Pores

This invention relates to pores having lumens, wherein a dendrimer is attached to the pore, and the use of such pores in separation devices, in devices for the sequencing of nucleic acids and proteins, and in pharmaceutical compositions, and also to a reagent and method for the structural analysis of proteins.

In materials science the engineering of the permeation properties of membranes comprising pores is an intensive area of research with applications, such as, in warfare sensing and the membrane-based separation of organic molecules and biomolecules (Bayley, H.; Martin, C. R. Chem Rev. 2000, 100, 2575-2594; Kohli, P.; Harrell, C. C; Cao, Z.; Gasparac, R.; Tan, W.; Martin, C. R. Science. 2004, 305, 984-986; Lakshmi, B. B.; Martin, C. R. Nature. 1997, 388, 758-760; Mara, A.; Siwy, Z.; Trautmann, C; Wan, J.; Kamme, F. Nano Letters. 2004, 4, 497-501; Li, J. L.; Gershow, M.; Stein, D.; Brandin, E.; Golovchenko, J. A. Nature Materials. 2003, 2, 611-615).

For example, the ability to develop membranes that allow selective passage of only particular species (for example molecules or ions) has many applications in separation technologies: these could be used to separate molecules of particular size or charge from those of differing size or charge.

Also desirable targets are membranes comprising pores that are designed as to allow the rate at which a particular species passes through the membrane to be chosen. For example, membranes that allow passage of a particular species at a slow rate may be useful to provide analytical data about that species as it passes through the membrane. Alternatively, a membrane developed so as to provide a constant controlled release of a species through that membrane at a given rate are also potentially useful.

There is, therefore, a need in the art for membranes that selectively control the ability of particular species to pass through that membrane.

Accordingly, the present invention provides a nanoscale pore having a lumen, wherein a dendrimer is attached to the pore.

The present invention will now be described with reference to the accompanying drawings, in which:

Figure 1 shows (A) the structure of polyamido amine dendrimer of generation 3 with 50% amine and 50% hydroxyl terminal groups and (B) the cross-sectional view of the heptameric αHL pore with cysteine substitutions K46C, K8C, and S106C. The model was generated using crystallographic data (Song, L.; Hobaugh, M. R.; Shustak, C;

Cheley, S.; Bayley, H.; Gouaux, J. E. Science. 1996, 274, 1859-1866.) and PyMoI. The internal diameters of the channel are: 2.9 nm, cis entrance; 4.1 nm, internal cavity; 1.3 nm, inner constriction; 2 nm, trans entrance of the β-barrel.

Figure 2 shows RP-HPLC analysis of the derivatization of PAMAM dendrimer with the heterobifunctional cross-linker SPDP. G3 -PAMAM before (A) and after (B) reaction with SPDP.

Figure 3 MALDI-ToF analysis of SPDP-modified PAMAM dendrimers to determine the average number of coupled pyridyldisulfide groups. G3 -PAMAM before (A) and after (B) reaction with SPDP.

Figure 4 shows (A) quantifying the yield of a G3-P AMAM-PDP preparation using SDS gel electrophoresis and Coomassie staining. Solutions of unmodified dendrimers of known concentration and SPDP-modified dendrimers of unknown concentrations were analyzed. The resulting electropherograms were subjected to densitometric analysis to determine the unknown concentrations. Lane 1, 67 μM G3-PAMAM; lane 2, 33 μM G3-PAMAM; lane 3, 17 μM G3-PAMAM; lane 4, G3-PAMAM-PDP, concentration determined to be 50 μM; and (B) Sulfhydryl-reactive G3-PDP couples specifically to monomeric αHL cysteine mutant K46C and causes a gel-shift in SDS-PAGE autoradiographs. Lane 1, K46C; lane 2, K46C treated with G3-PDP; lane 3, K46C treated with PEG-MAL 5 kD; lane 4, K46C with G3-PDP and excess reducing agent DTT; lane 5, K46C with PEG-MAL 5 kD and excess DTT.

Figure 5 shows coupling of a PAMAM-PDP dendrimer inside and outside the lumen of a protein channel. Homoheptameric αHL cysteine mutants were reacted with G5-PDP,

G3-PDP, and PEG-MAL and analyzed via gel electrophoresis and autoradiography. K46C ₇ (lanes 1 to 4) and SlOoC ₇ (lanes 5 to 7), were treated with PAMAM-PDP (lanes 2 and 5), PAMAM-PDP followed by PEG-MAL (lanes 3 and 6), or PEG-MAL (lanes 4 and 7). The modification reactions were performed with G5-PDP (panel A) or with G3- PDP (panel B).

Figure 6 shows representative single channel current traces of (A) SlOoC ₇ , (B) SlOoC ₇ - G3-PAMAM, (C) S106C ₇ -G2-PAMAM, and (D and F) K46C ₇ -G5-PAMAM. (E) All- points histogram of K46C ₇ -G5 -PAMAM (D) with a duration of 2 s. The recordings were performed in 1 M KCl, 20 mM TrisηCl pH 7.5 at a transmembrane potential of +100 mV ((A) to (D)) or -100 mV (F), with the chamber on the cis side of the protein pore grounded. The currents were filtered at 1 kHz and sampled at 5 kHz.

Figure 7 shows single channel current traces of (A) SlOoC ₇ , (B) SlOoC ₇ with 1 μM RNA oligonucleotide C ₃₀ at the cis side, (C) S106C ₇ -G3-PAMAM, and (D) SlOOC ₇ - G3-PAMAM with 1 μM RNA oligonucleotide C ₃₀ at the cis side, with a potential of +100 mV. The traces were filtered at 10 kHz and sampled at 50 kHz; and

Figures 8, 9 and 10 show examples of dendrimers that may be used in the invention.

The invention, therefore, provides a nanoscale pore, wherein the pore has a lumen, and a dendrimer is bound to the pore, and wherein the dendrimer functions to modify the ability of species to pass through the pore.

The dendrimer may be attached to the lumen of the pore. In this position, the dendrimer has the greatest effect on the passage of a species through the pore, because the lumen comprises the part of the pore with the smallest cross sectional area.

Alternatively, the dendrimer may be attached to an exterior surface of the pore.

This invention is not limited to any particular type of nanoscale pore. Any suitable pore may be used. However, in the present invention preferably the pore is an organic pore. The term organic takes its usual meaning in the art, therefore the organic pore

substantially comprises carbon and hydrogen, and also other elements, especially nitrogen, oxygen, sulfur, phosphorus and halogens.

More preferably the pore is a protein pore. A protein pore is a pore which is predominantly protein; however, other types of molecules may also be present.

Examples of protein pores suitable for use in the invention include alpha hemolysin, pneumolysin, outer membrane proteins such as porins, and other bacterial pore-forming toxins (Gilbert 2002) (Parker and Feil 2005) such as streptolysin O (Bhakdi, Tranum-

Jensen et al. 1985) or LukF (Olson, Nariya et al. 1999). The latter are oligomeric assemblies of protein subunits. The diameter of the lumens of protein pores depends on the type of pore and ranges from 1.2 nm for alpha hemolysin (Song, Hobaugh et al.

1996) to 26 nm for pneumolysin (Tilley, Orlova et al. 2005).

Particularly preferably the protein pore is a α-hemolysin (αHL) polypeptide. αHL is a bacterial toxin which self-assembles to form a heptameric protein pore. The X-ray structure of the αHL pore resembles a mushroom with a wide cap and a narrow stem, which spans the lipid bilayer (Fig. IB) (Song, L.; Hobaugh, M. R.; Shustak, C; Cheley,

S.; Bayley, H.; Gouaux, J. E. Science. 1996, 274, 1859-1866). The external dimensions of the heptameric αHL pore are 10 x 10 nm, while the central channel is 2.9 nm in diameter at the cis entrance and widens to 4.1 nm in the internal cavity (Fig. IB). In the transmembrane region, the channel narrows to 1.3 nm at the inner constriction and broadens to 2 nm at the trans entrance of the β-barrel. The defined structure of αHL has facilitated extensive engineering studies and has led to the development of tools for the targeted permeabilization of cells (Eroglu, A.; Russo, M. J.; Bieganski, R.; Fowler, A.; Cheley, S.; Bayley, H.; Toner, M. Nat Biotechnol. 2000, 18, 163-167) as well as new biosensor elements which permit the stochastic sensing of molecules (Bayley, H.;

Cremer, P. S. Nature. 2001, 413, 226-230).

Although organic pores are preferred, the invention is not limited to pores of this type. The invention includes any pore to which a dendrimer can be attached, and consequently affect the passage of differing molecules through the pore. Alternative embodiments of the invention exist wherein the pore is an inorganic pore. Preferably

the inorganic pore is composed of silica, silicon nitride, alumina, titanium, gold, platinum, zirconia, silicon nitride or a combination thereof.

Although the pore is not limited in relation to the material that it comprises, the invention is limited in relation to the size of the pore. The pore must be a nanoscale pore. By nanoscale is meant that the pore is one wherein the lumen has a diameter of less than 1 μm. Preferably the lumen has a diameter of less than 100 nm. More preferably the lumen has a diameter of 10 nm or less. Preferably the pore has a diameter of at least 1 nm. References to the diameter of the pore are to be interpreted as the diameter of the pore at its minimum value. In this regard please see fig. IB which illustrates that the diameter of the lumen may vary at different positions along its length.

In relation to the dendrimer, again the invention is not limited to dendrimers of a particular type. Dendrimers are highly branched (globular) nanoscale polymers

(Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.;

Ryder, J.; Smith, P. Polymer J. 1985, 17, 117-132; Newkome, G. R.; Yao, Z.; Baker, G.

R.; Gupta, V. K. J Org Chem. 1985, 50, 2003-2004; Tomalia, D. A. Prog Polymer

Science. 2005, 30, 294-324; Frechet, J. M.; Tomalia, D. A. Dendrimers and other dendritic polymers; John Wiley & Sons, 2002).

The IUPAC definition of a dendrimer is "A polymer having a regular branched structure" (Pure Appl. Chem., Vol. 71, No. 12, pp. 2349-2365, 1999). Their synthesis proceeds either via the divergent or the convergent route (Newkome, G. R.; Moorfield, C. N.; Vogtle, F. Dendrimers and Dendrons: Concepts, Syntheses, Applications; Wiley- VCH: Weinheim, 2001) and offers control over molecular mass, size, shape, the degree of branching, and type and number of terminal functional groups. Reflecting their special characteristics, many potential applications have been developed in materials science and nanotechnology for separation technology, surface coatings, and catalysis (Frechet, J. M. J. J Poly Science Part A. 2003, 41, 3713-3725) and in biological sciences for drug and gene delivery, vaccines, and bioimaging (Lee, C. C; MacKay, J. A.; Frechet, J. M.; Szoka, F. C. Nat Biotechnol. 2005, 23, 1517-1526).

As mentioned above, in the present invention the dendrimer is attached to the pore. The role of the dendrimer is to affect the ease with which other molecules may pass though the pore.

In order to achieve this goal, the size of the dendrimer is selected by consideration of the size of pore, and the degree to which it is desired to alter the ability of other species to a pass through the pore. When the dendrimer is attached to the lumen of the pore, the dendrimer is chosen by consideration of the dimensions of the lumen. Therefore the present invention is not limited in relation to type of the dendrimer. However, consideration must be given to the size of the pore to which the dendrimer is to be attached when determining the size of dendrimer to be used for a particular embodiment of the invention.

Preferably the dendrimer is a PAMAM dendrimer (tradename Starburst®) (Figs. IA and 8b), an aromatic polyether dendrimer (Fig. 9), a phenyl acetylene dendrimer (Fig. 10), a poly glutamic acid dendrimer (Fig. 8a), a poly (propylene imines) dendrimer (Fig. 8c), a polymelamine dendrimer (Fig. 8d), a polyester dendrimers (Fig. 8e) or any other dendrimer which is N-, C-, aryl-, Si-, adamantane-, N3P3-, pyridine-, ethano-, saccharide-, or cholic-acid-branched with aryl-, Si-, N-, amide-, ester-, P-,ether-, sulfone-, urea-, alkyl-, alkyne, ketone, alkene, oxazdiazole, siloxy, urethane, imide, ethyne, carbonate, thiol, phosphate, or thiourea-connectivity.or combinations thereof (see Dendrimers (Topics in Current Chemistry S.) (Hardcover) Springer- Verlag Berlin and Heidelberg GmbH & Co. K (30 JuI 1998) by Fritz Vogtle (Editor)). Methods of synthesis of these types of dendrimers are well known to the person skilled in the art.

PAMAM are an important sub-class of dendrimers (Figs. IA and 8b), and are particularly preferred for use in the present invention. They were historically the first dendrimers to be synthesized using the divergent strategy (Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Polymer J. 1985, 17, 117-132; Matthews, O. A.; Shipway, A. N.; Stoddart, J. F. Prog Polym ScL 1998, 23, 1-56; Tomalia, D. A.; Frechet, J. M. J. JPoIy Science Part A. 2002, 40, 2719- 2728). Their synthesis starts with a Michael addition of methyl acrylate to the ethylene diamine core followed by amidation of the tetra-ester with ethylene diamine yielding a

generation zero dendrimer. Subsequent Michael addition / amidation cycles provide dendrimers of increasing generation with the number of surface groups doubling each generation. The mass of the polymer also approximately doubles with each extension step until crowding between the terminal surface groups blocks further growth (Bauer, B. J.; Amis, E. J. In Dendrimers and other dendritic polymers; Frechet, J. M., Tomalia, D. A., Eds.; John Wiley & Sons, 2002, pp 255-284) leading to increasingly compact polymers (Li, D.; D. A. Tomalia In Dendrimers and other dendritic polymers; Frechet, J. M., Tomalia, D. A., Eds.; John Wiley & Sons, 2002, pp 285-307; Hierlemann, A.; Campbell, J. K.; Baker, L. A.; Crooks, R. M.; Ricco, A. J. J. Am. Chem. Soc. 1998, 120, 5323-5324). The compact structure of PAMAM polymers has been characterized with chemical (Islam, M. T.; Majoros, I. J.; Baker, J. R. J. Chromatogr. B. 2005, 822, 21-26; Peterson, J.; Allikmaa, V.; Subbi, J.; Pehk, T.; Lopp, M. Eur Poly Journal. 2003, 39, 33-42; Subbi, J.; Aguraiuja, R.; Tanner, R.; Allikmaa, V.; Lopp, M. Eur Poly Journal. 2005, 41, 2552-2558) and physicochemical techniques (Nourse, A.; Millar, D. B.; Minton, A. P. Biopolymers. 2000, 53, 316-328; Bosnian, A. W.; Janssen, H. M.; Meijer, E. W. Chem Rev. 1999, 99, 1665-1688).

As mentioned above, the diameters of the dendrimer used is chosen on a case-by-case basis by consideration of the size of the pore to be used in a particular embodiment, and therefore the present invention is not limited in relation to the size of dendrimer used. However preferably the dendrimer has a diameter of 20 nm or less and more preferably the dendrimer has a diameter of 10 nm or less. Preferably the dendrimer has a diameter of more than 1 nm.

The size of a dendrimer increases as the generation of the dendrimer increases. Preferably the dendrimer is a fifth-generation dendrimer, a third- generation dendrimer or a second-generation dendrimer. When the dendrimer is attached to the lumen of the pore, particularly preferably the dendrimer is a third-generation dendrimer or a second generation dendrimer.

The pore may be attached to any number of dendrimers. The number of dendrimers attached to the pore may be altered for example to vary the ease with which molecules may pass through the pore. A greater number of dendrimers will hinder the passage of

molecules through the pore, whereas a smaller number of pores will relatively ease the passage of molecules through the pore. Preferably, however, from one to ten dendrimers are attached to the pore. Embodiments of the invention are known in which one, two, three, four, five, six, seven, eight, nine or ten dendrimers are attached to the pore. More preferably from one to four dendrimers are attached to the pore. Most preferably only one dendrimer is attached to the pore.

The dendrimer and some part of the pore must be bound in some way. The invention is not limited in respect of how this binding is achieved. Preferably the dendrimer is attached to the pore by one or more linker groups. The linker group may be an organic group or an inorganic group, and may comprise a single atom or multiple atoms. In one embodiment of the invention, the linker group is thiopropanoyl, wherein the sulfur atom is bound to the pore and the acyl carbon is bound to the dendrimer.

Preferably the linker group is attached to the pore by one or more bonds. The bonds may be of any type, but examples that can be mentioned include cases in which the one or more bonds are one or more covalent bonds, electrostatic attractions, metal-chelate bonds, hydrophobic interactions or mixtures thereof. In addition, steric interactions between the linker and the pore may aid the binding of the linker to the pore.

Particularly preferably the linker group is attached to the pore by one or more covalent bonds, and most preferably the one or more covalent bonds are disulfide bonds or amide bonds. When the pore is a protein pore, the disulfide bond is formed by reaction of the linker group with a cysteine residue comprised within the pore.

Alternatively, rather than the linker group being an atom or multiple atoms the linker group may be one or more bonds, and therefore the dendrimer is bound directly to the pore. Again the one or bonds are not limited to any particular type, but examples of the one or more bonds are one or more covalent bonds, electrostatic attractions, metal- chelate bonds, hydrophobic interactions or combinations thereof. In addition, steric interactions may aid the binding of the dendrimer to the pore.

In a particularly preferred embodiment of the invention, the pore is αHL, the dendrimer is a third- generation PAMAM which is attached to the lumen of the pore via a linker which is thiopropanoyl.

In another particularly preferred embodiment of the invention, the pore is αHL, the dendrimer is a second- generation PAMAM which is attached to the lumen of the pore via a linker which is thiopropanoyl.

The invention also provides a method of attaching a dendrimer to a pore, although the invention is not limited to the pores when sythesised using this route.

Accordingly the invention provides a method of attaching one or more dendrimers to a nanoscale pore comprising a lumen, wherein an activated dendrimer is reacted with a nucleophilic group present in the pore.

The dendrimer is activated by the presence of an electrophilic group on the dendrimer.

This electrophilic group may be already present as a natural feature of the dendrimer; alternatively, the dendrimer may activated by derivatisation via attachment of an electrophilic group to the dendrimer.

The method requires the presence of a nucleophilic group on the pore. Any nucleophilic group may be used: for example a heteroatom such as oxygen, sulphur or phosphorus present within the pore.

If it is desired to attach the dendrimer externally to the pore, then the method is used wherein the nucleophilic group is located on an external surface of the pore. If it is desired to attach the dendrimer to the lumen of the pore, then the method is used wherein the nucleophilic group is located within the lumen.

In a preferred embodiment of the method, the pore is a protein pore and the nucleophilic group is the sulfur of a cysteine residue. As mentioned above, embodiments of the invention exist in which the dendrimer is attached externally to the pore, and embodiments of the invention exist wherein the dendrimer is attached to the

lumen of the pore. In order to provide an appropriate protein pore wherein the pore has a cysteine pore appropriately positioned to bind to the activated dendrimer, it is necessary to express an α-hemolysin single-cysteine mutant by in vitro transcription/translation.

α-hemolysin proteins were generated having cysteine residues at K46C7 (positioned at the cap of the pore forming a ring surrounding the cis entrance in the assembled pore - see Fig. IB). Use of pores of this type allows attachment of the dendrimer to the exterior of the pore.

In addition, α-hemolysin proteins were generated having cysteine residues at S106C7 (positioned within the lumen of the assembled pore - see figure IB) (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411-2416). Use of pores of this type allows attachment of the dendrimer to the lumen of the pore.

In addition, α-hemolysin proteins were generated having cysteine residues positioned at K8C7 (positioned within the cis entrance of the assembled pore - see figure IB) (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411 -2416).

As mentioned previously, it may be necessary to activate the dendrimer so it is electrophilic. Often the terminal groups of dendrimers are nucleophilic: for example amine and hydroxyl groups. Therefore a preferred method of synthesising the activated dendrimer is by reaction of one or more terminal amines of a dendrimer with a bis- electrophile wherein the amine reacts with a first electrophilic site of the bis- electrophile and a second electrophilic site of the bis-electrophile is unchanged.

By bis-electrophile is meant a species with a first electrophilic site and a second electrophilic site. The first electrophilic site reacts with a nucleophilic group, for example an amine, on the dendrimer. The second electrophilic site is left unchanged, and is therefore free to react with a nucleophilic site within the pore. Consequently, reaction of the first electrophilic site of the bis-electrophile with the dendrimer, and

then of the second electrophilic site of the bis-electrophile with a nucleophilic site on the pore, results in attachment of the dendrimer to the pore. In a preferred method the bis-electrophile is N-succinimidyl-3-(2-pyridyldithio)-propanoate (SPDP).

Using the method of the invention αHL protein pores carrying PAMAM bound both outside and inside the lumen can be synthesised.

In these embodiments G5 PANAM dendrimers were used having a hydrodynamic diameter of 6.2 nm and a hard-sphere diameter of 4.2 nm (value obtained from the solvent exclusion volume) (Nourse, A.; Millar, D. B.; Minton, A. P. Biopolymers. 2000, 53, 316-328; Dvornic, P. R.; Uppuluri, S. In Dendrimers and other dendritic polymers; Frechet, J. M., Tomalia, D. A., Eds.; John Wiley & Sons, 2002).

Further, G3 PAMAM dendrimers were used having hydrodynamic and hard-sphere diameters of are 4.1 and 2.9 nm respectively (Nourse, A.; Millar, D. B.; Minton, A. P. Biopolymers. 2000, 53, 316-328; Dvornic, P. R.; Uppuluri, S. in Dendrimers and other dendritic polymers; Frechet, J. M., Tomalia, D. A., Eds.; John Wiley & Sons, 2002). G2 IPAMAM dendrimers were used having hydrodynamic diameter of 2.9 nm (hard- sphere diameter not available).

Activated dendrimers were generated using G5 PAMAM with a mixed surface of terminal -OH and -NH ₂ groups at an average ratio of 90:10, G3 PAMAM with a mixed surface of 50:50 hydroxyl/amino groups, and G2 PAMAM with terminal NH ₂ groups. A mixed surface with hydroxyl groups avoids the formation of multimeric aggregates which can be found in purely amino-terminated but not in hydroxyl-functionalized G5 PAMAM dendrimers (Nourse, A.; Millar, D. B.; Minton, A. P. Biopolymers. 2000, 53, 316-328). The relative percentages of amine groups of the G3 and G5 PAMAM dendrimers were initially chosen to yield approximately the same number of sulfhydryl-reactive groups for each generation of dendrimer (13 for G5, and 16 for G3 and G2).

G5, G3 and G2 PAMAM dendrimers were reacted with the bis-electrophile SPDP according to the following scheme:

The N-succinimidyl activated ester of SPDP 1 couples to the terminal primary amines of the dendrimers 2 to yield amide-linked 2-pyridyldithiopropanoyl (PDP) groups 3.

Table 1 shows the properties of the activated dendrimers. The number of PDP groups coupled to PAMAM dendrimers was determined by MALDI-ToF and via photometric analysis which involved treatment of samples with excess reducing agent dithiothreitol (DTT) to cleave the disulfide bond of PDP, and the than detection of the cleavage product pyridyl-2-thione at 343 nm. The results of the photometric analysis yielded 5.0 ± 1.5, 11 ± 2, and 13 ± 2 PDP groups for G2-PDP, G3-PDP, and G5-PDP, respectively, which is similar to the numbers obtained from MALDI-MS (Table 1).

The yield of PAMAM-PDP dendrimers preparations was estimated using sodium- dodecylsulfate gel electrophoresis (SDS-PAGE) and Coomassie staining (Fig. 4A). The gel band intensities of PDP-modified dendrimers were compared with intensities of PAMAM dendrimers of known amount. Densitometric analysis indicated yields of approx. 15% for G2-PDP, 80% for G3-PDP and 50% for G5-PDP relative to the amount used as starting material (Table 1).

Table 1. Chemical Characteristics and Dimensions of G2-PDP, G3-PDP, and G5-PDP

number of number of PDP diameter [nm] [C] permeation modification terminal groups in through 2.9 of a highly amine PAMAM-PDP ^[b] nm entrance accessible groups ^[a] cysteine residue ^[d]

MALDI- photo- hydrodynamic hard¬

MS shell metric analysis analysis

G5-PDP 13 14 ± 3 13 ± 2 6.2 4.2

G3-PDP 16 13 ± 2 11 ± 2 4.2 2.6 +

G2-PDP 16 5.2 ± 5.0 ± 1.5 2.9 N.A. + N.A. 1.0

^[a] Obtained from ¹ H NMR data.

^[b] Average of three independent experiments on the photometric detection of pyridyl-2- thione at 343 nm released by the treatment of pydridyldithiopropanoyl-PAMAM with the reducing agent DTT.

^[c] Derived from sedimentation analysis (Nourse, A.; Millar, D. B.; Minton, A. P. Biopolymers. 2000, 53, 316-328; Dvornic, P. R.; Tomalia, D. A. In Polymer Data Handbook; Oxford University Press, 1999, pp 266-270).

10 ^[d] Permeation characteristics through a 2.9 nm pore represent the modification results with αHL heptamer SIO6C7 and K8C7. For the modification with accessible residues, heptamer K46C ₇ was used. Three independent experiments were performed for each modification reaction and gave the same results.

15 The activated PAMAM-PDP dendrimers were then reacted with the αHL protein. Unmodified αHL polypeptide migrated at a relative molecular mass of 30 kD (Fig. 4B, lane 1) while the major protein band after reaction with G3-PDP was up-shifted to 40

kD (Fig. 4B, lane 2). No up-shifted band was observed after the reaction mixture was treated with excess reducing agent DTT (Fig. 4B, lane 4) indicating that G3-PDP had coupled specifically to αHL through a disulfide bridge. By quantifying the intensity of the up-shifted αHL-G3 band relative to the unmodified αHL protein band (Fig. 4B, lane 2), it could be seen that 90% of the cysteine mutant was reacted. The same extent of modification was found for G2-PDP and G5-PDP (data not shown)

SPDP-derivatized PAMAM dendrimers were, therefore, prepared in a highly efficient manner, and the product mixture did not contain any SPDP or hydrolysis product 3-(2- pyridyldithio)-propanoic acid which might block coupling to cysteines.

A similar high extent of modification of 95% was obtained using the sulfhydryl-active reagent mono-methyl polyethylene glycol maleimide 5 kD (PEG-MAL 5 kD) (Fig. 4B, lane 3). In line with the formation of a thioether bridge (rather than a disulfide bond) between PEG and αHL, treatment with excess DTT did not cleave the conjugate (Fig. 4B, lane 5).

Dendrimers G5-PDP and G3-PDP were also reacted with the cysteine mutant K46C in an assembled αHL homoheptamer. The seven cysteine residues at the cap of the pore form a ring surrounding the cis entrance (Fig. IB). Based on the exposed position of the cysteine residues, both G3-PDP and G5-PDP were expected to react with hep tamer. To test the reactivity, radiolabeled homoheptamer K46C7 was generated, treated with G5- PDP, G3-PDP or, for comparison purposes, with PEG-MAL 5 kD, and analysed via SDS-PAGE and autoradiography to detect the appearance of up-shifted bands. αHL heptamers are not denatured in SDS-PAGE and therefore migrated as defined bands as seen for unmodified K46C ₇ (Fig. 5A & B, lane 1). Upon reaction with G5-PDP, an additional major up-shifted band appeared (Fig. 5 A, lane 2). This band represents the specific covalent coupling product of G5-PDP and the αHL cysteine mutant as no up- shifted band was observed with unmodified G5 (data not shown). Coupling of G5-PDP to K46C7 produced one but not a second up-shifted band strongly indicating that coupling of a second G5 dendrimer was disfavored. This could either be because of a steric clash or charge repulsion with the first tethered dendrimer, or because the first dendrimer coupled to a second or third cysteine residue thereby preventing reaction of a

second dendrimer. However, some residual cysteine residues of G5-PDP-treated heptamers could still couple to flexible and neutral PEG-MAL 5 kD to produce additional up-shifted bands (Fig. 5A, lane 3).

Further, a higher extent of modification was obtained when K46C ₇ was treated with PEG-MAL 5 kD alone (Fig. 5A, lane 4). The highly up-shifted and partly unresolved bands represent pore species with 5, 6, or 7 tethered PEG chains as observed in another study. ²⁴ It is worth mentioning that coupling of seven PEG chains with a total mass of up to 7 x 5 kD = 35 kD led to a dramatic gel up-shift of the αHL band (Fig. 5 A, lane 4) while addition of one G5 dendrimer molecule with a comparable mass of 29 kD produced a lesser gel shift (Fig. 5A, lane 2) even though addition of the positively charged and unfolded PAMAM polymer would have been expected to slow down the electrophoretic migration of the protein. This difference between PEG and PAMAM likely reflects the different ways in which flexible vs. compact polymers interact with the polyacrylamide gel meshwork. Most probably, PEG chains became entangled within the holes of the meshwork resulting in a retarded migration of the protein band while PAMAM is too compact to get intertwined.

Reaction of K46C7 heptamers with the smaller G3-PDP resulted in two up-shifted and closely migrating bands (Fig. 5B) which likely represent heptamers with one and two

G3 dendrimers. In line with the lower molecular mass of G3 (M _r 6.9 kD) both G3- heptamer conjugates migrated lower than the conjugate with G5-PDP (M _r 28.9 kD)

(Fig. 5A, lane 2). Interestingly, reaction of G3-modified K46C ₇ heptamers with PEG-

MAL 5 kD did not produce additional up-shifted bands (Fig. 5B, lane 3) which were however observed in the case of G5-modified heptamers (Fig. 5A, lane 3). This indicates that two G3-PDP dendrimers reacted with most of the cysteines thus blocking further coupling to the flexible PEG-MAL polymer.

Reaction of G3-PDP and G5-PDP dendrimers with K46C7 heptamers provide embodiments of the invention in which the dendrimer is attached externally to the pore. Similar coupling reactions between activated dendrimers and position S106C of the heptamer which is located inside the internal cavity of the αHL pore (Fig. IB) were also carried out. Reaction of the engineered cysteine residue with PEG-MAL 5 kD was

confirmed to have taken place by gel electrophoretic analysis of radiolabeled SIO6C7 heptamers revealing one up-shifted band (Fig. 5A, lane 7). A single, but not multiple up-shifted band, was obtained because the narrow pore accommodates only one PEG 5kD molecule as found in theoretical (Kong, C. Y.; Muthukumar, M. J Am. Chem. Soc. 2005, 127, 18252-18261) and other experimental studies (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411-2416; Movileanu, L.; Howorka, S.; Braha, O.; Bayley, H. Nat Biotechnol. 2000, 18, 1091-1095).

SIO6C7, based on the known hydrodynamic and hard-sphere diameters of the PAMAM dendrimers, would not be expected to couple to G5- PDP, due to the inaccessibility of the lumen. G3-PDP may or not be expected to couple to the cysteine residues in the internal cavity: the cis entrance is 2.9 nm wide, whereas the G3 hydrodynamic diameter is 2.9 nm and the G3 hard-sphere diameter is 4.1 nm.

Gel electrophoretic analysis showed that neither reaction with G3-PDP nor G5-PDP led to a major up-shifted band (Fig. 5 A and 5B, lane 5). The lack of up-shifted bands could have been due to a lack of coupling of the internal cysteine to the dendrimer, or alternatively, because although coupling occurred, an appreciable gel electrophoretic shift was not produced as the tethered dendrimer resided inside the lumen of the pore.

The two possibilities were distinguished by treatment of the SIO6C7 heptamer with G5- PDP and subsequently by PEG-MAL. Reaction with the flexible PEG chain is known to yield an up-shifted band and can therefore be used to probe whether S106C is accessible or whether the internal cavity is blocked by a dendrimer molecule. Gel electrophoretic analysis revealed that dual treatment by G5-PDP and then subsequently PEG-MAL 5 kD led to an up-shifted band (Fig. 5A, lane 6) implying that G5-PDP had not permeated the internal cavity, and had not bound. In contrast, reaction with G3- PDP and then subsequently PEG-MAL 5kD did not yield an up-shifted band (Fig. 5B, lane 6) indicating that G3 permeated into the pore and coupled to S106C thereby blocking subsequent modification with PEG-MAL.

A similar finding on the differential ability of G5-PDP, G3-PDP and PEG-MAL 5 kD to react with internal and more hindered binding sites was obtained by testing the reactivity of K8C7, which is positioned in the 2.9 nm wide cis entrance (Fig. IB &) with these activated dendrimers. These results demonstrate that again G3, but not G5, permeated through the cis entrance into the internal cavity. The finding that G5 did not enter the pore is remarkable considering that PEG-MAL 10 kD, whose hydrodynamic diameter of 6.2 nm is identical to G5 -PAMAM, permeated into the pore to couple to the even more hindered residue S106C. This highlights the different degrees of structural flexibility of dendrimers as opposed to flexible PEG polymers.

Reaction of G3-PDP dendrimer with K8C heptamers therefore provides an embodiment of the invention in which the dendrimer is attached to the lumen of the pore.

As mentioned previously, the dendrimer is inserted into the lumen of the pore in order to alter the ability of differing species to pass through the pore.

Consequently, one embodiment of the invention is a membrane comprising one or more pore as defined hereinabove. The membrane has a first side and a second side, and the pore provides a route, channel or path from the first side to the second side. By positioning a dendrimer in the pore, therefore, because the presence of the dendrimer changes the ease by which species can pass through the pore, the dendrimer changes the ease with which species may pass from the first side of the membrane to the second side of the membrane.

Preferably the membrane will comprise a number of pores, the lumen of which provide a number of channels between the first side of the membrane and the second side of the membrane. The precise number of pores on the membrane is not critical to the invention, and would need to be calculated on a case by case basis taking into account the size of the membrane and the required rate of passage of species between the first and second sides of the membrane. However, in a preferred embodiment the membrane has a density of pores of from 1 to 100 000 pores per μm ² .

The membrane, or barrier, itself is also not limited in type, and would need to be chosen by consideration of species it is desired to pass through the pore or prevent from passing through the pore, and therefore between the membrane's first and second sides. For example, the membrane must be stable in the presence of this species.

The membranes according to the invention are not limited to biological materials: biological membranes are only one example of membranes according to the invention.

However, preferably the membrane is organic. In one embodiment the organic membrane is an organic polymer, most preferably the organic polymer is a polycarbonate or polyterephtalate polymer. When the membrane is an organic polymer, preferably the density of pores is 1 to 100 000 pores per μm ² .

Most preferably the organic membrane is a lipid bilayer. In a particularly preferred embodiment of the invention, the membrane comprising a pore is formed by allowing mutant polypeptides K46C, K8C or S106C to assemble on rabbit erythrocyte membranes to form heptameric pores. When the membrane is a lipid bilayer, preferably the density of pores is 0.0001 to 100 pores per μm ² .

Alternatively the membrane is inorganic, wherein preferably the inorganic membrane is a gold-plated porous membrane prepared by the template synthesis method by electrolessly depositing gold along the pore walls of a polycarbonate template membrane (Martin, Nishizawa et al. 2001). When the membrane is inorganic, preferably the density of pores is 1 to 100 000 pores per μm ² .

In addition, preferably when the membrane is a lipid bilayer the pore is an organic pore, preferably an polymeric organic pore or a protein pore. Preferably when the membrane is an organic polymer the pore is an organic pore. Preferably when the membrane is an inorganic membrane, the pore is an inorganic pore.

In a particularly preferred embodiment of the invention, the membrane is a lipid bilayer, the pore is αHL, the dendrimer is a third- generation PAMAM, and the linker is thiopropanoyl.

As mentioned above, the dendrimer is attached to the pore in order to affect the ability of species to pass through the pore. It does this by being attached to the lumen of the pore, and therefore creating a blockage therein, or alternatively being bound to the exterior of the pore by a linker, wherein the linker is sufficiently flexible to allow the dendrimer to move between positions which make the entrance to the pore relatively more and less accessible.

The pores according to the invention are an important advance in the area of filtration membranes. Regulating and controlling the flow of matter across nanoscale porous membranes is an important topic in filtration-based separation technologies such as ultrafiltration, reverse osmosis, and dialysis in laboratory and industry. Existing synthetic or artificial membranes which permit the selective transport of material are composed of polymers or ceramics with nanometer-sized holes.

According to the invention, dendrimer molecules are attached to pores in organic and inorganic membranes in a new and useful way to engineer the permeation and filtration properties of semi-permeable membranes. It has been found that using dendrimers as molecular sieves in the pores of the membranes can be advantageous as the polymer is dense and thus capable of restricting the movement of small molecules. Apart from excluding the passage of molecules based on size, the dendrimer can also be tailored so the membrane acts as a filter via chemical affinity or electrostatic interactions with the species being filtered. Using dendrimers for different inorganic structures which can have different pore diameters can benefit from the availability of dendrimers in different sizes ranging form 3 to 10 nm. Finally, due to the modular character of the approach, the semi-permeability of the membrane can be changed by choosing a dendrimer of different composition while leaving the membrane composition itself unaltered.

Consequently, another embodiment of the invention is a filtration device comprising a membrane as described hereinabove, an inlet and an outlet.

The inlet and the outlet are not limited in type, style or size. However, the inlet and the outlet are in fluid communication, and the inlet is positioned at the first side of the membrane and the outlet is positioned at the second side of the membrane.

The filtration device can be used in a method of separating a first species and a second species, wherein a fluid comprising the first species and the second species is positioned in the inlet, passed through the membrane, and collected at the outlet.

The principle of the device is that the dendrimer attached to the pore affects the passage of each of a first species and a second species to a greater or lesser extent. Therefore the first species is substantially prevented from passing between the first and second sides of the membrane whereas the second species is substantially allowed to pass through. Therefore the first species and the second species are separated: the first species being on the first side of the membrane, and the second species being on the second side of the membrane.

The dendrimer and membrane of the device must be chosen so as to achieve the separation effectively. For example, the dendrimer must be chosen so as to prevent the first species from passing through the pore and therefore the membrane, whilst allowing the second species to pass through the pore and therefore the membrane. The membrane must be chosen so that it is essentially non-permeable to the first and second species it desired to separate, otherwise after selective separation of the first and second species by the pore, re-mixing of the first and second species would occur and resolution would be lost.

The invention is not limited with respect to the species it may be used to separate, and each species may be any chemical or other type of entity. Further, each species may comprise one component only or comprise multiple components. Therefore the method or device of the invention may be used to separate three components into an individual component and a mixture of two components, wherein the individual component is the first species and the two components are the second species as defined above.

The method may be used when the first and second species to be separated are ions or molecules or atoms, or mixtures thereof; that is the first species and second species may individually be a mixture of ions, molecules and atoms, or the first species may be, say, ions, whereas the second species is, say, molecules.

As mentioned previously, the dendrimer is chosen so as to affect the separation. One way of achieving this is by selecting a dendrimer of an appropriate size, and thereby the first species and the second species are separated by their differing sizes. The larger first species is prevented from passing though the lumen by the dendrimer, whereas the smaller, or more flexible second species is allowed to pass through the lumen.

Alternatively, the first species and the second species are separated by their differing charges or polarities. This can be achieved, for example, by synthesis of a positively charged dendrimer, which would prevent the passage through the lumen of a more positively charged first species and allow the passage through the lumen of a less positively charged second species.

However, it should be noted that other ways in which the passage of the first and second species through the lumen could be discriminated can be envisaged, and the invention is not limited to the embodiments described above.

The effect of the presence of the dendrimer on the ability of the pore to pass an ionic current - that is allow ions to pass through the pore - was investigated by single channel current recordings.

The recordings were performed with heptamers SIO6C7 (as a reference), SIO6C7-G3- PAMAM, SIO6C ₇ -G2-PAMAM, and K46C ₇ -G5-PAMAM. At a potential of + 100 mV (chamber on the cis side of the protein grounded), current traces of unmodified SIO6C7 exhibited a unitary conductance of 925 ± 73 pS (n = 7) and an rms noise of 1.82 ± 0.17 normalized to the noise at 0 mV (Fig. 6A). By contrast, traces of SIO6C7 with a single G3-PAMAM inside the lumen had a lower current (Fig. 6B) with an average conductance of 512 ± 63 pS and a normalized rms noise of 2.82 ± 0.33 (n = 8). The current blockade of 45 ± 6% relative to the open channel indicates that the presence of

the dense polymer inside the pore lumen largely blocked the permeation of ions either by steric or electrostatic effects, or a combination of both. Placing a smaller and less dense G2 dendrimer led to a less pronounced current reduction of 25 ± 5% (Fig. 6C) (conductance of 686 ± 114 pS, normalized rms noise of 4.35 ± 0.48, n = 5).

The altered channel properties could be seen to be specific to the presence of the dendrimer because by addition of excess reducing agent DTT, which cleaved the disulfide bond between the cysteine and the PAMAM polymer, the conductance rose to that of the open pore.

Therefore in the embodiment of the invention wherein the dendrimer is attached to the lumen, the invention provides a method of altering the ability of pores to pass an ionic current by positioning a dendrimer within the pore.

For pores comprising dendrimers just outside the lumen, for example G5 -PAMAM tethered to K46C, a trace with a conductance of 880 pS was decorated with fast downward current fluctuations (Fig. 6D). An all-points-current histogram generated from a trace of K46C ₇ -G5 -PAMAM with a duration of 2 s displays the current levels for the fluctuations (Fig. 6E). The peak at 88 pA corresponds to the open channel, while the peak at 68 pA and a minor peak at 53 pA reflect blockade level 1 and blockade level 2 respectively. Blockades to level 2 were not investigated further due to their low frequency. The dominating current blockades from the open channel to level 1 were characterized by an amplitude of 210 ± 23 pS and an average duration of 2.2 ± 0.4 ms. The events occurred at a high frequency of 150 ± 30 s ^"1 (n = 3, total of 2000 events). The current fluctuations of K46C ₇ -G5 -PAMAM only occur at a potential of +100 mV (chamber on the cis side of the protein grounded) but not at -100 mV (Fig. 6F - chamber on the trans side of the protein grounded) suggesting that the events were dependent on the movement of the charged PAMAM relative to polarity of the potential.

Larger current fluctuations indicate a movement of the dendrimer. The fluctuations oscillated at a rate of 150 per second between the conductance level for the open channel and a 22% blockade level. These fluctuations can be interpreted as a rattling

movement of the dendrimer between a state close to the pore entrance and a state further away. The molecular forces for the fast dynamic changes could be the voltage- driven movement of the positively charged dendrimer outside the pore (positive potential at trans side) and the entropy-driven contraction of the stretched dendrimer branches bringing the polymer back to the pore entrance. When the potential was reversed to -100 mV at the trans side, no large current fluctuations were observed. This is in line with the model on the voltage-triggered movement of the dendrimer. Accordingly, the positive PAMAM ball would be expected to move towards the negative pole at the trans side and remain at the pore entrance without subsequent movement out of the channel. The voltage-dependent dynamic behavior of the tethered PAMAM dendrimer indicates that it could function as a voltage sensing molecular-ball valve.

Therefore in the embodiment of the invention wherein the dendrimer is attached externally to the pore the invention provides a way to regulate the flow a matter across the membrane in response to an external stimulus such as voltage.

The ability of the pore comprising a dendrimer to act as an ion sieve was further investigated by determination of the permeability ratio Pci-/Pκ+ for heptamers SIO6C7, SIO6C ₇ -G2-PAMAM, and S106C ₇ -G3-PAMAM: that is pores wherein the . I-V curves were constructed for currents recorded under both cisltrans and translcis KCl gradients and charge selectivities were calculated from the reversal potential, V _r (Table 2).

Table 2. Permeability ratios for αHL heptamers modified with PAMAM dendrimers

Heptamer V _r [mV] ^LaJ P _α -/Pκ+ ^LaJ Pκ+ /Pa- TAT Properties of PAMAM inside protein pore

Hydro- Number of dynamic free amino diameter groups on surface ^[b]

SIO6C7 -4.5 ^[c] 1.47 ± 0.06 0.68 ± 0.03 N.A. N.A

5. 4M 1.59 ± 0.06 0.62 ± 0.2 N.A. N.A.

S106C ₇ -G2 -9 C)[c 2.41 ± 0.14 0.41 ± 0.02 2.9 11

1 2.69 ± 0.25 0.37 ± 0.03

S106C 7-G3 -4 C)[ ^c 1.52 ± 0.06 0.65 ± 0.03 4.2

6. ₆ [d] 1.77 ± 0.05 0.56 ± 0.02

^[a] Average of at least three independent experiments ± standard deviation.

^[b] Determined by subtracting the number of amide-coupled PDP groups from the total number of primary amino groups prior to the modification with SPDP.

^[c] 300 mM KCl (cis), 100 mM (trans)

^[d] 100 mM KCl (cis), 300 mM (trans)

For SIO6C7, this analysis yielded a permeability ratio of Pci-/Pκ+ ⁼ 1.47 ± 0.06 (n = 3) implying that αHL is a weakly anion selective channel in agreement with previous studies (Gu, L. Q.; Dalla Serra, M.; Vincent, J. B.; Vigh, G.; Cheley, S.; Braha, O.;

Bayley, H. Proc Nat Acad Sci U S A. 2000, 97, 3959-3964). For S106C ₇ -G2-

PAMAM, the permeability ratio was found to be 2.41 ± 0.14 (Table 2) suggesting that addition of PAMAM made the pore more anion selective. This preference for anions is most likely due to the 11 additional positively ionised primary amino groups on the surface of the dendrimer inside the channel lumen. By comparison, SIO6C7 pores modified with G3 -PAMAM having only three terminal primary amino groups had a

Pα-/Pκ+ ratio of 1.52 ± 0.06 (Table 2) which is similar to the value for the unmodified pore. The G3 dendrimer had only three primary amino groups because the other amines were converted into amide bonds in the course of the attachment to the PDP groups. The recordings of pores modified with G3 and G2-PAMAM clearly demonstrated a positive correlation between the number of positively ionisable groups on the surface of the dendrimer and the resulting change of the permeability ratio of the modified pore.

The permeability ratio of S106C ₇ -G2-PAMAM of 2.41 ± 0.14 is further enhanced by increasing the number of positive charges. Preferred embodiments of the invention relate to dendrimers with high numbers of surface primary amines, which lead to a bigger preference for anions and dendrimers derivatives which terminate with quaternary amines and are therefore positively charged (Lee, J. H.; Lim, Y. B.; Choi, J. S.; Lee, Y.; Kim, T. L; Kim, H. J.; Yoon, J. K.; Kim, K.; Park, J. S. Bioconjug. Chem. 2003, 14, 1214-1221)

In addition, the preferred passage of anions over cations can also be enhanced by changing the size of the dendrimer. Therefore in order to gain increased selectivity, embodiments are preferred in which the space between the dendrimer and the sides of the pore are minimised, so that the anion is forced to pass through, rather than around, the dendrimer.

Therefore in another embodiment of the invention, the pores according to the invention are used for stochastic sensing, wherein individual molecules are detected by their ability to modulate ionic current flowing through a single pore. This approach has been used for analytes such as toxic metal ions (Braha, O.; Walker, B.; Cheley, S.; Kasianowicz, J. J.; Song, L.; Gouaux, J. E.; Bayley, H. Chem Biol. 1997, 4, 497-505), drugs, (Gu, L. Q.; Braha, O.; Conlan, S.; Cheley, S.; Bayley, H. Nature. 1999, 398, 686-690), enantiomers (Kang, X. F.; Cheley, S.; Guan, X.; Bayley, H. J. Am. Chem. Soc. 2006, 128, 10684-10685), TNT (Guan, X.; Gu, L. Q.; Cheley, S.; Braha, O.; Bayley, H. Chembiochem. 2005, 6, 1875-1881) and nucleotides (Astier, Y.; Braha, O.; Bayley, H. J. Am. Chem. Soc. 2006, 128, 1705-1710). Stochastic sensing is also an attractive label-free strategy to study single-molecule kinetics of chemical transformations such as the cis-trans isomerisation of azobenzene (Loudwig, S.;

Bayley, H. J Am. Chem. Soc. 2006, 128, 12404-12405) or the multistep formation or breaking of covalent bonds (Shin, S. H.; Luchian, T.; Cheley, S.; Braha, O.; Bayley, H. Angew Chem Int Ed Engl. 2002, 41, 3707-3709; 3523; Luchian, T.; Shin, S. H.; Bayley, H. Angew Chem Int Ed Engl. 2003, 42, 3766-3771; Luchian, T.; Shin, S. H.; Bayley, H. Angew Chem Int Ed Engl. 2003, 42, 1926-1929).

An essential component in these sensors has been the covalent attachment of small molecules and linear polymers within the pore. For example, the tethering of DNA oligonucleotides to engineered pores enabled the sequence-specific detection of individual free DNA strands (Howorka, S.; Cheley, S.; Bayley, H. Nat Biotechnol. 2001, 19, 636-639; Howorka, S.; Movileanu, L.; Braha, O.; Bayley, H. Proc Nat Acad Sci USA. 2001, 98, 12996-13001; Howorka, S.; Bayley, H. Biophys J. 2002, 53, 3202- 3210). Organic polymers such as polyethylene glycol (PEG) were also tethered to pores via engineered cysteines. Single channel current recordings of these pores demonstrated that a single PEG chain modulated the ionic current passing through the pore. Based on the characteristic current modulations, differences in the conformational dynamics of individual linear polymers of different chain length could be observed (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411-2416). PEG-modified pores also led to the development of biosensor elements capable of detecting protein analytes at the single-molecule level (Movileanu, L.; Howorka, S.; Braha, O.; Bayley, H. Nat Biotechnol. 2000, 18, 1091- 1095).

Therefore another application of the nanoscale pores attached to dendrimers according to the invention is in polypeptide and nucleic acid sequencing. In this single molecule approach, a transmembrane potential drives individual DNA and RNA strands through a pore thereby causing characteristic blockades in ionic current. Based on different current blockades, blockade durations and current signatures, nanopore recordings successfully identified nucleic acid homopolymers of different composition and block- copolymers. In an extension of these achievements, the nanopore sequencing may read base-sequences from individual translocating DNA or RNA strands. Recently, base- specific detection was attained on non-translocating DNA strands and on ribonucleotides using cyclodextrin adapter molecules lodged inside the αHL channel.

However, achieving single-base resolution in translocating DNA strands has, however, remained elusive because the passage of DNA is too fast (a few μs per base) to resolve individual bases. Slowing the passage of DNA is therefore of great interest.

Therefore, one embodiment of the invention is a nucleic acid or polypeptide sequencing device comprising a membrane as described hereinabove, an inlet and an outlet.

The inlet and the outlet are not limited in type, style or size. However, the inlet and the outlet are in fluid communication, and the inlet is positioned at the first side of the membrane and the outlet is positioned at the second side of the membrane.

The device according to the invention can be used in a method of sequencing a nucleic acid or a polypeptide in which a fluid comprising the nucleic acid or a polypeptide and an ionic salt is positioned on the first side of a membrane as described hereinabove, a potential difference is applied across the membrane and a resulting current is measured over time. The purpose of measuring the resulting current over time is to infer the identity of the translocating DNA strand.

Preferably the method of the invention is used when the nucleic acid is DNA or RNA.

The dendrimer positioned within the pore is not limited in type, but must be chosen so as to decelerate the passage of DNA to an appropriate rate to achieve resolution of individual bases. A nanopore filled with a positively charged and hyperbranched polymer such as PAMAM may be used to decelerate the passage of DNA via electrostatic interactions and/or steric factors.

PAMAM dendrimers are particularly appropriate for several reasons. First, DNA and RNA have a high affinity to positively charged dendrimer molecules, which is exploited in gene delivery carriers. Second, the forces of attraction between DNA and dendrimer have been characterized at the single molecule level for example with optical tweezers. Third, the affinity of DNA to PAMAM can be finely tuned by varying the chemical composition and charge of the dendrimer. Finally, spherical dendrimers can be confined inside the inner cavity of the αHL pore thus leaving the

barrel and inner constriction largely free of polymer. This narrow part of the pore is usually used as sensor region to detect bases using e.g. circular cyclodextrin molecules. By controlling the precise position of components of the engineered pore, the spatial separation of the functionally important DNA sensing region from the molecular sieve region can be achieved. Slowing down the translocation of DNA strands has so far not been demonstrated with other approaches. Nucleic acids strands could be slowed down by DNA translocating enzymes tethered to the nanopore, via linear polymers, or via genetic modifications of the protein pore.

In one embodiment of the method, addition of 1 μM RNA oligonucleotide C30 to the cis side of SIO6C7 with a potential of +100 mV at the trans side, resulted in frequent short current deflections (Fig. 7B) which were, however, absent in recordings without RNA (Fig. 7A). The high-amplitude blockades had an average current amplitude of 87 ± 4 % of the open channel conductance, a dwell time of 133 ± 34 μs, and occurred at a frequency of occurrence of 6.5 ± 1.2 s ^"1 (n =3) which is similar to other reports. The high-amplitude events represent the translocation of nucleic acids strand from the cis to the trans side of the pore which temporarily block the movement of ions through the lumen of the pore (Kasianowicz, J. J. Nature Materials. 2004, 3, 355-356; Akeson, M.; Branton, D.; Kasianowicz, J. J.; Brandin, E.; Deamer, D. W. Biophys J. 1999, 77, 3227- 3233).

In contrast, current traces of heptamer SIO6C ₇ -G3-PAMAM before (Fig. 7C) and after addition (Fig. 7D) of 1 μM C30 did not show any apparent differences. Analysis of extended C30 traces revealed that the frequency of occurrence of high-amplitude events for C30 traces was 0.005 s ^"1 , which is similar to the background noise of the recordings without RNA. The absence of high-amplitude events suggests that the dendrimer blocked the passage of the nucleic acids strands. It can be ruled out that translocation events occurred but their current signatures was masked by the lower conductance level of SIO6C ₇ -G3-PAMAM pores. RNA translocation events with a residual conductance of approx. 120 pS should have given rise to highly distinctive current deflections in traces of SIO6C7-G3-PAMAM with a mean conductance of 512 ± 63 pS and a noise band from approx. 620 to 400 pS. Therefore, the translocation of RNA polymers (Mw 8614 D) was reduced by more than three orders of magnitude.

The present invention, therefore, also provides a method of determining the structure of a pore, comprising attaching a dendrimer to the pore and measuring variations of an ionic current through the pore. Preferably the dendimer is attached to the lumen of the pore. Preferably the information relates to the dimensions of the pore.

Current traces of dendrimer- modified pores provided insight into the biophysical nature of engineered constructs. For example, recordings from αHL pores with internal G2 and G3 dendrimers established a direct and positive dependence between the size of the polymer inside the pore lumen and the extent of the current blockade. In addition, comparison of G2 and G3 traces with those of G5 revealed information about the conformational freedom of the dendrimer located inside and outside the narrow channel lumen, respectively. In G2 and G3 pores, there were no large current steps indicating that large conformational rearrangements of the dendrimer did not occur most likely due to the steric restraints in the narrow lumen. However, the normalized rms noise of G2 (4.35 ± 0.48) and G3 (2.82 ± 0.33) traces was higher than that of unmodified pores (1.82 ± 0.17) suggesting some minor movement of the dendrimer relative to the pore and/or small conformational changes within the dendrimer structure.

The invention also relates to a sulfhydryl-reactive dendrimer according to formula I:

wherein D represents a dendrimer.

These activated dendrimers suitable for use in attaching a dendrimer to a pore. The present invention also provides use of the above sulfhydryl-reactive dendrimer in the substituted-cysteine accessibility method (SCAM).

The sulfhydryl-reactive dendrimers can be used in the substituted-cysteine accessibility method (SCAM) (Karlin, A.; Akabas, M. H. Methods Enzymol. 1998, 293, 123-145) which infers the surface accessibility of residues by determining how fast sulfhydryl-

reactive reagents couple to single-cysteine mutants of a protein. SCAM has been widely exploited to probe the structure of membrane proteins and ion channels in combination with patch clamp or lipid bilayer recordings (Karlin, A.; Akabas, M. H. Methods Enzymol 1998, 293, 123-145; Akabas, M. H.; Stauffer, D. A.; Xu, M.; Karlin, A. Science. 1992, 258, 307-310) and gel electrophoretic mobility shift assays (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411-2416; Movileanu, L.; Cheley, S.; Howorka, S.; Braha, O.; Bayley, H. J Gen Physiol. 2001, 117, 239-252; Walker, B.; Bayley, H. J Biol Chem. 1995, 276», 23065-23071; Lu, J.; Deutsch, C. Biochemistry. 2001, 40, 13288-13301; Howorka, S.; Sara, M.; Wang, Y.; Kuen, B.; Sleytr, U. B.; Lubitz, W.; Bayley, H. J Biol Chem. 2000, 48, 37876-37886).

Despite the wide range of sulfhydryl-active organic reagents or polymeric reagents, most are too small or too flexible(Movileanu, L.; Cheley, S.; Howorka, S.; Braha, O.; Bayley, H. J Gen Physiol. 2001, 117, 239-252) for membrane proteins with pore diameters of more than 2 to 3 nm, because they can either access all pore-lining residues, or their coupling to cysteine residues does not give rise to an appreciable current block or gel shift, making it difficult to detect successful modification. Sulfhydryl-reactive PAMAM dendrimers with diameters of several nanometers can overcome these constraints as shown in the present invention. In calibration experiments with the structurally defined membrane pore αHL, dendrimers exhibited a sharp size-dependent permeation cut-off, and readily identified a 2.9 nm wide pore entrance. PAMAM-PDP dendrimers with diameters from 2 to 10 nm are well suited to probe various nanometer-sized pore-forming proteins with important biomedical roles such as bacterial pore-forming toxins (Parker, M. W.; Feil, S. C. Prog Biophys MoI Biol. 2005, 88, 91-142; Menestrina, G. Pore-forming peptides and protein toxins; Taylor & Francis, CRC, 2003), pores of the complement system or dilating purinergic receptors (Surprenant, A.; Rassendren, F.; Kawashima, E.; North, R. A.; Buell, G. Science. 1996, 272, 735-738; Khakh, B. S.; North, R. A. Nature. 2006, 442, 527-532). In addition, the approach can also be applied to explore the molecular structure of interesting biological nanomaterials such as porous S-layer proteins (Howorka, S.; Sara, M.; Wang, Y.; Kuen, B.; Sleytr, U. B.; Lubitz, W.; Bayley, H. J Biol Chem. 2000,

48, 37876-37886; Sleytr, U. B.; Messner, P.; Pum, D.; Sara, M. Angew Chem Int Edit. 1999, 35, 1035-1054).

The invention also relates to new method to alter the properties of proteins via targeted chemical modification by attachment of a dendrimer. Preferably the dendrimer is attached using a method as described by the present invention. Coupling of non- branched linear organic polymers or biopolymers has been used in the past to modify or expand the natural characteristics of proteins such as in pharmacology to increase the half-life of therapeutic proteins via PEGylation (Veronese, F. M.; Pasut, G. Drug Discov Today. 2005, 10, 1451-1458), in molecular biology to introduce sequence specificity into nucleases via attachment of a DNA oligonucleotides (Corey, D. R.; Schultz, P. G. Science. 1987, 238, 1401-1403), or in microarray technology to achieve targeted immobilization of PNA-modified proteins onto DNA-microarrays (Winssinger, N.; Harris, J. L.; Backes, B. J.; Schultz, P. G. Angew. Chem. Int. Ed. Engl. 2001, 40, 3152-3155).

Another embodiment of the invention provides a pharmaceutical composition comprising an active ingredient and one or more excipients wherein the active ingredient and the one or more excipient are enclosed by a membrane as described hereinabove.

It is important to have methods to control the release of the active ingredient of a pharmaceutical excipient to a subject to which the composition has been administered. In this regard the invention also provides a pharmaceutical composition that provides a controlled release of the active ingredient.

In the pharmaceutical composition according to the invention, the active ingredient is enclosed behind the membrane according to the invention. Again the membrane is not limited, other than it must be pharmaceutically acceptable. The pore and the dendrimer are selected using the size and charge based criteria outlined hereinabove for filtration devices, so as to allow the passage of the active ingredient from the first side of the membrane to the second side of the membrane at an appropriate rate, at which second

side the active ingredient may be absorbed by the subject who has been administered the active ingredient.

In one embodiment of the pharmaceutical composition according to the invention, the rate of release of the active ingredient through the membrane is dependent on the pH of the solution exterior to the membrane. Therefore, for example, the rate of release may be such that an active ingredient may not be released when the pharmaceutical is in a low pH environment, such as the stomach, and then subsequently released when the pharmaceutical passes into the small intestine. Active ingredients favoured for use in this embodiment of the invention include active ingredients that are unstable in the stomach, for example bisphosphonates.

The present invention alters pore permeation properties by placing a spherical dendrimer inside the lumen. The use of hyperbranched dendrimers has specific advantages over other spherical materials such as quantum dots of similar size or functionality. First, it would be difficult to couple a solid quantum dot into a pore lumen of similar diameter because the hard, highly symmetrical sphere would likely clash with the corrugated protein surface of the less symmetrical, non-circular pore opening. By contrast, compact dendrimers have some residual degree of structural flexibility which can help overcome small steric permeation barriers. Indeed, G3 PAMAM with a diameter of 2.9 nm successfully passed the 2.9 nm wide opening of the αHL pore. Furthermore, the hyperbranched character of the dendrimer is important in controlling and tuning the flow or matter through the engineered pore while solid impermeable spheres would likely lead to much more drastic and less tuneable changes in the permeation properties. In summary, placing dendrimers into a pore lumen is a unique approach to introduce ion selectivity filters or molecular sieves. The approach is not only restricted to protein pores but can be applied to engineer the permeation properties of inorganic or metallic porous structures for the separation of biopolymers or linear polymers for purification or sensing purposes.

Examples

The present invention will now be described with reference to the following examples, which are not to be considered limiting.

Experimental Methods

In each example, all reagents used were of analytical grade and were purchased from Sigma Aldrich (Gillingham, UK) unless otherwise stated. Polyamido-amine (PAMAM) dendrimers of generation 5 with a mixed surface of 10% primary amino / 90% hydroxyl groups (PAMAM), and PAMAM of generation 3 with 50% amine and 50% hydroxyl terminal groups were purchased from Dendritech, Inc. (Midland, MI), while PAMAM of generation 2 with 100% amine groups were obtained from Sigma Aldrich. G5 dendrimers were obtained as a 3.8% w/w methanolic solution, G3 as a 32% w/w solution, and G2 as 20% w/w solution. Methoxy-polyethyleneglycol- ethylmaleimide 10 kD was obtained from Apollo Scientific (Stockport, UK). Qiaprep kits for the purification of plasmid DNA were purchased from Qiagen (Crawley, UK). E. coli T7 S30 extract for coupled in vitro transcription/translation from circular DNA (Product # Ll 130) was from Promega (Southampton, UK). [ ³⁵ S]Methionine (1,200 Ci/mmol) was supplied by GE Healthcare. RNA oligonucleotides were obtained from Dharmacon.

Example 1

Synthesis oO-(2-Pyridyldithio)-propanoic acid

3-(2-Pyridyldithio)-propanoic acid was synthesized as a precursor for the generation of λ/-succinimidyl 3-(2-pyridyldithio)-propanoate (SPDP) (Carlsson, J.; Drevin, H.; Axen, R. Biochem J. 1978, 173, 723-737). 2,2'-Dipyridyldisulfide (DPDS, 2.4 g, 0.011 mol) was dissolved in anhydrous DCM and mixed with 3-mercapto-propanoic acid (0.577 g, 0.474 ml, 5.4 mmol). The solution instantly turned bright yellow and was left stirring at room temperature over night. The solvent was removed under vacuum to yield a viscous yellow oil and the product was purified by column chromatography (Al ₂ O ₃ ; dichloromethane : ethanol; 6:4). Once the yellow band corresponding to the thione by-

product had eluted, 4 ml acetic acid per 100 ml solvent was added to elute the product. Fractions containing the desired compound were pooled and the solvent removed under vacuum. The resulting viscous oil was placed on a high vacuum line for 5 hours to remove all traces of acetic acid to yield the title compound (1.06 g, 91% yield). Calculated mass: 215.01. ES ⁺ m/z - 216.0 (MH ⁺ ), 238.1 (M + Na ⁺ ). ¹ H NMR (d ₆ - DMSO) δ/ppm 12.43 (IH, broad s, -OH), 7.20 - 8.46 (4H, m, aromatic), 3.0 (2H, t, J = 6.91, 13.68, -CH ₂ -), 2.6 (2H, t, J = 6.81, 13.67, -CH ₂ -); ¹³ C NMR (δ/ppm) 172.5, 158.9, 149.6, 137.7, 121.2, 119.2, 33.4, 33.3.

Example 2

Synthesis of SPDP

3-(2-Pyridyldithio)-propanoic acid (194 mg, 0.902 mmol) in anhydrous DMF was cooled to 0 ⁰ C in an ice bath. 7V-hydroxysuccinimde (116 mg, 1.01 mmol) was added followed by dicyclohexylcarbodiimide (206 mg, 1.01 mmol), both in anhydrous DMF, and the solution was brought back up to room temperature and left stirring overnight. Afterwards, the solution was filtered under vacuum from the white urea-containing precipitate and the solvent was removed from the filtrate under high vacuum to yield a yellow oil. The crude material was purified via column chromatography (ethyl acetate : petrolether (40 - 60) = 3:7). The chromatography fractions containing the product were combined and the solvent removed to yield a colorless oil that crystallized overnight at -20 ⁰ C (141 mg, 50% yield). Calculated mass: 312.36. ES ⁺ m/z - 313.11 (MH ⁺ ), 335.10 (M + Na ⁺ ). ¹ H NMR (de-DMSO) δ/ppm 7.23 - 8.5 (4H, m, aromatic), 3.12 (4H, m, - CH ₂ -CH ₂ -), 2.8 (4H, s, -CH ₂ -CH ₂ -). ¹³ C NMR (de-DMSO) δ/ppm 170.0, 167.4, 158.5, 149.7, 137.8, 121.4, 119.5, 32.8, 30.2, 25.4.

Example 3

Synthesis of activated dendrimers by reaction with SPDP

For the preparation of PAMAM-PDP generation 5, SPDP (50 mg, 0.16 mmol) in DMSO (100 μl) was added to a solution of 0.3 M MOPS buffer pH 7.4 (2 ml) and G5 dendrimer (10% NH ₂ surface groups, 3.78% w/w in MeOH) (2.4 ml, 2.3 μmol). For G3-P AMAM-PDP, SPDP (5 mg, 16 μmol) in DMSO (100 μl) was mixed with a

solution of 0.3 M MOPS buffer pH 7.4 (2 ml) and G3 dendrimer (50% NH ₂ surface groups, 32.7% w/w in MeOH) (62 μl, 2.5 μmol). For G2-PDP, SPDP (1 mg, 3.2 μmol) in DMSO (100 μl) was added to a solution of 0.3 M MOPS buffer pH 7.4 (100 μl) and G2 dendrimer (100% NH ₂ surface groups, 20% w/w in MeOH) (18 μl, 0.93 μmol). The molar ratio of SPDP to PAMAM amine groups was five for G5 to ensure complete modification, but 0.5 and 0.25 for G3 and G2, respectively, to provide sufficient modification whilst preventing loss of product due to the precipitation of highly modified dendrimer. After reaction at RT for 1 hour, dendrimers were precipitated by the addition of ice-cold acetone (20 ml) and centrifugation at 21,000 g for 1 minute. The supernatant was removed and the resulting pellet re-suspended in ddH ₂ O (1.5 ml). Extraction of unreacted SPDP and its hydrolysis product 3-(2-pyridyldithio)-propanoic acid present in solution and possibly encapsulated (Beezer, A. E.; King, A. S. H.; Martin, I. K.; Mitchel, J. C; Twyman, L. J.; Wain, C. F. Tetrahedron. 2003, 59, 3873- 3880) within the dendrimer core was conducted using DCM. The extraction was repeated 8 times with each 2 ml to remove SPDP and 3-(2-pyridyldithio)-propanoic acid as judged by TLC analysis (1:4 MeOH: DCM; stained with KMnO ₄ ; Rf for 3-(2- pyridyldithio)-propanoic acid, 0.64; Rf for SPDP, 0.88). The extracted dendrimer solution was precipitated by the addition of ice-cold acetone (10 ml) and centrifugation at 21,000 g. The supernatant was removed and the pellet re-suspended in ddH ₂ O. The solution was then centrifuged under vacuum for 5 minutes at room temperature to remove traces of acetone and stored at -80 ⁰ C.

Example 4

Characterisation of activated dendrimers

Reversed Phase HPLC Analysis of PAMAM-PDP

The chemical derivatization of PAMAM dendrimers with SPDP was analyzed via reversed phase HPLC using a Varian ProStar system with a Module 410 autosampler, Model 210 solvent delivery module, and a Model 320 UV detector. A Discovery Bio Widepore C5 column (250 x 4.6 mm, 5 μm beads) was loaded with 10 μl of a 10 mg/ml solution of PAMAM dendrimers at a flow rate of 1 ml/min. The mobile phase was a linear gradient beginning with 90:10 water (0.1% TFA) /acetonitrile to either 50:50 or 70:30 water (0.1% TFA) /acetonitrile over 30 min followed by 2 min at the

final gradient concentration. Analysis was conducted using Star Chromatography Workstation software Version 5.51.

MALDI-ToF Mass Spectrometric Analysis of PAMAM-PDP Samples were analyzed with a Kratos Analytical Axima CFR Matrix Assisted Laser Desorption Ionisation Time of Flight Mass Spectrometer. 2,5-Dihydroxybenzoic acid (DHB) was used as matrix at a final concentration of 10 mg/ml in the sample.

Gel Electrophoretic Analysis of Dendrimers Dendrimers were analyzed via SDS-polyacrylamide gel electrophoresis (20% acrylamide gels for G2 and G3, and 12% acrylamide gels for G5) and Coomassie Blue Staining. The intensities of gel bands were determined using Scion Image (Scion Imaging, Frederick, MA).

Example 5

Expression of α-Hemolysin Single-Cysteine Mutants by In Vitro Transcription/Translation

The generation of mutants K8C and S106C of the semisynthetic gene αHL-RL2-D8 has been published elsewhere (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411-2416). αHL-RL2-D8 contains the conservative replacements Val-124 → Leu, Gly-130 → Ser, Asn-139 → GIn, He- 142 → Leu, which were introduced to facilitate cassette mutagenesis. In mutant S106C αHL-RL2-D8, Lys at position 8 was replaced by Ala to prevent adventitious proteolysis. D8 encodes for a C-terminal extension of eight aspartates. The mutant K46C was obtained by site-directed mutagenesis using in vivo recombination PCR ³¹ of the gene αHL-RL2-D8. None of the mutagenic changes alter the electrical properties of the pore as shown by single channel current recordings.

³⁵ S-labeled αHL polypeptides were generated from plasmid DNA harboring mutant αHL genes by coupled in vitro transcription/translation (IVTT) using the E. coli S30 T7 IVTT kit (Promega number Ll 130) (Cheley, S.; Braha, O.; Lu, X.; Conlan, S.; Bayley, H. Protein Sci. 1999, 8, 1257-1267). Briefly, the reaction volume contained the amino

acid mixture minus methionine (1.25 μl), premix (5 μl), and S30 mixture (3.75 μl) supplemented with 1 μg/ml rifampicin, and 7 μCi of [ ³⁵ S] methionine (GE Healthcare, 1,200 Ci/mmol; corresponding to a 15 mM solution) (1 μl). Supercoiled plasmid DNA (500 ng) was added to give a final reaction volume of 12.5 μl. The reactions were incubated for 1 h at 37 ⁰ C and then centrifuged for 5 min at 21,000 x g. The concentration of free thiols in the IVTT mix was 10 mM.

Example 6

Binding of PAMAM dendrimer and PEG polymer to single-cysteine mutant of αHL

Translation mix of K46C αHL monomer (0.25 μl; 10 mM thiol groups) was diluted into buffer containing 10 mM MOPS-NaOH, pH 7.4, 150 mM NaCl, 0.5 mM EDTA (ME buffer) (3 μl), and immediately reacted with 2 mM G3 -PAMAM-PDP or G5 -PAMAM- PDP (2 μl), or 2 mM G2-P AMAM-PDP (4 μl) for 20 min at 20 ⁰ C. The diluted αHL translation mix was also reacted with 50 mM PEG-MAL with a MW 5,000 (2 μl). The samples were analyzed with SDS-PAGE and autoradiography as described (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 722, 2411-2416).

Example 7

Binding of PAMAM-PDP and PEG-MAL to heptameric αHL protein pores

Mutant αHL polypeptides K46C, K8C, and S106C were allowed to assemble on rabbit erythrocyte membranes to form heptameric pores as described (Walker, B.; Krishnasastry, M.; Zorn, L.; Kasianowicz, J.; Bayley, H. J Biol Chem. 1992, 267, 10902-10909). Briefly, translation mix (1 μl) was diluted in ME buffer (5 μl) supplemented with 10 mM DTT (3 μl). Assembly was initiated by adding a suspension of rabbit erythrocyte membranes (1 mg protein /ml, 2.5 μl). The mixture was incubated for 1 hour at 20 ⁰ C with gentle resuspension of the membranes every 10 minutes. The suspension was centrifuged and the supernatant discarded. The pellet was washed with 1 x ME buffer (50 μl) and taken up in 1 x ME buffer (10 μl). αHL pores were then subjected to chemical modification by mixing the suspended membranes containing heptamers with 2 mM PAMAM-PDP (2 μl). After incubation for 10 minutes, unreacted

cysteine residues were quenched by the addition 1 M N-maleoyl-β-alanine (2 μl). Heptamers were also treated with PAMAM-PDP followed by the addition of 50 mM PEG-MAL 5 kD (2 μl), solely by reaction with 50 mM PEG-MAL 5 kD (2 μl) without dendrimers, or with 50 mM PEG-MAL 10 kD (2 μl). After 10 minutes, the suspensions were centrifuged for 5 minutes at 21,000 g. Supernatants were removed, pellets containing the heptamers were resuspended in ME buffer (10 μl), mixed with 2x Laemmli buffer and analyzed with SDS-PAGE at 100 mV followed by autoradiography. For single channel current recordings, modified heptamers were eluted from SDS-PAGE gels as described (Howorka, S.; Movileanu, L.; Lu, X.; Magnon, M.; Cheley, S.; Braha, O.; Bayley, H. J Am. Chem. Soc. 2000, 122, 2411- 2416).

Example 8

Binding of PAMAM-PDP and PEG-MAL to inorganic membranes

Disulfide-derivatized PAMAM dendrimers are coupled inside the pore lumen of gold- plated porous membranes. Disulfide-derivatized PAMAM dendrimers are prepared by coupling 7V-Succinimidyl 3-(2-pyridyldithio)-propanoate (SPDP) to amino-terminated PAMAM dendrimers of generation 5. SPDP (50 mg, 0.16 mmol) in DMSO (100 μl) is added to a solution of 0.3 M MOPS buffer pH 7.4 (2 ml) and G5 dendrimer (10% NH ₂ surface groups, 3.78% w/w in MeOH) (2.4 ml, 2.3 μmol). After reaction at RT for 1 hour, the dendrimer is precipitated by the addition of ice-cold acetone (20 ml) and centrifugation at 21,000 g for 1 minute. The supernatant is removed and the resulting pellet re-suspended in ddH ₂ O (1.5 ml). Extraction of unreacted SPDP and its hydrolysis product 3-(2-pyridyldithio)-propanoic acid present in solution and possibly encapsulated (Beezer, King et al. 2003) within the dendrimer core is conducted using DCM. The extraction is repeated 8 times with each 2 ml to remove SPDP and 3-(2- pyridyldithio)-propanoic acid as judged by TLC analysis (1:4 MeOH: DCM; stained with KMnO ₄ ; Rf for 3-(2-pyridyldithio)-propanoic acid, 0.64; Rf for SPDP, 0.88). The extracted dendrimer solution is precipitated by the addition of ice-cold acetone (10 ml) and centrifugation at 21,000 g. The supernatant is removed and the pellet re-suspended in ddH ₂ O. The solution is then centrifuged under vacuum for 5 minutes at room temperature to remove traces of acetone and stored at -80 ⁰ C.

The gold-plated porous membranes are prepared with the template synthesis method by electrolessly depositing gold along the pore walls of a polycarbonate template membrane (Martin, Nishizawa et al. 2001). The template is a commercially available filter, 6 μm thick, with cylindrical, 30-nm-diameter pores and 6 x 10 ^s pores per cm ² of membrane surface area. The inside diameters of the gold pores deposited within the holes of the template are controlled by varying the deposition time. The gold pores of membranes used for the PAMAM coupling have an inside diameter of 10 nm, as determined by electron microscopy and ion flux measurement (Martin, Nishizawa et al. 2001).

For the coupling of disulfide-terminated PAMAM dendrimers into the lumen of the gold pores, a 2 mM solution of SPDP-PAMAM in 50 mM MOPS buffer pH 7.4 (3 ml) is added to a sample of gold-plated porous membranes and incubated for 1 h. Unbound dendrimer is removed by washing with 50 mM MOPS buffer pH 7.4 (3 ml) and ddH^O.

Example 9

Single-Channel Current Recording

Single-channel current recordings were performed by using a planar lipid bilayer apparatus as described (Braha, O.; Walker, B.; Cheley, S.; Kasianowicz, J. J.; Song, L.;

Gouaux, J. E.; Bayley, H. Chem Biol. 1997, 4, 497-505). Briefly, a bilayer of 1,2- diphytanoyl-sft-glycerophosphocholine (Avanti Polar Lipids) was formed on an aperture (80 μm in diameter) in a Teflon septum (Goodfellow Corporation, Malvern,

PA) separating the cis and trans chambers of the apparatus. Each compartment contained 1.0 ml of 1 M KCl, 20 mM TrisηCl pH 7.5. Gel-purified heptameric αHL protein (final concentration 0.01-0.1 ng/ml) was added to the cis compartment, and the electrolyte in the cis chamber was stirred until a single channel inserted into the bilayer.

Transmembrane currents were recorded at a holding potential of +100 mV (with the cis side grounded) by using a patch-clamp amplifier (Axopatch 200B, Axon Instruments, Union City, CA). For analysis, currents were low-pass filtered at 1 kHz and sampled at

5 kHz by computer with a Digidata 1200 A/D converter (Axon Instruments), as described (Movileanu, L.; Howorka, S.; Braha, O.; Bayley, H. Nat Biotechnol. 2000,

18, 1091-1095). Single-channel conductances were determined by fitting the peaks in amplitude histograms to Gaussian functions.

Example 10 Ion selectivity measurements

Ion selectivity measurements were performed using asymmetrical conditions with one chamber (cis or trans) containing 300 mM KCl and the other chamber containing 100 mM KCl, supplemented each with 10 mM Tris-HCl pH 7.5. After the measurements, the lipid bilayer membrane was broken to determine the value of electrode junction potentials (normally around 0.5 mV). The permeability ratios (Pκ+/Pα-) were calculated from reversal potentials by using the Goldman-Hodgkin-Katz equation.

* _κ + _ L ^a ci itrans l ^a _a Jew ^

P \a 1 e ^{VrF IRT} - \a 1

where V _r is the reversal potential (i.e., the electrical potential giving zero current), ax is the activity of ion X, subscripts c and t represent the cis and trans compartments, R is the gas constant and F the Faraday constant. The temperature was 23 ± 1 ⁰ C. V _r was obtained by a polynomial fit of the current- voltage (I- V) data near zero current.