April 15, 1998
Pool immobilization is a common procedure in in vitro selection. Generally, random oligonucleotide pools aggregate, especially in the presence of divalent cations such as Mg++, Ca++. To prevent this, pools may be annealed to column-immobilized primers. Primers are generally immobilized as biotinylated oligonucleotides on immobilized avidin or streptavidin. Pool immobilization was first used by David Bartel in his RNA ligase selection(1). In his procedure, he annealed a biotinylated oligo to the 3' constant region of the pool RNA. This mixture was then incubated with avidin-agarose beads. An alternate procedure, the one presented here, is to bind the primer to immobilized streptavidin, then anneal the pool RNA.
Preparation of the Oligonucleotide Primer
Generally primers used for immobilization of pools contain a pairing region, a spacer and a biotinylated nucleotide such as Bio-TEG or Biotin-dT. The pairing region is generally 18 or so bases long and complimentary to a constant region in the pool. The spacer is generally 3 bases long, such as dTdTdT. Its function is to provide a flexible tether between the pairing region and the biotinylated nucleotide (Figure 1).
Pools may be immobilized with 5' or 3' biotinylated oligos, although 5' biotinylated oligos are used more frequently. For 3' biotinylated oligos, there exist DNA synthesis columns prepared with such nucleotides. For 5' biotinylated oligonucleotides, Biotin-dT (Glen Research) is especially useful because it can be phosphorylated with ATP-g32P and T4 Kinase. Since coupling yield of biotin-dT to synthetic oligonucleotides is generally less than 50%, these oligos must be purified by gel electrophoresis. This is generally accomplished on a 20% polyacrylamide gel. The separation may be facilitated by adding 3 or 4 more bases after the Biotin-dT during synthesis. Because coupling of bases to the 5' hydroxyl of biotin-dT is efficient, what results is a biotinylated oligo 4 or 5 bases longer than the capped, unbiotinylated impurities. After elution from the gel, the purity of oligos with 5' terminal biotin-dT may be checked by one of two methods. One can radioactively label the oligos using T4-kinase purify, then measure how much radioactivity is retained on immobilized streptavidin after incubation. Alternatively, a gel can be run with two lanes containing oligo, one lane containing streptavidin, and one lane not. The fraction of the oligo gel-shifted by the streptavidin is a measure of fraction biotinylated.
Preparation of Immobilized Streptavidin
A 100 無 streptavidin column is prepared by adding 200 無 immobilized streptavidin slurry to a plastic BioRad minicolumn. The immobilized streptavidin is rinsed with 6 mL (60 column volumes) wash buffer(2). This rinse removes, among other things, unbound streptavidin molecules and is necessary to insure that the biotinylated oligo will be immobilized. The column is then capped on the bottom.
Immobilization of biotinylated oligo on Streptavidin
23 無 100 然 biotinylated oligo (2.3 nmol) is added to the capped column along with 100 無 wash buffer. This is about 1/3 of the theoretical capacity of the gel, 68 nmol/mL. The mixture is agitated 1 hour, then allowed to settle. Centrifugation of the column for a brief time at 400 RPM in the swinging bucket centrifuge can facilitate this. The column is then washed with 4 mL 0.5x wash buffer to remove unbound oligo.
Immobilization of Pool RNA
0.46 nmols 32P-labeled pool RNA in 50-500 無 25 mM HEPES pH 7.4, 1 mM EDTA is prepared for gel immobilization by heating it for 3 minutes, then allowing it to cool 3 minutes. It is then added to the immobilized biotinylated-oligo/streptavidin column along with an equal volume of wash buffer. The column is then capped and the mixture is agitated for 1 hour, then centrifuged briefly at 400 RPM to settle column material. The gel is washed with 2 mL (20 volumes) selection buffer(3) or whatever is the appropriate buffer for the next step of the selection. A survey meter may be used to estimate the fraction of pool that stuck to the column. When the biotinylated oligo is very pure, this procedure gives >65 % pool immobilization for RNA pools around 288 bases long. For shorter pools such as an 86 base pool, greater than 90 % of the RNA is immobilized.
Immobilized RNA can be eluted from the column by first washing the column with 40 volumes wash buffer to remove divalent cations that may degrade RNA at high pH. Then the RNA is eluted by addition of two 1 mL fractions of 10 mM NaOH 1 mM EDTA to the column and collecting the eluent in two 100 mL aliquots of 250 mM HEPES acid to immediately neutralize the base once the RNA is eluted. The pH of the resulting solution is around 7.4.
The choice of whether to use avidin or streptavidin is based on several factors. Avidin is much less expensive than streptavidin. The two proteins bind biotin with similar affinities, Ka=1015M-1, but avidin is glycosylated, and therefore more hydrophilic than streptavidin. Furthermore, the pI of avidin is about 10 whereas that of streptavidin is about 5. This means that at neutral pH, avidin will be positively charged whereas streptavidin will be slightly negatively charged. Because of its positive charge and hydrophilic nature, avidin is likely to have a higher degree of non-specific RNA binding than streptavidin. However, successful RNA selections have been performed with both immobilized proteins.
Denaturation of the RNA before binding to the column is done in low ionic strength buffer. This disfavors formation of rigid structures. Annealing by contrast must be done in the presence of salt. Efficient annealing occurs in 0.5 x wash buffer.
The Pierce technical bulletin for ImmunoPure Immobilized Streptavidin(4) states that there is 1-2 mg streptavidin per mL agarose gel. This corresponds to a 17-34 nmol protein per mL agarose (17-34 然), or 68-136 nmol binding sites per mL agarose (68-136 然). The amount of primer added to the streptavidin agarose is 23 nmol/mL agarose (23然), roughly 1/3 to 1/6 of the theoretical capacity. The amount of RNA annealed, 4.6 nmol RNA per mL agarose is 1/5 of the theoretical capacity based on complete binding of the oligo.