FIGURE 2.

Designs of RNA pools used to isolate the original nucleotide synthase ribozymes and generate improved variants of the a15 isolate. The RNA pools were transcribed from double-stranded DNA pools. Regions of random or degenerate sequence are shown as boxes; regions of fixed sequence are shown as lines. (A) The initial random-sequence pool (Unrau and Bartel 1998). This pool had three 76 nt random-sequence regions (N76), joined by StyI and BanI restriction sites. A population of ~1.5 × 10¹⁵ different sequences was used in the initial selection and yielded three nucleotide synthase families. (B) The degenerate-sequence pool based on a 5′ truncated variant of a15, the prototype isolate of the family A nucleotide synthase ribozymes. This pool was constructed to generate ribozyme variants with sequence diversity suitable for identifying critical residues and inferring secondary structure by comparative analysis. It was made from two pieces linked by a BanI site. Most residues of the a15 ribozyme were within the 73 and 93 nt degenerate regions (M73 and M93), which were mutagenized such that on average 80% of the residues would be the same as the starting sequence and 20% would be substituted by one of the other three possibilities. A 10 nt random-sequence region (N10) was inserted into the terminal loop of stem I. Note that in its RNA form this pool has variation extending to its 3′ terminus. (C) The pool with variation in both sequence and length. This pool was based on the a.6.10 ribozyme isolated from the degenerate-sequence pool (panel B). Two pools were synthesized, each containing five regions (R1–R5) in which a modified coupling cycle was used to introduce both point deletions and substitutions. The number of positions in each mutagenized region varied from 7 to 25, as indicted by the numbers prefixed by a D (D15, D25, D8, D10, D11, and D7; note that R4 spans both the D10 and D11 boxes). Each position subjected to the modified coupling cycle incurred a 20% substitution frequency and a deletion frequency of either 17% or 25%, depending on the pool. The remaining sequence was kept fixed to form the ribozyme helices (stems I to VI), the arms of which are indicated by Roman numerals. Equimolar amounts of the two pools were combined to generate a starting pool with ~5 × 10¹³ unique deletion mutants.