Biomod/2013/Aarhus/Results And Discussion/Chemical Modification
Introduction to chemical modification
In this section, the synthesis of the two different chemical modifications of the plate origami is described. In the chemically modified version of the plate, the origami plate itself will have a significant function along with the attached cholesterol and the photosensitizer, indium pyropheophorbide a chloride (In(PPa)Cl).
When the functional base plate is exposed, the cholesterols will bind the plate to the cell membrane, thereby getting the photosensitizer In(PPa)Cl close to the cell. This will create a high local concentration of the photosensitizer between the baseplate and the cell. When irradiated with light the produced singlet oxygen may damage the cell the baseplate is bound to, without damaging the cells nearby. If the same damage should be obtained with the In(PPa)Cl in solution there would be a larger amount of unspecific collateral damage.
In the functionalization of the base plate, only one cholesterol or In(PPa)Cl was supposed to be attached to each strand, with a total of 19 cholesterols and 5 photosensitizers on each plate. To accomplish this, a modified 2′,3′-dideoxyuridine triphosphate (ddUTP) was used in the enzymatic labeling. The results in this section are about the synthesis and conjugation of the three molecules to DNA strands. The photosensitizer and cholesterol were subsequently used in cell experiments to investigate if the design would work, which is described in the system in action section.
Introduction to cholesterol
Cholesterol (3) (Figure 27) is used to attach the origami plate to the cell surface by incorporation of the cholesterol in the cell membrane. Cholesterol is a natural and important steroid found in many living organisms. The steroid structure of cholesterol is the precursor for a number of important hormones. Besides being the precursor for other compounds, cholesterol has important functions as a lipid. In cells cholesterol is imbedded in the lipid bilayer to control the permeability and the fluidity of the cell membrane.
When cholesterol is embedded into the membrane, the alcohol group is oriented in order to interact with the head groups of the phospholipids, while the rigid quattro cyclic system and the tail are parallel to the fatty acid chains. These proporties can be exploited by using cholesterol to direct functionalities to the cell surface. The idea has been proven viable in experiments by Langecker et al. where a lipid membrane channel is bound to the cell surface using the affinity of the cholesterol for the lipid membrane (Figure 28). 
The ability of the cholesterols to attach to cell membranes is also exploited in this project, in order to bind the origami plate to the cell surface, thereby ensuring a high local concentration of the photosensitizer at the cell surface. In order to conjugate cholesterol to DNA staple strands, the cholesterol was modified with an NHS-ester handle. This NHS-ester could be used in an amide coupling reaction with a beforehand amine modified DNA strand. In total, 19 staple strands are modified with cholesterol for use in the origami plate.
Synthesis of cholesterol derivative
The cholesterol derivative 6 was synthesized in a three step procedure (Scheme 4). The linker and cholesterol are envisioned to react in a simple SN2 reaction. The oxyanion, from the deprotonated alcohol, performs a nucleophilic attack on the α-carbon with bromide as the leaving group. The second step is an acidic deprotection of the ester and the third step is an activation of the acid with N-hydroxysuccinimide (NHS).
The synthesis of 4 resulted in a low yield (8%), when using the procedure described by Simeone et al. and an optimization was as a consequence tried.  However, the optimization did not result in a significant improvement as the highest obtained yield was 12% in a time-consuming reaction procedure.
The synthesis of 5 is a deprotection of the tert-butyl ester with acid. Compound 4 was dissolved and left to react in formic acid at room temperature. After the reaction was deemed to be finished by TLC, the formic acid was evaporated under reduced pressure after which the product could be extracted from the organic phase with aq. saturated solution of sodium hydrogen carbonate. Hereafter the water phase was made acidic with 2 M hydrochloric acid and the product was back-extracted using dichloromethane (DCM). This method of purification resulted in the pure product in 76% yield after evaporation in vacuo.
To heighten the reactivity of the carboxylic acid this was converted into an NHS-ester, compound 6. The acid 5 was mixed with the coupling reagent N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide (EDC) and NHS after which the reaction was left to stir for 3 h. An NHS-ester is easily hydrolyzed as NHS is a very good leaving group. The inherent reactivity, which makes it useful as a coupling intermediate, also renders it hard to handle without the functionality decomposes. Therefore, the work-up had to be performed quickly. The desired product 6 was synthesized in 80% yield, which was deemed successful.
Introduction to photosensitizer3, 4] The photosensitizer produces singlet oxygen as illustrated in Figure 5 when irradiated with near infrared light. In addition singlet oxygen is known to induce apoptosis, however, the mechanism for this is not fully known.
A study by Mitsunaga et al. investigated the effect of photosensitizers bound to a cell surface.  Although a different photosensitizer was used in this project, the mechanism of cell death induced by singlet oxygen remains the same. In this study, the photosensitizer is bound to an antibody targeting specific receptors on the cell surface. It was concluded from the experiments that the photosensitizer could work from the outside as well as the inside of the cell. Additionally, it was observed that the singlet oxygen produced by the photosensitizer in solution was not necessary to induce cell death. For this reason, the system with the photosensitizer-modified baseplate, should work when the base plate is bound to the cell surface.
Synthesis of the photosensitizer
The photosensitizer can be synthesized in five steps starting from pheophytin a, a natural compound that can be extracted from spirulina powder (Scheme 6). The first four steps were performed according to literature. By activating indium(pyropheophorbide a)chloride as an NHS ester it can be conjugated with an amine modified DNA strand.  After the photosensitizer was synthesised and conjugated with DNA the ability of the photosensitizers to induce apoptosis in cells was analysed in a test setup.
In order to extract pheophytin a (7) from the spirulina powder it was necessary to lyse the cells, which was obtained by performing a cryogenic extraction. The extracted amount of pheophytin a corresponded with the 1 %w/w of spirulina powder reported in the literature. The measured melting point interval was lower and broader than reported in the literature, which indicated a contaminated product.  Impurities were also found in the NMR spectra. When performing the extraction again, further purification of the product unfortunately proved to be unsuccessful. Since it was possible to obtain a pure product at a later point in the synthesis of In(PPa)Cl, these impurities were concluded to be of less importance.
The first synthetic step was a trans-esterification of pheophytin a (7). Compound 7 was dissolved in allyl alcohol before conc. sulphuric acid was added. This reaction afforded the desired product 8 in 79% yield.
The second step was performed to remove the methyl ester group. In order to accomplish this, compound 8 was dissolved in 2,4,6-collidine and heated to 170 °C. Compound 9 was obtained in a high yield (91%).
The subsequent reaction was insertion of indium into the porphyrin ring, which led to a change in color from black to dark blue/green, and the insertion was confirmed by a UV-Vis spectrophotometry. The UV-Vis spectrum of In(PPa allyl ester) Cl (10) (appendix) was compared with a UV-Vis spectrum of PPa allyl ester (9) (appendix). Pyropheophorbide a allyl ester had a maximum absorbance at 274 nm and moderate absorbance at 410 nm. In the UV-Vis spectrum for In(PPa allyl ester)Cl no significant absorbance below the maximum absorbance at 423 nm is observed. In the 1H-NMR spectrum the amine protons were no longer present, and a downfield shift for two of the methine bridge protons was observed, as expected upon metalation of the porphyrin ring. This reaction afforded the indium compound in 45% yield.
The last step was a formation of an activated acid in form of an NHS-ester. This reaction was performed using a general procedure. 
This reaction afforded compound 12 in 91% yield without purification, as this treatment would decompose the labile NHS-ester. The NHS ester was used to conjugate the photosensitizer with a DNA strand. After conjugation the product was purified. The NHS ester was used in such excess that impurities were of less importance.
Synthesis of 5-propargylamino-ddUTP
For the chemical modification of DNA staple strands, the chosen method involved the enzyme terminal deoxynucleotidyl transferase, which was used to add a modified nucleotide to the 3′-end of the DNA strand. The chosen nucleotide was modified with a terminal amine, which could subsequently be used in an amide bond formation. In order to only add one nucleotide to the DNA strand a dideoxynucleotide was used. The nucleotide, which had already proved viable with the TdT method, was 5-propargylamino ddUTP (20).  Compound 20 is ideal for labeling a large number of different DNA strands with terminal amines, and therefore it was decided to synthesize the compound instead of buying it commercially, as it is very expensive (598 euro/mg). The synthesis was carried out over seven steps, starting from 2′-deoxyuridine (13) (5.6 euro/g). The first part of the strategy has previously been reported by McGuigan et al., however, modifications of the procedures were necessary. [12, 13]
The overall strategy (Scheme 7) was to remove the 3′-hydroxy group, functionalize the 5-position with propargylamine and functionalize the 5′-alcohol with a triphosphate chain. This was accomplished by converting the hydroxy groups into good leaving groups, followed by two SN2 reactions, giving the bicyclic system in 15. An elimination reaction affords 16, where the oxetane ring is opened and a double bond is formed. A hydrogenation of the double bond followed by an iodination of the 5-position in uridine leads to compound 18. A Sonogashira cross coupling reaction is used to incorporate a protected amine, which after the triphosphate synthesis can be deprotected.
The two first steps were carried out without purification of 14 as described in the literature. In the first reaction it was important to control the temperature because the sulfene species is unstable at higher temperature.  So the solution of 13 in pyridine was cooled to 0°C before MsCl was added. The first reaction resulted in brown sticky oil, which was subsequently dissolved in MeOH and NaOH, heated to 80°C and left for the oxetane ring to be formed. These two reactions yielded 15 in 48% over two steps.
The next step was a ring opening by an elimination reaction giving the double bond and homoallylic alcohol in compound 16. After purification the desired compound was obtained in an acceptable yield 58%.
The subsequent step was a hydrogenation of 16 resulting in 2′,3′-dideoxyuridine (17). This was accomplished with a catalytic amount of palladium over carbon and hydrogen gas. This reaction resulted in a quantitative yield.
The iodination of the base moiety was performed using iodine and ceric ammonium nitrate (CAN) as in the procedure by Asakura et al.  The nucleoside, iodine and CAN were dissolved in glacial acetic acid and then heated to 80 °C. After 37 min the reaction was complete and the solvent was evaporated. This reaction afforded the product in a good yield 72%.
The Sonogashira cross coupling was inspired by literature and performed under inert conditions with pre-dried reagents, however, the product was only obtained in 41% yield after purification. 
The triphosphate synthesis was performed using a procedure by Caton-Williams et al.  The reagents were dried overnight before they were dissolved in dry DMF. The tributylammonium pyrophosphate reagent was mixed with salicyl phosphorochloridite for half an hour to make the reactive species (Scheme 8), after which the mixture was added to the nucleoside for further reaction.
After 3 h an oxidative iodine solution was added followed by addition of water in order to hydrolyze the cyclic intermediate (Figure 29).
Crude 1H- and 31P-NMR were acquired which showed peaks corresponding to the desired product. However, the product could not be obtained after purification by preparative TLC. Because of time constraints the last reaction could not be performed again and another purification technique could not be tested.
Labeling with cholesterol
The labeling of DNA strands with cholesterol was carried out over two steps. The first step was a labeling reaction using the enzyme terminal deoxynucleotidyl transferase (TdT), where the nucleotide 5-propargylamino-ddUTP was added to the 3’-end of the DNA strand. The second step was the amide bond formation where the free amine will react with the NHS-ester. The outline of the two steps can be seen in Scheme 9 below. The procedure for the first step was inspired by the article by Jahn et al.  As 5-propargylamine-ddUTP was not successfully synthesized due to time constraints a brought version of the nucleotide was used in these labeling procedures.
The experiment was at first performed on 6 of the 19 DNA staple strands, from the origami baseplate, targeted for modification. 0.5 nmol of each strand were mixed together and to the solution the co-factor CoCl2, the reaction buffer containing the potassium cacodylate, the 5-propargylamino-ddUTP (20) and the enzyme was added. The volume was diluted to 100 μL and incubated at 37 °C for 15 min. The reaction was stopped using EDTA and an ethanol precipitation was performed to isolate the DNA. After the liquid was discarded the residue was dissolved in H2O and a solution of 6 in DMF was added. MeCN was added together with TEA and the reaction was incubated overnight at room temperature. After another ethanol precipitation the product was purified by RP-HPLC using a gradient (10% to 70% MeCN in TEAA buffer over 30 min). Even though there were 6 different strands it was not a problem separating the cholesterol modified from the ones without. The yield was calculated to 17% by UV absorbance. The remaining staple strands were modified using the same procedure.
Labeling with photosensitizer
The labeling of DNA strands with In(PPa)Cl was carried out over two steps (Scheme 10) as in the labeling of cholesterol onto DNA. The first step was the labeling reaction using the enzyme terminal deoxynucleotidyl transferase (TdT), where the nucleotide 5-propargylamino-ddUTP was added to the 3’-end of the DNA strand. The second step was the amide bond formation where the free amine will react with the NHS-ester of In(PPa)Cl.
The modified DNA was obtained as a light green solid. The conjugation with the DNA strand was apparent in the RP-HPLC chromatogram. The UV-Vis spectrum of the conjugated product corresponds well with the modified DNA strand: Absorption around 260 nm corresponds with the DNA and absorption at approximately 425 nm and 660 nm corresponds with the absorption at 423 nm and 657 nm observed in the UV-Vis spectrum of (10). The yield of the light green product was estimated to 1.6 nmol (65%).
Both the cholesterol NHS-ester and the In(PPa)Cl NHS-ester were successfully synthesized and conjugated to DNA staple strands. The precursor nucleoside 5-[3-(Trifluoroacetyl amino)-prop-ynyl]-2’,3’-dideoxyuridine was synthesized in six steps, however the last step was not completed due to time limitations and the 5-propargylamino-ddUTP was therefore not obtained.
- Langecker, M. et al. Synthetic lipid membrane channels formed by designed DNA nanostructures. Science 338, 932–936 (2012)
- Simeone, L. et al. Cholesterol-based nucleolipid-ruthenium complex stabilized by lipid aggregates for antineoplastic therapy. Bioconjugate Chem. 23, 758–770 (2012) 
- Arian, D. et al. A nucleic acid dependent chemical photocatalysis in live human cells. Chem. Eur. J. 16, 288–295 (2010).
- Cló, E. et al. DNA-programmed control of photosensitized singlet oxygen production. J. Am. Chem. Soc. 128, 4200–4201 (2006).
- Kochevar, I. E. et al. Singlet Oxygen , but not Oxidizing Radicals , Induces Apoptosis in HL-60 Cells. Photochem. Photobiol. 72, 548–553 (2000). 
- Mitsunaga, M. et al.' Cancer cell-selective in vivo near infrared photoimmunotherapy targeting specific membrane molecules. Nat. Med. 17, 1685–1691 (2011). 
- Wang, H.M. et al. Chemical constituents from the leaves of Nelumbo nucifera Gaertn. cv. Rosa-plena. Chem. Nat Comp. 47, 316-318 (2011)
- Frost, C. G. et al. Selectivity in Palladium Catalysed Allylic Substitution. Tetrahedron: Asymmetry. 3, 1089-1122 (1992)
- Funder, E. et al. Synthesis of Dopamine and Serotonin Derivatives for Immobilization on a Solid Support. J. Org. Chem. 77, 3134-3142 (2012)
- Jahn, K. et al. Functional patterning of DNA origami by parallel enzymatic modification. Bioconjugate Chem. 22, 819–823 (2011).
- McGuigan, C. et al. Discovery of a new family of inhibitors of human cytomegalovirus (HCMV) based upon lipophilic alkyl furano pyrimidine dideoxy nucleosides: action via a novel non-nucleosidic mechanism. J. Med. Chem. 47, 1847–1851 (2004). 
- Asakura, J. et al. Cerium (IV) catalyzed iodination at C5 uracil nucleosides. Tetrahedron Lett. 29, 2855–2858 (1988).
- Tahmassebi, D. et al. Substituent Effects on the Stability of Sulfenes. Phosphorus, Sulfur Silicon Relat. Elem. 181, 2745–2755 (2006).
- Jäger, S. et al. A versatile toolbox for variable DNA functionalization at high density. J. Am. Chem. Soc. 127, 15071–15082 (2005). 
- Caton-Williams, J. et al. Protection-Free One-Pot Synthesis of 2’-Deoxynucleoside 5'-Triphosphates and DNA Polymerization. Org. Lett. 13, 4156–4159 (2011).
SITEMAP | BIOMOD 2013 NANO CREATORS | Aarhus University
Copyright (C) 2013 | BIOMOD Team Nano Creators @ Aarhus University | Programming by: Peter Vium Skaarup.