Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
Hollow microcapsules built by layer by layer assembly for the encapsulation and sustained release of curcumin S. Manju, K. Sreenivasan ∗ Laboratory for Polymer Analysis, Biomedical Technology Wing, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Poojapura, Trivandrum 695012, India
a r t i c l e
i n f o
Article history: Received 30 April 2010 Received in revised form 9 September 2010 Accepted 11 October 2010 Available online 15 October 2010 Key words: Layer by layer assembly Polyelectrolyte Microcapsules Curcumin Sustained release
a b s t r a c t Hollow microcapsules fabricated by layer-by-layer assembly (LbL) using oppositely charged polyelectrolytes have figured in studies towards the design of novel drug delivery systems. The possibility of loading a fair amount of active component of poor aqueous solubility is one of the encouraging factors on the wide spread interest of this emerging technology. Curcumin has potent anti-cancer properties. Clinical application of this efficacious agent in cancer and other diseases has been limited due to poor aqueous solubility and consequently minimal systemic bioavailability. LbL constructed polyelectrolyte microcapsules based drug delivery systems have the potential for dispersing hydrophobic agent like curcumin in aqueous media. Here we report the preparation of LbL assembled microcapsules composed of poly(sodium 4-styrene sulfonic acid) and poly(ethylene imine) one after another. The microcapsules were characterized using various analytical techniques. Curcumin was encapsulated in these microcapsules and the efficacy of the released curcumin was studied using L929 cells. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Considerable interest has been focused in the design and elaboration of delivery systems capable of providing sustained release of bioactive components [1–4]. Such systems comparing to conventional drug formulations, have several advantages such as lowered toxicity, prolonged action and greater convenience to the patients. Driven by these factors, past few decades have seen a surge of activities in the development of new systems for drug delivery applications [5–7]. Materials fabricated by layer-by-layer (LbL) assembly have unique opportunities in the context of designing new systems for the controlled release of therapeutics [8–12]. LbL assembly is an emerging technique to fabricate novel materials for a wide variety of biomedical applications including drug delivery [8]. Hollow microcapsules created by LbL assembly of oppositely charged polyelectrolytes on a sacrificial template have been explored recently for drug encapsulation and delivery [13,14]. Therapeutics can accumulate inside the microcapsules in the free form which is an additional advantage of this technology. To date, a wide range of diverse chemical entities including drugs, dyes and biopolymers have been included into hollow microcapsules. These studies are aimed to deliver drugs into cell components, fluorescence activated cell sorting and photodynamic therapy [15,16]. Layer by layer assembly is an easy to use method for the designing of multicomposite films as well as modification of the surfaces
∗ Corresponding author. Tel.: +91 471 2520248; fax: +91 471 2341814. E-mail address:
[email protected] (K. Sreenivasan). 0927-7765/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2010.10.021
of polymeric devices. In this approach, solid charged substrate is immersed in a solution of oppositely charged polyelectrolyte solution and the polyion is adsorbed via electrostatic attraction. The adsorption of each layer of polyelectrolyte is due to the charge overcompensation leading to charge-reversal of the surface. The charge reversal primarily regulates the adsorption process and the formation of the layers of oppositely charged polyelectrolyte. Cyclic repetition of this process of alternating adsorption of oppositely charged polyelectrolyte directs the formation of the thin multilayer architectures [17]. In the layered construction, a wide range of synthetic polyelectrolyte, biopolymers, lipids and inorganic particles have successfully been employed to fabricate materials by taking advantage of electrostatic interaction between oppositely charged species in their stepwise adsorption from an aqueous solution [8]. Polyelectrolyte shell with controlled thickness and composition was fabricated by step wise adsorption of oppositely charged polyelectrolyte on to partially cross linked melamine formaldehyde colloidal particle. Subsequent removal of core leads to the formation of very stable hollow structure [18,19]. The resulting hollow capsules are of spherical shape with narrow distribution. Curcumin, a yellow polyphenol extracted from the rhizome of turmeric (Curcuma longa), is a bis-␣, unsaturated  diketone with a molecular weight of 368.38. For centuries turmeric has been used as a spice and coloring agent in Indian food, as well as therapeutic agent in traditional Indian medicine. A volume of literature has established that free curcumin induces cell cycle arrest and apoptosis in human cancer cell lines derived from a variety of solid tumors including colorectal, lung, breast, pancreatic and
S. Manju, K. Sreenivasan / Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
589
prostate carcinoma [20–25]. The wide spread application of this effective agent has, however, been limited due to its poor water solubility and the resultant reduced systemic bioavailability. One approach that has been used widely to circumvent such limitation is the encapsulation of the drug in polymers in format like nano particles. Recently polymeric formulations containing curcumin have been reported in diverse applications including the design of drug eluting stents [26–28]. Here in we report the fabrication of microcapsules using LbL assembly loaded with curcumin. We used poly(sodium 4-styrene sulfonic acid) (PSS) and poly(ethylene imine) (PEI) on positively charged partially cross linked MF colloidal particle (∼2.5 m) by layer by layer construction and the MF was removed by acid leaching. The bioactivity of the released drug is demonstrated by cytotoxicity studies using L929 cell lines.
2.2.4. In vitro release kinetics of curcumin A known amount of microcapsules (20 mg) loaded with curcumin was dispersed in PBS (pH 7.4) and kept in darkness in an orbital shaker at 37 ◦ C at shaking speed of 200 rpm. A calibration plot was constructed between the optical density (OD) of curcumin solutions in ethanol at 426 nm versus concentrations. At definite time intervals, the amount of curcumin released was estimated from the corresponding OD of the solution. The percentage of curcumin released was determined from the equation
2. Materials and methods
2.2.5. Fourier transform infrared (FT-IR) studies of microcapsules FTIR spectrums were recorded in the range 4000–400 cm−1 using a Nicolet 5700 FTIR Spectrometer with a horizontal ATR accessory containing diamond Crystal (Madison, USA). The numbers of scans were 50.
2.1. Materials Poly(sodium 4-styrene sulfonic acid) MW 70,000, poly(ethylene imine) 50% (w/v) aqueous solution, micro particles based on melamine formaldehyde as templates and MEM supplemented with Fetal Bovine Serum were purchased from Sigma Chemical Co., Bangalore, India. Curcumin (biocurcumax), Arjuna Natural Extracts Ltd., Kerala, was used as received. The other reagents or chemicals were obtained from Merck, Mumbai, India. L929 (mammalian cell line that has demonstrated reproducible results) was obtained from National Centre for Cell Sciences, Pune, India. Phosphate buffered saline (PBS, pH 7.4) was prepared by dissolving 17.97 g of disodium hydrogen phosphate, 5.73 g of monosodium hydrogen phosphate and 9 g of sodium chloride in 1 L distilled water. 2.2. Methods 2.2.1. Preparation of hollow microcapsules Polyelectrolyte multilayer capsules were prepared according to the method reported earlier [16]. Briefly, melamine formaldehyde (MF) templates were alternatively coated with six double layers of poly(sodium 4-styrene sulfonic acid) (PSS) and poly(ethylene imine) (PEI). Hollow micro capsules were obtained by dissolving the MF cores using hydrochloric acid (pH <1.5). After core dissolution, the content was collected by centrifugation (1274 rcf, 10 min) and the collected capsules were washed thoroughly with water. The hollow microcapsules were freeze dried from the aqueous suspension for subsequent use. 2.2.2. Loading of curcumin Curcumin solution was prepared in ethanol (0.01 wt%). Dried hollow microcapsules were dispersed in 5 mL of the solution and kept in an orbital shaker (37 ◦ C, 200 rpm) for 12 h in darkness. Microcapsules filled with curcumin were collected by three repeated centrifugation (1274 rcf, 10 min)/washing cycles in water. The drug loaded microcapsules were then freeze dried. In ethanol the layers get opened and the drug is diffused into the capsules [16]. 2.2.3. Entrapment efficiency The entrapment efficiency of curcumin loaded in the hollow microcapsules was determined as follows. Initial concentration of curcumin added and the amount of non accumulated curcumin in the supernatant liquid was measured by UV spectroscopy at 426 nm. The amount of curcumin accumulated in the interior of the capsule was determined by the difference between the initial concentration of curcumin added and the amount of non accumulated curcumin remaining in the solvent.
Release (%) =
Ct × 100 C0
(1)
where Ct is the concentration of released curcumin collected at time t and C0 is the total amount of curcumin released. The measurements were made in triplicate and the average was taken.
2.2.6. Transmission electron microscopy (TEM) TEM pictures of polymeric particle were taken in a Hitachi H600 TEM Instrument operating at different magnification. A finely focused illuminating probe can be scanned across the specimen and a standard Everhart–Thornley detector collects the secondary electrons. The specimen holder with rotational grid facility was used with the microscope. 2.2.7. Energy dispersive (EDS) X-ray analysis Energy dispersive X-ray spectroscopy analysis was performed using an EDX model 6051 SP (Oxford Instruments, UK) attached to the scanning electron microscope. Prior to the measurement, a thin layer of gold was coated on its surface. 2.2.8. Particle size The particle size distribution of the core and layer by layer constructed core shell polymer particle was determined using photon correlation spectroscopy (PCS) on a Zetasizer nano ZS(Malvern Instruments, Malvern, UK). 2.2.9. Zeta potential potential values were obtained using Malvern Zetasizer Nano ZS with a He–Ne laser beam. All measurements were done at a wavelength of 633.8 nm. Zeta potential was measured by applying an electric field across the dispersion. Particles within the dispersion with a zeta potential will migrate toward the electrode of opposite charge with a velocity proportional to the magnitude of the zeta potential. The frequency shift or phase shift of an incident laser beam caused by these moving particles is measured as the particle mobility, and this mobility is converted to the zeta potential by inputting the dispersant viscosity, and the application of the Smoluchowski or Huckel theories. 2.2.10. Cytotoxicity studies Cytotoxicity evaluation of hollow microcapsules and curcumin loaded microcapsules was carried out by the test on Extract method with monolayer of L929 mouse fibroblast cells according to ISO standards (ISO 10993-5, 1999). Briefly, L929cells were subcultured from stock culture (National Centre for Cell Sciences, Pune, India) by trypsinization and seeded in to multi-well tissue culture plates. Cells were fed with Dulbecco’s minimum essential medium supplemented with bovine serum and incubated at 37 ◦ C in 5% carbon dioxide atmosphere. The extract was prepared by incubating the test material with culture medium with serum at 37 ± 2 ◦ C for 24 h
590
S. Manju, K. Sreenivasan / Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
at an extraction ratio of 0.1 mg/mL. The extract was filtered and adjusted the pH. 100% extracts of the test samples was diluted to get concentration of 50% and 25% with media. Different dilutions of extracts of test samples, negative control and positive control in triplicate were placed on sub confluent monolayer of L929 cells. After incubation of cells with extract of test materials and controls at 37 ± 2 ◦ C for 24 h, cell culture was examined microscopically for cellular response using a phase contrast inverted microscope (Leica, WLD MPS32, Germany). The morphology of the cells was assessed in comparison with negative control (ultra high density polyethylene) and positive control (dilute phenol). 3. Results and discussion Polyelectrolyte multilayer films (PSS/PEI)6 were prepared on partially crosslinked MF of ∼2.5 m diameter by alternate coating of six layers each of PSS and PEI. Despite water solubility of poly(styrenesulfonic acid) and poly(ethylene imine), electrostatic attraction retains the polyelectrolyte multilayers PSS/PEI on the surface of core particle even after extensive washing. In the initial step of polyelectrolyte multilayer construction poly(sodium 4-styrenesulfonicacid) was adsorbed on positively charged MF core particle and as a result the net surface charge became negative. Charge overcompensation results in the adsorption of oppositely charged poly(ethylene imine) on microcapsules. Here we controlled the layer thickness by six (PSS/PEI) layers on MF micro particle. Layer by layer construction of each layer; polycation (PEI)/polyanion (PSS) on MF was quantified by potential using Zetasizer ZS. Zeta potential value changes from negative to positive on the adsorption of PSS followed by PEI. Fig. 1 shows the zeta potential variation of PSS/PEI system after each polyelectrolyte deposition. On the adsorption of PEI on MF microparticle the surface potential became positive (+59.9 mV). PSS adsorption rendered the surface negative (−73.7 mV). Charge reversals observed after each new polyelectrolyte deposition and thus confirm the multilayer formation. In this case the overall charge a particle attain in aqueous suspension was >50 mV on each adsorption and hence shows high potential stability of core shell micro particle in the suspension. After adequate coating MF core particles were dissolved using HCl (pH <1.5) without affecting the integrity of the capsule wall. Fig. 2 shows the size distribution profile of melamine formaldehyde, layer by layer constructed core shell micro particle and empty polyelectrolyte core shell. Average particle size of layer by layer constructed polyelectrolyte core shell micro particle was about 2827.313 nm and that of melamine formaldehyde was about 2780.82 nm. The shell thickness of the microcapsules was calculated from the difference of these values and was 46.49 nm. However, the average particle size of microcapsules after core dissolution was about 2200 nm reflecting the availability of inner voluminous space for drug encapsulation. Microcapsules showed a negative zeta potential with magnitude 59.1 mV, ruling out the possibility of forming
Fig. 1. Charge reversal on alternate adsorption of polyelectrolyte on MF micro particle by Zeta potential versus (PSS/PEI) adsorption.
Fig. 2. Particle size and size distribution of core shell micro particles, layer by layer constructed micro particles and empty polyelectrolyte microcapsules.
aggregates leading to the precipitation of the colloidal suspensions in aqueous media. EDS trace depicted in Fig. 3 shows the peak associated with nitrogen, sodium, sulfur and oxygen showing the presence of poly(ethylene imine) and poly(sodium 4-styrene sulfonic acid) in the capsule wall. Spherical morphology together with a nearly uniform size distribution is apparent from TEM images shown in Fig. 4. By close observation of TEM images shown in Fig. 4B and C central portion is empty confirming the effective removal of MF core particle with HCl (pH <1.5) without adversely affecting the (PSS/PEI)6 multilayer shell. Curcumin was loaded into hollow capsules as depicted earlier. In ethanol, the polyelectrolyte layers are opened and allow the diffusion of curcumin into the capsule. Encapsulation efficiency of curcumin was 4.5 g/mg of microcapsules. Fig. 5 shows the IR spectrum of curcumin loaded micro spheres. The spectrum shows a broad peak at 3411.5 cm−1 characteristic of OH stretching frequency in curcumin. Two small peaks at 2924 cm−1 and 2849 cm−1 associated with the stretching of C–H in the ring further confirm the presence of curcumin in the capsules. The bands around 1610 cm−1 show NH stretching of poly(ethylene imine). C–N stretching at 1116 cm−1 of CH2 –NH2 moiety further confirms PEI on the layer. The spectrum also shows C–S stretching band at 1230 cm−1 showing the presence of poly(styrene sulfonate). Thus microcapsule spectrum after encapsulation undoubtedly shows the presence of included curcumin in the hollow microcapsules. To evaluate the drug release profile of curcumin, micro capsules were immersed in PBS at 37 ◦ C and time dependent drug release profile was measured under physiological condition. Fig. 6 shows in vitro release profile of curcumin from the microcapsules. An initial release of 1.11% (0.26 g/mL) of total drug encapsulated was observed in 24 h, followed by a sustained release for a period of 1 week. Within one week a total release of 2.77% was acquired in PBS indicating the potential of curcumin loaded micro capsules as a sustained drug delivery vehicle. The release profile of curcumin loaded micro capsules is controlled by diffusion mechanism as a result of partitioning between polymeric micro capsules and surrounded aqueous phase. Majority of the drug still remains inside micro capsules indicating drug adsorption and stabilization on the inner layer through non covalent interactions. Cytocompatibility of hollow micro capsules and the bioactivity of the released curcumin from the capsules were studied using in vitro cytotoxicity evaluation. Morphology of mouse fibroblast cell on contact with the extract of empty polyelectrolyte micro capsules in culture medium with serum is depicted in Fig. 7C. There is no considerable change in the morphology of cells even after contact with the extract for 24 h indicating non cytotoxicity of empty microcapsules reflecting that the components used in the fabrica-
S. Manju, K. Sreenivasan / Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
591
Fig. 3. Energy dispersive X-ray spectrum of the microcapsules constructed by LBL assembly.
Fig. 4. TEM images of (A) layer by layer constructed core shell micro particle (B) and (C) core dissolved polyelectrolyte microcapsules.
tion of the capsules are not eliciting any toxic response. However, on incubating the mouse fibroblast cells for 24 h with extract of curcumin loaded polyelectrolyte microcapsules shows severe cytotoxicity (Fig. 7D). Completely distracted cell morphology of mouse fibroblast cells confirms the toxic effect of curcumin extracted from the polyelectrolyte microcapsules. It reveals that the bioactivity of curcumin is unaltered during the processes of encapsulation and extraction. The amount of curcumin extracted in to the cell cul-
ture medium was 0. 25 g/mL. 100 L of this extract was placed on subconfluent monolayer of L929 cells which showed severe cyctotoxicity (Fig. 7D). The quantity of curcumin in 100 L is 0.025 g indicating that low quantity of curcumin can inflict damage to the cells. The morphology of cell on incubating with negative and positive control is shown in Fig. 7A and B. The stable hollow microcapsules after the dissolution of MF lead to a central empty portion with sufficient diameter. There is no
Fig. 5. FTIR spectrum of curcumin loaded micro capsules.
592
S. Manju, K. Sreenivasan / Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
Fig. 6. Release kinetics of curcumin from microcapsules in phosphate buffered saline.
significant shrinking of capsules even after core dissolution and hence high percentage of space is available for drug encapsulation. The core dimension plays an important role in controlling the capsule size. Thus nano particles or micro particles as sacrificial colloidal system lead to capsules of appropriate dimensions and such systems are emerging as new tools in the field of targeted drug delivery. In spite of the wide availability of several polymeric systems for drug delivery, the quest to develop efficient vehicles with a dual role of protection of the drug and modulating the release of the encapsulated drug are still pursuing vigorously. Much attention has recently been focused on hollow polyelectrolyte microcapsules as drug delivery modules. Apart from meeting many of the requirements for an ideal delivery vehicle, the simplicity in designing the capsules together with the feasibility of encapsulating the delicate bioactive components is probably the factor for recent surge of interest in this technology. The release of encapsulated drug depends on the permeability or the break down of the layered architecture. The release profile shown in Fig. 6 favors that release
is due to the permeation of curcumin out of the capsules. Breakdown of the layers would have resulted in burst release of the drug. The strong electrostatic interaction among the layers, presumably, ensures the stability of the capsules. In a recent work of Markarian et al. the role of ion pairing together with hydrogen bonding has been highlighted in the stabilization of polyelectrolyte layers composed of homopolynucleotides and PEI [29]. The effect of different stimuli in modulating the drug release from capsules has been figured recently. Chung et al. have demonstrated the pH modulated release of methylene blue from hollow capsules constructed using layers of PAA and PAH [30]. Several recent studies have shown the role of temperature, ionic strength etc. in controlling the release profile from microcapsules [31–34]. Our study, on the other hand, was carried out in PBS at physiological pH and it seems that the release is diffusion controlled. Diffusion of, in particular, small substrates from capsules constructed by LbL approach has been recently demonstrated by Kreft et al. [35]. Our data reflect that, since the integrity of the capsules is retained, system of this kind would be useful for an extended period of release. After a period of one week, only 2.77% of the included drug is released reflecting that a system of this kind could be suitable for stretching the release to a prolonged period. The amount of curcumin released within 24 h is 0.26 g/mL (0.026 g/100 L) into PBS and cytotoxicity data show that 0.025 g/100 L is sufficient to induce toxic response as reflected in Fig. 7D. The in vitro release data and cytotoxicity results strongly favor that low quantity of drug is required to get the response and it seems that microcapsules can provide the dose for a prolonged period. Further the severe cytotoxicity of the extract of drug encapsulated microcapsules on L929 cells strongly indicate that bioactivity of the drug is retained even after encapsulation. Our preliminary results show that curcumin, a potential anticancer component, can be loaded inside microcapsules constructed by LbL approach with the possibility of releasing the potent agent in a sustained fashion for prolonged period.
Fig. 7. Fibroblast cell after contact with (A) ultra high molecular weight polyethylene (UHMWPE), (B) dilute phenol, (C) the extract of empty polyelectrolyte micro capsules and (D) the extract of curcumin loaded micro capsules. A phase contrast inverted microscope (Leica, WLD MPS32, Germany) was used for obtaining the optical micrographs.
S. Manju, K. Sreenivasan / Colloids and Surfaces B: Biointerfaces 82 (2011) 588–593
4. Conclusion We have demonstrated that partially soluble (in water) naturally occurring anticancer agent; curcumin can be accumulated in hollow microcapsules having polyelectrolyte multilayer shells (PSS/PEI)6 . The microcapsules were found to be cytocompatible while the extract of capsules loaded with curcumin showed severe cytotoxicity on the mouse fibroblast cell indicating released curcumin is active. The high stability of polyelectrolyte microcapsules in the aqueous medium depicts that these drug carriers are suitable for drug delivery applications. Our future studies will be directed to modify this delivery format to acquire targeting ability towards cancer cells. Acknowledgements We are grateful to Dr. C.P. Sharma and Willy Paul for Zeta potential and particle size distribution analysis and to Dr. T.V. Kumari for assistance on cell culture studies. Ms. Manju thanks Department of Biotechnology, New Delhi, India, for financial support. References [1] R.C. Mundargi, et al., J. Controlled Release 125 (2008) 193. [2] S.H. Hu, et al., Langmuir 24 (2008) 239. [3] L.R. Moroni, et al., Biomaterials 27 (2006) 5918.
[4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35]
593
W. Hua, et al., Biomaterials 28 (2007) 99. H. Ai, et al., J. Controlled Release 84 (2002) 122. F.R. Caruso, et al., Science 282 (1998) 1111. J.B. Zhang, et al., Biomaterials 26 (2005) 3353. Z.Y. Tang, et al., Adv. Mater. 18 (2006) 3203. Q. Gan, T. Wang, Colloids Surf. B: Biointerfaces 59 (2007) 24. G. Decher, et al., J. Thin Solid Films 211 (1992) 831. G. Decher, C. Fuzzy, Science 277 (1997) 1232. K. Onda, et al., Biotechnol. Bioeng. 51 (1996) 163. K. Ariga, et al., J. Am. Chem. Soc. 119 (1997) 2224. G.B. Sukhorukov, et al., Colloids Surf. A: Physiochem. Eng. Aspects 137 (1998) 253. X. Qiu, et al., Langmuir 17 (2001) 5375. K. Wang, et al., J. Mater. Chem. 17 (2007) 4018. A. Antipov et al., US Patent No. 20070224264, 2007. S. Ye, et al., J. Biomater. Sci. Polym. Ed. 16 (2005) 909. H. Huang, et al., J. Am. Chem. Soc. 121 (1999) 3805. S. Stewart, G. Liu, Chem. Mater. 11 (1999) 1048. B.B. Aggarwal, et al., Anticancer Res. 23 (2003) 363. K. Mehta, et al., Anticancer Drugs 8 (1997) 470. C. Ramachandran, W. You, Breast Cancer Res. Treat. 54 (1999) 269. J.L. Arbiser, Mol. Med. 4 (1998) 376. G.B. Mahady, et al., Anticancer Res. 22 (2002) 4179. S. Aggarwal, et al., Int. J. Cancer 111 (2004) 679. A. Duvoix, Cancer Lett. 223 (2005) 181. K.T. Nguyen, et al., Biomaterials 25 (2004) 5333. J. Ch Pan, et al., J. Controlled. Release 116 (2006) 42. M.Z. Markarian, et al., Biomacromolecules 8 (2007) 59. A.J. Chung, M.F. Rubner, Langmuir 18 (2002) 1176. S.T. Dubas, et al., J. Am. Chem. Soc. 123 (2001) 5368. J.F. Quinn, F. Caruso, Langmuir 20 (2004) 20. T. Serizwa, et al., Angew. Chem. Int. Ed. 2 (2003) 1115. O. Kreft, et al., Angew. Chem. Int. Ed. 46 (2007) 5605.