OLIGONUCLEOTIDE-BASED TUNING OF PORE-FORMING PEPTIDES FOR INCREASING PORE SIZE, MEMBRANE AFFINITY, STABILITY, AND ANTIMICROBIAL ACTIVITY

Abstract

Pore-forming peptides or proteins modified utilizing DNA nanotechnology, which provides definition and control of pore size, an increase in stability when inserted in a lipid membrane, and membrane affinity. Chemical modifications are easily made on the compound through hybridization to the oligonucleotide attached to the peptide or protein. The compound can hybridize to a DNA template thereby defining the number of monomers assembled to a pore and thus the size of the formed pore. The DNA template can range from a unique single strand composed of multiple hybridization sites separated by flexible linkers to a complex rigid DNA nanoconstruct, such as a DNA origami-based ring, serving as a scaffold for pore formation. Hydrophilic modification at the transmembrane segment or terminus of the peptide provides long-lived pores and keeps the compound in a membrane-spanning conformation.

Claims

1. A hybrid pore-forming compound, comprising: a pore-forming peptide or protein having a first terminus and a second terminus, wherein a first oligonucleotide is linked to the first terminus, wherein the first oligonucleotide is derived from DNA, RNA, LNA, BNA, or PNA, wherein a first functional moiety is linked to the second terminus, and wherein the first functional moiety is hydrophilic.

2. The compound according to claim 1, wherein a second functional moiety is linked to a second oligonucleotide that is hybridized to the first oligonucleotide bonded to the first terminus of the peptide, wherein the second oligonucleotide is derived from DNA, RNA, LNA, BNA, or PNA, wherein the second functional moiety is a hydrophobic membrane binder, receptor, drug molecule, antibody, aptamer, metabolite, or fluorescent marker.

3. The compound according to claim 2, wherein the hydrophobic membrane binder is present and comprises a cholesterol moiety.

4. The compound according to claim 1, wherein the first functional moiety bonded to the second terminus of the peptide is a third oligonucleotide or fluorophore, wherein the third oligonucleotide is derived from DNA, RNA, LNA, BNA, or PNA.

5. The compound according to claim 1, wherein a template strand having a plurality of complementary hybridization sites is hybridized to a portion of the first oligonucleotide linked to the first terminus.

6. The compound according to claim 1, wherein a plurality of the pore-forming peptides or proteins are present with each first oligonucleotide linked to the first terminus hybridized to the template strand, such that the compound forms a larger pore.

7. The compound according to claim 1, wherein a plurality of the pore-forming peptide or proteins are present with each first oligonucleotide linked to the first terminus hybridized to the template strand such that the compound has a tetrameric, hexameric, octameric or dodecameric pore conformation.

8. The compound according to claim 1, wherein the pore-forming peptide is present and is Ceratotoxin A (CtxA).

9. The compound according to claim 1, wherein the first functional moiety is the third oligonucleotide, and wherein a second membrane binder is linked to the third oligonucleotide to aid in stabilizing a pore formed by the compound.

10. The compound according to claim 1, wherein the peptide or protein is functionalized differently at the first terminus as compared to the second terminus.

11. A method for forming the hybrid pore-forming compound according to claim 1, comprising the steps of: obtaining i) the pore-forming peptide or protein, ii) the first oligonucleotide and iii) the first functional moiety; and forming the pore-forming compound by self-assembly of i), ii), iii).

12. The method according to claim 11, further including the step of reacting in solution the template strand having a plurality of complimentary hybridization sites with an excess of the pore-forming compounds comprising the i), iii); removing excess unreacted pore-forming compounds, preferably by high pressure liquid chromatography.

13. The method according to claim 12, further including the step of hybridizing the first oligonucleotide to the second oligonucleotide, wherein the second oligonucleotide is present in an excess amount as compared to the first oligonucleotide.

14. A membrane, comprising a substrate and the compound according to claim 1, wherein the compound provides a pore between a first side of the substrate and a second side of the substrate.

15. The compound according to claim 1, wherein an extended, preferably rigid, nucleic acid nanostructure serves as a template for assembling a plurality of, pore formers into a pore, wherein the nucleic acid nanostructure contains the second functional moiety directly attached thereto, and wherein the nucleic acid nanostructure is a DNA origami structure or a single stranded tile assembly or a RNA origami structure.

16. The compound according to claim 3, wherein the first functional moiety bonded to the second terminus of the peptide is a third oligonucleotide or fluorophore, wherein the third oligonucleotide is derived from DNA, RNA, LNA, BNA, or PNA, and wherein a template strand having 4, 6, 8, or 12 complementary hybridization sites is hybridized to a portion of the first oligonucleotide linked to the first terminus.

17. The compound according to claim 16, wherein a plurality of the pore-forming peptides or proteins are present with each first oligonucleotide linked to the first terminus hybridized to the template strand, such that the compound forms a larger pore, or wherein a plurality of the pore-forming peptide or proteins are present with each first oligonucleotide linked to the first terminus hybridized to the template strand such that the compound has a tetrameric, hexameric, octameric or dodecameric pore conformation.

18. The compound according to claim 17, wherein the pore-forming peptide is present and is Ceratotoxin A (CtxA), wherein the first functional moiety is the third oligonucleotide, and wherein a second membrane binder is linked to the third oligonucleotide to aid in stabilizing a pore formed by the compound, and wherein the peptide or protein is functionalized differently at the first terminus as compared to the second terminus.

19. A method for forming the hybrid pore-forming compound according to claim 18, comprising the steps of: obtaining i) the pore-forming peptide or protein, ii) the first oligonucleotide and iii) the first functional moiety; and forming the pore-forming compound by self-assembly of i), ii), iii).

20. A membrane, comprising a substrate and the compound according to claim 18, wherein the compound provides a pore between a first side of the substrate and a second side of the substrate.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The invention will be better understood and other features and advantages will become apparent by reading the detailed description of the invention, taken together with the drawings, wherein:

[0037] FIG. 1 illustrates a schematic representation of DNA-assembled peptide pores with programmable pore size, (a) Assembly scheme mediated by DNA hybridization with a schematic representation of the planar lipid bilayer setup used to record the pore-forming activity and single channel conductance of CtxA and of template-assembled versions of CtxA-DNA. The pore-forming peptide CtxA bears on its N-terminus a covalently attached single stranded DNA with two domains. The terminal domain binds to a template strand that presents 4, 6, 8 or 12 complementary hybridization sites. An additional DNA strand binds to the compound, carrying a hydrophobic moiety for increasing the membrane affinity. An additional hydrophilic poly-thymine segment is bonded to C-terminus of the peptide to keep the peptide in a membrane-spanning conformation, leading to long-lived pores. Four thymine bases, acting as flexible linkers, separate the DNA segments. (b-e) Schemes, projections and top views of possible special arrangements of (b) a tetrameric, (c) a hexameric, (d) an octameric pore or (e) a dodecameric pore, with fully occupied template strands containing four, six, eight or twelve binding sites for the CtxA-DNA monomers.

[0038] FIG. 2 illustrates increased pore-forming activity of CtxA-DNA after addition of a cholesterol strand. Experiments at 5 nM peptide concentration with (a) native CtxA, (b) ssDNA-CtxA and (c) dsDNA-CtxA did not exhibit frequent pore formation. (d) Strongly increased pore formation occurs for Chol-dsDNA-CtxA carrying a cholesterol moiety, which increases the affinity to the lipid membrane. (e) Expanded view of the current recording in (d) showing stepwise current fluctuations typical of native CtxA. All recordings are performed at an applied potential difference of +180 mV.

[0039] FIG. 3 displays a conductance versus time recording of the typical pore-forming activity of native CtxA. The recording shows well-defined, stepwise increases and decreases of the conductance that correlate with the uptake or release of a peptide monomer from the assembled pore. All recordings were performed at an applied potential difference of +180 mV and in presence of 20 nM peptide.

[0040] FIG. 4 shows the ill-defined conductance fluctuations induced by ssDNA-CtxA. The top panel corresponds to a continuous 5-minute recording in presence of 5 nM ssDNA-CtxA and shows little pore-forming activity as expected by the low concentration used. The other panels show different zoomed sections, illustrating the noisy and unstable conductance level instead of the usual well-defined conductance levels of native CtxA. All recordings were performed at an applied potential difference of +180 mV.

[0041] FIG. 5 is a scheme illustrating the DNA-modified CtXA peptide pores. When the DNA part of the peptide is single-stranded, the oligonucleotide can move freely in the solution, potentially covering the pore or being pulled into it. When the DNA is double-stranded, however, the increased rigidity prevents the DNA strands to cover or to be pulled into the pore.

[0042] FIG. 6 illustrates templating DNA-modified CtxA peptides leads to preferential pore sizes. Conductance traces versus time are shown, next to the corresponding histograms. (a) Typical CtxA conductance levels observed for Chol-dsDNA-CtxA. (b) In the presence of the 4-mer template, quick fluctuations to the conductance value of a 4-mer pore (02) occur. Using the 8-mer template (c) and the 12-mer template (d), current steps directly to the conductance value for an 8-mer pore (06) and even larger fluctuations in the case of the 12-mer, are observed. The colored boxes in the background of the conductance traces of each panel correspond to 4-mer (02), 8-mer (06) and 12-mer (010) conductance levels. The values were calculated by taking the average conductance values from multiple experiments with native CtxA (for 02 and 06) and from an extrapolation of the values of native CtxA conductance open states to find an estimation of the 12-mer conductance. All recordings were carried out with a CtxA-dsDNA-chol concentration of 5 nM (a, b, c) or 20 nM (d) and with applied potential differences of (a and d)+180 mV, (b)+160 mV and (c)+140 mV. Experiments in a, b and c were carried out in 1 M NaCl, 10 mM HEPES in water while experiments in d were carried out in 3 M CsCl, 10 mM HEPES in a 30/70 v/v glycerol/water mixture to increase the stability of the membranes. The conductance values for the latter case have been scaled down to correspond to the conductance values that would have been obtained in 1 M NaCl in water in order to compare efficiently the different experiments.

[0043] FIG. 7 is a 60-second current trace of an experiment in which the template strand-4-mer was added in a solution containing Chol-dsDNA-CtxA. The zoom of the dotted rectangle is shown in FIG. 6b. The recording was performed at an applied difference of potential of +160 mV and at a Chol-dsDNA-CtxA concentration of 5 nM. In inset 1 regular Chol-dsDNA-CtxAcan be observed while inset 2 shows a distinct increase in conductance from the baseline to the second open state, corresponding to incorporation into the bilayer of a tetrameric pore, with all four binding sites of the template strand-4-mer occupied. Inset 3 shows an increased pore forming activity that does not correspond to the usual CtxA open state conductances and that could not be defined.

[0044] FIG. 8 is a 60-second current trace of an experiment in which the template strand-8-mer was added in a solution containing Chol-dsDNA-CtxA. The zoom of the dotted rectangle is shown in FIG. 6c. The recording was performed at an applied difference of potential of +140 mV and at a Chol-dsDNA-CtxA concentration of 5 nM. In inset 1 and 2 we can see repetitive and distinct increases in conductance from the baseline to the sixth open state, corresponding to incorporation into the bilayer of an octameric pore, with all eight binding sites of the template strand-8-mer occupied.

[0045] FIG. 9 represents traces of current corresponding to experiments with non-fully occupied template strand-8-mer. (A) Direct insertions in planar lipid bilayers of templated pores comprised by four (dotted line) or five (dashed line) monomers. (B) Direct insertions in planar lipid bilayers of templated pores comprised by six monomers (dotted line). An increase in the current is observed, potentially arising from a seventh monomer joining the templated assembly, resulting in a bigger pore (dashed line). Both traces of current correspond to experiments in which the template strand-8-mer was added in a solution containing Chol-dsDNA-CtxA at a concentration of 0.1 nM. The recordings were performed at an applied potential difference of +200 mV.

[0046] FIG. 10 illustrates the effect of a purification process on the assemblies shown for a hexameric assembly. Conductance traces versus time are shown, next to the corresponding histograms. Conductance variations observed for CtxA-dsDNA-chol in presence of 6-mer template (a) without purification or (b) after purification of the assembly. (a) Without purification different conductance fluctuations can be observed, leading to smaller or bigger pore structures. Inset: zoom showing direct increase in conductance to open state 04 followed most likely by fluctuation of one monomer in and out of the membrane. (b) After purification, the removal of free monomers leads to one main population of pores with a size close to the value expected from the use of a 6-mer template. The recordings were carried out with a CtxA-dsDNA-chol concentration of (a) 1 nM and (b) 4 nM and applied potential differences of +200 mV.

[0047] FIG. 11 shows that Addition of 8-mer template in presence of Chol-dsDNA-CtxA leads to high pore formation. This representative trace of current shows intermittent pore formation by Chol-dsDNA-CtxA (added at a concentration of 0.1 nM) followed by an increase in conductance reaching saturation of the amplifier (dashed line, >16 nA). The inset shows a zoom of the beginning of this trace, in which the first three open states of Chol-dsDNA-CtxA can be observed. This recording was performed at an applied potential difference of +180 mV.

[0048] FIG. 12 shows a DNA origami-based template ring. (a) Three-dimensional model of the DNA nanostructure, templating 20 compounds. DNA double helices are depicted as cylinders. Double stranded DNA connections (large radial cylinders) connect the pore forming peptides (thin cylinders in the ring center) which are constituting the final pore. The block structure connected to the ring does not have any function for the pore formation process, but is required to avoid a long single stranded scaffold loop that might interfere with the pore. (b) TEM positive stain images of the DNA origami rings.

[0049] FIG. 13 is a current versus time recording of a DNA origami templated CtxA pore. After insertion of the pore into the lipid membrane, the pore current first fluctuated until reaching a stable value of around 5 nA after about 15 minutes. The pore remained in the bilayer even after reversal of the polarity with the same conductance value. This pore was stable for more than one hour. Two zoomed traces show the insertion process of the pore into the membrane in detail. The current increased with several discrete levels to a maximum of 15 nA. In this experiment, 20 DNA strands each carrying 2 CtxA monomers were attached to the DNA origami ring. A concentration of 20 pM of origami construct was used in that experiment corresponding to a peptide concentration of 0.8 nM.

[0050] FIG. 14 illustrates a long-lived pore formation by DNA-double modified CtxA peptides. (a) Current-voltage relationship of a long-lived single channel formed by DNA-double modified CtxA monomers. The red curve represents a linear fit resulting in a conductance G=6.4 nS. (b) Apparent single-step insertion of a single pore from DNA-double modified peptides at an applied potential of −180 mV (I=−1512±108 pA). (c) This same pore as in (b) remains in the membrane after reversal of the voltage polarity from −20 mV to +20 mV. (d) A 5-min current trace shows the presence of the pore formed by DNA-double modified CtxA monomers (+20 mV) with a measured current of I=124±7.55 pA.

[0051] FIGS. 15A-15G display the chemical structures of the different compounds used. FIG. 15A Commercially synthesized ssDNA-CtxA; FIG. 15B ssDNA-CtxA obtained after an overnight click chemistry reaction between azide-CtxA and DBCO-ssDNA; FIG. 15C ssDNA-CtxA-12T obtained after an overnight click chemistry reaction between commercially synthesized ssDNA-CtxA-azide and a 12T-DBCO oligonucleotide; FIG. 15D ssDNA-CtxA-peg4-12T obtained after an overnight click chemistry reaction between commercially synthesized ssDNA-CtxA-azide and a 12T-peg4-DBCO oligonucleotide; FIG. 15E ssDNA-CtxA-12T obtained after an overnight click chemistry reaction between azide-CtxA-Thiol and a DBCO-ssDNA. The resulting ssDNA-DNA-CtxA-Thiol later reacted overnight with a 12T-maleimide oligonucleotide via a thiol-maleimide reaction; FIG. 15F ssDNA-CtxA for the origami structure experiments obtained after an overnight click chemistry reaction between azide-CtxA and a DBCO oligonucleotide; and FIG. 15G Double-labelled ssDNA-CtxA for the origami structure experiments obtained after an overnight click chemistry reaction between azide-CtxA and an oligonucleotide possessing two DBCO moieties.

[0052] FIG. 16 corresponds to schematic drawings of different macromolecular constructs based on the disclosed compound. Left panel: insertion process of templated Chol-ds-DNA-CtxA-T.sub.12. The cholesterol strand increases the membrane affinity of the construct by inserting into the bilayer. In the transmembrane conformation, the hydrophilic poly-T segment hinders the peptide to flip back out of the membrane. (1) Templated Chol-ds-DNA-CtxA, (2) templated Chol-ds-DNA-CtxA-T.sub.12, (3) CtxA-T.sub.12, (4) templated ss-DNA-CtxA, (5) templated ds-DNA-CtxA, (6) Chol-ds-DNA-CtxA-T.sub.12 templated from both sides. The strand that links the trans DNA segment to the template on the trans side (i.e. the strand that is shown to have a 90 degree kink) may or may not contain a lipid anchor (such as cholesterol) to stabilize the assembly.

[0053] FIG. 17 corresponds to schemes of proposed constructs based on the disclosed compound. (a) Permanently locking of the peptide in a transmembrane conformation by hybridizing a complementary sequence to the second oligonucleotide on the transmembrane terminus. The oligonucleotide to be hybridized can be functionalized with a membrane binder like cholesterol to localize the oligonucleotide on the membrane. (b) Using a receptor-binding moiety as an affinity-increasing agent to target membranes of cells expressing this specific receptor. (d) Affinity moieties can also be attached to the templating strand or structure. We created an embodiment of this principle by attaching cholesterol moieties to the DNA-origami templates. (d) A compound containing multiple pore formers covalently attached to a single DNA strand, allowing creating pores with higher numbers of constituent peptides with the same number of DNA strand. We disclosed this concept already within the DNA-origami based embodiment, where we employed a compound consisting of two CtxA peptides linked to one DNA strand. (e) A compound carrying no linker region but only a template-binding region. Such a simplified compound was used in the DNA-origami based embodiment.

[0054] FIG. 18 illustrates another schematic representation of DNA-assembled peptide pores with programmable pore size that are templated from both sides of the membrane (i.e. from cis and from trans) to increase the pore stability. The design can be the same as was used in previous figures or it can be slightly different, with a shorter oligonucleotide covalently linked to the peptide to force the DNA template sterically to lay flat on the membrane instead of extending above the pore. A flat orientation parallel to the membrane plane would circumvent possible transient or permanent blockade of the pore entrance by DNA strands from the template. Such a blockade, if it were to occur, may reduce the probability of analytes to translocate through the pore. (a) Side view, (b) top view and (c) front view of an octameric CtxA pore, templated from both sides (similar double-templating approaches can be pursued with different pore sizes, e.g. tetrameric pores, dodecameric pores and larger). (d) Assembly scheme showing the hybridization of the DNA-peptide pores from both sides.

[0055] FIG. 19 shows current versus time recordings that illustrate the pore forming behavior of two different peptide-DNA structures. Panels (a) and (b) represent the insertion of CtxA peptides modified with DNA on both termini (dsDNA-CtxA-T.sub.12) in the presence of the 8-mer template only on the cis side, with an applied potential of +180 mV (a) or +50 mV (b). While the recorded current showed current fluctuations between the 7-mer and the 8-mer levels at the high voltage used (+180 mV), reduction of the voltage to +50 mV resulted in improved stability of this octameric pore at the expected current (i.e. conductance) level of an octameric pore (as indicated by the shaded region). The electrolyte solution consisted in 1 M NaCl, 10 mM HEPES, pH 7.3. Panels (c) and (d) illustrate the improvements made with regard to the stability of such DNA-peptide pores when an 8-mer template is present on both sides of the membrane. Although the applied voltage (+100 mV) is higher than the voltage on panel (b), the resulting pore displays a dramatically improved stability with almost no fluctuations of current during hours of recording. This pore remained open for more than 4 h. The electrolyte solution consisted in 150 mM NaCl, 10 mM HEPES, pH 7.3.

[0056] FIG. 20 shows resistive pulses that are consistent with translocation events of macromolecules through DNA peptide pores templated from both sides. The macromolecular analytes consist of a mixture of PEG 4000, PEG 1500, PEG 200 molecules and dextrane sulfate 8000 molecules and were added to a final concentration in the recording chamber of 461 μM, 922 μM, 69 μM and 31 μM respectively. (a) The measured current suggests the presence of an octameric pore incorporated into the lipid membrane. After addition of the macromolecular analyte mixture, multiple current decreases can be observed. The relatively unstable baseline could be due to the template on the trans side not being connected to the pore, leading to unstable pores such as shown is FIGS. 19 a, b. Another explanation could be frequent translocation of small PEG 200 molecules leading to a seemingly unstable baseline. The horizontal dashed line corresponds to 0 pA current. (b) The current value suggests the presence of multiple pores simultaneously in the bilayer. After the last incorporation step and in presence of the mixture of analytes, resistive pulses start to appear, which are consistent with the translocation of macromolecular analytes. This experiment was performed with the same analyte mixture as used in panel a),

[0057] FIG. 21 shows the influence of the DNA templates on the cytotoxic activity of ssDNA-CtxA peptides on A549 cells. The plots show the change in confluence of A549 cells over time as a function of the concentration of templated and non-templated CtxA peptides. For native CtxA peptides, concentrations of 50 to 100 μM were necessary to reduce cell viability. In the case of the DNA-templated pores (tetramers, hexamers, octamers and dodecamers), less than 2 to 5 μM of total CtxA concentration was sufficient to completely stop the replication of the same cells and hence a more than 20-fold lower concentration compared to native CtxA peptides.

DETAILED DESCRIPTION OF THE INVENTION

[0058] The invention encompasses hybrid pore-forming compounds comprising various functional groups that allow tailoring of pore size, stability and membrane affinity, among other benefits. The hybrid pore-forming compound comprises a peptide or protein having covalently attached nucleic acid oligonucleotides on both termini of the peptide or protein. Methods for making hybrid pore-forming compounds are straightforward. Many advantages are provided by the hybrid pore-forming compounds of the invention.

[0059] One particular advantage of the invention is the straightforward attachment of chemical modifications. The compounds presented here do not require further chemical synthesis and can be produced from off the shelf components.

[0060] In one embodiment, cholesterol is linked to the construct as a hydrophobic moiety to increase the affinity of the compound to a lipid bilayer of a membrane. Pore formation is observed at concentrations more than ten times lower than in absence of the cholesterol. The result was achieved without changing the peptide sequence which is one way to increase the potency of an antimicrobial peptide. However, if desired, peptide sequence can be modified as well.

[0061] Other modifications can be accomplished through hybridization to DNA strands with, by way of nonlimiting example, attached fluorescent markers, drug molecules, receptors, antibodies, aptamers, metabolites, etc. These modifications can be useful for drug delivery as well as targeted cell killing.

[0062] Modification of the peptide or protein by covalently attaching ligands to specific sites allows for the creation of biosensors. Such sensors based on resistive pulse sensing allow the detection of specific target molecules. Due to the modularity of the invention, additional functionalizations are straightforward to add or exchange.

[0063] An additional advantage provided by the hybrid pore-forming compounds is an increased lifetime of pores from peptides forming short-lived channels. Long-lived pores are required for resistive pulse sensing in order to allow collection of enough data for subsequent analysis. By providing hybrid compounds having long-lived pores efficiency of the molecule to kill pathogens cells or for drug delivery is increased.

[0064] As described herein, a single stranded DNA on the part of the peptide that is inserted into the membrane, which successfully increases the time during which the peptide is in a transmembrane conformation, namely from tens of milliseconds up to a few seconds or even several minutes.

[0065] Another important advantage of the compounds is the tailored pore size obtained by using DNA templates. The DNA template can range from a unique single strand composed of multiple hybridization sites separated by flexible linkers to complex rigid DNA nanoconstructs, such as a DNA origami-based ring, serving as a scaffold for pore formation. Rigid, large nanostructures used as templates, like the demonstrated origami ring, enable the formation of drastically larger peptide pores. Disclosed herein are simple “one pot” assembly of template structures with a large number of peptide monomer binding sites. The relatively rigid template disclosed leads to more stable peptide pores, as compared to a flexible template from double stranded DNA.

[0066] The compounds of the present invention offer specific benefits for resistive pulse sensing. For example, stable pores are provided, wherein pores are stable general for several minutes. Data analysis is impractical if in addition to translocation events, current fluctuations are caused by changes in the pore size. Increased affinity eases the formation of single pores. DNA assemblies allow simple tuning of the pore diameter and large pore diameters. Larger pores enabled by the compounds of the invention can be utilized to analyze larger target molecules.

[0067] The compounds of the invention also offer various benefits for targeted cell killing and drug delivery. The compounds are pore formers with increased affinity and are active at low concentrations. Pathogen-specific binding moieties, such as receptor binding moieties, antibodies, aptamers, etc. allow targeting selected cells, such as pathogen or other target cells. DNA assemblies in general allow for combination of several different binding motifs and stimulus response compounds with little effort. The compounds are useful for applications such as delivering macromolecules, such as siRNA, DNA, proteins, carbohydrates, etc., wherein minimum pore sizes are required, larger than the analyte size. Larger pores are also more efficient in killing cells because they disrupt cellular homeostasis.

[0068] The compounds of the present invention take advantage of the barrel-stave assembly mechanism of a pore-forming peptide, such as CtxA, to a pore, which is advantageous because various peptides such as CtxA are intrinsically able to form pores from a wide range of number of monomers and pore diameters.

[0069] When a peptide is used to create the compound that has an alpha-helical structure, such as CtxA, design flexibility is present for the assembly and reconfigurations, which can add functionality in order to stabilize the pores, for example. Typically, for instance in natural ion channel proteins, alpha helical proteins are more flexible and have more functionality than pores made from beta-sheet structures.

[0070] The compounds take advantage of a peptide that can be modified on its terminus, one or both of its C-terminus and N-terminus depending on the peptide utilized, without impeding its ability to self-assemble into well-defined pores. The C-terminus modification can be orthogonal in terms of chemical reactivity to the modification on the N-terminus, which means that both ends can be modified and functionalized differently with selective chemical reactions.

[0071] In an embodiment where the C-terminus modification of the peptide includes a thiol group (—SH), the compound is incorporated more readily into a membrane than a native peptide, such as CtxA, and assembles to pores, which demonstrates that the incorporation can be facilitated and controlled.

[0072] Chemical reactions of physical interactions of chemicals or reagents with the C-terminus of a peptide, such as CtxA, on the trans side of the membrane, i.e. the side opposite to the cis side where the compounds were added, can add functionality and stabilize the pores.

[0073] The hybrid pore-forming compounds take advantage of peptides that can self-incorporate into lipid or polymer membranes. Combination with a membrane is quite practical as methods for reconstitution into the membranes are not necessary.

[0074] When peptides such as CtxA are utilized, advantage of voltage-dependent incorporation can be used to form pores on demand, which also can be switched off on demand.

[0075] It is also been demonstrated herein that templating increases the local concentration of pore-forming peptides and therefore the probability of forming pores. This effect biases the assembly towards pores with large diameters. The compounds designed with a spacer strand, template strand, and short, flexible thymidine regions between these segments enables self-assembly of intended pore sizes reducing steric problems. Alternatively, neutral PEG linkers can be inserted. This is no small task since many alternative designs for templating may have steric constraints, low yield of complete assemblies and assemblies of inactive pores, etc.

[0076] The open design of the compound in a scaffold makes it possible to template and assemble pores without adding to the electrical resistance of the pore.

[0077] The assembly of pores can occur either before the experiments combined with purification of assembled pores or pore assembly can be carried out in situ by step-wise addition of the various components and molecules for the assembly. This flexibility in use of the molecules is advantageous to mitigate possible issues of limited solubility of either the individual components or the full assembly. Sequential, in-situ assembly from one or both sides of the membrane) also makes it possible to assemble pores with complex functionality.

[0078] For example, various pore-forming peptides are suitable for use in the present invention. Examples include, but are not limited to, ceratotoxin A, alamethicin, MelP5, melittin, magainin, cecropin, etc., or any synthetically evolved variant of these peptides. Preferably, pore-forming peptides that do not contain post-translational modifications or unnatural amino acids can be used, as they are simple to synthesize and to modify chemically to attach the DNA oligonucleotides.

[0079] Oligonucleotide can be made of nucleotide sequences including but not limiting to DNA, RNA, PNA, LNA, BNA, as well as unnatural nucleic acids. N-terminus oligonucleotides devoid of spacer region could be used, although the template should have a different geometry to account for steric hindrance. Such templates could present longer linkers between each hybridization site, while presenting hydrophobic moieties, receptors, etc. directly to the membrane surface. In another variant, multiple pore-forming peptides could bind to one oligonucleotide, allowing to multiply the number of monomers in a pore while keeping the template strand short.

[0080] The moieties on the spacer strand can be anchoring moieties such as diacyl chains, cholesterol, receptor-binding agents, aptamers, antibodies, metabolites, etc. These moieties can also be elements enhancing the detection of the construct like fluorescent dyes, proteins (or genes encoding for them) or aptamers, metabolic nanoparticles, radioactive markers, quantum dots, etc.

[0081] Pore formation and disassembly can be triggered by chemical, mechanical, light or other signals mediated by the DNA hybridization (toehold mediated strand displacement, light-switchable bases, aptamers, etc).

[0082] C-terminus moieties can be added to obtain long-lived pores. These moieties are hydrophilic or become hydrophilic once reaching the trans side of the membrane (for example due to pH change). These groups can be oligonucleotide sequences (DNA, RNA, PNA, LNA, BNA, as well as unnatural nucleic acids), amino acids or any other hydrophilic group (thiol, azide, biotin, etc.). The anchoring can also be realized by templating the pore-forming peptides on the trans side of the membrane.

[0083] The membrane can be a lipid membrane (phospholipid membrane, archaea membrane, etc.), a synthetic membrane (block copolymer, etc.), can be charged, uncharged or zwitterionic. The pore can also be employed for forming solid-state hybrid pores. Therefore, the pore is inserted into an existing solid-state nanopore to change its properties.

[0084] Here we describe a modular DNA-assembly system based on Ceratotoxin A (CtxA) peptides, for increased membrane affinity, geometrical arrangement of monomers and increased pore lifetime. Like alamethicin, the 36-amino acid peptide CtxA forms short, barrel-stave pores that increase or shrink in diameter as a consequence of association or dissociation of peptide monomers. To this end, we covalently link ssDNA to the N-terminus of CtxA peptides for hybridization of complementary oligonucleotide sequences.

[0085] Linking the peptide-DNA hybrid to DNA-coupled hydrophobic moieties increases membrane affinity. By adding a templating DNA strand with a defined number of binding sites, we bias the pore diameter towards 4-, 6-, 8- or 12-mer assemblies. We demonstrate that the introduction of a hydrophilic DNA domain on the C-terminus side of CtxA, part that is thought to insert into the lipid bilayer.sup.1, stabilizes the hybrid monomers in a bilayer-spanning conformation. This modification turns short-lived CtxA pores with dynamically fluctuating diameters into long-lasting pores with well-defined diameters, which could be beneficial for sensing applications.

[0086] One feature of this construct is the absence of solid, ion-blocking structures, similar to the large αHL cap domain containing a vestibule. In contrast, the assembly we present is a short membrane-spanning channel, geometrically arranged by a loosely packed, open DNA structure, that circumvents additional contributions to resistance from a vestibule or a tightly packed templating structure. Additional distinct advantages of this minimalistic design of an artificial ion channel are the use of solid-phase synthesis of all components with the potential for large-scale production and the possibility to modify the sequence and number of amino acids of the peptide at any position with great flexibility for high-throughput screening. DNA-assembled peptide pores could also be used for applications other than sensing, such as drug delivery or targeted killing of pathogens.

[0087] Results and Discussion

[0088] Assembly of CtxA-DNA hybrid constructs. We design a new peptide-DNA hybrid based on the 36-amino acid peptide CtxA. A 55-base long ssDNA oligonucleotide is attached with its 5′ end to the N-terminus side of the peptide, as shown in FIG. 1. Its sequence consists of two parts: a spacer region and a template-binding domain. Flexible linkers consisting of four thymine (T) bases connect all domains of the construct. A spacer strand hybridizes to the 25-nucleotide-long spacer region to form a rigid duplex. To increase its affinity to the lipid membrane, this spacer strand is modified with a cholesterol moiety on its 3′ end, henceforth referred to as the cholesterol strand. The template-binding region composed of 18 nucleobases allows binding of the peptide-DNA hybrid to a single stranded DNA template with four, six, eight or twelve hybridization sites, separated by linkers of four thymine bases.

[0089] To obtain a more rigid and more defined DNA template, we constructed a ring-shaped origami structure with a diameter of 33.4 nm (inner diameter 20.6 nm). The ring was formed by bending the DNA helices to a circle. A 82×15×11 nm block structure is connected to the ring. The block does not serve the pore assembly but uses up the remaining scaffold to avoid an influence of a large single stranded DNA loop in close vicinity to the pore. 20 staples distributed on the bottom side of the ring carry extensions, ready to hybridize DNA conjugated peptides with a 20 nucleotide-long complementary attachment sequence (instead of the 55-base-long sequence of the other design). As a 20-meric peptide would form a much smaller pore of approximately 8.7 nm diameter, linkers are required. Therefore, peptide-DNA conjugates are hybridized in a distal conformation to the structure with the connecting DNA duplex acting as a 6.1 nm linker between origami structure and peptide. Two additional thymine bases were inserted on both sides of this double stranded attachment, serving as flexible linkers. To further avoid steric and electrostatic repulsion of the double stranded DNA linkers that might impede peptide pore assembly, we connected the peptide via an uncharged PEG linker to the attachment oligonucleotide. Alternatively, a strand carrying two reactive DBCO groups attached via a PEG linker was used to form up to 40-meric peptide pores.

[0090] Similar to the previously described design, additional functional moieties can easily be attached via staple extensions. We added eight further extensions with an orthogonal attachment sequence to the bottom of the ring that can carry cholesterol moieties to increase the membrane affinity of the origami structures to the membrane.

[0091] We use the CtxA peptide for this study as pores assembled from small peptides have attractive characteristics such as forming barrel-stave pores with well-defined conductance levels, and forming channels with a pore length ≤5 nm, potentially increasing the resolution of sensing applications. Its low molecular weight allows large-scale synthesis of the peptide as well as chemically functionalizing the peptide. Moreover, as the peptide is positively charged, it readily binds stronger to the negatively charged outer membranes of many pathogens compared to mammalian cells. This feature makes CtxA-based pore assemblies attractive for targeted cell killing applications

[0092] FIGS. 1b-e illustrate the loose and open molecular design of the DNA-assembled peptide pore we designed. Ions can flow between the strands of the DNA-based scaffold with the advantage that the scaffold does not significantly increase the access resistance to the transmembrane segment of the pore. This design focuses the voltage drop to the membrane-spanning peptide part of the pore rendering it sensitive to conductance changes and thus increasing the signal to noise ratio for applications such as resistive pulse sensing.

[0093] Hydrophobic functionalization for increased pore forming activity. We introduce a cholesterol moiety to increase the affinity of the DNA-assembled peptide pore to the membrane. FIG. 2 compares the activity of different DNA-assembled peptide pores with pores from native CtxA peptides recorded at the same total peptide concentration of 5 nM in single channel recordings with planar lipid bilayer experiments. Typically, native CtxA, CtxA-DNA without any additional strands (ssDNA-CtxA) and CtxA-DNA with the spacer strand (dsDNA-CtxA) exhibit little to no pore formation (FIGS. 2a-c) at these low peptide concentrations. In contrast, CtxA-DNA hybridized to the cholesterol strand (Chol-dsDNA-CtxA) shows frequent pore-formation events (FIG. 2d). A magnification of the current recording from this molecule shows well-defined conductance steps similar to the ones observed with native CtxA at concentrations above 10 nM (FIG. 3).

[0094] These results indicate that N-terminal dsDNA modification of the CtxA peptide does not inhibit its insertion in membranes but rather favors it when this modification exposes a hydrophobic group. Moreover, the conductance behavior of the resulting pores is also very similar to the one from native CtxA, with values increasing non-linearly, with the pore size. In addition to this non-linear increase in conductance, the zoom of the current trace induced by Chol-dsDNA-CtxA, shown in FIG. 2e, also displays stepwise increases of equal conductance values, indicating the simultaneous presence of multiple pores of the same size in a transmembrane conformation.

[0095] While a minimum concentration of 10-20 nM is required for native CtxA to display significant pore formation, concentrations as low as 1 nM are sufficient to observe frequent pore formation events by Chol-dsDNA-CtxA.

[0096] Interestingly, control experiments with ssDNA-CtxA added at concentrations above 10 nM produce ill-defined conductance fluctuations instead of the usual well-defined behavior of native CtxA (FIG. 4). The N-terminally dangling ssDNA may cover the pore or enter its lumen, leading to variable conductance levels. Addition of the complementary cholesterol strand restores the well-defined stepwise conductance behavior as shown in FIGS. 2d, e, possibly because the increased stiffness of the resulting dsDNA strand at the N-terminal entrance of the pore reduces the probability that it covers the pore or enters into the pore (FIG. 5).

TABLE-US-00001 TABLE 1 Comparison of the open state conductance values for native CtxA and Chol-dsDNA-CtxA, measured at an applied potential difference of +180 mV. Conductance (pS) Chol-dsDNA- Area-equivalent Open states Native CtxA CtxA diameter (nm).sup.b O1 (3-mer)   50 ± 25   35 ± 20 0.1 ± 0.1 O2 (4-mer)  510 ± 30  445 ± 155 0.5 ± 0.3 O3 (5-mer) 1645 ± 200 1515 ± 320 1.0 ± 0.4 O4 (6-mer) 3125 ± 305 3100 ± 170 1.4 ± 0.3 O5 (7-mer) 4630 ± — 5200 ± 600 1.9 ± 0.6 O6 (8-mer) 6600 ± — 6960 ± 120 2.3 ± 0.3 O7 (9-mer) 9500 ± 400 2.8 ± 0.5 O8 (10-mer) 11640.sup.a 3.3 ± — O9 (11-mer) 14140.sup.a 3.8 O10 (12-mer) 16760.sup.a 4.3 .sup.aThe conductance values of O8 to O10 are extrapolated from a fitting function to the experimental data of the first 7 open states. This function assumes a geometrical arrangement of the monomers, a peptide diameter of 1.2 nm and a pore length of 2.7-4 nm. .sup.bThe diameter is estimated based on the measured and extrapolated conductance values G of Chol-dsDNA-CtxA. Errors are calculated from standard deviation of the mean. The equation we used was shown by Cruickshank et al., Biophysical Journal 73, 1925-1931 (1997) and takes into account two access resistances (one for entrance and one for exiting the pore) and the channel resistance: [00001]$d=\frac{\rho G}{\pi }\times \left(\frac{\pi }{2}+\sqrt{\frac{{\pi }^2}{4}+\frac{4\pi l}{\rho G}}\right)$where the channel length (/) is comprised between 2.7 nm, the shortest channel length based on the amino acid sequence of the transmembrane segment of the peptide, and 4 nm, assuming the transmembrane channel spans the whole bilayer, a buffer resistivity (ρ) of 0.1130 Ω m (from a measured conductivity σ = 8.85 S m.sup.−1). The experiments with templated 12-mer were carried out in 3 M CsCl, 10 mM HEPES in 30% (v/v) glycerol in water to increase the stability of the lipid membranes (measured conductivity σ =13.99 S m.sup.−1).

[0097] Table 1 shows that the conductance values of the first six open states (O1 to O6) for native CtxA and Chol-dsDNA-CtxA are not significantly different (paired-sample t-test, p=0.36). This result implies that the addition of dsDNA on the N-terminal side of CtxA does not significantly change the diameter of the formed pores.

[0098] Conventionally, pore-forming peptides are engineered by modifying the amino acid sequence of peptides to increase their pore-forming activity at low concentrations.sup.3-5. We show here that the addition of DNA strands with hydrophobic cholesterol moieties achieves the same effect without the need for sequence alterations, which will affect peptide function.

[0099] Formation of pores with defined sizes by DNA templating.

[0100] Further exploring the potential of combining a DNA strand and a pore-forming peptide, we introduce a DNA template strand consisting of a repeating sequence, complementary to the template-binding region sequence of CtxA-DNA. We aim to form pores comprised of a defined and constant number of monomers that compares to the number of hybridization sites on the template. To demonstrate feasibility, we show the influence of template strands containing four, six, eight or twelve hybridization sites for CtxA-DNA (‘4-mer template’, ‘6-mer template’, ‘8-mer template’ and ‘12-mer template’). As shown in Table 1, such tetra-, hexa-, octa- and dodecameric assemblies result in pores with 0.5±0.3, 1.4±0.4, 2.3±0.5 nm and 3.9±0.5 nm diameter, based on the measured O2, O4, and O6 conductance values of the Chol-dsDNA-CtxA and an extrapolation of these measured values to find an estimated O10 conductance value.

[0101] FIG. 6a shows a current trace across a planar lipid bilayer recorded at a constant applied potential difference of +180 mV in the presence of 5 nM Chol-dsDNA-CtxA before addition of a template strand. Dynamically changing conductance levels with discrete and recurring step sizes are consistent with a barrel-stave model of pore formations and are represented by different peaks in the corresponding histogram. After addition of a 4-mer template, the relative frequency for observing conductance levels changes dramatically: FIG. 6b shows a strong bias of the distribution towards the conductance of the 4-mer as expected. At the time resolution of our recordings (˜20 μs), we observe single-step conductance changes of 330±26 pS, to a level that corresponds to the expected tetrameric pore (FIG. 6b and FIG. 7). The O2 level is only interrupted by very short decreases in conductance. We hypothesized that when a monomer changes from a membrane-spanning conformation to a membrane-laying conformation, it cannot escape the assembly and quickly inserts back into the lipid membrane due to its high local concentration. The conductance rarely increases to levels above the tetrameric pore, further suggesting that this pore comprises four monomers linked together, since free monomers lead to dynamic fluctuations of conductance levels as shown in FIG. 6a.

[0102] Dynamic fluctuations of pore diameters however still occurred (FIG. 7, panel 1), with the conductance increasing sequentially from the baseline to O1, O2 and O3. We attribute these results to the presence of free Chol-dsDNA-CtxA hybrids that interact with themselves to form trimers or interact with the templated CtxA-DNA assembly to form off-target pore sizes.

[0103] Further increasing the length of the template to eight or twelve hybridization sites allows the formation of larger pores. We show a two-second recording of such an octameric assembly in FIG. 6c (full 60 s current trace in FIG. 8) and a two-second recording of a dodecameric assembly in FIG. 6d. When using an 8-mer template, we can discern several discrete transitions to conductance values of 8470±700 pS corresponding to insertion in the bilayer of large pore structures. Despite the low concentration (5 nM) of DNA-assembled peptides, we however also observe conductance variations of smaller pores (average conductance value: 1460±270 pS) leading to a non-stable current baseline resulting in discrete conductance steps with a difference of 7010±700 pS. This value is consistent with a pore comprised by eight monomers moving in and out of the bilayer. The histogram of the full 60 s current trace shows this dominant population which corresponds to the 8-mer (O6) as expected by the use of template strand-8-mer. With a template possessing 12 hybridization sites, we observed even larger transitions to conductance values that were closest to the 11- and 12-mer based on the Table 1 values.

[0104] While the current traces we show in FIG. 6 display the formation of pores having the expected size, this result was not obtained on each attempt and we often observed fluctuations of the conductance values that might be due to smaller or bigger CtxA open states. We attribute the smaller pores to structures with template strands not fully occupied by the peptide-DNA hybrids. Alternatively, a template strand fully occupied but having one or more monomers not adopting a transmembrane conformation would also lead to smaller-than-expected pores. In FIG. 9, we show two examples of direct insertion into the bilayer of pores comprised of four, five and six monomers when we use the 8-mer template. The occasional observation of conductance increases from one predominant open state to a higher state may correspond to free peptide-DNA monomers joining a templated pore. The current trace we show in FIG. 9b may reflect this possibility. Peptide monomers from different template strands, fully occupied or not, can also aggregate to form larger structures. Two templated pores could for example combine, forming a bigger pore. Using a large excess of DNA-modified monomers compared to the template strand and removal of unbound monomers by purification could potentially reduce the possibility of free peptide-DNA monomers joining templated pores.

[0105] We compare in FIG. 10 two representative experiments displaying pore-forming activity upon addition of Chol-dsDNA-CtxA in presence of 6-mer template, before and after purification of the assemblies. The current trace shown in FIG. 10a is recorded under the same experimental conditions as in FIG. 6, without purification. The initial current transitions seem to correlate with a pentameric activity followed by fluctuations of the conductance between O3 (5-mer) and O4 (6-mer) open states, as shown in the inset of FIG. 10a. As the first increase in conductance was from baseline to O4 conductance value, we attribute these events to a fully templated 6-mer template with one monomer fluctuating in and out of the bilayer. The conductance value later increases by ΔG=1±0.1 nS—which might be due to formation of a separate pore—but the fluctuations between 5-mer and 6-mer continue to occur as described by the dotted lines. FIG. 10b displays a 1-second portion of a current trace of the same construct but after purification of the fully assembled 6-mer templates. We now observe one predominant conductance state with a value comprised between the conductance value of a pentamer and the one of a hexamer. While we also observed tetrameric activity, the purification of the templated assemblies circumvents the possibility for free monomers to join the constructs, resulting in more defined structures. The smaller assemblies observed here are most likely induced by monomers that do not adopt the transmembrane conformation.

[0106] In addition to biasing the size of the formed pores towards the desired size, the presence of a template strand also leads to an increased pore formation at lower concentration than without the template. In close to 50% of the experiments, addition of the 8-mer template results in incorporation of many pores in parallel, sometimes exceeding the amplifiers current limit (˜16 nA). We present in FIG. 11 an example of increased pore-forming activity in the presence of the 8-mer template in a solution containing Chol-dsDNA-CtxA (0.1 nM). We observe regular Chol-dsDNA-CtxA conductance states with varying levels at the beginning of this current trace, followed by a rapid increase in the conductance exceeding the amplifier limit while the bilayer remains intact. In absence of the template strands, we rarely observe similarly extensive pore-forming activity even at peptide concentrations that are 200-fold higher (20 nM). This high activity clearly demonstrates a strong enhancement of the template and the cholesterol strands on pore formation, especially considering the low peptide-DNA concentration (0.1 nM) used in these experiments.

[0107] Formation of pores with defined sizes by DNA templating.

[0108] For the assembly of size tuneable pores for resistive pulse sensing, discrete and well-defined DNA objects are required. By employing several partly complementary strands, only flexible constructs with limited complexity can be formed as shown by two recent papers from Henning-Knechtel et al..sup.7 and Spruijt et al..sup.6. In contrast, the DNA origami technique allows the assembly of large, relatively rigid and geometrically well-defined templates with desired shape and atomistic precision. Readily available tools for simple design, production and purification ease the development of these macromolecular constructs. Consisting of one scaffold strand of about 8000 bases and about 200 staples, origami structures can provide a large number of attachment sites for pore formers and other functional groups, for example by affinity moieties, enhancing binding to membranes or moieties to capture target analytes for enhanced detection. Similar extended nucleic acid structures can also be obtained using other methods like single stranded tile assembly or RNA origami.sup.8,9. FIG. 12 shows a 3D representation together with a TEM image of such DNA origami-templated CtxA pores. FIG. 13 shows a current trace of a that compound across a planar lipid bilayer. This pore was formed by an origami ring carrying 20 DNA binding sites. Each attached DNA oligonucleotide carries two CtxA monomers attached via PEG linkers, leading to a maximum of 40 templated CtxA peptides. The insertion of the pore occurred with several discrete levels that may be due to the presence in a transmembrane conformation of different number of peptide monomers. The first level, with a current of 0.45±0.25 nA, corresponds to the insertion of 5-8 CtxA monomers is rapidly followed (after around 10 ms) by a second level with a current of 3.6±0.8 nA. This level, corresponding to around 16 CtxA monomers assembled to form a pore, was interrupted by levels of smaller current (1.3±0.4 nA) that we interpreted as a few monomers flipping out of the pore leading to a smaller pore consisting of about 13 monomers. The current kept rising, reaching after 15 seconds a maximum value of 13.8±0.6 nA corresponding to about 30 monomers. 15 seconds later, the pore current started to decrease to finally stabilize at a current value of 5.4±0.1 nA corresponding to about 19 peptide monomers in the pore. This pore was then stable for more than an hour and the voltage could be decreased to 80 mV and even reversed to −100 mV without losing the pore from its inserted state in the membrane. In this experiment, we did not use a cholesterol anchor to increase the affinity of the constructs as they were already active at very low concentrations. In other experiments, with and without cholesterol anchors, we observed long-lived formation of large pores from single-labelled DNA-CtxA conjugates.

[0109] We estimated the number of monomers in a pore extrapolated with a polynomial fit to the experiments data from the conductance value of the first six open states of native CtxA.

[0110] Hydrophilic modification of the CtxA on the C-terminus results in long-lived open states.

[0111] Applications such as sensing require long-lived, stable pores in lipid or polymer membranes. To design CtxA-DNA hybrids that form longer-lasting pores, we added an additional hydrophilic segment consisting of 12 thymine bases to the C-terminal part of the peptide, yielding a CtxA covalently attached on both ends to a DNA strand. We hypothesized that this segment might trap CtxA in a transmembrane conformation once the C-terminal section is transmembrane. This transmembrane DNA sequence allows hybridization of a complementary 12 adenine-bases oligonucleotide that is added on the trans side of the membrane. After hybridization, this double stranded segment is even less likely to flip back to the cis side of the membrane and potentially further stabilizing the pore. Modifying this 12 adenine-based oligonucleotide with a cholesterol moiety could anchor the polyA sequence to the bilayer, increasing the probability for hybridization to the 12 thymine bases of the C-terminal part of the peptide.

[0112] FIG. 14a shows a current-voltage curve of a single channel inserted in a membrane with a relatively linear behavior. This behavior is unusual for a voltage-gated peptide like CtxA, for which pore formation usually only occurs above a threshold voltage. This result corroborates the assumption that the peptide is stabilized in its transmembrane conformation by the hydrophilic C-terminal modification with T.sub.12. FIGS. 14b-d show insertion of a pore formed by DNA-double modified CtxA monomers into a lipid membrane (b) and its presence in the bilayer at voltages as low as −20 mV (c). We hypothesize that the hydrophilic T.sub.12 tail hinders the peptide from returning to a conformation that is associated peripherally to the membrane like native CtxA. As FIGS. 14c-d show, the nanopore furthermore stayed open within the lipid membrane after reversal of the polarity and remained open for at least 20 minutes without modification of its conductance value (d). In contrast, native CtxA or DNA-assembled CtxA pores without the C-terminal T.sub.12 modification typically resulted in pores with lifetimes in the range of hundreds of milliseconds about three orders of magnitude shorter.

[0113] To explore programmed assembly of CtxA peptides with C-terminal T.sub.12 modification to long-lived pores of the desired size, we repeated the experiments with an 8-mer template (FIGS. 19a,b). While the recorded current showed current fluctuations between the 7-mer and the 8-mer levels at the high voltage used (+180 mV, FIG. 19a), reduction of the voltage to +50 mV resulted in improved stability of this octameric pore (FIG. 19b). We obtained such long-lasting octamers in 30% of the experiments involving the 8-mer template (N>10); in the remaining experiments, we observed long-lasting pores comprising fewer monomers than expected. Occasionally, we also observed brief, transient fluctuations in conductance, which we again attribute to one or a few peptide monomers leaving and joining the assembly.

[0114] Altogether, these results support the hypothesis that the hydrophilic C-terminal modification with T.sub.12 stabilized the peptides in their transmembrane conformation. Adding a template strand additionally allowed defining the size of the pores. Modifying the transmembrane part of the CtxA peptide, hence provides a strategy for prolonging the open state of pores from pore-forming peptides, which otherwise fluctuate dynamically between conductance levels.

[0115] Taking advantage of the T.sub.12 tail present on the trans side of the membrane, we designed another oligonucleotide sequence possessing two hybridization regions, one to bind to the template and the other one to bind to the T.sub.12 tail of CtxA. We first prepared and purified structures composed of an 8-mer template and eight of these new oligonucleotide sequences (see FIG. 18d, bottom part), referred to as the trans-templating structure. After pore formation by DNA-CtxA-T.sub.12, the trans-templating structure hybridized to the CtxA peptides through its T.sub.12 tail (FIG. 18d) and led to an improved pore stability (FIGS. 19c,d). The pore showed in example in FIGS. 19c,d remained in an open state in the membrane for more than four hours.

[0116] By modifying the transmembrane part of the CtxA peptide, we provide a pathway for formation of long-lived pores, from pore-forming peptides which otherwise lead to dynamic fluctuations between conductance levels.

[0117] After obtaining long-lived and stable pores, we added a mixture of analytes (PEG 4000, PEG 1500, PEG 200 and Dextrane Sulfate 8000, with final concentrations of 461 μM, 922 μM, 69 μM and 31 μM respectively) on the compartments on both sides of the membrane. We show in FIG. 20 current traces from two different experiments after addition of the analytes to octameric pores formed by CtxA peptides modified with DNA on both ends. FIG. 20a shows multiple decreases in current from a baseline corresponding to an open octameric pore. FIG. 20b also displays several current decreases that could correspond to translocations of molecules. The current first increased in multiple steps which we attributed to the insertion of multiple pores in the membrane. The potential translocations, however, only started to appear after the last current step which would suggest that the analytes could only pass through this pore. Another hypothesis is that the pore grew in steps and the translocations appeared when the pore was big enough to allow the molecules to pass through the pore.

[0118] Assembling CtxA pores with the 4-mer, 6-mer, 8-mer or 12-mer templates prevented the replication of cancer cells at lower concentrations than native CtxA.

[0119] DNA templating of CtxA pores could serve to enhance toxicity of pores in applications of targeted cell killing. Biologically, the antimicrobial peptide CtxA is produced by the Medfly Ceratitis capitata, in order to protect its eggs.sup.10. As cytotoxicity of pore-forming antimicrobial peptides correlates with pore size and pep-tide concentration.sup.11, we hypothesized that templating CtxA would kill pathogen or cancer cells at lower concentrations than native CtxA, because templating results in a high local concentration and in larger pores than non-templated CtxA.sup.11.

[0120] To investigate the cytotoxic activity of templated DNA-peptides, we monitored the growth of the epithelial lung cancer cell line A549 upon addition of 4-mer, 6-mer, 8-mer or 12-mer templated CtxA peptides compared to adding native CtxA as control. As a metric for cell viability, we determined the change in confluence over time for each peptide concentration. FIG. 21 shows that 50 to 100 μM of native CtXA were necessary to reduce cell viability. In the case of the DNA-templated pores of CtxA, less than 5 μM of total CtxA concentration was sufficient to completely stop the replication of the same cells. Similar toxicity required 100 μM of native CtxA, and hence a more than 20-fold higher concentration compared to templated CtxA peptides.

CONCLUSION

[0121] A modular assembly platform based on the 36-amino acid pore-forming peptide CtxA using DNA nanotechnology is disclosed. Covalently linking the peptide to a single-stranded DNA oligonucleotide also allows simple functionalization of the peptide-DNA hybrid with a cholesterol-bearing DNA strand, drastically increasing the peptide's affinity to the membrane as well as its propensity to form well-defined conductance states. Employing DNA template strands with a defined number of binding sites for the DNA-modified peptide monomers preferentially leads to pores of predefined size. We demonstrate tetrameric, hexameric, octameric and dodecameric assemblies with estimated inner diameters ranging from approximately 0.5 to 3.9 nm based on the conductance values. While these templated assemblies still formed short-lived pores like native CtxA, the addition of a hydrophilic DNA domain to the transmembrane side of the CtxA peptide traps the peptide monomers in a membrane-spanning conformation. This modification results in long-lived pore formation, which can last as long as 30 minutes at one stable and constant conductance level. Alternatively, rigid DNA origami structures allowed templating more monomers (up to 40 in this presented attempt) and thus forming larger pores. These origami pores were also stable for longer times than native CtxA, reaching open pore duration of more than 80 minutes.

[0122] DNA-mediated assembly of pore-forming peptides provides several distinct advantages compared to pores traditionally used in nanopore sensing: (i) sub-nanometer fine-tuning of the pore diameter with single monomer increments by using templating structures, (ii) the presence of a loose and open DNA-based frame, without additional resistance caused by a vestibule, (iii) sequence-specific orthogonal hybridization chemistry for targeting defined positions in the structure, (iv) straight-forward design with three different species, (v) potential to open new pathways for adding specific functions such as receptors, antibodies or aptamers, and (vi), possibility to kill cancer cells at more than 20-fold lower total peptide concentrations compared to non-templated CtxA peptides. This result, in principle, makes it possible to use the addition of a DNA template to trigger pore formation in situ by using a DNA-CtxA peptide concentration that is too low for killing before addition of the template. Future extensions of this work may target specific cell types.sup.12, pathogens or analytes by hybridizing ssDNA-tagged ligands, antibodies or aptamers to the construct.

[0123] In summary, DNA engineering of CtxA monomers with a straight-forward design makes it possible to form long-lasting, well-defined pores with constant single-channel conductance that can potentially be useful for sensing, drug delivery and pathogen cell killing applications.

[0124] Methods

[0125] DNA-Modified Peptides Preparation

[0126] The compounds are commercially synthesized from a CtxA derivative bearing the DNA oligonucleotide (ssDNA-CtxA) or reacted in-house. Therefore, an CtxA containing an azide group (Azide-CtxA) on its N-terminus reacts overnight via click-chemistry to dibenzocyclooctyne (DBCO)-containing oligonucleotide of the desired sequence. CtxA with a hydrophilic tail is formed from ssDNA-CtxA modified with an azide group on its C-terminus (ssDNA-CtxA-Azide) or with an Azide-CtxA having a thiol group on its C-terminus (Azide-CtxA-Thiol). ssDNA-CtxA-Azide and Azide-CtxA-Thiol then react with an excess 12T-DBCO or 12T-maleimide to form ssDNA-CtxA-12T. High Pressure Liquid Chromatography (HPLC) is then used to remove the excess, unreacted species, ssDNA-CtxA-12T monomers are added in large excess to a solution containing the template strand and HPLC is used to collect the full assemblies and remove the excess ssDNA-CtxA-12T monomers. We confirmed the successful conjugations by HPLC peptide mapping, SEC or SAX. The chemicals were solubilized either in pure water or in TE buffer. FIG. 15 shows the chemical structure of the linkers used for the different compounds we used.

[0127] Lipids Preparation/Planar Lipid Bilayer Formation

[0128] The lipid composition for the bilayers consisted of POPC, DOPE and POPS with a 7:3:1 (w/w) ratio (PC/PE/PS) or for the origami structures in diphytanoyl phosphatidylcholine (DiPhyPC). We dissolved the resulting lipid mixture in pentane (total lipid concentration: 10 mg/mL). The buffered electrolyte solution consisted of 1M NaCl, 10 mM HEPES, pH 7.4 or 3M CsCl, 10 mM HEPES, pH 7.4 in 30% glycerol.

[0129] We designed Teflon chambers with one big compartment (containing a maximum volume of 1.5 mL) and one small compartment (containing a maximum volume of 200 μL). We pretreated Teflon films (Eastern Scientific LLC, Rockville, Md.) with apertures of 50 μm by pipetting 1 μL of pre-treatment solution—hexadecane in hexane 2.5% (v/v)—onto both sides of the aperture of the Teflon film. We then mounted the Teflon film in a Teflon chamber using high-vacuum grease (Dow Corning Corporation), separating the two compartments of the Teflon chamber. The only connection between the compartments was the aperture in the Teflon film over which we formed virtually solvent-free planar lipid bilayers using the technique described by Montal and Mueller.sup.10. Briefly, we added electrolyte solution to both compartments (1.2 mL in the bigger compartment and 160 μL for the smaller one) and spread 1-2 μL of lipid solution onto the surface of the buffered electrolyte solution. After the solvent evaporated, a lipid monolayer (Langmuir film) formed at the air-water interface. We raised and lowered the electrolyte solution until we measured a baseline current (−3 pA<|<3 pA) indicating that a bilayer had formed. We then thinned the membrane by lowering and raising the electrolyte solution in one compartment, until we measured a capacitance of 60±10 pF. To monitor capacitance, we applied a triangular voltage and the capacitance was either calculated by the amplifier or determined visually using the manual capacitance compensation of the amplifier. Prior to adding the peptide, we checked the stability of the bilayer (absence of leak currents, expected noise level) by applying transmembrane voltages of up to 200 mV for 5 min at both polarities. We used one of two different amplifiers, an EPC7 (HEKA Instruments Inc., Holliston, Mass., USA) or a BC-535 (Warner Instruments Hamden, Conn., USA)) and connected both compartments to the amplifier through two Ag/AgCl pellet electrodes (Warner Instruments Hamden, Conn., USA). We placed the Teflon chambers inside Faraday cages, on BM4 vibration isolation platforms (Minus K® Technology, Inc., Inglewood, Calif., USA). We carried out all experiments at room temperature (22±1° C.) and tested all setups with different model cells to confirm proper functionality of the amplifiers and function generators. We sampled currents at 50 kHz and low-pass 10 kHz. We analysed data using OriginLab (OriginLab Corporation, Northampton, Mass., USA) and pClamp (Molecular Devices, Sunnyvale, Calif., USA) software.

[0130] Flexible Template Experiments

[0131] We performed the experiments with the peptide-DNA hybrid in two different ways: (i) we added all components sequentially—ssDNA-CtxA first, then the template strands, one after the other—or (ii) we mixed the ssDNA-CtxA with the spacer strand (with or without cholesterol) and the template strand in a DNA LoBind tube (Eppendorf Tubes, Hamburg, Germany) and let them react overnight at room temperature (pre-incubation experiments). We also performed the pre-incubation experiments with and without HPLC-SEC purification. In the case of unpurified assemblies, we added the DNA-modified peptides and the spacer strand stoichiometrically and then added the n-mer template (with n equal to 4, 6 or 8) at a ratio of 1 mole for n moles of DNA-modified peptide. When purifying the assemblies, we added the peptide-DNA monomers in excess compared to the template strand (at least 5 n moles of peptide for 1 mol of n-mer template) and removed the excess monomers using HPLC-SEC purification (Agilent SEC 3 column, 300 mm, 300 nm pore size, 4.6 mm internal diameter). We then stored the collected assemblies at +4° C. for later use and add the spacer strand in excess prior to the experiment. When a cholesterol moiety is bound to the spacer strand, we heat the strand for 5 minutes at +60° C. to avoid aggregation of the cholesterol moieties. For the experiments with sequential addition, we added the various components at concentrations in the nanomolar range, while for pre-incubation experiments, we added all the components in micromolar concentrations and later diluted the samples prior to adding them to the buffered electrolyte solution.

[0132] Origami Structure Assembly Protocol

[0133] To fold the origami structures, we mixed 50 nM scaffold with a 4-fold molar excess of staple strands in a 20 mM MgCl.sub.2 solution containing TE buffer. We heated the structures to 90° C. and cooled them down slowly from 57° C. to 49° C. within 18 hours. Subsequently, we removed the excess staples from the folded structure by PEG precipitation as described elsewhere.sup.11. Within the PEG purification, we changed the buffer to TE containing 1M NaCl.

[0134] For attachment of CtxA to the origami constructs, we added the peptide-oligonucleotide conjugates to the structures with a five times molar excess and incubated for at least one hour at room temperature. We finally removed unbound peptides and peptide-DNA conjugates by HPLC size exclusion chromatography (Agilent SEC 3 column, 300 mm, 300 nm pore size, 4.6 mm internal diameter).

[0135] We heated the cholesterol oligonucleotides to 60° C. for 5 minutes before attachment to the origami structures. When required, we added the cholesterol strands with a final concentration of 1 μM to the origami constructs directly before use, without further purification.

[0136] Cell Culture Experiments.

[0137] We grew lung epithelial A549 cells in RPMI 1640 medium containing 25 mM HEPES, 1% L-glutamine, 10% fetal bovine serum and added 1% penicillin-streptomycin for prevention of bacterial contamination of cell cultures. We seeded the cells in 96-well plates (TPP, Trasadingen, Switzerland) with 10′000 cells per well, in a total volume of 50 uL per well. We placed the plates in the IncuCyte Zoom imaging system (Essen Bioscience, Ann Arbor, Mich., USA) placed inside of an incubator set to 37° C. and 5% CO2 and monitored the growth of the cell by monitoring the change in confluence over time. We seeded the cells and left them to adhere and replicate for 20 h before adding the templated CtxA-DNA assemblies or native CtxA as control. We then measured the cell confluence every 2 h. To estimate the effect of the different peptide assemblies on the change in confluence over time, we fitted the observed confluence levels after peptide addition linearly.

[0138] DNA and Amino Acid Sequences

[0139] FIG. 15 furthermore illustrates the linker chemistry of the compound.

[0140] CtxA's amino acid sequence is shown in SEQ ID N01.

[0141] The nucleotide sequences of the oligonucleotides used with the flexible templates and shown in SEQ ID NO2 to 4.

[0142] Sequences are provided from 5′ to 3′ as generated from the design software cadnano.

[0143] The sequences of the block component staples are shown in SEQ ID NO5 to 148. The sequences of the unmodified ring staples are shown in SEQ ID NO149 to 165. The cholesterol strand sequence and the sequence of the strand used to attach the CtxA peptides are shown in SEQ ID NO166 and 167. The sequences of the peptide attachment sites are shown in SEQ ID NO168 to 188. The sequences of the cholesterol attachment sites are shown in SEQ ID NO189 to 196. The scaffold sequence is shown in SEQ ID NO197.

[0144] For the avoidance of doubt, the compositions of the present invention encompass all possible combinations of the components, including various ranges of said components, disclosed herein. It is further noted that the term “comprising” does not exclude the presence of other elements. However, it is to also be understood that a description of a product or composition comprising certain components also discloses a product consisting of said components. Similarly, it is also to be understood that a description of a process comprising certain steps also discloses a process consisting of the steps.

[0145] In accordance with the patent statutes, the best mode and preferred embodiment have been set forth; the scope of the invention is not limited thereto, but rather by the scope of the attached claims.

REFERENCES

[0146] 1 Langecker, M. et al. Synthetic lipid membrane channels formed by designed DNA nanostructures. Science 338, 932-936, doi:10.1126/science.1225624 (2012). [0147] 2 Krishnan, S. et al. Molecular transport through large-diameter DNA nanopores. Nat Commun 7, 12787, doi:10.1038/ncomms12787 (2016). [0148] 3 Burns, J. R., Stulz, E. & Howorka, S. Self-assembled DNA nanopores that span lipid bilayers. Nano Lett 13, 2351-2356, doi:10.1021/n1304147f (2013). [0149] 4 Burns, J. R. et al. Lipid-bilayer-spanning DNA nanopores with a bifunctional porphyrin anchor. Angewandte Chemie (International ed. in English) 52, 12069-12072, doi:10.100.sup.2/anie.201305765 (2013). [0150] 5 Burns, J. R., Seifert, A., Fertig, N. & Howorka, S. A biomimetic DNA-based channel for the ligand-controlled transport of charged molecular cargo across a biological membrane. Nature Nanotechnology 11, 152, doi:10.1038/nnano.2015.279 https://www.nature.com/articles/nnano.2015.279#supplementary-information (2016). [0151] 6 Spruijt, E., Tusk, S. E. & Bayley, H. DNA scaffolds support stable and uniform peptide nanopores. Nature Nanotechnology, doi:10.1038/s41565-018-0139-6 (2018). [0152] 7 Henning-Knechtel, A., Knechtel, J. & Magzoub, M. DNA-assisted oligomerization of pore-forming toxin monomers into precisely-controlled protein channels. Nucleic Acids Research, gkx990-gkx990, doi:10.1093/nar/gkx990 (2017). [0153] 8 Ke, Y., Ong, L. L., Shih, W. M. & Yin, P. Three-Dimensional Structures Self-Assembled from DNA Bricks. Science 338, 1177-1183, doi:10.1126/science.1227268 (2012). [0154] 9 Wei, B., Dai, M. & Yin, P. Complex shapes self-assembled from single-stranded DNA tiles. Nature 485, 623, doi:10.1038/nature11075 https://wvvw.nature.com/articles/nature11075#supplementary-information (2012). [0155] 10 Marchini, D., Marri, L., Rosetto, M., Manetti, A. G. & Dallai, R. Presence of Antibacterial Peptides on the Laid Egg Chorion of the Medfly Ceratitis capitata. Biochemical and biophysical research communications 240, 657-663 (1997). [0156] 11 Saint, N., Marri, L., Marchini, D. & Molle, G. The antibacterial peptide ceratotoxin A displays alamethicin-like behavior in lipid bilayers. Peptides 24, 1779-1784, doi:http://dx.doi.orq/10.1016/j.peptides.2003.09.015 (2003). [0157] 12 Majd, S. et al. Applications of biological pores in nanomedicine, sensing, and nanoelectronics. Current Opinion in Biotechnology 21, 439-476, doi:http://dx.doi.orq/10.1016/j.copbio.2010.05.002 (2010). [0158] 13 Montal, M. & Mueller, P. Formation of bimolecular membranes from lipid monolayers and a study of their electrical properties. Proceedings of the National Academy of Sciences 69, 3561-3566 (1972). [0159] 14 Stahl, E., Martin, T. G., Praetorius, F. & Dietz, H. Facile and Scalable Preparation of Pure and Dense DNA Origami Solutions. Angewandte Chemie International Edition 53, 12735-12740, doi:doi:10.1002/anie.201405991 (2014).