WP3 - Container supported reaction / production
The WP objectives are to develop chemical reactions at the interface of micro- and nanoscale compartments such as to allow the demonstration of a powerful new platform that uses both (bio)chemistry and IT and interface them via MEMS technology to achieve new degrees of regulating chemistry (alteration of the chemistry on demand, parallelization) and computer control (new programmability). Such an idea requires that we implement reaction types that would have an input/output that could be controlled by a computer. Thus, we have decided to develop a photochemical production of amphiphiles as proof-of-concept and apply it such as to obtain an observable outcome, e.g., the production of amphiphiles that would induce morphogical changes of nano- and microscale structures, so called chemtainers. In addition, the platform development and its intended application (iterative chemistry) needed the precise addressing and readdressing of chemtainers, which would be achieved through the tethering of recognition molecules, DNA tags, and their photochemical alterations during chemical reactions.
Amphiphile production located on chemtainers
As amphiphiles, fatty acids were chosen, as they can be synthesized from non-surfactant molecules, here picolynium esters, in a single-step photochemical reaction by various organic and metal complexes. Our reaction is driven by a photosensitizer ruthenium trisbipyridine (Figure 3.1 and 3.2) that absorbs visible light energy (lambdamax = 465 nm).
Figure 3.1: Schematic representation of the photoexcitation and redox reactions. Step I: The energy associated with the light absorption is manifested by excitation of an electron of the metal core and its transfer to one of the ligands (MLCT). Step II: This MLCT is subsequently stabilized by an electron donation from an 8-oxoguanine. Step III: This secondary excited state transfers an electron to the precursor molecule (pL) thereby cleaving it and producing the amphiphile molecule (L). Step IV: The 8-oxoguanine is then regenerated by dihydrophenylglycine (H-atom donor) before it re-enters a photocatalytical round. Additionally, the picolinium radical is neutralized by the H-atom donor
Due to the various containers proposed in the MATCHIT project, the feasibility of the production of fatty acids (FA) by a photo-fragmentation reaction was successfully demonstrated in three self-assembled chemtainers or on their surface: FA vesicles, oil-in-water droplets stabilized by a monolayer of FA (o/w droplets) and reverse-micelles of FA in isooctane and 1-octanol (9:1).
To test the effects of the interfaces on the photocatalyzed fragmentation of the FA precursors, several systems were envisioned and the corresponding molecules synthesized (Figure 3.2).
Figure 3.2. Catalyst configurations. Catalyst I: aqueous intermolecular catalyst, 9-butyl-8-oxoguanine [or 2-amino-9-butyl-1H-purine-6,8(7H,9H)-dione] (1) and bis-(2,2’-bipyridyl)-4,4’-dimethyl-[2,2’]bipyridinyl)-ruthenium chloride = diMeRu (2); Catalyst II: Aqueous intramolecular catalyst, bis-(2,2’-bipyridyl)-(9-N-[4-(4’-methyl[2,2’]bipyridinyl-5-yl)-butyl]-8-oxoguanine)-ruthenium chloride = 8-oxoGRu (3); Catalyst III: Lipophilic intermolecular catalyst, 9-undecyl-8-oxoguanine (4) and [Ru(II)(bpy)2(4-decyl,4’-methyl-bpy)] chloride = RuC10:0 (5); Catalyst IV: lipophilic intramolecular catalyst, bis-(2,2’-bipyridyl)-(9-N-[4-(4’-decyl-[2,2’]bipyridinyl-5-yl)-butyl]-8-oxoguanine)ruthenium chloride = 8-oxoGRuC10:0 (6).
diMeRu (Figure 3.2, 2), which is highly soluble in aqueous media, was shown to electrostatically interact with the negatively charged interfaces of the vesicles/chemtainers. The derivatization with a hydrocarbon side chain (Figure 3.2, 4-5) resulted in a very extensive association (over 95% for the ruthenium complex Figure 3.2, 5) with the container interfaces, which was independent of the surface charge density.
Amphiphile production without containers
In the absence of any chemtainer, the precursor molecules (pL) form a separate phase in water, as small droplets. In these samples, the photofragmentation reaction rates were observed to increase according to the following sequence, Catalyst I. < Catalyst III < Catalyst II < Catalyst IV (Figure 3.2) with an approximate 1:20:45:90 rate ratio. The rates of the photochemical reactions were therefore dependent on both the catalyst-precursor association and the ruthenium-nucleobase covalent linkage. The picture obtained was clearly consistent with the general catalysis rules: intramolecular reactions are preferred to intermolecular reactions, as is the close proximity of the substrate to the catalyst (i.e., the non-covalent association of the catalyst with the precursor).
Amphiphile production supported by FA vesicles
The intermolecular reaction with Catalyst I (Figure 3.2) was very slow in the presence of FA vesicles due to a poor association of the 9-butyl-8-oxoguanine with the amphiphile structure. The inclusion of a covalent linkage of the ruthenium complex with the nucleobase resulted in a significant rate increase. In contrast, the rate difference between intra- and intermolecular reactions (Catalyst IV versus III) for the lipophilic cases became almost insignificant. The rates achieved in intermolecular lipophilic configuration (Catalyst III) were also comparable to those of intramolecular reactions where the catalyst interacted electrostatically with the bilayers (Figure 3.3 Catalyst III vs. II).
Figure 3.3: Rate dependence on catalyst configuration in the presence of preformed FA vesicle chemtainers. The reaction mixture was composed of 0.1 mM of Ru-catalyst, 5 mM precursor, 10 mM decanoic acid, and 15.75 mM H-source at pH=7.0, which was adjusted using NaOH. The columns (from left to right) correspond to Catalyst I, II, III and IV in Figure 3.2.
The presence of vesicles can alter the conversion rates expected by general rules of catalysis. The conversion rate became less dependent on the ruthenium complex-nucleobase catalyst configuration as long as both the ruthenium and the oxoguanine are associated with the bilayers. Indeed, at the relatively high local concentrations of catalysts resulting from the insertion in the interface, the advantage of the covalent linkage almost completely disappeared. This result indicated that the bilayer, i.e., the container, is sufficient for the co-location needed for efficient electron transfer between the nucleobase and ruthenium [Maurer et al. 2011].
The dynamic nature of decanoate/decanoic acid vesicles presents a significant obstacle in the preservation of chemtainer identity, as it prevents the determination of the origin of the structures produced during the growth and division of the chemtainers. To overcome this problem a new method to stabilize the structures was adapted to the system, where the bilayer hosting the photo-catalytic amphiphile production is anchored to a silica bead. This setup yields a first generation of chemtainers with a well-defined size (5 µm in size, although 40-200 nm beads should work as well) and composition, making it easy to differentiate first generation chemtainers from those of subsequent generations. The system is stable for weeks and does not inhibit the photochemical reaction [Albertsen et al. 2013].
Other chemtainers (oil in water and water in oil emulsions) were also tested for amphiphile production (collaboration with WP 2) and it was established the photochemistry can work with any of the proposed amphiphile chemtainers provided a slight reaction composition optimization is undertaken.
Alterations of tags using photochemistry
Development of the photochemical DNA tag alteration:
Two possible approaches to photochemical alterations of DNA tags were examined in detail: i) the alteration of the DNA tags is a direct consequence of the photochemistry, e.g., a photochemical ligation between two oligomers or ii) the photochemistry is used to deprotect an oligomer that only then could be used in a ligation reaction, hence for the alteration of DNA tags. In this case, we decided to follow the ii) approach.
To photochemically induce template-directed DNA-tag alterations (Figure 3.4), the existing DNA tag is activated on its 3′-end phosphate with an imidazole. The incoming DNA strand has a 5′-amino group that cannot ligate with a 3′-activated oligomer because it is “protected” by a picolinium carbamate. The protecting group has to be removed in a two-step procedure: photocleavage of the picolyl group followed by spontaneous decarboxylation to yield the free amine.
Figure 3.4: XNA ligation schematics. a) Activation of 3′-phosphates of oligomers with imidazole (Im-3′-p), and protection of 5′-deoxy-5′-amino-modified oligomers with picolinium carbamate. (5’-NH2-pic) b) Light driven deprotection with Ru(bpy)3/ascorbic acid by cleavage of the picolinium ester and successive decarboxylation. c) Schematic of a template directed ligation process using the above described chemistry.
For light dependent ligations, i.e. DNA-address alteration, several architectures are possible. The concept was successfully demonstrated (Cape et al., 2012) using a 55-mer hairpin sequence with a 3′-phosphate and a 30 nt 5´-terminal overhanging to react with a picolinium carbamate 13-mer that can hybridize to the 5′-overhang of the hairpin. The product was a 68 mer. In a hybridized state, the resulting free 5´-amine oligomer reacted with the adjacent 3′-phosphate imidazole, displacing the imidazole leaving group rapidly. The ligation reaction was completed almost quantitatively within 10 minutes (Figure 3.5, Left).
Figure 3.5: (Left) The kinetic profile of the reactants and products: Diamonds: Ligated products; triangles: picolinium carbamate oligomers. (Left) Hairpin with templating overhang and (Right) without templating overhang).
The success of this reaction was shown to strongly depend on a template, as it did not proceed when using a hairpin with only a single nucleotide overhang (Figure 3.5, Right).
The specificity of the ligation was tested in the presence of mutants that have a mismatch close to the reaction site. The results showed a preferential ligation when the template and the oligomers were complementary, even in cases where the mutant (HP-mx, x = 1 to 3) to complementary template (HP1) ratio is as high as 7.5:1. However, the effect of these mismatches is mainly a kinetic one, instead of a total inhibition. Thus, the greatest selectivity would be achieved in the early reaction stages (Figure 3.6).
Figure 3.6: Ligation yield with respect to activated starting material. (Left) Independent ligation reactions. The sample injected at 5 minutes, was the earliest possible HPLC injection after mixing reagents of the four reactions. Blue circles: HP1; Red squares: HP-m1; Green triangles HP-m2 and Purple diamonds: HP-m3. (Right) Mixed ligations with all templates in a pot. Red squares: Percentage of mutant hairpins ligated with the Oligomer; Blue circles: Percentage of correct hairpin ligated with the Oligomer.
After 3h incubation periods, 39% of imidazole activated HP1 reacted to form the correct product HP1-Oligomer, meanwhile only 7% of the mismatch hairpins had ligated (Figure 3.6 right).
Due to the excess of the mismatch hairpins, the ratio [HP1-Oligomer] vs. [HPmn-Oligomer]total was 4:5 in this assay, showing, that the selectivity is not optimal, still, taken into account the stoichiometric excess of the mutant strands, the selectivity of Oligomer towards the correct sequence was calculated to be 6:1. To improve this selectivity, future experiments will be carried out at higher reaction temperatures, i.e. closer to the Tm of the Oligomer/HP1 or even the intramolecular HP1 hybridization instead of at 20 °C. The fully matched ligation reaction was shown to be effective at temperatures of up to 60 °C.
Insertion of DNA tags in dynamic chemtainers
The realization of the WP objectives requires the durable attachment of DNA tags onto the surfaces of chemtainers. Amphiphilic DNA conjugated to 1 or 2 hydrocarbon chains as long as one layer thickness in a DA bilayer were tested for their interaction with chemtainers (provided by Prof. A. Herrmann at University of Groningen, ECCell project, FP7). These amphiphilic DNA strands interacted poorly with FA bilayers. Thus, we decided to use transmembrane amphiphiles, called bola-amphiphiles, which we previously have tested as membrane building blocks (Caschera et al., 2011, Collaboration with WP2 and WP1).
Figure 3.7. Two possible configurations of the bola-amphiphile anchor. The anchor could be designed such as to span the membrane, having the 5’-phosphate of the DNA as one headgroup and the alcohol as the other (left, Type A). However, in this configuration the triazole ring would be located within the hydrophobic core of the membrane. The type B anchor (right) was designed to span the whole membrane having the triazole ring as one headgroup and, e.g., an alcohol as the other.
The bola-amphiphile derivatized DNAs were synthesized using a click chemistry pathway to investigate two possible configurations of transmembrane anchors (Figure 3.7). The insertion of the DNA strands derivatized with amphiphilic anchors was performed by adding them to preformed chemtainers and determining their location in the chemtainer suspension with fluorescently labeled complementary DNA strand (Figure 3.8). Only the Type B bola-amphiphile anchor performed satisfactorily demonstrating the possibility to functionalize fatty acid bilayers with polyelectrolyte molecules, i.e. DNA recognition strands (Wamberg et al. 2013).
Figure 3.8: Insertion of the proposed derivatized DNA. The location of the DNA is established by hybridization of a fluorescent complementary strand. (Alexa 488) First row, In the two micrographs, the vesicles are visible as shadows indicating the absence of accumulation of the derivatized DNA (Type A) , while the membranes of the chemtainers are clearly highlighted in the second row by the fluorescent DNAs indicating the accumulation of Type B molecules in the chemtainer bilayers (last picture). For all pictures, same bar = 10 µm.
An alternate route for the synthesis of transmembrane anchors that would permit a tagging and de-tagging of chemtainers was also explored in collaboration with WP1.
The simultaneous amphiphile production and DNA reprogramming were successfully carried out as follows: the reactions were started in a sample containing amphiphile and DNA precursors (i.e., picolylcarbamate oligomers), but no preformed chemtainers. Upon irradiation, it was shown that both reactions proceeded simultaneously.
Albertsen, A.N. and Monnard, P.-A. "Bead supported protocell life cycles." in preparation.
Caschera, F., Bernadino de la Serna, J., Löffler, P.M. G., Rasmussen, T.E., Hanczyc, M.M., Bagatolli L.A. and Monnard, P.-A.*: (2011) “Stable Vesicles Composed of Mono- or Dicarboxylic Fatty Acids and Trimethylammonium Amphiphiles.” Langmuir, 27, 14078-14090.
Cape, J. L., Edson, J. B., Spencer, L. P., DeClue, M. S., Ziock, H.-J., Maurer, S. E., Rasmussen, S., Monnard, P.-A. and Boncella, J. M.: (2012) “Photo-triggered DNA phosphoramidate ligation in a tandem 5´-amine deprotection/ 3´-imidazole activated phosphate coupling reaction.” Bioconj. Chem., 23, 2014-2019.
Maurer, S. E., DeClue, M.S., Albertsen, A. N., Dörr, M., Kuiper, D. S., Ziock, H., Rasmussen, S., Boncella, J. M., and Monnard, P.-A.*: (2011) “Interact-ions between catalysts and amphiphile structures and their implications for a protocell model.” ChemPhysChem, 12, 828-835.
Wamberg, M.C., Wieczorek, R., Lykke Pedersen, P., Kwak, M., de Vries, J.W., Herrmann, A. and Monnard, P.-A.*, “Attaching nucleic acid to membranes using bola-amphiphiles synthesized by Click-chemistry. in preparation.