaced with a 37.5 °C solution of 0.05% Trypsin-EDTA in DMEM to detach adherent cells. The cell suspension was transferred to a centrifuge tube and centrifuged for 10 minutes at 1000 rpm to pellet the cells. The supernatant was aspirated off and replaced with 5–8 mL of 0.075M KCl hypotonic solution, and cells were incubated for 20 minutes. Ten drops of fixative solution (methanol: glacial acetic acid, 3:1, v/v) was added to the hypotonic solution and incubated for 10 minutes at room temperature. The solution was gently mixed and centrifuged for 10 minutes at 1000 rpm. Supernatant was aspirated off and 5–8 mL of fixative solution was added, gently inverted, and incubated at room temperature for 30 minutes. After incubation, the tube was centrifuged for 10 minutes at 1000 rpm, supernatant was aspirated off, and cell pellet was resuspended in 5–8 mL of fixative solution. This centrifugation and resuspension in fixative solution was repeated 3 times. The fixed cells were used immediately or stored at 4 °C for up to two weeks.

Fluorescence in situ hybridization

After cell culture and nuclei preparation, the fixative solution was removed. The cell pellet was resuspended in fresh fixative solution. Glass microscope slides (Fisher Scientific) were dipped in distilled water and, while holding slides at a 45° angle, an aliquot of the somatic nuclei suspension (3–5 μL or enough to cover the slide) was dropped onto the slide and allowed to run down the length of the slide. Slides were then allowed to dry at a ~20° angle in an environmental chamber set 28 °C and 38% relative humidity to fix nuclei and evaporate solvents.

An aliquot of the labelling solution (100 μL, 0.5–1.0 nM Cy3-labeled Invader probe solution, in deionized water) and 100 μL of PCR buffer ([Tris-HCl] = 10 mM, [KCl] = 50 mM, pH 8.3) were dropped on the slides. The slide was placed in a glass culture dish, covered with a lid and moved to a 37.5 °C incubator for ~3h. Slides were then washed with 37.5 °C TE buffer ([Tris] = 10 mM, [EDTA] = 1 mM, pH 8.0) for 3 min by submersion and gentle pipetting, rinsed with autoclaved water by pipetting, and allowed to dry at room temperature. An aliquot of Gold SlowFade Plus DAPI (3 μL, Invitrogen) was placed on the slide and a round glass coverslip was mounted and sealed with nail polish. The slides were then viewed on a Nikon Eclipse Ti-S/L100 fluorescence microscope equipped with a SOLA SMII LED light source system and Cy3 and DAPI filter cubes at 60X magnification. Images of fluorescently labelled nuclei were captured using a 14-bit Cool SNAP HQ2 cooled CCD camera and processed with NIS Elements BR 4.20 imaging software. The qualitative assessment of signal strength was determined through qualitative comparison of the signal brightness of all counted nuclei (on average 72 nuclei were counted per probe).