Zeptomole detection of a viral nucleic acid using a target-activated ribozyme

Abstract

We describe a strategy for the ultra-sensitive detection of nucleic acids using "half" ribozymes that are devoid of catalytic activity unless completed by a trans-acting target nucleic acid. The half-ribozyme concept was initially demonstrated using a construct derived from a multiple turnover Class I ligase. Iterative RNA selection was carried out to evolve this half-ribozyme into one activated by a conserved sequence present in the hepatitis C virus (HCV) genome. Following sequence optimization of substrate RNAs, this HCV-activated half-ribozyme displayed a maximal turnover rate of 69 min(-1) (pH 8.3) and was induced in rate by approximately 2.6 x 10(9)-fold by the HCV target. It detected the HCV target oligonucleotide in the zeptomole range (6700 molecules), a sensitivity of detection roughly 2.6 x 10(6)-fold greater than that previously demonstrated by oligonucleotide-activated ribozymes, and one that is sufficient for molecular diagnostic applications.

FIGURE 1.

A half-ribozyme derived from the Class I ligase. (A) A half-ribozyme based on the sequence of the 207t Class I ligase (black) was derived from the constitutive ligase by introducing a nick (dashed line) so that a trans-acting target nucleic acid (green) will associate and allow ligation (red arrow) of a substrate RNA with a 3′ cis-diol (S_OH, red) to a substrate RNA with a 5′-triphosphate (pppS, blue). Note that pppS and S_OH form a bimolecular substrate RNA complex that interacts with the half-ribozyme-target complex. Nomenclature for paired regions (P1–P7) is analogous to that used for the constitutive ligase. (B) Observed rate of multiple turnover ligation promoted by the 207t half-ribozyme in the presence (closed red circle) versus the absence (open red circle) of bound target relative to that observed without half-ribozyme and target (blue square). Reaction conditions were as follows: buffer (30 mM Tris-HCl at pH 7.5; 60 mM MgCl₂), 1.0 μM half-ribozyme, 0.1 μM target, 25 μM each S_OH and pppS.

FIGURE 2.

Half-ribozyme that interacts with HCV sequences. (A) Conserved sequence of HCV 5′-UTR used as target oligonucleotide (green) and minor sequence variants (red) as a sequence logo (Schneider and Stephens 1990). The height of each letter represents its absolute number of occurrences in ~1500 GenBank entries. (B) Half-ribozyme (black) was engineered to associate with target nucleic acid (green) and to direct the ligation of a substrate RNA with a 3′ cis-diol (S_OH, red) to a substrate RNA with a 5′-triphosphate (pppS, blue).

FIGURE 3.

HCV-activated half-ribozyme from initial selection. (A) Half-ribozyme sequence library (black and yellow) that performs autoligation for use in iterative RNA selection carried non-parental sequence at 62 positions (yellow) and can interact with the HCV target (green) to direct ligation of an extended S_OH substrate RNA that allows PCR (extension not shown). Helix P7 was extended by 17 nucleotides relative to the 207t half-ribozyme (Fig. 1A ▶) and the initial HCV half-ribozyme (Fig. 2B ▶) to provide a nonrandomized sequence of sufficient length to allow reverse transcription. (B) Sequence of the clone 8/7 half-ribozyme isolated from iterative selection. Input sequence (panel A) is black and sequence changes relative to it are indicated (yellow). Three different multiple turnover versions of the clone 8/7 half-ribozyme were constructed by 5′ truncation (numbered blue arrows). Two guanosine residues were added to the 5′ end of each multiple turnover half-ribozyme to allow efficient transcription by T7 RNA polymerase (not shown). (C) Observed rate of autoligation promoted by the clone 8/7 half-ribozyme as a function of pH. Reaction conditions: buffer (30 mM MES at pH 6.0 and 6.5 or Tris-HCl at pH 7.0–8.3, 60 mM MgCl₂), 1 μM half-ribozyme, 1.1 μM target, trace S_OH. (D) Observed rate of autoligation as a function of the concentration of MgCl₂ of the clone 8/7 half-ribozyme (red circles) and the final half-ribozyme sequence library obtained from a second iterative RNA selection (blue squares; see text and Fig. 4 ▶). Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5; MgCl₂ as indicated), 1 μM half-ribozyme, 1.1 μM target, trace S_OH.

FIGURE 4.

Efficient HCV sequence-activated multiple turnover half-ribozymes. (A) Four sequence libraries (libraries 1–4) based on the autoligation version of the clone 8/7 half-ribozyme (black) maintained the nucleotide changes in the clone 8/7 half-ribozyme sequence relative to the 207t half-ribozyme (yellow). Library 1 carried completely randomized sequence in all other single-stranded positions (blue highlight). Libraries 2–4 carried seven positions that were randomized at a 6% frequency (boxed) and an inserted random sequence domain (blue boxes). The four libraries were 3′ truncated relative to the clone 8/7 half-ribozyme and were used with an appropriately truncated HCV target sequence (green) because a reverse transcription primer could adequately anneal to a defined sequence internal to that used in the initial iterative RNA selection. S_OH was extended at its 5′ end to allow for PCR amplification (not shown). (B) The clone 21 half-ribozyme isolated from iterative selection maintained all of the sequence changes produced in the clone 8/7 half-ribozyme isolated from the initial iterative RNA selection (yellow) and carried an inserted region that, on examining all members of this sequence family, had a conserved portion (blue) and a variable portion (purple). Two conserved changes were present in J3/4 (blue). An alternate P3 helix allowed by the inserted region is shown and was tested using mutations (pink) in the 3′ side of P3 (M1) and sequence, allowing either the alternate (M2) or original (M3) P3 base-pairing arrangement. Three different multiple turnover versions of the clone 21 half-ribozyme were constructed by 5′ truncation (numbered blue arrows). Two guanosine residues were added to the 5′ end of each multiple turnover half-ribozyme to allow efficient transcription by T7 RNA polymerase (not shown). Joining regions J1/3 and J3/4 are indicated (gray). (C) Time course of autoligation of the clone 21 half-ribozyme (red circles) or variants either carrying mutations M1 and M2 (blue squares) or mutations M1 and M3 (green triangles). Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5, 10 mM MgCl₂), 1 μM half-ribozyme, 1.1 μM target, trace S_OH. (D) Observed rate of multiple turnover ligation promoted by the clone 21 half-ribozyme in the presence (solid red circle) or absence (open red circle) of bound target relative to that observed without half-ribozyme or target (blue square). Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5, 60 mM MgCl₂), 1 μM half-ribozyme, 0.01 μM target, 10 μM each S_OH and pppS.

FIGURE 5.

Characterization and optimization of clone 21 in multiple turnover configuration 3. (A) The rate of the initial catalytic cycle of multiple turnover configuration 3 (blue) was compared with the rate of autoligation (red) when their respective substrate RNAs were preincubated as indicated. Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5, 60 mM MgCl₂). Autoligation reactions contained 1 μM half-ribozyme, 1.1 μM target, and trace S_OH; reactions using the multiple turnover half-ribozyme contained 10 μM half-ribozyme, 10 μM and 1 μM each S_OH and pppS. (B) Turnover rates of configuration 3 afforded by mutant S_OH and pppS substrate RNAs expressed relative to the initial substrate RNA pair (red and blue). P2 interaction with target nucleic acid (black line) and potential wobble base pairs (green highlight) are indicated. Circles indicate relative rates of mutations further characterized. Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5, 100 mM KCl, 60 mM MgCl₂), 1 μM half-ribozyme, 0.01 μM target, 10 μM each S_OH and pppS. (C) Turnover rate of configuration 3 afforded by original (red circles), C8U/flip-13 (blue squares), or C8U/flip-13/A5G (green diamonds) substrate RNA pairs as a function of substrate RNA concentration. Reaction conditions were the same as panel B, except the concentration of S_OH and pppS varied, as indicated. (D) Turnover rate of configuration 3 using the C8U/flip-13/A5G substrate RNA pair as a function of MgCl₂ and KCl concentrations. Reaction conditions: buffer (30 mM Tris-HCl at pH 7.5, MgCl₂, and KCl concentration varied as indicated), 1 μM half-ribozyme, 0.001 μM target, 1 μM each S_OH and pppS. (E) Maximum turnover rate of multiple turnover configuration 3 using the C8U/flip-13/A5G substrate RNA pair as a function of pH inferred from Lineweaver-Burk analysis (closed red circles) compared with direct measurement of rate in the absence of target nucleic acid at identical pH values (open red circles).

FIGURE 6.

Limit of detection of HCV sequence. (A) Calculated LOD was determined by solving Equation 5 (text) after substrate RNA concentration was converted to number of molecules using kinetic parameters established for clone 21 half-ribozyme in configuration 3 and substrate RNA pair C8U/flip-13/A5G. Calculated LOD is expressed as a function of substrate RNA concentration and pH. The intersection of lines on the surface of the graph indicates specific substrate RNA concentrations and pH values. The surface corresponding to 10⁴ and 10⁵ target molecules is indicated (dashed white line). The condition used to experimentally determine LOD is indicated (blue circle). (B) Ligation product from duplicate reactions, examining product formation as a function of HCV target copy number. The minor species migrating more rapidly than the major species observed in some lanes is derived from N₋₁ pppS substrate RNA generated from in vitro transcription and was not used in quantification. The large amount of unligated S_OH carrying radiolabel was well separated from ligated product (not shown). Reaction conditions: buffer (30 mM MES at pH 6.0, 150 mM MgCl₂, and 0.9 M KCl), 1 μM half-ribozyme, target as indicated, 0.1 μM each S_OH and pppS. (C) Quantification of product formation as a function of HCV target copy number from four half-ribozyme reactions: two reactions each from two independent serial dilutions. Dashed red line with arrowhead indicates LOD extrapolated from a power function fit to signal from 10⁷ to 10⁴ HCV copies to signal observed in the absence of HCV target. Standard deviation from four separate trials amounted to <10% of the average half-ribozyme activity in the presence of 10⁷ to 10⁴ molecules and <15% in reactions containing 0, 10², and 10³ target molecules. Power function fit to signal observed with 10⁷ to 10⁴ HCV molecules with R² = 0.99946.