Engineering Dendritic Aptamer Assemblies As Superior Inhibitors of Protein Function
Youngmi Kim,Donn M. Dennis,Tim Morey,Liu Yang,Weihong Tan
DOI: https://doi.org/10.1002/asia.200900421
2009-01-01
Chemistry - An Asian Journal
Abstract:Branching out: A new method has been developed for designing superior inhibitors of protein function by assembling heterogeneous aptamers into dendritic aptamer assemblies composed of two different functional domains: an anchor and a dendrimer-like inhibitor (27Apt and 15Apt, respectively). The engineering of this heterogeneous ligand assembly enhances molecular function for thrombin binding, and provides novel advantages and further improves inhibitory efficacy. The concept of molecular assembly is ubiquitous in nature. Standard examples include amino acids, nucleic acids, and carbohydrates, all of which may assemble in different sequential orders or geometries, resulting in the creation of unique functions. In bioengineering, assembling molecules that possesses distinctive functions in a new structure is highly advantageous in creating new properties or amplifying existing ones. The enhancment of binding properties, for instance, can be achieved most easily by implementing the phenomenon of multivalency, or polyvalency.1 Multivalency is defined as the “cooperative” effect that takes place among multiple binding domains when two molecules generate multiple and simulateous interactions. It is such interaction that results in the dramatic increase of binding affinities. Particularly, the design of inhibitors takes advantage of polyvalency by linearly assembling capture molecules, such as repetitive epitopes of antibodies, sugars, or small molecules, that recognize specific cell receptors. Although efforts in this direction have all shown progress,2 demand still remains high for better molecular recognition elements and strategies. Therefore, we report herein a DNA aptamer based molecular assembly and demonstrate its improved inhibitory efficacy to protein activity. As a new class of binding elements, DNA aptamers have outstanding potential as future pharmaceutical agents owing to their important advantages over antibody epitopes or carbohydrates in molecular assembly.3–5 For instance, DNA aptamers possess relatively small molecular weights (making multiple-molecule assembly feasible) and have high binding affinities. Structurally, DNA aptamers are easily modifiable with robust structural restoration under ambient conditions.6 One of the most well-known examples is thrombin and its two binding aptamers, 15Apt binding to exosite 1 and 27Apt recognizing exosite 2.7, 8 Their binding specificities are very high, but their affinities are rather different, sub-nanomolar Kd for 27Apt and sub-micromolar Kd for 15Apt.9, 10 Exocite 1 is a fibrin-binding exocite that can be recognized by 15Apt, which modulates fibrinogen binding and thrombin-mediated clotting. As a potential anticoagulant, only the 15Apt should have the enzymatic inhibitory functions required for thrombin-mediated coagulation because it occupies the fibrinogen-binding exosite 1. Although a series of trials demonstrated the efficacy of 15Apt, its insufficient binding affinity has often been a significant drawback. To solve this problem, preliminary studies reported that the linear assembly of 15Apt and 27Apt11, 12 can improve the binding strength, but the final results were far beyond what we would consider to be the limits of inhibitory potency for aptamer assemblies. Therefore, we took a dendritic approach to the design of an effective multiligand assembly and tested its inhibitory efficacy against the target protein, thrombin. We have designed assemblies composed of three functional groups: 1) 27Apt that possesses high binding affinity to thrombin, 2) dendritic 15Apts composed of multiple 15Apts, and 3) a linker that couples the dendritic 15Apts and 27Apt. The schematic structure of this assembly is shown in Figure 1. Briefly, the dendrimer-like configuration of 15Apt functions as a highly branched inhibitory domain that is composed of multiple 15Apts.14 Recalling that 15Apt binds to exosite 1, the engineering of this heterogeneous ligand assembly activates the “cooperative” effect, as described above. Specifically, the process results in a net gain of extra ligands and, although the concentration of inhibitors around exosite 1 is artificial, it is still increased. Also important to the design structure is an anchoring molecule which is highly beneficial in overcoming the low binding affinity of 15Apt.11, 12 The anchor domain (27Apt) ensures that the dendrimer-like assembly of 15Apts remains proximate to the thrombin so that the effective local concentration is extremely high compared to nonlinked 15Apts.1, 11, 12 Dendritic aptamer assemblies. Each assembly contains an anchor domain, which is 27Apt. In the middle of the linker region, a dendritic modifier is introduced to covalently link the inhibitory aptamer (15Apt). In the absence of a dendritic modifier, only one 15Apt is linked to 27Apt, which we term Mono-15Apt assembly. In contrast, doubler or trebler dendritic modifiers allow two and three 15Apts to be linked to 27Apt with the identical distance. These are called Dual-15Apt and Triple-15Apt, respectively. (S)n gives the number of linkers (see notation in Table 1). To design the assemblies, an eight-spacer phoramidite linker between dendritic 15Apt and 27Apt has been engineered.11 Then we applied two different dendritic modifiers, a doubler, and a trebler to fulfill different numbers of 15Apt conjugation (Figure 1). In other words, the new assemblies contained either two or three 15Apts and are named Dual-15Apt assembly or Triple-15Apt assembly, respectively. New assemblies were then based on the optimal design of the Mono-15Apt assembly, with each dendritic modifier inserted between the third and fourth hexaethylene glycol (HEG) unit. This site was randomly chosen because the inhibitory efficiency of the assemblies is not dependent on the position of the modifier, but rather the number of assembled 15Apts (data not shown). In addition, to make use of high-pressure liquid chromatography (HPLC) as the purification method, each dendritic assembly was modified with a Dabcyl quencher at 5′ to make it more hydrophobic. This was done because a long polyethylene glycol linker often makes the oligonucleotides too hydrophilic to withstand purification by reverse-phase liquid chromatography. After HPLC purification, the molecular weight of the final products was verified (Figure 2). According to the gel image, we were able to ascertain that the molecular weight of each assembly was increased by the number of 15Apts. Thus, we concluded that the desired assemblies of DNA aptamer probes had been successfully synthesized and purified. Agarose gel image with ethidium bromide (EB) staining. To verify the final products after HPLC purification, the length of each DNA probe was estimated using agarose gel electrophoresis. The target protein, thrombin, is a serine protease (EC 3.4.21.5) that hydrolyzes soluble fibrinogen into insoluble strands of fibrin, as well as catalyzes many other coagulation-related reactions. To evaluate the inhibitory potency of the newly designed probes, prothrombin time (PT) was determined by using human plasma samples.15 PT measures the clotting time upon activation of the extrinsic pathway. Thus, determining PTs with varying concentrations of each probe can provide an effective way of comparing inhibitory potency (Figure 3). PTs of 27Apt alone were not reported because 27Apt would not affect clotting time.11 The results showed that the Mono-15Apt assembly could delay the PT twofold over the 15Apt alone.8 The Dual-15Apt assembly showed about 260 % and 130 % longer PT compared to 15Apt and Mono-15Apt assembly, respectively. Such improvement is more clearly shown in the difference of IC200 value. The IC200 value is defined as the concentration needed to delay clotting twice as long which is a frequently used measure of inhibition in the blood clotting test. The IC200 of the Dual-15Apt assembly was 0.2 μM, which was five times lower than that of Mono-15Apt assembly at 1 μM. This dramatic improvement of anticoagulation is quite surprising because, even when the concentration of 15Apt or Mono-15Apt assembly is doubled, their inhibition of thrombin activity cannot be as efficient as the Dual-15Apt assembly. Thus, the results partially support the success of the dendrimeric approach in assembling highly efficient functional aptamers. However, we could not observe any significant improvement over the Dual-15Apt assembly when compared to the Triple-15Apt assembly. This can be explained by the steric hindrance and heavy molecular weight of such an assembly. Briefly, the heavy molecular weight may reduce the diffusion time, which is one of the main factors affecting kon in aptamer binding with thrombin, likely resulting in lower inhibition than expected. Moreover, when three or more 15Apts are configured together, the resulting density hampers the ability of each one from reaching the binding site. Therefore, according to the anticoagulation comparisons, Dual-15Apt assembly is the optimal design by which to maximize anticoagulation and minimize structural complexity. Dose-dependent prolongation of prothrombin time (PT). A human plasma sample was used to determine the anticoagulant efficacy of each assembly. The Dual-15Apt and Triple-15Apt Assemblies show the best anticoagulant efficacy, although Triple-15Apt did not show much improvement when compared to that of Dual-15Apt assembly. The inset box summarizes PT maximum and IC200, which is the concentration to prolong the clotting time twice as long. The data of 15Apt and Mono-15Apt assembly are obtained from the previous report.11 In the case of Dual- and Triple-15Apt assemblies, it could be argued that the enhanced anticoagulation effect simply arises from the doubled or tripled concentration of 15Apt in the mixture of samples, rather than significant increase of 15Apt’s local concentration around thrombin. To investigate such a possibility, we used a simplified testing system that contained physiologically conditioned buffer, thrombin, and substrate. The clotting process was monitored by a change in optical density. To minimize the trimolecular interactions involving two thrombin proteins and single Dual-15Apt assembly, we chose a concentration of 1 nM for each component, far below the Kd of 15Apt. Thus, the observed result should only reflect the bimolecular interaction. The results, as shown in Figure 4, indicate limited inhibition of 1 nM, 2 nM, and 3 nM of 15Apt-treated sample. The reaction rate of each sample was compared to the rate of the blank sample (VB). As the concentration of 15Apt is increased, the initial rate is slowed to 72, 68, and 44 % of VB. At the same time, the rates are 5.2 % and 0.92 % of VB for the Mono-15Apt and Dual-15Apt assemblies, respectively. Thus, it can be seen that the inhibition of the assemblies is greatly increased. Interestingly, Dual-15Apt assembly has about a sixfold slower reaction rate than Mono-15Apt. We believe that this finding results from the enhanced local concentration of 15Apts and not only from the doubled or tripled actual concentration of 15Apt in the mixture of the samples. Our initial hypothesis suggested that molecular assemblies can increase binding affinity by taking advantage of multivalent molecular interactions. In the case of Dual-15Apt, we have demonstrated that, by the addition of one extra 15Apt, we can achieve increased effective concentration of 15Apts proximate to thrombin exosite 1. Real-time monitoring of thrombin activity. When treated with each inhibitor, the reaction rate is compared. Thrombin activity, pre-incubated with each inhibitor, was monitored after the substrate was added. 15 Apt alone at various concentrations; Mono-15Apt and Dual-15Apt assemblies at 1 nM concentration. In conclusion, we have shown that the dendritic approach for molecular assembly is an effective way to significantly enhance molecular function. By designing a series of assemblies and testing them against each other using human plasma samples, we have shown that Dual-15Apt assembly is the best inhibitor for thrombin because it possesses 1) a higher, and therefore effective, local concentration and 2) a more flexible steric space for inhibition. Aside from the proof-of-concept demonstrated by this antithrombin model, the actual experimental results demonstrate the efficacy of the molecular assembly as a novel treatment modality for many diseases involving abnormal coagulation. In addition, recent studies have reported positive anticoagulation therapy outcomes in patients whose tumor growth and hematosis were linked to thrombosis.16–18 Therefore, dendritic molecular assembly of aptamers can be one of the most effective approaches for use in aptamer-based drug development. Probe synthesis: All DNA synthesis reagents were from Glen Research. All DNA probes (Table 1) were synthesized with an ABI3400 DNA/RNA synthesizer using standard procedures. Spacers were coupled for a total of 15 min to ensure high coupling yield. For complete cleavage and deprotection, overnight incubation with ammonia was used. After ethanol precipitation, the precipitates were then redissolved in 0.5 mL of 0.1 M triethylammonium acetate (TEAA, pH 7.0) for further purification with high-pressure liquid chromatography (HPLC). The HPLC was performed on a ProStar HPLC Station (Varian, CA) equipped with a fluorescent and photodioarray detector. A C-18 reverse phase column (Alltech, C18, 5 μM, 250×4.6 mm) was used. Name Sequence (5′-3′) 15Apt GGTTGGTGTGGTTGG 27Apt GTCCGTGGTAGGGCAGGTTGGGGTGAC Mono-15Apt assembly Dabcyl-TGGTTGGTGTGGTTGG-(S)8-GTCCGTGGTAGGGCAGGTTGGGGTGACT Dual-15Apt assembly Dabcyl-TGGTTGGTGTGGTTGG-(S)5-(doubler)-(S)3-GTCCGTGGTAGGGCAGGTTGGGGTGACT Dual-15Apt assembly 2 Dabcyl-TGGTTGGTGTGGTTGG-(S)3-(doubler)-(S)5-GTCCGTGGTAGGGCAGGTTGGGGTGACT Triple-15Apt assembly Dabcyl-TGGTTGGTGTGGTTGG-(S)5-(trebler)-(S)3-GTCCGTGGTAGGGCAGGTTGGGGTGACT Real-time monitoring of coagulation process: A well-known mechanism during the clotting process is the turbidity of sample, which is dramatically increased owing to the cleavage of fibrinogen to fibrin. Thus, UV/Vis measurement is used to conduct real-time monitoring of the clotting process. A Cary 100 Bio UV/Vis spectrophotometer (Varian Inc., Palo Alto, CA) was used. Briefly, a mixture of 200 microliters of physiological buffer (25 mM Tris-HCl at pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, and 5 % (v/v) glycerol) containing 1 nM of thrombin (Haematologic Technologies, Inc.) was incubated with each inhibitor for half an hour. Fibrinogen (Sigma Aldrich, MO), was then added to the sample mixture while the absorption was recorded at 500 nm. PT measurement: To evaluate anticoagulation efficacy, we determined the PT for each ligand using human plasma sample, following the manufacturer’s instructions. For PT determination, 50 μL of Universal Coagulation Reference Plasma (UCRP, Fisher Scientific, PA) was pre-incubated at 37 °C with different amounts of each ligand for 2 min, and then 50 μL of thromboplastin-L was added to initiate the extrinsic clotting cascade. Finally, the scattering signal was monitored until the signal was saturated. To calculate PT, the end time was determined at the point where scattering signal reached half maximum between lowest and maximum points. It was repeated twice, and each set of experiments was conducted with one batch of plasma. This work is supported by NIH, ONR, and NSF.