A Novel Acidophilic RNA Motif That Recognizes Coenzyme A

Abstract

Specific recognition of nucleotide cofactors by RNA may be important in engineering new RNA enzymes (ribozymes). Although in vitro selections (SELEX) have identified nucleic acid motifs ("aptamers") that bind a variety of adenosine cofactors, none of these recognizes coenzyme A (CoA), the primary biological cofactor used in acyltransfer reactions. We used SELEX experiments with two random RNA pools to identify aptamers that bind CoA. Functional boundary determination and extensive comparative sequence analysis (including reselection of a mutagenized, circularly permuted RNA) led to the identification of a 52 nucleotide minimal aptamer ("min52"). The RNA structural motif contains a large internal loop with 26 unpaired nucleotides flanked by helices of any base-paired sequence. Twenty loop nucleotides are specifically required for binding activity, 12 of which are derived from the original primer binding sequences. Specificity studies with CoA analogues demonstrated that the aptamer recognizes many adenosine analogues, including ATP, and that recognition is predominantly through the Höogsteen face of adenine. Binding activity is greatest at acidic pH (optimum near 5.0), in low or no monovalent salt, and at high concentrations of either Mg2+ or Mn2+. Strong binding activity (86% of maximum) is observed at pH 4.0, suggesting that at least some extreme conditions (acidic pH) may be compatible with RNA World theories of the origin and early evolution of life. In the presence of 10 mM Mg2+, binding is unaffected by the addition of 1 mM Ca2+, but it is mildly inhibited by 1 mM Zn2+ or Co2+ or by 0.1 mM Cu2+ or Ni2+. The dissociation constant (Kd) for the association of min52 RNA with ATP in solution was measured to be 2.4 +/- 0.4 microM under the conditions of the selection and 0.5 +/- 0.1 microM under optimized conditions. Finally, we show that the selected CoA aptamer populations contain other RNAs at low frequencies that preferentially recognize intact CoA and are not eluted from the resin by AMP alone.

Recombination, RNA evolution, and bifunctional RNA molecules isolated through Chimeric SELEX

“…To test the effect of arbitrarily recombining RNA aptamers, individual 70N and 80N isolates from the populations described above were joined using an overlap extension reaction shown in Figure 2+ RNAs in each populations contain specific primer binding sequences at their 59 and 39 ends to facilitate pool amplification+ The sequence at the 39 ends of the 70N populations is identical to that at the 59 ends of the 80N populations+ When 70N and 80N RNAs are reverse transcribed, converted to double-stranded DNA, mixed, and denatured, the 70N top strand can reanneal with the 80N bottom strand+ The recessed 39 ends can then be extended with any DNA polymerase to yield 150N, chimeric molecules that contain the binding elements from each parental sequence+ The two domains have been, in effect, recombined through their common sequence elements+ Table 1 shows the 22 pairwise combinations generated through the overlap extension reaction+ Aptamers from several different selections were fused in order to minimize artifacts that could be introduced from any particular set+ Series "Q" combines 70Cm and 80CoA individuals, series "R" reverses the order (70CoA with 80Cm), and series "S" combines 70Cm with 80S individuals+ These particular aptamers were chosen because they were among the most active in binding to and eluting from their respective cognate resins Burke & Hoffman, 1998)+ Each RNA was assayed for its ability to be retained on the corresponding affinity resins and to be eluted by free target molecules under conditions similar to those used in the original selections+ The total amount of RNA that could be recovered in this way was generally reduced to between 10% (indistinguishable from background) and 70% (equivalent to original activity within experimental error) of that of the parental aptamer (Fig+ 3)+ Not all combinations were affected to the same degree+ For instance, 80Cm#30 and 80Cm#33 gave similar yields when assayed alone, but when combined with the same three 70CoA domains, the 80Cm#30-derived chimerae (R#2, R#5, and R#8) lost considerably more activity than did the 80Cm#33-derived chimerae (R#3, R#6, and R#9)+ The loss of binding activity is most likely caused by the formation of alternative base pairing patterns that preclude folding into the active conformation ["alternative conformation Hell" (Uhlenbeck, 1995)], such that the active conformation constitutes a smaller fraction of the population of chimeric molecules than it does in the parental 70N or 80N RNA+ Alternatively, subtle conformational changes may distort the binding site and directly reduce the binding affinity+ Either way, loss of affinity resinbinding activity directly reduces the selective fitness of these RNAs+ To test whether misfolding is responsible for reduced resin-binding activity, ATP-binding activities of several RNAs were measured in solution using spin filtration (Jenison et al+, 1994)+ Various concentrations of refolded RNA were incubated with [a-32 P]ATP and filtered through 30-kDa molecular weight cutoff membranes+ Free ATP passes through the membrane, whereas RNA-bound ATP does not+ Dissociation constants (K d s) were determined graphically by plotting the fraction of the ra-FIGURE 3. Effect of chimerae formation on binding activity, as measured by the fraction of all counts recovered from a cognate affinity resin that eluted upon addition of free target+ Affinity resins were derivatized with (A) coenzyme A, (B) 59 AMP (C-8 linkage), or (C) chloramphenicol+ Solid bars, the activity of the aptamer alone; hatched bars, activities of the corresponding chimerae (compare Table 1)+ dioactivity retained above the membrane as a function of RNA in the assay and taking K d to be equal to the ATP concentration at half the calculate...…”

“…Percentage of input RNA eluted from cognate resins during each cycle of the reselection first for Cam-binding (left), then for AMP-or CoA-binding (right)+ Numbers on the x-axis indicate SELEX cycles+ Triangles, population "F" (70CoA/80Cm); circles, population "I" (70Cm/80CoA); squares, population "P" (70Cm/80S)+ Aggregate activities of the parental, monofunctional RNA pools are shown by horizontal lines at the side+ FIGURE 6. Sequences of the bifunctional aptamers from chimeric SELEX, divided into populations F (70CoA/80Cm), I (70Cm/80CoA), and P (70Cm/80S)+ Sequences corresponding to the motifs shown in Figure 1 are underlined+ Three 70CoA (Burke & Hoffman, 1998) or 80Cm aptamers that are homologous to the respective domains of particular chimeric isolates are shown for comparison+ Nearly identical domains are aligned with gaps indicated with dashes+ Overlapping sequences used in fusing the two domains are shown in lowercase except where mutated (uppercase for base changes, dashes for deletions)+ 59 and 39 primer binding sites are shown above+ "n" refers to unreadable sequence positions+ other isolates or in aptamer sequences identified in the original CoA or Cam selections Burke & Hoffman, 1998)+ The 70Cm domains of I#103 and I#130 are identical, and three point mutations distinguish the 70CoA domains of isolates F#218 and F#230, both of which conform to the motif shown in Figure 1C+ The 70CoA domains of F#105, F#120, and F#134 differ by four to five point mutations from each other and from isolate 70CoA#68, which does not share the canonical motif in Figure 1C+ Similarly, the 80Cm domain of F#115 differs from aptamer 80Cm#46 in five positions, whereas the 80Cm domain of F#118 differs from aptamer 80Cm#47 in three positions+ Chimeric isolate I#1 was the only one that appeared to have both CoA-and Cam-binding sequences within one domain, suggesting that the original 70Cm isolate may have also contained the canonical CoA-binding element found in the 70CoA population+ The functional significance of this apparent evolutionary convergence was not tested directly+ The presence of 21 point mutations (including single-nucleotide insertions or deletions) within the overlapping primer binding sequences used to fuse the two domains suggests a net background mutation rate of approximately 1+9% per position subsequent to chimerae formation+…”

“…To test this possibility, we determined the 59 and 39 boundaries of the essential binding elements using deletion/selection experiments (Wallis et al+, 1995;Burke & Hoffman, 1998)+ End-labeled RNA was partially alkaline hydrolyzed and applied to the appropriate affinity resins+ Fractions collected during the wash or elution portions were pooled, normalized for radioactivity, and size-separated by gel electrophoresis+ Typical data are shown in Figure 9A and summarized in Figure 9B+ In every case, the binding elements are fully contained within the expected parental 70N or 80N domain+ Furthermore, in those chimerae containing previously recognized binding sequences, the observed functional boundaries correspond to the edges of the sequences required to form the expected binding element (70CoA domains of F#7 and F#11 for CoA-binding, 80S domain of P#42 for AMP-binding)+…”

“…The "Chimeric SELEX" method described here simulates random recombination among functional RNAs derived from RNA populations with 70 or 80 positions of random sequence+ Specifically, we make use of previously described RNA populations that bind chloramphenicol [Cam ], adenosine , or coenzyme A [CoA (Burke & Hoffman, 1998)]+ These aptamers were selected from random initial pools that contained either 70 ("70N") or 80 ("80N") positions of random sequence+ Among the diverse sequences from the Cam-binding populations ("70Cm" and "80Cm"), there were some that contained the motif in Figure 1A, which has a weak but suggestive similarity to the portions of 23S rRNA involved in Cam binding and peptide bond formation )+ All of the isolates from the adenosine-binding population (80S), which was selected to recognize S-adenosyl methionine (SAM), conformed to the sequence motif shown in Figure 1B+ This element has also been isolated in selections for RNAs that bind ATP (Sassanfar & Szostak, 1993) and NAD ϩ (Burgstaller & Famulok, 1994), and it recognizes many other adenosine derivatives+ Of the two CoA-binding populations, the "80CoA" sequences were highly diverse, whereas the "70CoA" population was dominated by a motif that recruited secondary structural elements and 12 specifically required unpaired nucleotides from the primer binding sites (Fig+ 1C)+ This RNA element also binds many adenosine derivatives, but its specificity differs significantly from that of apta-FIGURE 1. Schematic representation of some of the aptamer modules used in this study+ A: One of the Cam-binding aptamer motifs present in both population 70Cm and 80Cm + B: Conserved adenosine-binding aptamer motif present in population 80S (Sassanfar & Szostak, 1993;Burgstaller & Famulok, 1994;+ C: Conserved CoA-binding aptamer motif found in population 70CoA (Burke & Hoffman, 1998)+ N, any nucleotide; n, its base pairing complement; x, variable number of nucleotides of any sequence+ Nucleotides that were intolerant of mutations or were highly conserved are in bold print+ mers from the 80S population+ The 70CoA and 80CoA populations also contained low frequencies of RNAs that required intact CoA for elution from the CoA resin and were not eluted by AMP alone (Burke & Hoffman, 1998)…”

“…Schematic representation of some of the aptamer modules used in this study+ A: One of the Cam-binding aptamer motifs present in both population 70Cm and 80Cm + B: Conserved adenosine-binding aptamer motif present in population 80S (Sassanfar & Szostak, 1993;Burgstaller & Famulok, 1994;+ C: Conserved CoA-binding aptamer motif found in population 70CoA (Burke & Hoffman, 1998)+ N, any nucleotide; n, its base pairing complement; x, variable number of nucleotides of any sequence+ Nucleotides that were intolerant of mutations or were highly conserved are in bold print+ mers from the 80S population+ The 70CoA and 80CoA populations also contained low frequencies of RNAs that required intact CoA for elution from the CoA resin and were not eluted by AMP alone (Burke & Hoffman, 1998)…”