|Home | About | Journals | Submit | Contact Us | Français|
Alzheimer's disease (AD) is a progressive, age-dependent, neurodegenerative disorder with an insidious course that renders its presymptomatic diagnosis difficult1. Definite AD diagnosis is achieved only postmortem, thus establishing presymptomatic, early diagnosis of AD is crucial for developing and administering effective therapies2,3.
Amyloid β-protein (Aβ) is central to AD pathogenesis. Soluble, oligomeric Aβ assemblies are believed to effect neurotoxicity underlying synaptic dysfunction and neuron loss in AD4,5. Various forms of soluble Aβ assemblies have been described, however, their interrelationships and relevance to AD etiology and pathogenesis are complex and not well understood6. Specific molecular recognition tools may unravel the relationships amongst Aβ assemblies and facilitate detection and characterization of these assemblies early in the disease course before symptoms emerge. Molecular recognition commonly relies on antibodies. However, an alternative class of molecular recognition tools, aptamers, offers important advantages relative to antibodies7,8. Aptamers are oligonucleotides generated by in-vitro selection:systematic evolution of ligands by exponential enrichment (SELEX)9,10. SELEX is an iterative process that, similar to Darwinian evolution, allows selection, amplification, enrichment, and perpetuation of a property, e.g., avid, specific, ligand binding (aptamers) or catalytic activity (ribozymes and DNAzymes).
Despite emergence of aptamers as tools in modern biotechnology and medicine11, they have been underutilized in the amyloid field. Few RNA or ssDNA aptamers have been selected against various forms of prion proteins (PrP)12–16. An RNA aptamer generated against recombinant bovine PrP was shown to recognize bovine PrP-β17, a soluble, oligomeric, β-sheet-rich conformational variant of full-length PrP that forms amyloid fibrils18. Aptamers generated using monomeric and several forms of fibrillar β2-microglobulin (β2m) were found to bind fibrils of certain other amyloidogenic proteins besides β2m fibrils19. Ylera et al. described RNA aptamers selected against immobilized monomeric Aβ4020. Unexpectedly, these aptamers bound fibrillar Aβ40. Altogether, these data raise several important questions. Why did aptamers selected against monomeric proteins recognize their polymeric forms? Could aptamers against monomeric and/or oligomeric forms of amyloidogenic proteins be obtained? To address these questions, we attempted to select aptamers for covalently-stabilized oligomeric Aβ4021 generated using photo-induced cross-linking of unmodified proteins (PICUP)22,23. Similar to previous findings17,19,20, these aptamers reacted with fibrils of Aβ and several other amyloidogenic proteins likely recognizing a potentially common amyloid structural aptatope21. Here, we present the SELEX methodology used in production of these aptamers21.
Initially, the protein used for SELEX is pretreated with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) to obtain homogeneous, aggregate-free preparations, as described previously23. This step is necessary because pre-formed aggregates induce rapid aggregation of amyloidogenic proteins, resulting in poor experimental reproducibility24, and are undesirable for selection of aptamers for unaggregated, non-fibrillar forms of the protein.
Before using the protein for SELEX, desalting is performed to remove the cross-linking reagents used for PICUP. This PICUP reaction mixture contains the cross-linking reagents and ~55 μM protein (nominal concentration).
The synthetic ssDNA library used here for SELEX included 49 randomized nucleotides (A:T:G:C=25:25:25:25%) flanked by constant regions comprising cloning sites (BamHI, EcoRI) and a T7 promoter, as described previously26.
To remove the unincorporated nucleotides, use two G-50 columns according to the manufacturer's instructions.
First, the RNA and protein are incubated in solution and then the RNA sequences that bind to the protein are separated from the non-binders. As SELEX cycles progress, filter binding will give an indication of protein RNA binding enrichment.
RNA is extracted from the filters to obtain the sequences that bind to the protein. These sequences are amplified for the next SELEX cycle.
To proceed to the next cycle of SELEX, RNA has to be reverse-transcribed to DNA and amplified by PCR.
In SELEX experiments, the nature of Aβ40 oligomers used as the target, the quality of RNA used for each cycle, and successful RNA extraction and amplification after each cycle are important. We used PICUP to generate an oligomeric Aβ40 mixture for SELEX after purification and removal of the cross-linking reagents. The desalting experiments described in Part 2 usually lead to 50–55% protein loss. The protein amount and quality can be assessed using absorbance measurements (λ=280) and SDS-PAGE (Figure 1). The average absorbance profile of Aβ40 eluates from 5 individual experiments overlaying a typical SDS-PAGE profile of Aβ40 eluted in one of those experiments are shown in Figure 1. The data show that the protein consistently elutes off the column in fractions 3–5 and the cross-linking reagents elute after fraction 6 (increased absorbance in fraction 7, Figure 1). SDS-PAGE shows the typical Aβ40 oligomer distribution22. This distribution was reproducible after the protein fractions were lyophilized (2.11), treated with HFIP (2.11), resolubilized (7.1), and re-analyzed by SDS-PAGE.
Integrity of RNA for each SELEX cycle is also important for iterative progression of SELEX, especially when nuclease-susceptible ribo-oligonucleotides are used. After RNA amplification and labeling (Part 4), the quality of labeled RNA can be assessed by TBE-urea-polyacrylamide gel electrophoresis. A typical profile of an intact labeled RNA product before and after G-50 purification (Part 5) is shown in Figure 2.
After each SELEX cycle, RNA is extracted from the membrane after filter binding (Part 8) and reverse-transcribed (Part 9) to DNA for PCR amplification (Step 9.5). The DNA template from a previous cycle is then used to generate 32P-labeled RNA (Part 4) for the next cycle. If some contaminating DNA template from a previous cycle persists in the labeled RNA product after in-vitro transcription reactions (Part 4), the efficiency of SELEX cycles will be reduced, demanding more cycles. To control for this, after each SELEX cycle and the corresponding reverse-transcription PCR reaction, agarose electrophoresis is performed (9.12). Absence of DNA product in the negative-control reaction tubes (9.4) indicates successful removal of the DNA template originating from a previous SELEX cycle (Figure 3). If DNA amplification by PCR is observed in the negative-control tube, it is advised that the duration of incubation with RQ1 DNase (4.3) be prolonged. The manufacturer-recommended incubation duration with RQ1 DNase is 15 min, however, we found that longer incubations (4–5 h) were required to remove the template DNA completely (Figure 3).
The starting point of the SELEX process is synthesis of a random oligonucleotide library typically containing 1012–1015 sequences. In DNA SELEX, this library is used directly after a ssDNA pool is generated, whereas in RNA SELEX, demonstrated here, the ssDNA library is converted first to an RNA pool enzymatically by in-vitro transcription. Then, SELEX is performed iteratively whereby each cycle comprises exposure and binding of oligonucleotides to the intended target, partitioning of binders from non-binding sequences, and elution of binding sequences. In later cycles, the stoichiometry of the target and RNA and/or the number of washes can be altered, or competitive inhibitors can be added in the binding or wash buffer to increase the stringency of the SELEX conditions. Filter binding is a classic, simple, and fast method used for SELEX10, though numerous methods have been described28 for binding and partitioning. Filter binding is ideal for capture and selection of RNA-target interactions in solution. When the selection targets are small molecules or peptides, the pore size of the membrane used in filter binding becomes a limiting factor and an important consideration.
Following elution, the target-binding oligonucleotides are amplified and used for further cycles. When using DNA SELEX, the enriched pool can be amplified by PCR directly and used for the next cycle after digestion of the complementary sequence and DNA labeling. When using RNA SELEX, this pool has to be converted to DNA by reverse-transcription and then amplified by PCR, in-vitro transcribed, and labeled again for continuation to the next cycle. Usually, 8–20 such cycles are required to obtain an enriched aptamer pool, which is cloned subsequently and individual aptamers are sequenced and characterized. In studies using amyloid proteins, characterization of aptamers is particularly important because the inherent affinity of oligonucleotides for fibrillar amyloid structures potentially hinders development of aptamers specific for non-fibrillar amyloid proteins under physiological conditions21. This inherent, sequence-independent affinity of oligonucleotides may have led to generation of fibril-cross-reactive aptamers in studies aiming to generate aptamers for non-fibrillar amyloidogenic proteins17,19–21. Recently, Takahashi et al. reported generation of RNA aptamers against an oligomeric model of Aβ40 and showed reactivity with monomeric Aβ with micromollar affinity29. However, cross-reactivity of these aptamers with fibrillar Aβ40 or with other fibrillar amyloidogenic proteins was not determined.
This work was supported by grants AG030709 from NIH/NIA and 07-65798 from the California Department of Public Health. We acknowledge Margaret M. Condron for peptide synthesis and amino acid analysis, Dr. Elizabeth F. Neufeld for helping and supporting the initial steps of the project, Dr. Chi-Hong B. Chen for providing support and reagents, and Dr. Andrew D. Ellington for helpful discussions.
Disclosures No conflicts of interest declared.